ABCD1 deletion-induced mitochondrial dysfunction is corrected by SAHA: implication for adrenoleukodystrophy

Abstract

X-linked Adrenoleukodystrophy (X-ALD), an inherited peroxisomal metabolic neurodegenerative disorder, is caused by mutations/deletions in the ABCD1 gene encoding peroxisomal ABC transporter adrenoleukodystrophy protein (ALDP). Metabolic dysfunction in X-ALD is characterized by the accumulation of very long chain fatty acids (VLCFAs; ≥ C22:0) in the tissues and plasma of patients. Here, we investigated the mitochondrial status following deletion of ABCD1 in B12 oligodendrocytes and U87 astrocytes. This study provides evidence that silencing of peroxisomal protein ABCD1 produces structural and functional perturbations in mitochondria. Activities of electron transport chain-related enzymes and of citric acid cycle (TCA cycle) were reduced; mitochondrial redox status was dysregulated and the mitochondrial membrane potential was disrupted following ABCD1 silencing. A greater reduction of ATP levels and citrate synthase activities was observed in oligodendrocytes as compared to astrocytes. Further, most of the mitochondrial perturbations induced by ABCD1 silencing were corrected by treating cells with SAHA (suberoylanilide hydroxamic acid), an HDAC inhibitor. These observations indicate a novel relationship between peroxisomes and mitochondria in cellular homeostasis and the importance of intact peroxisomes in relation to mitochondrial integrity and function in the cell types that participate in the pathobiology of X-ALD. These observations suggest SAHA as a potential therapy for X-ALD.

Introduction

X-linked adrenoleukodystrophy (X-ALD) is the most common peroxisomal disorder, occurring in one out of every 17,000 births (Singh et al. 1984a, Moser et al. 2001b). X-ALD is caused by the various mutation or deletion of the ABCD1 gene which codes for a peroxisomal ATP-binding cassette transporter type D (ABCD1), also known as adrenoleukodystrophy protein (ALDP) (Mosser et al. 1993, Dubois-Dalcq et al. 1999) (www.x-ald.nl). X-ALD is considered as a neurometabolic disease because of very long chain fatty acids (VLCFA)-mediated progressive demyelination in the central nervous system (CNS), axonopathy in the spinal cord, and adrenal insufficiency (Moser et al. 2001a). Mutations in ABCD1 can lead to a variety of clinical phenotypes ranging from the relatively benign adult disease of adrenomyeloneuropathy (AMN) to a fatal childhood cerebral ALD (cALD). cALD is characterized by progressive cerebral demyelination with a strong inflammatory response in the white matter leading to neurodegeneration and death often before the patient reaches adolescence (Moser et al. 2001a, Singh & Pujol 2010). AMN affects adults (second to fourth decade) and is characterized by a pure myelopathy and peripheral neuropathy. The same mutation could result in cALD or AMN phenotype (Berger et al. 1994). Molecular mechanisms of VLCFA-induced inflammatory disease in cALD versus the milder phenotype in patients with AMN may involve modifier genes or environmental/epigenetic mechanisms but not well understood at present (Korenke et al. 1996, Smith et al. 1999, Berger & Gartner 2006).

The biochemical “hallmark” of X-ALD and its less severe adult-onset form, adrenomyelonueropathy (AMN), is excessive accumulation of VLCFAs (≥C22:0) which is used as a diagnostic test for X-ALD (Moser et al. 2001a). Studies from our laboratory have reported that VLCFA are β-oxidized in peroxisomes (Singh et al. 1984a) and that VLCFA accumulation in X-ALD/AMN is caused by its defective catabolism in peroxisomes (Singh et al. 1984b, Poulos et al. 1986, Wanders et al. 1988). The ABCD1 gene product, a peroxisomal membrane transporter protein (adrenoleukodystrophy protein, ALDP) is described to participate in the translocation of VLCFA into peroxisomes (van Roermund et al. 2012, van Roermund et al. 2008). The transport of VLCFA as compared with their CoA derivatives into peroxisomes has been the subject of debates in subsequent studies. Previously, we reported that unlike mitochondrial fatty acid transport, VLCFA are transported into peroxisomes as free fatty acids (Singh et al. 1992) and are converted inside the peroxisome to their CoA derivatives by VLCFA-acyl-CoA synthase for their β-oxidation (Singh et al. 1992). Studies using over-expressed human ABCD1 in Saccharomyces cerevisiae concluded that VLCFA are transferred into peroxisomes as their CoA derivatives (van Roermund et al. 2012). However, a recent study reported that ABCD1 has a thioesterase activity to degrade VLCFA-CoA to free VLCFA prior to VLCFA translocation into peroxisomes where it is again activated in the lumen to VLCFA-CoA by peroxisomal VLCFA-CoA synthase for its further β-oxidation (De Marcos Lousa et al. 2013). These findings are consistent with our earlier studies documenting transport of free VLCFA as compared to its CoA-derivatives into peroxisomes and that VLCFA-acyl-CoA synthase is associated with the luminal surface of peroxisomal limiting membrane (Lazo et al. 1990, Singh et al. 1992, Smith et al. 2000). Three such transporters have been identified in peroxisomes - ABCD1, 2, and 3 - with significant sequence homology (Kamijo et al. 1990, Kamijo et al. 1992, Lombard-Platet et al. 1996, Holzinger et al. 1997). The correction of the metabolic defect in X-ALD fibroblasts following transfection with ABCD1 established its function in the peroxisomal VLCFA β-oxidation (Cartier et al. 1995, Braiterman et al. 1998). Further, correction of VLCFA metabolism following transfection of ABCD2 or ABCD3 into X-ALD cells suggest promiscuous activity among these transporters and thus possible correction of metabolic defect in X-ALD disease (Kemp et al. 1998, Netik et al. 1999) is proposed to have a certain functional redundancy of ABCD2 with ABCD1 (Netik et al. 1999, Kemp et al. 1998). Studies from our laboratory and others reported the normalization of metabolic defect by pharmacological induction of ABCD2 in ABCD1 silenced astrocytes and oligodendrocytes, ALD mice as well as in X-ALD derived skin fibroblasts (Singh et al. 2011, Singh et al. 2013b, Gondcaille et al. 2014, Singh et al. 2013a).

Studies of X-ALD post-mortem brains (Gilg et al. 2000, Powers et al. 2005), skin fibroblasts (Vargas et al. 2004), and lymphocytes (Uto et al. 2008) from X-ALD patients documenting oxidative and nitrosylated proteins implied that reactive oxygen species (ROS) and reactive nitrogen species (RNS; ex: ONNO-) mediated mechanisms participate in ALDP-loss-induced pathologies in different cell types. The relationship between ABCD1-loss-induced VLCFA accumulation and cellular oxidative stress was further established by the observed ROS in Abcd1/Abcd2 silenced astrocytes (Singh et al. 2009b), oligodendrocytes (Singh et al. 2013a, Baarine et al. 2012a) and C6 cells (Baarine et al. 2012b). Furthermore, a study using autopsy samples from X-ALD patients and ABCD1/ABCD2 silenced astrocytes also described a correlation between the accumulation of VLCFA and an induced lipotoxic response via activation of 5-lipoxygenase in X-ALD pathology (Khan et al. 2010). Also, the increase in oxidative markers (glutamic semialdehyde (GSA), aminoadipic semialdehyde (AASA) and others) in fibroblasts derived from X-ALD patients (Fourcade et al. 2008) supported the relationship between VLCFA and increased cellular oxidative stress. These observations indicate that VLCFA-induced oxidative stress may play a role in the pathobiology of X-ALD.

Singh and Pujol (Singh & Pujol 2010) proposed a “three-hit hypothesis” model as a framework to better understand the molecular mechanisms associated with the disease pathogenesis of X-ALD. The metabolic derangements in VLCFA caused by the loss of ALDP (first hit) generates oxidative stress (second hit) which in turn, with the participation of environmental, stochastic, genetic, or epigenetic factors, may induce inflammatory disease and a subsequent generalized loss of peroxisomes or peroxisomal function (third hit), creating a vicious cycle resulting in cell loss and progressive inflammatory demyelinating disease. Recent studies using cell culture models and X-ALD mice show that ALDP dysfunction and VLCFA accumulation in vitro and ex vivo triggers free radical production, mitochondrial depolarization, increased intracellular Ca²⁺, and LDH release (supporting an activation of cell death) (Fourcade et al. 2010, Hein et al. 2008, Baarine et al. 2012a, Baarine et al. 2012b, Galino et al. 2011). A recent functional genomic analysis study of mouse model of ALD disease and human autopsy tissue reported mitochondrial dysregulation, insulin desensitization, and an NF-κB-mediated proinflammatory disease (Schluter et al. 2012). Mitochondria, a dynamic organelle often referred as “powerhouse of the cell” and “ATP reservoir”, are essential for cellular energy requirements. Therefore, any defect in brain mitochondrial function may lead to energy deficiency as well as increased generation of reactive oxygen species (ROS) and ultimately to neurodegeneration (Chaturvedi & Flint Beal 2013). Peroxisomal alterations influence mitochondrial functions (Schrader & Fahimi 2006, Fourcade et al. 2008, Hein et al. 2008, Baarine et al. 2009, Muller et al. 2011). However, clinical significance of mitochondrial dysfunction will only be relevant to X-ALD in cell types that accumulate VLCFA and in turn participate in the disease pathology.

The brain is the major site for cALD pathology, where inflammatory mechanisms are upregulated in astrocytes and microglia and cell death signaling is upregulated in oligodendrocytes (Eichler et al. 2008, Schluter et al. 2012, Hein et al. 2008, Singh et al. 2013a). Therefore, we investigated VLCFA derangement induced effects on mitochondrial homeostasis, function and biogenesis in B12 rat oligodendrocyte and U87 human astrocyte cell lines stably silenced for ABCD1, as a cell culture model of X-ALD pathology. We observed that peroxisomal ALDP absence induces mitochondrial dysfunction including mitochondrial biogenesis dysregulation and loss of electron transport chain enzymatic activities, resulting in dysregulation of mitochondrial redox and mitochondrial membrane potential in both U87 and B12 cells. These observations describe a novel relationship between peroxisome metabolism and mitochondrial integrity and functions. Interestingly, consistent with our previous report describing correction of metabolic defects in X-ALD cells by suberoylanilide hydroxamic acid (SAHA), an HDAC inhibitor, SAHA also corrects the mitochondrial perturbations indicating its therapeutic value in X-ALD disease (Singh et al. 2013a).

Materials and methods

Cell cultures and treatments

Rat B12 oligodendrocytic cells (Sigma-aldrich, MO) and human U87 astrocytes (ATCC, VA) silenced for ABCD1 (Singh et al. 2013a). Briefly, a set of 3 human (SK-009605-00-10) and rat (SK-098142-00-10) specific SMART vector 2.0 lentiviral shRNA particles (10⁸ TU/ml) for ABCD1 were purchased from Thermo Fisher Scientific Dharmacon (CO, USA). The vector had an hCMV promoter, a TurboGFP reporter gene and a puromycin selection gene. Human U87 astrocytes and rat B12 oligodendrocytes were cultured in DMEM with 10% FBS in the presence of antibiotic, and viral particles (ABCD1 and non-targeting control) were added with a multiplicity of infection (MOI) of 2.5 and 3.0 respectively for U87 astrocytes and B12 oligodendrocytes. The selected cells were maintained in culture media containing puromycin at 0.5 μg/ml until further use. ABCD1 silencing was observed by Western blot and mRNA quantification (Singh et al. 2013a). ABCD1 silenced U87 and B12 cells were treated with SAHA at 5 μM every 24 h for 72 h.

RNA extraction, gene expression analysis by RT-PCR and protein analysis by western blot

Following total RNA extraction using TRIzol (Invitrogen, NY), cDNA was synthesized from total RNA by using RT² First Strand kit and protocol (Qiagen, CA). Real time PCR was conducted using Bio-Rad iCycler (iCycler iQ Multi-Color Real Time PCR Detection System; Bio-Rad, CA) real time system. The expression of human and rat mitochondrial energy metabolism genes involved in mitochondrial respiration was profiled by using Mitochondrial Energy Metabolism plus RT² Profiler PCR Array (Qiagen). Data analysis was performed by using the Web-based software for cataloged and custom arrays developed by Qiagen (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php). Protein extraction and western blot analysis using U87 and B12 cells were performed as described earlier (Baarine et al. 2009) using various antibodies raised against PGC-1α (Abcam 106814, MA), PGC-1β (Santa-Cruz, TX), ABCD1 (EMD Millipore, MA), β-actin (Cell Signaling, MA), and GAPDH (Abcam).

Measurement of transmembrane mitochondrial potential with JC-1

Variations in the mitochondrial transmembrane potential (ΔΨ_m) were measured with 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide (JC-1; Invitrogen) as described earlier (Baarine et al. 2009). Briefly, JC-1 stock solution was prepared at 1 mg/ml in DMSO, stored in the dark at −20°C. For mitochondrial depolarization, cells (1×10⁶) were incubated with JC-1 (1μg/ml final concentration) for 15 min at room temperature in culture medium. JC-1 (λex =530nm) fluorescent was quantified at 590 nm (λem) using a Max Pro spectrofluorometer (Molecular Devices, Sunnyvale, CA).

Mitochondria characterization by staining with Mitotracker Red and MitoSOX

MitoTracker Red CMXRos (Invitrogen) (λex ~579 nm and λem ~600 nm) localises to mitochondria with healthy membrane potential, and fades as mitochondria lose membrane potential. Cells were incubated for 30 min with 100 nM of mitotracker red at 37°C before fixing (4% para-formaldehyde) for microscopy. Images were taken by using an inverted fluorescent microscope (Olympus BX60 equipped with Olympus DP70 camera. Corrected total cell fluorescence (CTCF) was calculated as follows: CTCF = Integrated Density - (Area of selected cell X Mean fluorescence of background readings) using ImageJ software. This analysis was performed in at least 60 cells per measurement. MitoSOX^™ Red reagent (λex ~510 nm and λem ~580 nm) permeates live cells and is selectively targeted to the mitochondria. MitoSOX was used to evaluate the mitochondrial superoxide anion (O₂^·−) as described earlier (Baarine et al. 2009). Briefly, cells (1×10⁶) were incubated with MitoSOX (5 μM) in PBS for 30 min at 37°C followed by microscopic and spectrophotometric analysis. The fluorescence was detected at 580 nm using a Max Pro spectrofluorometer (Molecular Devices, Sunnyvale, CA).

Adenosine triphosphate measurement and NAD+/NADH

ATP levels were measured using ATP Colorimetric Assay Kit from Biovision. Briefly, cell (1 × 10⁶) homogenate was deproteinized and ATP levels were measured in supernatant using a kit from Biovision according to the manufacturer’s protocol. ATP levels were normalized by number of cells and then to the control and presented as percent of control. NAD+, NADH and NAD+/NADH ratio was measured by using the NAD/NADH Quantitation Colorimetric Kit from biovision. Briefly, cells (2 × 10⁵) disrupted by two freeze/thaw cycles were centrifuged at 14000 rpm for 5 min at 4°C and supernatant was analyzed for total NAD (NADH and NAD) and NADH measurements according to the manufacturer’s protocol.

Mitochondrial enzymatic activities

The specific activities (Units/mg of proteins) of mitochondrial OXPHOX complexes (I, II, IV, and V) were measured by spectrophotometry at 37°C in cell fractions with an adapted protocol (Malgat et al. 1999, Flamment et al. 2008, Guillet et al. 2011, Guillet et al. 2010). Mitochondria are routinely isolated by differential centrifugation as described previously by our laboratory (Singh et al., 1993) with minor modifications. Briefly, 5 to 7 million cells homogenised in buffer A (100 mM sucrose, 50 mM KCl, 50 mM Tris base and 5 mM EGTA, pH 7.4 with HCl) were centrifuged at 450 × g for 5 min at 4°C to remove unbroken cells and nuclei. Supernatant was further centrifuged at 12,000× g for 20 min at 4°C and the pellet was suspended in Buffer A for enzyme activity measurements.

Complex I (NADH ubiquinone reductase)

Activity was measured in a reaction medium containing 80 mM phosphate buffer (pH 7.4), 1 mg/ml fatty acid-free bovine serum albumin, 100 μM decylubiquinone, 1 mM KCN, 75 μM 2,6-dichlorophenol-indophenol (DCPIP), and 5 μg of mitochondrial proteins with or without 10 μM rotenone (rotenone is an inhibitor of complex I). After 3 min of incubation at 37°C, the reaction was initiated by adding 0.3 mM NADH (final concentration). The activity was measured at 600 nm by monitoring the reduction of DCPIP (oxidation of NADH).

Complex V (ATP synthase)

Activity is measured by a coupled assay using lactate dehydrogenase and pyruvate kinase as the coupling enzymes. Mitochondrial fraction prepared above were disrupted by 2 freeze-thaw cycles followed by 6 cycles of sonication of 5 s each. Ten μg of disrupted mitochondria was incubated in a buffer containing 50 mM Tris, 5 mg/ml fatty acid-free bovine serum albumin, 5 mM MgCl2, 10 mM KCl, 2 mM phosphoenolpyruvate (PEP), 0.5 mM ATP, 0.5 μg/ml antimycine A, 3 μM carbonyl cyanide p (trifluoromethoxy) phenylhydrazone (FCCP), 20 mU/μl lactate dehydrogenase, and 20 mU/μl pyruvate kinase, pH 8. The reaction was initiated by adding 100 μM NADH and the activity was measured at 340 nm by monitoring the reduction of NADH as described previously (Malgat et al. 1999, Flamment et al. 2008, Guillet et al. 2011, Guillet et al. 2010). The nonspecific reduction of NADH was measured under the same conditions after the addition of 10 μg/ml oligomycin.The specific activity of complex V was calculated by subtracting the oligomycin-insensitive activity.

Citrate Synthase

Activity was measured in a reaction medium consisting of 0.15 mM 5,5′-dithiobis(2-nitrobenzoic acid) (regenerated with Tris buffer, 1 M, pH 8.1), 0.5 mM oxaloacetate, 0.3 mM acetyl-CoA, and 0.1% Triton X-100 following initiation of reaction by addition of 10 μg of mitochondrial proteins, from freeze/thaw disrupted mitochondria fraction. The change in optical density was recorded at 412 nm for 1.5 min as described earlier (Flamment et al. 2008).

MnSOD activity

The SOD activity was measured in freshly isolated mitochondria using a superoxide dismutase assay kit according to the manufacturer’s procedure (Cayman Chemical, MI) in a fraction of isolated mitochondria (Singh et al., 1993). SOD specific activity was expressed units per milligram of proteins.

Measurement of cellular respiration using a seahorse XF-96e analyser

Measurement of the oxygen consumption rate of WT, NT, ABCD1 silenced cells or SAHA treated cells were measured in real-time, simultaneously, in an XF96 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA) as described by Nadanaciva (Nadanaciva S 2012). Briefly, cells were seeded in XF96-well plates at 10,000 cells/120μL culture medium/well and incubated in a 37 °C, 5 % CO2 humidified atmosphere for 24 h. The XF-96 sensor cartridge containing 4 reagent delivery ports (ports A, B, C and D) for injecting test compounds were placed in XF calibration buffer. The following day, the culture medium from the cell plates was aspirated and the cells rinsed three times in pre-warmed XF assay medium modified DMEM. The cells were then maintained in 150 μL/well of XF assay medium modified DMEM at 37 °C for 30 min to allow the temperature and pH of the medium to reach equilibrium before the first rate measurement. Modulator molecules (25 μL of oligomycin, FCCP, Rotenone and antimycin) were then pre-loaded into reagent delivery port A, B and C respectively, of each well in the XF96 sensor cartridge to provide the final concentration of each reagent.

The sensor cartridge and the calibration plate were then loaded into the XF96 Extracellular Flux Analyzer to calibrate the cartridges. When the calibration was complete, the calibration plate was replaced by plates with cells. Four baseline rate measurements of the oxygen consumption rate (OCR) of cells were made using a 2 min mix, 3 min measure cycle. The compounds were then injected pneumatically (by the XF96 Analyzer) into each well, and 4 OCR measurements were made using the 2 min mix, 3 min measure cycle. The instrument measures the extracellular flux changes of oxygen and protons in the medium surrounding the cells.

Statistical analysis

Statistical analyses were performed on data from at least three independent experiments. Statistical significance was determined using the non-parametric Mann and Whitney test. Data were expressed as mean ± SD of n determinations. P value less than 0.05 was considered statistically significant.

Results

ABCD1 silenced oligodendrocytes and astrocytes display mitochondrial membrane potential dysfunction

ABCD1 silencing in B12 and U87 cells manifested by decreased levels of ABCD1 protein and RNA message (Fig. 1A-C), a reduction of peroxisomal VLCFA β-oxidation and by accumulation of VLCFA (Not shown) (Singh et al. 2013a). To assess mitochondrial health and morphology in these cells, we measured the mitochondrial membrane potential (ΔΨ_m) (Fig. 1D, E). ΔΨ_m is a key indicator of cellular viability, as it reflects the pumping of hydrogen ions across the inner membrane during the process of electron transport and oxidative phosphorylation, the driving force behind ATP production. Silencing of ABCD1 induced 30% and 20% decrease in ΔΨ_m of B12 and U87, respectively (Fig. 1D, E). Data are expressed as percent of control (RFU/10⁶ cells). The absolute RFU (/10⁶ cells) or relative fluorescent unit measured using JC-1 dye in U87 astrocytes (7995) is twice as high as that seen in B12 oligodendrocytes (4290).

Figure 1. Effects of ABCD1 silencing on the mitochondrial membrane potential and redox status.

ABCD1 was silenced using a set of three human (SK-009605-00-10) and rat (SK-098142-00-10) specific SMART vector 2.0 lentiviral shRNA particles (108 TU/ml) for ABCD1. SMARTvector 2.0 non-targeting shRNA control particles (NT) (108 TU/ml, Thermo Fisher Scientific) designed so that no known gene targeted in human, mouse or rat were used as negative controls (A-C). The transmembrane mitochondrial potential (ΔΨ_m) was measured with JC-1 dye after 72 h of culture (D, E). For O₂^·− measurement, cells were stained with MitoSOX and then observed by fluorescent microscope (F) or the fluorescence was quantified by spectrofluorometry (G). Data shown are mean ± SD from three independent experiments performed in duplicate. For MnSOD activity (H), mitochondria were isolated from B12 and U87 cells and MnSOD activity was biochemically determined (H). Activity is presented as specific activity (SA). ABCD1 gene expression was normalized to the housekeeping gene RPLP0. Data are represented as mean ± SD (n = 3). ^*P < 0.05, ^**P < 0.005, ^***P < 0.001 versus WT.

Mitochondrial redox balance is altered in ABCD1 silenced cells

The effects of ABCD1 silencing on mitochondrial anti-oxidant defenses were determined by measuring the enzymatic activity levels of MnSOD in ABCD1 silenced astrocytes and oligodendrocytes. As depicted in Figure 1, loss of ABCD1 in both cell lines induces a reduction in MnSOD enzymatic activity (Fig. 1H) by about 25%. MnSOD enzymatic activity in U87 WT cells (2U/mg of protein) is approximately three times that of B12 WT cells (0.75U/mg of protein). Decrease in MnSOD enzymatic activity in ABCD1 silenced cells (Fig. 1H) is associated with an increase in mitochondrial superoxide anions (O₂^·−) measured using the selective MitoSOX dye. As shown in Fig. 1F, silencing of ABCD1 in both cell lines results in increased mitochondrial O₂^·− production. Further, quantification of O₂^·− shows higher O₂^·− production in ABCD1 silenced cells as compared to WT cells in both cell lines (Fig. 1G). The increase of O₂^·− is more pronounced in U87 cells as compared to ABCD1silenced B12 cells (Fig. 1G). The increased levels of superoxide anions O₂^·− and loss of ΔΨ_m in ABCD1 silenced cells is related at least in part, to the loss of antioxidant defenses such as MnSOD.

ABCD1 silenced oligodendrocytes and astrocytes display perturbations in mitochondria biogenesis and expression of genes related to energy metabolism

To assess the impact of ABCD1 silencing on mitochondrial homeostasis, we evaluated the number of mitochondria, expression of mitochondrial genes, and levels of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) and 1-beta (PGC-1β), transcription factors for expression of mitochondrial genes in ABCD1 silenced B12 and U87 cells (Fig. 2). To assess the mitochondrial mass in silenced cells, the mitochondria were stained with Mitotracker red dye (Fig. 2A) and quantified the mitochondrial mass using ImageJ software (Fig. 2B). As shown (Fig. 2A) the Mitotracker red fluorescence in ABCD1 silenced cells was less as compared to WT cells. Quantification of mitochondrial mass by image analysis using ImageJ software confirms this observation (Fig. 2B). Reduction in number of mitochondria (Fig. 1S shown as supplementary data) is related to mitochondria biogenesis dysfunction. PGC-1, a master regulator of mitochondria biogenesis, therefore we investigated the status of PGC-1α and PGC-1β in ABCD1 silenced cells. Both PGC-1a and PGC-1b have low expression levels in ABCD1 silenced cells as compared to WT cells (Fig. 2D). Reduction of PGC1a expression is more pronounced in B12 cells as compared to U87 ABCD1 silenced cells. The decrease in mitochondrial biogenesis regulator expression (Fig. 2D) is correlated with decrease of mitochondrial number (Fig. 2A, B) as well as a decrease in the expression of mitochondrial genes such as mt-ND1 and mt-Cytb in U87 and B12 cells, respectively (Fig. 2E). Mt-ND1 and mt-Cytb, two mitochondrial genes, are investigated as markers for mitochondrial genome activity. Further, we also evaluated the expression of 84 key genes related to mitochondrial respiration, including genes encoding components of the electron transport chain and oxidative phosphorylation complexes (Fig. 2C). ABCD1 silencing in B12 oligodendrocyte, induces dysregulation in the expression of 17 genes whereas in astrocytes, 21 genes are affected (Fig. 2C). The dysregulated genes in B12 cells are associated with activities of complexes I, IV, and V (Fig. 2C). Some genes implicated in the activity of complex III are also dysregulated in U87 cells (Fig. 2C). Perturbation in oxidative phosphorylation (Oxphox) gene expression was further evaluated by RT-qPCR and profiler PCR array (Qiagen) (Fig. 2C). Figure 2C shows the number of genes whose expression was altered more than 2 fold (up or down regulation).

Figure 2. Effects of ABCD1 silencing on the mitochondrial biogenesis and oxidative metabolism gene expression.

Mitochondrial density was monitored by staining mitochondria in B12 and U87 cells with MitoTracker Red CMXRos (Invitrogen) (A). Cells were incubated for 30 min with 100 nM of MitoTracker Red at 37°C before being fixed (4% para-formaldehyde) and mounted for microscopy. Mitochondria mass (B) was quantified by using ImageJ software. Data are represented as mean ± SD (n = 3). ^* P<0.05, ^**P<0.005 versus WT or NT cells. Mitochondrial biogenesis in B12 and U87 cells was analyzed by western blotting (D). Levels of mt-ND1 and mt-Cytb mitochondrial gene transcripts for U87 and B12 respectively were standardized with mRNA level of the GAPDH and RPLP0, respectively (E). mt-ND1 and mt-Cytb, two mitochondrial genes, were investigated as markers for mitochondrial genome activity. Results are expressed as mean values ± SD (n=3). The expression of 84 key genes involved in mitochondrial respiration was measured using the Mitochondrial Energy Metabolism plus RT² Profiler PCR Array from Qiagen in WT, NT and Abcd1 silenced B12 and U87 cells. The number of genes showing a two-fold or greater change in expression (up or down) is shown (C).

ABCD1 silencing induces dysfunctions in mitochondrial oxidative phosphorylation systems and ATP production

The NADH⁺, H⁺ and FADH₂, produced by the intermediate metabolism, are oxidized by the mitochondrial respiratory chain to establish an electrochemical proton gradient for F1F0-ATP synthase (complex V) to produce ATP (Koopman et al. 2013). Since the expression of OXPHOS genes are perturbed in both cell lines (Fig. 2C), we investigated the effects of ABCD1 silencing on the enzymatic activities of complexes I and V. The data in Figure 3A shows reduced activity of complex I in ABCD1 silenced B12 and U87 cells, and a greater reduction was observed in B12 than U87 cells (Fig. 3A). These data are presented as percent of control of the specific activity. Basal activity of complex I in B12 cells (0.25 mU/mg of protein) is four times higher than that of U87 cells (0.06 mU/mg of protein). Complex V or the ATP synthase activity in both ABCD1 silenced B12 cells (0.04 mU/mg of protein) is 20% less than that of U87 cells (0.05 mU/mg of protein) (Fig. 3B).

Figure 3. Mitochondrial enzymatic activities of OXPHOS complexes and ATP production in B12 and U87 cells.

Enzymatic activity of complexes I (A) and V (B) was monitored in WT, NT, ABCD1 silenced B12 and U87 cells. Specific activity (nmol/min/mg of protein) was determined as described in Methods. Results were normalized with WT and expressed as percent of control as mean values ± SD (n=3). ^* P<0.05. ATP levels (C) were normalized with WT level (μmol/10E06 cells) and presented as percent of control as mean ± SD. ^* P<0.05. ABCD1 silenced cells compared to WT or NT cells (A, B, C).

To assess whether or not the observed membrane potential dysfunction (Fig. 1D, E), dysregulation of expression of the OXPHOS genes (Fig. 2C), and the perturbations of the oxidative phosphorylation enzymatic activities (Fig. 3A, B) lead to a defect in energy homeostasis, we quantified the amount of ATP produced by WT, NT and ABCD1 silenced cells. As shown in Figure 3C, loss of peroxisomal ABCD1 in B12 or U87 cells resulted in decreased ATP levels by 35 and 25%, respectively (Fig. 3C). The absolute quantity of ATP in U87 cells (440μmol/1 million cells) is nine times as high as that of the B12 cells (50μmol/1 million cells).

ABCD1 loss induces downregulation in the TCA cycle (citrate synthase enzymatic activity) and NAD/NADH ratio in B12 oligodendrocytes

Citrate synthase (CS), the first enzyme in the TCA cycle, linking the glycolytic and TCA pathway, is responsible for “coupling” of glycolytic and TCA cycle metabolism to meet the cellular bioenergic needs. Fig. 4A shows that loss of ABCD1 in oligodendrocytes induces down regulation of mitochondrial citrate synthase activity by 25% while in U87 astrocytes this activity remains intact (Fig. 4A). The absolute basal activity of CS in B12 WT is almost two times (95 nmol/min/mg of protein) that of U87 WT cells (53 nmol/min/mg of protein). The tricarboxylic acid cycle (TCA) generates NADH which is in turn used by the mitochondrial oxidative phosphorylation system for the biosynthesis of ATP. The NAD/NADH ratio, which plays an important role in regulating the intracellular redox state, is increased in ABCD1 silenced B12 cells while it decreases in U87 cells (Fig. 4B). Increase in NAD/NADH ratio in Abcd1 silenced B12 cells along with the decrease in ATP levels indicate that rupture in energy homeostasis is a key downstream effect of ABCD1 absence in B12 and U87 cells. A complete bioenergetic profile using a Seahorse XF96 Analyzer on both cell lines confirm the bioenergetic failure in ABCD1 silenced cells (Fig. 5). We measured the basal respiration, ATP synthesis, and the maximal respiration or the spare respiratory capacity (Fig. 5A). This is accomplished by sequentially adding specific modulators of mitochondrial respiration and measuring the oxygen consumption rate (OCR) as an indicator of respiration. After oligomycin injection OCR (Oxygen Consumption Rate) decreases as a result of ATP synthesis block at mitochondrial complex 5. FCCP a known uncoupling agent, treatment shows maximum possible respiratory rate in the absence of electron transport chain limitation. Finally, rotenone and Antimycin A inhibit mitochondrial respiration. Rotenone inhibits mitochondrial complex I, resulting in decreased flow of electrons in the electron transport chain leading to decreased oxygen consumption. Antimycin A inhibits the complex III activity by blocking the transfer of electrons between cytochrome b_H and coenzyme Q. The basal respiratory rates observed in B12 ABCD1 silenced cells were 30% less than WT cells. In U87, the basal respiration level was increased in ABCD1 silenced cells (Fig. 5C). ATP synthesis was reduced in B12 Abcd1 silenced cells (Fig. 5B) and in U87 ABCD1 silenced cells by approximately 50% (Fig. 5C). Maximal respiratory rate under condition of uncoupling respiration (FCCP) or respiratory capacity was reduced by 60% in B12 ABCD1 silenced cells (Fig. 5B) and 25% in U87 ABCD1 silenced cells (Fig. 5C).

Figure 4. Effects of ABCD1 silencing on citrate synthase (TCA cycle) in B12 and U87 cells.

CS specific activity (nmol/min/mg of protein) were determined in B12 and U87 cells (A) (as described in Methods section). Results were normalized to WT activity and expressed as percent of control as mean ± SD. (B) NADH, NAD⁺ levels and NAD⁺/NADH ratio. Data are presented as mean ± SD (n=3). ^*P < 0.05 versus WT or NT.

Figure 5. Complete bioenergetic profile by measuring the OCR of B12 and U87 cells using a Seahorse XF96.

Cell respiratory measurement in WT, NT and ABCD1-silenced cells were performed using Seahorse X96e. Measuring the Oxygen Consumption Rate (OCR) was used as an indicator of respiration (A). Cells were seeded at 10,000 cells/well for 24 h. Multiple parameters of oxidative metabolism (basal respiration, ATP synthesis and the maximal respiration or the spare respiratory capacity) were measured. This was accomplished by sequentially adding specific modulators of mitochondrial respiration (Cells were exposed sequentially to oligomycin (oligo), FCCP, and rotenone/Antimycin A) (A). Non-mitochondrial respiration after the final addition was subtracted from the other values. (A) Representative cell respiratory experiment of B12 WT, NT or ABCD1 silenced cells. (B) and (C) quantification of ATP production, basal respiration and respiratory activity in B12 and U87 cells, respectively. Data are presented as mean ± SD (n=2).

SAHA corrects mitochondrial dysfunction induced by ABCD1 loss in B12 oligodendrocytes and U87 astrocytes

SAHA, a potent class I and II HDAC inhibitor (Dokmanovic et al. 2007), was shown to correct the ABCD1 loss-induced VLCFA accumulation in B12 and U87 cells (Singh et al. 2013a). This efficacy of SAHA was mediated by upregulation in the expression of ABCD2 and ability of ABCD2 to complement the function of ABCD1 in X-ALD cells (Netik et al. 1999, Pujol et al. 2004). In order to evaluate the effects of SAHA in the mitochondrial dysfunction induced by ABCD1 silencing, ABCD1 silenced cells were treated with SAHA (5 μM, 72 h). We investigated mitochondrial homeostasis, such as effects of SAHA on mitochondrial mass (Fig 6A-B), mitochondrial health (Fig 6D), mitochondrial antioxidant MnSOD activity (Fig 6C) and ATP production (Fig 6E). Interestingly, SAHA induced mitochondrial biogenesis by increasing the mitochondria mass per cell in both cell lines (Fig 6A-B) leading to an increase in the mitochondria antioxidant activity (Fig 6C), corrected the mitochondria transmembrane potential (Fig 6D), and increased the levels of ATP (Fig 6E). These observations provide evidence that SAHA corrects the mitochondrial dysfunctions observed in ABCD1 silenced astrocytes and oligodendrocytes.

Fig 6. Effects of SAHA on mitochondrial dysfunction induced by ABCD1 loss in B12 and U87 cells.

(A and B) beneficial effects of SAHA on mitochondrial biogenesis. (C) MnSOD activity. (D) The transmembrane mitochondrial potential (ΔΨ_m). (E) ATP levels. Data shown are mean ± SD (n=3). *P < 0.05, **P<0.005 versus WT or NT.

Discussion

X-ALD is a complex disease with various phenotypes affecting different cells/organs even with the same mutation in ABCD1. The different degree of VLCFA accumulation in different cell types suggest that VLCFA-induced pathobiology should be studied in the cells that accumulate VLCFA and the cells that participate in disease pathology (e.g CNS cells, adrenal and testis cells) (Knazek et al. 1983, Ho et al. 1995, Whitcomb et al. 1988, Jang et al. 2011, Powers & Schaumburg 1974, Singh et al. 1981). Because of this, we investigated the VLCFA-induced pathologies in ABCD1-silenced astrocyte and oligodendrocyte cell lines. Our studies report the following: 1) ABCD1 silencing alters the cellular and mitochondrial redox (mitochondrial oxidative state; production of superoxide (Fig 1F, G) and its detoxification by MnSOD (Fig 1H)) in both astrocytes and oligodendrocytes; however, U87 astrocytes have higher MnSOD activity than B12 oligodendrocytes. 2) Silencing of ABCD1 in both astrocytes and oligodendrocytes results in dysfunction of mitochondria in terms of its proteome activity and ATP levels, and greater ATP reduction was observed in ABCD1-silenced oligodendrocytes than ABCD1-silenced astrocytes. 3) Silencing of ABCD1 in astrocytes and oligodendrocytes reduces the mitochondrial mass. 4) ABCD1 silencing resulted in decreased activity of citrate synthase, an enzyme linking the glycolytic and TCA pathways for energy production, in B12 oligodendrocytes but not in U87 astrocytes. 5) Consistent with our previous report, showing the normalization of VLCFA by treatment with SAHA, a class I and II HDAC inhibitor, and thus protection against VLCFA-induced pathological changes (Schluter et al. 2012), SAHA treatment of ABCD1-silenced astrocytes and oligodendrocytes corrects mitochondrial dysfunctions, as mitochondrial mass, and mitochondrial activity in terms of redox and ATP biosynthesis in both astrocytes and oligodendrocytes. Mitochondrial activity is known to be critical for cell survival and these studies provide proof of the direct relationship of peroxisomes and mitochondria for mutual survival sometimes referred to as “big brother and little sister” relationship (Schrader & Yoon 2007). Secondly, observed differences in mitochondrial energetics between B12 oligodendrocyte and U87 astrocytes may account for the loss of oligodendrocytes and survival of astrocytes in ABCD1-silenced cells in culture (Baarine et al. 2012a) and in X-ALD brain (Moser et al. 2001a, Singh & Pujol 2010).

VLCFA are considered as pathological “hallmark” of X-ALD. In addition to the abnormality of the peroxisomal β-oxidation by loss of ABCD1 (Moser et al. 2001a, Singh et al. 1984a), the observed increased VLCFA synthesis via FA elongation (Tsuji et al. 1981, Rizzo et al. 1984, Kemp et al. 2005) is also likely to contribute to overall VLCFA cellular load. ELOVL1 and 3 have chain length specificity toward VLCFA (Ohno et al. 2010, Ofman et al. 2010, Westerberg et al. 2004) and are therefore likely to play a role in the pathophysiology of X-ALD. Silencing of ABCD1 in oligodendrocytes resulted in increased expression of ELOVL1 but the same outcome was not observed in astrocytes (Singh et al. 2013a). Similarly, increased expression of ELOVL1 was also reported in oligodendrocytes derived from induced pluripotent stem cells (IPSCs) generated from fibroblasts from X-ALD patients (Jang et al. 2011). Induction of ELOVL1 following ABCD1 silencing indicates that VLCFA derangement induced mechanisms participate in the induction of ELOVL1. Therefore, the observed differential activities of VLCFA synthesis by FA chain elongation and VLCFA catabolism in peroxisomes may account for different amounts of VLCFA found in different cell types.

Studies reported in this manuscript describe that deficiency of ABCD1 in ABCD1-silenced cells alters the mitochondria redox as well as alters the mitochondrial proteome and its activity for synthesis of ATP. Previous animal studies with ABCD1 null mice (12 months) have reported reduced mitochondrial DNA along with downregulation of mitochondrial biogenesis pathway driven by PGC-1a/PPARg and reduced expression of mitochondrial proteins cytochrome c, NDUFB8 and VDAC (Morato et al. 2013), even though these mice accumulate relatively low amount of VLCFA than human cells. Mitochondrial dysfunction was also reported in cells incubated with exogenous VLCFA (Lopez-Erauskin et al. 2013). Reduced mitochondrial DNA and mitochondrial protein levels were also reported in the white matter of autopsy brain from patients with X-linked adrenoleukodystrophy (Morato et al. 2013, Fourcade et al. 2008, Singh & Pujol 2010, Lopez-Erauskin et al. 2012). In addition to X-ALD, other neurodegenerative diseases are also associated with mitochondrial pathologies (Lin & Beal 2006, Martinez et al. 2010, Pratico 2008, Stack et al. 2008, Zhou et al. 2008, Dai et al. 2014). Previous studies from our laboratory also reported that inflammatory disease mediators reduced the number and function of peroxisomes in neurological disorders such as cALD, multiple sclerosis, Krabbe disease and endotoxemia (Singh et al. 2009a, Singh et al. 2004, Khan et al. 2005, Paintlia et al. 2008, Haq et al. 2006). These observations indicate that the disease condition with altered peroxisomes can induce secondary mitochondrial changes. Secondly, while mitochondrial activity is required for cellular survival, the peroxisomal dysfunction can lead to chronic mitochondrial disease.

Recent in vitro studies described increased production of ROS, decreased glutathione (GSH) levels, lipoxidation and glycoxidation, and loss of mitochondrial transmembrane potential in X-ALD cell models (Fourcade et al. 2008, Hein et al. 2008, Baarine et al. 2012a, Baarine et al. 2012b). Fibroblasts from X-ALD patients showed mitochondrial changes only when treated with additional exogenous VLCFA. Studies using astrocytes, neurons (Zarrouk et al. 2012) and oligodendrocytes also investigated VLCFA toxicity by adding exogenous VLCFA (C22:0, C24:0, C26:0) and reported the loss of both astrocytes and oligodendrocytes with greater effect on oligodendrocytes in Ca²⁺ dysregulation (Hein et al. 2008) and mitochondria dysregulation (Baarine et al. 2012b). Young ALD mice, generated by deletion of ABCD1, shows only the metabolic phenotype of excessive VLCFA without clinical phenotype (Forss-Petter et al. 1997, Lu et al. 1997), however, a recent study reported oxidative stress associated bioenergetic/mitochondrial dysfunctions processes in old ALD mice (≥8 month) (Lopez-Erauskin et al. 2013). Studies of VLCFA-induced pathology so far point to possibly a multiple phase disease process in X-ALD pathology. First phase of metabolic disease characterized by accumulation of VLCFA; second phase, VLCFA induced oxidative stress and oxidative injury; third phase VLCFA induced inflammatory disease and in fourth phase inflammatory disease mediators induce reduction in PPARs and peroxisomes, Thus, propagating a vicious cycle of these phases (Singh et al. 2009a). First two phases (VLCFA abnormality and oxidative stress) are observed in AMN and in Abcd1-mice and thus ALD mice may serve as a model for AMN disease. On the other hand all the four phases, from VLCFA abnormality to reduction in peroxisomes in observed is inflammatory lesions of cALD brain.

Previous studies from our laboratory using U87 astrocyte and B12 oligodendrocyte cell lines stably silenced for ABCD1 established the role of ABCD1 deletion-induced VLCFA derangement in inflammatory mechanisms in astrocytes and cell death in oligodendrocytes without the need for addition of exogenous VLCFA (Singh et al. 2013a). While ABCD1-silenced astrocytes are able to maintain cellular homeostasis of antiapoptotic proteins (Bcl2, Bcl-Xl) and survival proteins (p-ERK1/2), the loss of ABCD1 in oligodendrocytes upregulated the expression of proapoptotic proteins (bad, bim, bax, bid) and activation of caspase 3/9 as well as inhibition of p-ERK1/2 suggests for the involvement of mitochondrial cell death dependent mechanisms in Abcd1 deletion induced loss of oligodendrocytes. On the other hand, ABCD1 silencing induced an inflammatory responsees as upregulation of 5-Lox, activation of NFkB and expression of cytokines and iNOS (Singh et al. 2011). As mitochondrial dysfunction leads to apoptotic cell death, we investigated the mitochondrial homeostasis in ABCD1-silenced astrocytes and oligodendrocytes. The data described in this manuscript document that silencing of ABCD1 in both oligodendrocyte and astrocyte cell lines results in decreased transmembrane mitochondrial potential as measured by JC-1 (Fig. 1), alterations in OXPHOS gene expression (Fig. 2) and reduction of enzymatic activities of mitochondrial electron transport chain (ETC) complexes (Fig. 3). Relatively, the mitochondrial activity (Δψm, complex I, complex V or ATP synthase) and levels of ATP and mitochondrial number are higher in ABCD1-silenced astrocytes than in oligodendrocytes. NADH level was decreased in B12 oligodendrocytes but not U87 astrocytes (Fig. 4B). The NAD+/NADH ratio which reflects an abnormal redox status was increased in B12 and decreased in U87 cells. Furthermore, ABCD1 silencing decreased the citrate synthase activity (an indicator for the TCA cycle activity) in ABCD1-silenced B12 cells and not U87 cells (Fig 4A). Together, these data indicate that energy homeostasis failure (production of ATP) in B12 cells may be a contributing factor for the vulnerability of oligodendrocytes. The defects in these mitochondrial enzymatic activities may be due to a functional damage (oxidation, carbonylation) of these proteins induced by VLCFA-induced oxidative stress (Fourcade et al. 2008, Galino et al. 2011). Our studies also report that silencing of ABCD1 in astrocytes and oligodendrocytes induces overproduction of mitochondrial superoxide (Fig. 1F, G) associated with a decrease in activity of mitochondrial MnSOD (Fig. 1H) leading to increased oxidative injury (Fig. 1F, G).

The evaluation of mitochondrial activity by measuring the oxygen consumption rate of B12 and U87 cells under treatment with different mitochondrial activity modulators confirmed the mitochondrial dysfunction in ABCD1-silenced cells (Fig. 5A). In both cell lines ATP production and respiratory capacity were reduced (Fig. 5A, B). These observations indicate that in astrocytes and oligodendrocytes, the cell types that participate in X-ALD pathology, VLCFA derangement induced by silencing of ABCD1 was sufficient to induce mitochondrial/oxidative injury as compared to fibroblasts derived from X-ALD patient that required treatment with additional high amounts of VLCFAs. This establishes a causal relationship between ABCD1 and mitochondrial dysfunction in X-ALD disease.

Therapeutic options for X-ALD include Lorenzo’s oil as competitive inhibition of VLCFA elongation, hematopoietic stem cell transplantation, antioxidant therapy and recently described lentivirals therapy, lentiviral gene therapy (Aubourg et al. 1990, Singh et al. 1998, Moser et al. 2005, Deon et al. 2008, Cartier et al. 2009, Mastroeni et al. 2009, Fourcade et al. 2010). Lorenzo’s oil inhibits the ELOVL1 activity for synthesis of VLCFA (Sassa & Kihara 2014). However, treatment seems to have little effect on the neurological disease as the disease progresses despite treatment (Moser et al. 2005). Based on the possible functional redundancy of peroxisomal ABCD transporters, several studies described the potential of pharmacological induction of ABCD2 or ABCD3 (Netik et al. 1999, Singh et al. 2013a, Flavigny et al. 1999) and suggested it as a therapeutic approach for X-ALD. Recently, we described that use of SAHA, an HDAC inhibitor, decreases the VLCFA load in X-ALD cells by increasing the VLCFA degradation via SAHA-induced induction of ABCD2 as well as inhibition the ELOVL1 activity (Singh et al. 2013a). Moreover, SAHA treatment also inhibits the VLCFA-induced proapoptotic signaling mechanisms in ABCD1-silenced oligodendrocytes and VLCFA-induced proinflammatory responses in ABCD1-silenced astrocytes (Singh et al. 2013a). Studies described in this manuscript document that SAHA treatment also blocks/attenuates the Abcd1 silencing mediated mitochondrial dysfunction and energetics failure (Fig 6) via PGC1a- and PGC1b-induced biogenesis of mitochondria and thus correction of mitochondrial/cellular homeostasis. SAHA is known to cross the blood brain barrier (Hockly et al. 2003) and its ability to induce ABCD2 in human biopsy brain tissue, safety in long term treatment (Singh et al. 2013a) and correction of the metabolic defect in brain of X-ALD mice (Singh et al. 2013a) provide support for use of SAHA as a potential therapy for X-ALD patients. SAHA is also reported to be effective in correcting the phenotype of the other metabolic disorders including Gaucher’s disease, Niemann-pick type C and spinal muscular atrophy (SMA) (Wiech et al. 2009). Since mitochondrial dysfunction is common to a number of neurological disorders including ALS, MS, and Alzheimer, correction of mitochondrial defects by SAHA itself or in combination with respective diseases drug treatments may provide additional benefits to patients with these diseases.

Abbreviations

CS

citrate synthase

X-ALD

X-linked Adrenoleukodystrophy

ABCD1/2/3

ATP binding cassette transporter D1/2/3

ALDP

Adrenoleukodystrophy protein

AMN

Adrenomyeloneuropathy

VLCFA

very long chain fatty acid

SAHA

suberoylanilide hydroxamic acid

HDAC

Histone deacetylase

cALD

childhood cerebral ALD

ROS

reactive oxygen species

RNS

reactive nitrogen species

LDH

lactate dehydrogenase

MnSOD

Manganese superoxide dismutase

OCR

oxygen consumption rate

ELOVL

Elongation of very long chain fatty acids

Oxphos

Oxidative phosphorylation

ONOO-

peroxynitrite