X-linked adrenoleukodystrophy (X-ALD) is a fatal neurodegenerative disorder, characterized by progressive cerebral demyelination cerebral childhood adrenoleukodystrophy (CCALD) or spinal cord neurodegeneration (adrenomyeloneuropathy, AMN), adrenal insufficiency and accumulation of very long-chain fatty acids (VLCFA) in tissues. The disease is caused by mutations in the ABCD1 gene, which encodes a peroxisomal transporter that plays a role in the import of VLCFA or VLCFA–CoA into peroxisomes. The Abcd1 knockout mice develop a spinal cord disease that mimics AMN in adult patients, with late onset at 20 months of age. The mechanisms underlying cerebral demyelination or axonal degeneration in spinal cord are unknown. Here, we present evidence by gas chromatography/mass spectrometry that malonaldehyde–lysine, a consequence of lipoxidative damage to proteins, accumulates in the spinal cord of Abcd1 knockout mice as early as 3.5 months of age. At 12 months, Abcd1− mice accumulate additional proteins modified by oxidative damage arising from metal-catalyzed oxidation and glycoxidation/lipoxidation. While we show that VLCFA excess activates enzymatic antioxidant defenses at the protein expression levels, both in neural tissue, in ex vivo organotypic spinal cord slices from Abcd1− mice, and in human ALD fibroblasts, we also demonstrate that the loss of Abcd1 gene function hampers oxidative stress homeostasis. We find that the α-tocopherol analog Trolox is able to reverse oxidative lesions in vitro, thus providing therapeutic hope. These results pave the way for the identification of therapeutic targets that could reverse the deregulated response to oxidative stress in X-ALD.
X-linked adrenoleukodystrophy (X-ALD: McKusick no. 300100) is a neurometabolic genetic disorder characterized by progressive demyelination within the CNS, adrenal insufficiency and a pathognomonic accumulation of saturated very long-chain fatty acids (VLCFA, C>22:0) in plasma and tissues. It is one of the most frequent peroxisomal disorders with a minimum incidence of 1 in 17 000 males. The two main neurological phenotypes are the severe childhood cerebral form (CCALD), which is rapidly progressive and associated with an inflammatory response in the brain white matter, and the slowly progressive adult adrenomyeloneuropathy (AMN), which presents with distal axonopathy in spinal cord and peripheral neuropathy (1). The disease is caused by mutations in the ALD gene (official nomenclature ABCD1) that lead to loss of function of the ALD protein (ALDP) (2). ALDP is an half ATP-binding cassette (ABC) transporter, an integral peroxisomal membrane protein whose putative role is the transport of VLCFA or VLCFA–CoA esters into the peroxisome for degradation by β-oxidation, by analogy to its yeast homologs Pxa1p and Pxa2p (3). The most divergent clinical phenotypes can occur within the same family; thus, there is no phenotype–genotype correlation (4), what suggests the existence of modifier genes. The mouse model, a classical knockout of the Abcd1 gene, does not reproduce the phenotypic variability observed in X-ALD patients, since it exhibits a late onset neurodegenerative phenotype with axonopathy in spinal cords and peripheral nerves, resembling a mild AMN phenotype (5,6). Also, there is no correlation between the ALD phenotype and the accumulation of VLCFA in plasma and fibroblasts from ALD patients, although a recent study suggests a link between the levels of VLCFA in white matter and disease severity (7). The underlying mechanisms that lead to cerebral demyelination, axonal degeneration in spinal cord and adrenal insufficiency are unknown, and the role played by the main accumulated VLCFA product, the hexacosanoic acid C26:0, remains to be elucidated. VLCFA incorporate in complex lipids in cell membranes of white matter and adrenal cortex, and are thought to destabilize and break myelin sheaths by occupying the lateral chains of proteolipid protein, gangliosides and phospholipids (1). ALD cells have been reported to have an increase in membrane microviscosity (8), produced by C26:0 and accounting for a defective response to trophic factors in adrenocortical cells (9). As an alternative or additional etiopathogenic mechanism, we propose that VLCFA could generate oxidative stress as other fatty acids do (10,11). Oxidative stress is a common finding to a growing number of neurodegenerative diseases, although its role as causative factor needs still to be ascertained for most of them (12). The high content of the highly peroxidizable polyunsaturated fatty acids in brain membranes, the elevated oxygen consumption and the relatively poor expression of enzymatic antioxidant defenses in nervous system support the high susceptibility of these organs to oxidative damage (13). Indeed, in X-ALD patients’ plasma and fibroblasts, indications for lipid peroxidation (by TBA-RS) and decrease of non-enzymatic antioxidant systems have been found (14,15). These results can be obscured since the measures are non-quantitative and are subjected to technical limitations mainly due to lack of specificity of the chemical reagents (16). In X-ALD brain of patients, immunohistochemical evidences for oxidative stress and damage have been found, mainly resulting from lipid peroxidation and mostly occurring in the inflammatory demyelinative lesions of ALD or AMN/ALD, as well as in adrenal cortex (17,18). In contrast, no direct evidence of oxidative stress or damage could be evidenced in adrenal cortex, kidney or brain in the disease mouse model, at 12 weeks of age, in spite of an increased immunohistochemical reaction against MnSod (Sod2) in adrenal cortex and kidney (18).
Thus, the reported analyses to date suggest that oxidative stress could be a hallmark of the disease, without providing information on a temporal sequence of events, thus not allowing discriminating between a causative versus a secondary role in disease physiopathogenesis. To address this important issue we set out to investigate: (i) whether oxidative stress could be evidenced in spinal cord of mice and if yes, (ii) whether this phenomenon could be found earlier than neuropathological degeneration and clinical impairment. For this, we used highly sensitive and selective state-of-the-art mass spectrometry-based techniques [gas chromatography/mass spectrometry (GC/MS)] to detect and measure the modification of protein structures by oxidative stress (19–21). Our results demonstrate oxidative damage to proteins in vivo in the mouse model of X-ALD, as early as 3.5 months of age, well before the neurological symptoms are established at 20 months of age (6) and before first signs of neuropathology in spinal cords appear at around 16 months of age (5,6). Identical types of oxidative lesions were detected in vitro in patient fibroblasts, thus leading to the proposal that oxidative modification of proteins is a major contributing factor to the pathogenesis of the disease. The oxidative lesions can be reversed using the lipid-phase antioxidant Trolox, which offers therapeutic hope. We also show that excess of VLCFA triggers oxidative stress in vitro and ex vivo in organotypic spinal cord slice culture system. Moreover, Abcd1 loss of function leads to a blunted oxidative stress response, which may play a major role in the neurodegenerative disease-leading process.
Direct oxidative, glicoxidative and lipoxidative damage to proteins in Abcd1− mouse spinal cords and human X-ALD fibroblasts
Emerging evidence indicates that oxidative stress causes specific protein modifications that may lead to a change in the structure and/or function of the oxidized protein (19). This can have a wide range of downstream consequences and may be the cause of subsequent cellular dysfunctions and tissue damages. We set out to investigate specific and direct oxidative damage to proteins in the mouse model of the disease and in human fibroblasts. For this we used specific and well-characterized markers such as glutamic semialdehyde (GSA), which derived from the metal-catalysed oxidation of proline and arginine, and aminoadipic semialdehyde (AASA), which resulted from lysine oxidation (20). Besides, the effects of free radicals in proteins can involve third-party molecules which amplify the damage, such as carbohydrates and lipids, in processes termed glycoxidation and lipoxidation (21). Both carbohydrates and polyunsaturated fatty acids, when reacting with free radicals, generate highly reactive dicarbonyl compounds, such as glyoxal, methylglyoxal, 4-hydroxynonenal and malondialdehyde, among others. These reactive carbonyl compounds can generate specific non-enzymatic adducts when reacting with proteins, such as (Nε-(carboxymethyl)-lysine) (CML), Nε-carboxyethyl-lysine (CEL) and Nε-malondialdehyde-lysine (MDAL) (21). Here, we quantified the concentration of selected markers of each pathway of protein oxidative damage (GSA and AASA for carbonylation, CML and CEL for glycoxidation/lipoxidation and MDAL for lipoxidation) by GC/MS. We evaluated spinal cord lysates as the main target of pathology in the Abcd1− mouse, and found indeed marked increases in MDAL, CEL, AASA and GSA levels at 12 months of age (Fig. 1A). This is well in advance of both neuropathological and neurological symptoms (5,6). The protein adduct derived from fatty acid peroxidation MDAL (22) is the first marker to appear altered, as early as 3.5 months of age (Fig. 1B), while no signs of oxidative lesions are found at 18 days (Fig. 1C). Thus, oxidative damage appears very early in mouse life, suggesting a role as disease-causative agent rather than a late consequence of other primary, non-identified, pathogenic events. Moreover, patient fibroblasts belonging to both CCALD (n = 4) and AMN (n = 4) phenotypes also show marked increases in MDAL, CEL, CML, AASA and GSA levels (Fig. 1D). These findings indicate that oxidative lesions to proteins are pathogenic phenomena common between mouse and X-ALD patients’ fibroblasts, irrespective of their phenotype.
Altered expression of enzymatic antioxidant defenses in the mouse
Oxidative stress occurs when the net flux of free radicals production exceeds the antioxidant defenses of the cell, in charge of maintaining the balance in the continuous cycle of ROS (reactive oxygen species) generation and inactivation. In nervous system, the main detoxifying enzymes that scavenge ROS include the superoxide dismutases (SOD), the glutathione peroxidases (GPX), peroxiredoxins (PRX) and catalase. To ascertain whether the observed oxidative damage could be related to a disturbed antioxidant response, we have quantified the protein levels of catalase, Gpx1, Gpx3, Gpx4, Prx 1 to 6, Sod1 (Cu/ZnSod) and Sod2 (MnSod), at different time points in disease progression. We find that Gpx1 is induced roughly 1.8-fold, reflecting a physiological response to increased free radicals, whereas both the cytosolic (Sod1) and mitochondrial (Sod2) SOD are 1.7-repressed in Abcd1− spinal cord at 3.5 months of age. Identical results were obtained at 12 months of age. This indicates a disturbed response of the first barrier against superoxide anion, concomitant with the observed oxidative lesions in this organ (Fig. 2A). No significant dysregulation of the above-cited enzymes was found in mouse brain cortex (Fig. 2B), or in spinal cords at earlier stages (18 days, data not shown).
Specific induction of catalase in Abcd1− motorneurons
To investigate whether the changes in the expression levels of the antioxidant enzymes observed in whole spinal cord lysates are the result of a general phenomena occurring in all cells or in contrast, involve specific cell populations, we performed immunohistochemistry against catalase, Gpx1, Gpx3, Gpx4, Prx 1 to 6, Sod1 and Sod2. Surprisingly, catalase induction was detected in Abcd1− motorneurons as early as 3.5 months; this induction was sustained until 17 months of age (Fig. 2C). This cell-type-specific upregulation occurs in the absence of peroxisomal proliferation as evidenced by the peroxisomal membrane protein Pex14 (Fig. 2D). At 3.5 months of age, no other signs of histological abnormalities, as ascertained with staining against lectine, GFAP, α-synaptophysin or Sudan Black, were encountered. For the rest of antioxidant enzymes, no difference between genotypes could be evidenced (data not shown). Because glial cells are the major contributors to protein mass in spinal cord, it is to assume that the changes detected in western blot experiments observed here, regarding Gpx1, Sod1 and Sod2 protein levels, respond mainly to change in glial cells or both neurons and glial cells. These changes appear beyond detection limits of the immunohistochemical technique.
Oxidative stress is caused by C26:0 in vitro and ex vivo
Hexacosanoic acid (C26:0) has been suggested to exert toxic effects on cells in culture, mainly increase in membrane microviscosity with the consequent functional impairment in adrenocortical cells (9). We tested the hypothesis of a direct implication of C26:0 as likely causative agent of the observed oxidative lesions and enzymatic dysregulations, using an ex-vivo culture model system, the organotypic slice culture of spinal cord (OSCSC) (23,24). Once recovered from the traumatic procedure of the section and after 2 weeks of culture, we analyzed the slices by western blotting. Indeed, expression of antioxidant proteins is dysregulated in the Abcd1− samples in the same manner than at 3.5 m of age in spinal cord: Gpx1 is 3.3-fold induced and both SOD (Sod1 and Sod2) are 1.5-downregulated in Abcd1− OSCSC (Fig. 3A). These results recapitulate the disease-associated features identified in vivo and validate OSCSC as cellular model to study physiopathogenesis or to assay experimental treatments. We next addressed the question of capability of C26:0 to generate oxidative stress. We incubated the slices with doses of 10–100 µm, taking into account that concentrations in patients can vary between 5 (in plasma (9)) and 60 µm (in brain demyelinating plaques (25)). After 72 h of incubation with C26:0 (10–100 µm), we found that the slices accumulated a maximum of 1.3-fold of VLCFA at the higher dose (not shown). This was sufficient to lead to an increase in Gpx1 and catalase protein levels in the control OSCSC at 100 µm, suggesting that the VLCFA might generate an enzymatic antioxidant response. The increase in catalase level occurs in the absence of peroxisomal proliferation as determined by western blot against the peroxisomal membrane protein Pex14. Interestingly, the Abcd1− slices cultures do not respond to C26:0 treatments, even at the higher dose of 100 µm what might indicate a blunted response to oxidative stress (Fig. 3A). Experiments were reproduced with 100 µm C18:1 and C20:0, without effect on protein expression (not shown). None of the doses of fatty acids used triggered visible toxicity signs, as examined by light microscopy and protein quantification.
In patient fibroblasts, we have also observed an antioxidant response when incubating the fibroblasts and their controls with 10–100 µm of C26:0. Here, the uptake seemed more efficacious and the intracellular doses reached moved between 10- and 20-fold over the basal levels prior treatment, without any signs of toxicity (not shown). Western blot analysis demonstrates that SOD2 was induced by this specific VLCFA in human fibroblasts after 2 days of treatment, in absence of induction of SOD1, catalase or GPX family of enzymes (Fig. 3B and data not shown). Further, we quantified protein non-enzymatic oxidative modifications in fibroblasts incubated with a high dose (100 µm) of C26:0 for 7 days. All markers, lipoxidative (MDAL), glicoxidative/lipoxidative (CEL, CML) and protein oxidative (GSA; AASA), roughly doubled their level in the fibroblasts samples derived from X-ALD patients. Interestingly, the control fibroblasts tolerated the high dose of C26:0 without accumulating oxidative lesions (Fig. 3C), suggesting that the mechanisms that should buffer the oxidative stress caused by C26:0 work less efficiently in X-ALD fibroblasts, or alternatively, the C26:0-dependent induction of oxidative stress is lower in healthy individuals, who can properly metabolize the VLCFA.
Molecular effects of C26:0 in fibroblasts: oxidative stress and decrease of mitochondria membrane potential (ΔΨ)
The different types of observed protein oxidative damage and the induction of antioxidant response suggest that in X-ALD general mechanisms triggering oxidative stress, such as free radical production, could take place (26). To date, neither a direct cytotoxic effect of C26:0 nor a role of C26:0 in free radical production has been reported. We addressed this specific question using the well-known carboxy-H2DCFDA probe (27). The acetate group of this probe is cleaved by esterases upon cell entry, leading to intracellular trapping of the nonfluorescent 2′,7′-dichlorofluorescein. Subsequent oxidation by ROS, particularly H2O2, superoxide anions (O2-) and hydroxyl radicals (OH·), yields the fluorescent product DCF. Under basal conditions, we found an increased DCF formation of around 1.5-fold in X-ALD patient’s fibroblasts when compared with control fibroblasts of matched passage number. The experiments were performed in two different groups of controls and patient fibroblasts: the first having had around 12–15 passages, and the second between 19 and 23 passages. The results were similar between both groups; representative samples are depicted in Fig. 4A. Doses of C26:0 from 10 µm upwards induced ROS increases in X-ALD fibroblasts, whereas doses of 50 µm are required to start producing ROS in control fibroblasts.
Free long-chain fatty acids have been reported to produce ROS by blocking the electron transport chain and modifying mitochondrial inner membrane potential (ΔΨ) by acting as protonophores (28). We investigated the ΔΨ status of X-ALD fibroblasts using the dye rhodamine 123, which takes benefit of the differential membrane potential to get selectively accumulated in mitochondria (29). Although basal levels were identical between both genotypes, 10 µm of C26:0 was able to significantly decrease ΔΨ by 15%, strongly suggesting that C26:0 is able to uncouple mitochondria (Fig. 4B).
The tripeptide glutathione (γ-l-glutamyl-l-cysteinylglycine or GSH) is the most abundant and important nonprotein thiol in mammalian cells, reaching concentrations up to 10 mm in the cytoplasm and about 1 mm in blood (30). GSH plays a central role in protecting cells of all organs, including the brain, against damage produced by free radicals, oxidants and electrophiles. As C26:0 is able to generate ROS, and to induce an enzymatic antioxidant response in spinal cords involving Gpx1, we measured the reduced glutathione (GSH) levels in fibroblasts using the cell-permeable substrate monochlorobimane, which is essentially non-fluorescent until conjugated to thiols. Although the levels of GSH appeared equivalent between X-ALD and control fibroblasts under basal conditions, addition of 10 µm C26:0 to the medium produced a marked decrease in X-ALD fibroblasts, and to a lesser extent in control samples (Fig. 4C). All experiments on this section were also performed with 10 µm of C18:1 and C20:0, with results not different from fibroblasts treated with vehicle only (not shown).
Increased sensitivity of X-ALD fibroblasts to oxidative stress
Given the increased amounts of oxidative lesions produced in X-ALD fibroblasts upon C26:0 incubation, we hypothesized that human X-ALD fibroblasts could be more sensitive to the damage induced by oxidative stress than control fibroblasts. As the glutathione peroxidase system is induced in vivo under basal conditions and responds to C26:0 ex vivo, together with the observed depletion of GSH upon C26:0 treatment in vitro, we have chosen an experimental approach that induces ROS levels by depleting cellular GSH (31,32). We incubated control and X-ALD fibroblasts with buthionine sulfoximine (BSO), an inhibitor of γ-glutamylcysteine synthetase, the enzyme that synthesizes GSH. Indeed, incubation with 500 µm BSO for 24 h dramatically compromised patients fibroblasts viability. Cell death was quantified by measuring the levels of lactate dehydrogenase (LDH) activity, an enzyme released in the medium by damaged membranes (33). LDH activity was much higher in X-ALD fibroblasts (ranging from 13 to 59%) than in fibroblasts from healthy individuals (from 0 to 11%) (Fig. 4D). This indicated that X-ALD cells were more sensitive to GSH depletion and, hence, more prone to undergo cell death due to oxidative stress-induced damage than their passage-matched control fibroblasts. Alternative, a higher basal level of ROS might cause the detrimental effect observed.
Trolox prevents oxidative lesions and SOD2 induction produced by C26:0
X-ALD cells have been reported to have an increase in membrane microviscosity (8), an alteration that is produced upon culture in the presence of VLCFA (9). α-Tocopherol is a well-known lipid antioxidant whose preferential site of action is within cell membranes, forming complexes with free saturated and unsaturated fatty acids and therefore neutralizing the detergent-like properties of the hydrolytic products that would otherwise disrupt membrane stability (34,35). Given its mechanism of action, we hypothesized that α-tocopherol could be of help in scavenging lipid peroxyl radicals produced by an excess of VLCFA in membrane phospholipids, and used therefore the well-known analog of α-tocopherol Trolox (36). Human fibroblasts were cultured in the presence of both high doses of C26:0 (100 µm) and 2 mm Trolox for 4 days. We observed that the above-described induction of SOD2 upon C26:0 treatments could be prevented by Trolox, strongly suggesting that superoxide anion radicals responsible for SOD2 upregulation were neutralized (Fig. 5A). Most importantly, Trolox was able to substantially ameliorate the levels of oxidative lesion markers exhibited by X-ALD fibroblasts, as quantified by the GC/MS technique. The normalization is complete in the case of glycoxidative and lipoxidative markers (Fig. 5B). These results open up important perspectives toward the use of antioxidants as adjuvant therapies for X-ALD.
Increased oxidative damage is a common pathogenetic component identified in a growing number of major neurodegenerative diseases, of particular relevance in Alzheimer’s disease, Parkinson’s disease and amyotrophic lateral sclerosis (ALS) (37–40). The discrimination between a primary or causative versus a secondary role of oxidative stress and damage in neuropathogenesis has, however, proven a difficult matter with few exceptions, such as the SOD1 mouse, a model of ALS (41), and the toxic models, for instance the induction of Parkinson through treatment with the mitochondrial toxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (42). Mouse models can be powerful tools providing answers to this capital issue, of direct relevance for the rational design of therapeutic approaches. We formerly reported that Abcd1 knockouts present the first signs of neuropathology at 15–16 months of age (6), as assessed by electron microscopy of peripheral nerves and spinal cords; followed by and an overt clinical phenotype at 20 months of age, as demonstrated by locomotor, behavioral and electromyographical approaches (5,6). Here, we present compelling evidence of oxidative damage to proteins in spinal cord of Abcd1− mice, as early as 3.5 months of age. This is based on MDAL values, which can be taken as first quantitative and specific disease marker, thus highlighting lipid peroxidation-associated processes. At 12 months of age, Abcd1− mice accumulate additional oxidative modifications of proteins, arising from metal-catalyzed oxidation (GSA, AASA), and from glycoxidation/lipoxidation (CML, CEL). This is still before onset of neuropathological signs and well in advance of clinical disease that appears at 20 months in the mouse. Altogether, this data indicate that oxidative stress and damage are early events in the pathogenic cascade that leads to spinal cord degeneration. Since the Abcd1− mice do not develop cerebral demyelination, it is not possible to demonstrate in the mouse model that oxidative damage could also trigger cerebral demyelination. However, since impairment of peroxisomal functions in oligodendrocytes can result in demyelination or axonal damage (43), this possibility remains open.
The pattern of protein oxidative damage observed in X-ALD support both diversity and specificity, thus aligning X-ALD with Alzheimer’s disease or with ALS, and differentiating it from Parkinson’s or dementia with Lewy’s bodies. In the latter, no oxidative lesion markers at all, or raised levels of MDAL only, have been found (40). As the targets of protein molecular damage have been shown to be disease-specific among the most common neurodegenerative diseases (40), finding ALD-specific protein targets for oxidative, glycoxidative and lipoxidative damage will be a major focus of future studies.
The X-ALD mouse presents microglia activation, together with astrogliosis, very late in life (5). Here, we show a very early motorneuronal involvement, based on the increase in catalase immunoreactivity, a sign of defense against ROS. It is tempting to speculate about a scenario comparable with ALS, in which different cell types contribute to different phases of disease progression—motorneurons in early phases of disease onset and microglia in later phases (44). More experimental results are required in order to identify the primary ROS-producing cell types and also, the most vulnerable cells to the consequent unbalanced redox status in our experimental system.
An interesting aspect revealed by our experiments is that Abcd1 loss of function is correlated with defective oxidative stress homeostasis: (i) X-ALD fibroblasts fail to compensate the noxious effects of VLCFA excess, as showed by the upcoming of different types of oxidative damage to proteins upon application of VLCFA; (ii) X-ALD fibroblasts show a higher sensitivity to glutathione depletion, which leads eventually to death; (iii) X-ALD fibroblasts produce higher levels of ROS than control fibroblasts upon VLCFA excess; (iv) Abcd1− spinal cords exhibit lower basal levels of Sod1 and Sod2 and (v) Abcd1− OSCSC exhibit a blunted antioxidant response upon VLCFA excess. This defective antioxidant response could ultimately constitute a major component leading to oxidative damage and pave the way for the cascade of events that might lead to neurodegeneration.
Further, we provide here direct evidence of a main role for VLCFA in the generation of ROS, leading to GSH consumption, activation of enzymatic antioxidant defenses at the protein expression levels, both in neural tissue ex vivo and in vitro in human fibroblasts, and ultimately causing specific oxidative lesions in proteins, which are identical to the ones found in spinal cords at 12 months of age, or in X-ALD fibroblasts under basal conditions. Because we ignore at this point whether additional substrates to VLCFA/VLCFA–CoA are imported into peroxisomes via Abcd1, we cannot formally conclude that accumulation of VLCFA is the cause for the elevated basal levels of ROS and oxidative lesions in X-ALD fibroblasts versus control samples. The precise molecular mechanisms by which an excess of C26:0 might produce ROS remain however elusive. Long-chain fatty acids have been shown to produce ROS through inhibition of the mitochondria respiratory chain and a change in membrane fluidity, at least in purified mitochondria (45). In our study, the increase in ROS production is concomitant with a mild decrease of ΔΨ of the inner mitochondrial membrane. This may be due to a protonophoric uncoupling action of fatty acids, as we did not have evidence of any change in the expression of mitochondrial uncoupling proteins in response to VLCFA (data not shown). Usually, such mild reduction in ΔΨ in mitochondria is not associated to increased ROS production but the opposite (46), thus suggesting that extra-mitochondrial sources of enhanced ROS production are more likely to be activated in response to VLCFA. Indeed, a few examples of fatty acids producing ROS via extramitochondrial sources include: (i) the saturated palmitic acid involving activation of NAD(P)H oxidase (47); (ii) arachidonic acid conversion via lipoxygenase (48,49); (iii) the degradation of VLCFA in peroxisomes, which could account for a part of observed ROS production and the catalase induction (50,51). However, our results in fibroblasts involving stimulation of mitochondrial Sod2 expression in the absence of Sod1 increase upon C26:0 treatments, does not allow us to exclude mitochondria involvement in the production of the observed ROS. Indeed, ultrastructural alterations have been found in X-ALD mitochondria from spinal cords and adrenocortical cells (52,53), warranting a careful assessment of mitochondria functionality. Also, the determination of fatty acid composition of mitochondria or mitochondria membranes in nervous system could shed light into the matter.
Taken together, our data suggests that impairment GSH homeostasis could be instrumental in X-ALD pathogenesis, as GSH acts as a co-factor in the reduction of ROS and lipid hydroperoxides (54,55). The observed depletion of GSH upon VLCFA application and the upregulation of Gpx1 in spinal cord as well as the higher sensitivity of X-ALD fibroblasts to GSH depletion suggest an underlying mechanism for the observed increased levels of protein lipoxidative damage-derived from build up of bifunctional electrophiles that may escape from Gpx activity. In its turn, lack of GSH would induce further sensitization to ROS, such as hydrogen peroxide, a notion sustained by the increased catalase expression. Also, GSH depletion could induce loss of glyoxalase activity, which might contribute to the reported increased CEL contents, and most importantly, it could lead to a diminished potential for cellular detoxification of ROS.
As deduced from the present study and from the recent reports in the literature (14,15), an early and tailored antioxidant intervention could be of benefit for X-ALD patients. Although most of the large clinical trials on α-tocopherol failed to show effects on decreasing mortality, or in reducing risk of cerebro- or cardiovascular accidents (56–59), in our study Trolox is able to induce reversion of oxidative lesions in ALD fibroblasts, most effectively the ones derived from lipid peroxidation and glycoxidation, together with the VLCFA-dependent SOD2 upregulation. Thus, the reversibility of damage has positive implications for therapy for α-tocopherol itself or for other lipid-phase last generation antioxidants, although this needs to be demonstrated in vivo in Abcd1− mice. Deciphering the exact pathways that result in the production of intracellular ROS in X-ALD may also help to define additional therapeutic targets.
MATERIAL AND METHODS
The following chemicals were used: 6-carboxy-2′,7′-dichlorodihydrofluorescein diacetate, diacetoxymethyl-ester (H2-DCFDA), monochlorobimane (Molecular Probes), Rhodamine 123, hexacosanoic acid (C26:0) and free fatty acid BSA (Sigma).
The following antibodies were used for western blots: anti-rabbit catalase, dilution: 1/200 [200–4151 (Rockland)], anti-rabbit Gpx1, dilution: 1/200 [LF-PA0019 (Labfrontier)], anti-rabbit Gpx3, dilution: 1/1000 [LF-PA0054 (Labfrontier)], anti-rabbit Gpx4, dilution: 1/200 [LF-PA0054 (Labfrontier)], anti-rabbit Prx1, dilution: 1/5000 [LF-PA0001 (Labfrontier)], anti-rabbit Prx2, dilution: 1/5000 [LF-PA0007 (Labfrontier)], anti-rabbit Prx3, dilution: 1/5000 [LF-PA0030 (Labfrontier)], anti-rabbit Prx4, dilution: 1/5000 [LF-PA0009 (Labfrontier)], anti-rabbit Prx5, dilution: 1/5000 [LF-PA0010 (Labfrontier)], anti-rabbit Prx6, dilution: 1/5000 [LF-PA0011 (Labfrontier)], anti-rabbit Cu/ZnSod (Sod1), dilution: 1/1000 [SOD-101 (Stressgen)], anti-mouse Sod1, dilution: 1/1000 [556360 (BD Pharmigen)], anti-mouse MnSod (Sod2), dilution: 1/1000 [611581 (BD Pharmigen)], anti-rabbit Pex14, dilution: 1/1000 (a kind gift of Dr Marc Fransen) and anti-mouse γ-tubulin, dilution: 1/5000 [T6557, clone GTU-88 (Sigma)]. Goat anti-rabbit IgG linked to horseradish peroxidase, dilution: 1/10000 [81–6520 (Invitrogen)] and Goat anti-mouse IgG linked to horseradish peroxidase, dilution: 1/10000 [81–6120 (Invitrogen)] have been used as secondary antibodies.
The generation and genotyping of Abcd1− mice have been previously described (6,60). Mice used for experiments were on a pure C57BL/6J background. Animals were sacrificed and tissues were recovered and conserved at −80°C. All methods employed in this work are in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publications No. 85-23, revised 1996), and with the ethical committee of IDIBELL and the Generalitat de Catalunya.
Cell culture and treatments
Control and X-ALD human fibroblasts were grown in DMEM containing 10% of fetal bovine serum, 100 U/ml penicillin, 100 mg streptomycin, at 37°C in humidified 95% air/5% CO2. After the growing period, the medium was changed to serum-free medium supplemented with 10 or 100 µm BSA (free fatty acid)-bound fatty acid for 2 or 7 days at a C26:0 fatty acid/BSA ratio of 2:1. Untreated cells received an amount of BSA equal to that brought with the BSA–fatty acid complex. Unless otherwise stated, experiments have been carried out with cells at 95% of confluence which had a number of passages range 12–15. Same results have been obtained when we used fibroblasts with a number of passages range 19–23.
Cell death measurements
Human fibroblasts were cultured in 24-well plates up to 90% of confluence. Cells were then treated with BSO (500 µm) for 24 h and cell death was quantified using CytoTox-ONE™ Homogeneous Membrane Integrity Assay (Promega) according to manufacturer’s instructions.
Organotypic spinal cord slice culture (OSCSC)
The spinal cords from 18-day-old mice were immediately removed and placed in ice-cold dissecting media (pH 7.15). Next, the spinal cord was cut into 350-μm thick slices using a McIlwain tissue chopper to generate the organotypic slice cultures and placed into a sterile petri dish with Grey’s balanced salt solution. Spinal slices were transferred onto Millicell-CM cultured plate inserts (Millipore, MA, USA). The inserts were placed into wells of a six-well plate containing 1.0 ml of medium containing 50% MEM with Earl’s salts and glutamine, 25% Hanks balanced salt solution and 25% horse serum supplemented with 20 mm of HEPES acid–salt and d-glucose (6 mg/ml) (Gibco-BRL). Slices were incubated at 37°C for different period of times and media were changed twice a week (23,24). After 2 weeks of culture, OSCSC have been treated with free fatty acids dissolved in ethanol.
Evaluation of intracellular ROS
Intracellular ROS levels were estimated using the ROS-sensitive H2DCFDA probe as described (27,61). Following incubation with 10 µm H2DCFDA for 30 min, cells were washed twice with PBS and scraped into water. The fluorescence of H2DCFDA-stained cells was measured with a spectrofluorimeter (excitation wavelength 493 nm, emission wavelength 527 nm). Major changes in ROS levels observed in experimental settings were confirmed by direct examination of cells using fluorescence microscopy.
Determination of GSH
The intracellular content of GSH was determined using monochlorobimane, a thiol-reactive probe (62). Following incubation for 30 min with 100 µm monochlorobimane, cells were washed twice with PBS and scraped into water. The fluorescence of monochlorobimane stained cells was measured with a spectrofluorimeter (excitation wavelength 380 nm, emission wavelength 460 nm).
Determination of mitochondrial membrane potential by uptake of Rhodamine 123
The uptake of Rhodamine 123 by cells was determined under the conditions described by Johnson et al. (63). Briefly, cells were incubated with 5 µm Rhodamine 123 for 30 min in PBS. After the incubation, the cells were washed twice with PBS and scraped into water. The fluorescence of Rhodamine 123 was measured with a spectrofluorimeter (excitation wavelength 520 nm, emission wavelength 543 nm).
Measurement of GSA, AASA, CML, CEL and MDAL
GSA, AASA, CML, CEL and MDAL concentrations in total proteins from spinal cord homogenates were measured by GC/MS (39). Samples containing 500 µg of protein were delipidated using chloroform/methanol (2:1, v/v), and proteins were precipitated by adding 10% trichloroacetic acid (final concentration) and subsequent centrifugation. Protein samples were reduced overnight with 500 mm NaBH4 (final concentration) in 0.2 m borate buffer, pH 9.2, containing one drop of hexanol as an anti-foam reagent. Proteins were then re-precipitated by adding 1 ml of 20% trichloroacetic acid and subsequent centrifugation. The following isotopically labeled internal standards were then added: [2H8]lysine (d8-Lys; CDN isotopes); [2H4]CML (d4-CML), [2H4]CEL (d4-CEL) and [2H8]MDAL (d8-MDAL), prepared as described (22,64); and [2H5]5-hydroxy-2-aminovaleric acid (for GSA quantization) and [2H4]6-hydroxy-2-aminocaproic acid (for AASA quantization) prepared as described (20). The samples were hydrolyzed at 155°C for 30 min in 1 ml of 6 N HCl, and then dried in vacuo. The N,O-trifluoroacetyl methyl ester derivatives of the protein hydrolysate were prepared as described previously (20). GC/MS analyses were carried out on a Hewlett-Packard model 6890 gas chromatograph equipped with a 30-m HP-5MS capillary column (30 m × 0.25 mm × 0.25 µm) coupled to a Hewlett-Packard model 5973A mass selective detector (Agilent, Barcelona, Spain). The injection port was maintained at 275°C; the temperature program was 5 min at 110°C, then 2°C/min to 150°C, then 5°C/min to 240°C, then 25°C/min to 300°C, and finally hold at 300°C for 5 min. Quantification was performed by external standardization using standard curves constructed from mixtures of deuterated and non-deuterated standards. Analytes were detected by selected ion-monitoring GC/MS. The ions used were: lysine and d8-lysine, m/z 180 and 187, respectively; 5-hydroxy-2-aminovaleric acid and d5–5-hydroxy-2-aminovaleric acid (stable derivatives of GSA), m/z 280 and 285, respectively; 6-hydroxy-2-aminocaproic acid and d4–6-hydroxy-2-aminocaproic acid (stable derivatives of AASA), m/z 294 and 298, respectively; CML and d4-CML, m/z 392 and 396, respectively; CEL and d4-CEL, m/z 379 and 383, respectively; and MDAL and d8-MDAL, m/z 474 and 482, respectively. The amounts of products were expressed as the ratio of micromole of GSA, AASA, CML, CEL or MDAL per mole of lysine.
Tissues were removed from euthanized mice and flash-frozen on liquid nitrogen. Frozen tissues were homogenized in RIPA buffer boiled for 5 min and centrifuged. We measured protein concentration of the supernatant with a bicinchoninic acid Protein Assay Kit (Pierce). Ten to hundred micrograms were loaded onto each lane of 10% polyacrylamide gels for 60 min at 120 mV. Resolved proteins were transferred to nitrocellulose. Proteins were detected with enhanced chemiluminescence (ECL) western blotting analysis system (Amersham Biosciences). Photos and quantifications have been generated with the Bio-Rad Molecular Imager™ VersaDoc™ MP 4000 System which is a versatile and quantitative CCD-based imaging system.
Spinal cords were harvested from ALD male mice and their control littermates at 3.5 and 17 months of age, after perfusion with PFA4% basically as described (5,52). Spinal cords were embedded in paraffin and serial sections, 5 µm thick, were cut in a transversal plane. The sections were stained with the antibodies used for western blots, at 1:500 dilutions for catalase and Pex14. After overnight incubations at 4°C with the primary antibody, peroxidase-based detection of the primary antibody was accomplished with the modified labeled streptavidin (LSAB) technique (DAKO LSAB2 System Peroxidase). After counterstaining with hematoxylin for 20 s, sections were dehydrated and mounted with PDX-mounting solution (Fluka Biochemika, Switzerland). Specificity of the peroxidase reaction was controlled by treatment of sections not exposed to the primary antibodies. At least three sections were analyzed per animal and per stain.
This study was supported by funds from the European Commission (contract no. LSHM-CT2004-502987), the Association Française contre les Myopathies (AFM) (Project no. 9315), the European Leukodystrophy Association, the Asociación Española contra la Leucodistrofia (ALE-ELA España), the Fondo de Investigación Sanitaria (FIS, Ministerio de Sanidad Español, Instituto de Salud Carlos III, grant no. 01/1667, PI051118, P052241 and RD06/0013/0012), the Spanish Ministry of Education and Science (grants no. BFU2006-14495/BFI and AGL2006-12433), the regional government of Catalonia (grant no. 2005SGR00101), ‘La Caixa’ Foundation to M.P.O, and Spanish Research Network REDEMETH (no. G03/054). S.F. was a fellow of the ELA (grant no. ELA 2007-018F4) and the European Commission. J.G. was a fellow of the IDIBELL program of PhD-student grants, and J.L.-E. was a fellow of the Department of Education, Universities and Research of the Basque Country Government, no. BFI07.126.
We are indebted to Professor Isabel Fabregat for helpful scientific discussions and to M. Carmona, R. Blanco, J.J. Martínez and T. Martín for priceless technical assistance. The antibody for Pex14 was a kind gift of Professor Marc Fransen.
Conflict of Interest statement: none declared.
- oxidative stress
- vitamin e
- fatty acids
- adrenal gland hypofunction
- coenzyme a
- gas chromatography-mass spectrometry
- membrane transport proteins
- mice, knockout
- nerve degeneration
- nerve tissue
- neurodegenerative disorders
- retrograde degeneration
- spinal cord diseases
- spinal cord
- symptom onset