Copper In Lateonset Neurodegenerative Diseases

Neurodegenerative diseases that are characterized by late-onset in life offer a very interesting matter of analysis as far as risks related to copper are concerned, insofar as aberrant copper chemistry rather than excessive or deficient metal availability seems to be implicated as a causative factor of neurodegeneration, again strongly associated to overproduction of oxygen free radicals.

3.1. Amyotrophic Lateral Sclerosis 3.1.1. Molecular Basis of the Disease

Amyotrophic lateral sclerosis (ALS) is a progressive lethal disease, leading to death in a few years after the early symptoms. The target of neurodegeneration is quite specific, because only cortical and spinal motoneurones are affected, with consequent progressive impairment of all muscular activities, eventually leading to respiratory compromise and death. ALS occurs both as a sporadic (SALS) and a familial (FALS) disorder, with inherited cases accounting for about 10% of patients (44). The only gene conclusively linked to FALS so far is the one coding for the dimeric enzyme Cu,Zn superoxide dismutase (SOD1) (21q22) showing autosomal dominant mutations (45,46). Linkage studies have revealed that mutations in SOD1 are responsible for 10-15% of FALS (47). In spite of this intricate

Table 4

Redox Reactions of SOD1

A: With Superoxide (O2-) SOD1 • Cu(II) + O2- ^ SOD1 • Cu(I) + O2 SOD1 • Cu(I) + O2- + 2H+ ^ SOD1 • Cu(II) + H2O2

B: With Hydrogen Peroxide (H2O2)

SOD1 • Cu(II) + H2O2 ^ SOD1 • Cu(I) + O2- + 2H+ SOD1 • Cu(I) + H2O2 ^ SOD1 • Cu(II) + OH + OH-

C: With Peroxynitrite (ONOO)

SOD1 • CuO - NO2+ + H-tyr-protein ^ SOD1 • Cu(II) + OH" + NO2-tyr-protein etiopathological pattern, the phenomenology of the disease is identical in all sporadic and familial cases. This has stimulated a great deal of interest in studying the relationships between the enzyme involved in the few genetically determined cases and the neurodegeneration typical of the disease.

SOD1 is a very efficient enzyme, considered to be a primary defense barrier in oxidative stress because it catalyzes the disproportionation of two molecules of O2- into dioxygen and H2O2 at a rate very near the diffusion-controlled limit. The site of enzyme activity resides on the copper, which is involved in a redox cycle with the superoxide substrate (48). To date, there are more than 60 SOD1 point mutations reported in FALS families, distributed in all the 5 exons of the gene (49). Interestingly enough, mutants associated to the most severe cases of FALS are fully active. Conversely, mutants with very low enzymatic activity are associated to mild forms of the disease. For instance, mutant H46R, which is almost completely inactive since affecting the active site (50), is reported typical of a "mild" Japanese form of FALS (average duration 17 yr) (51). Conversely, G93A mutants are fully active but are associated with a severe form of the disease (52). The most frequent mutation, A4V, is found in almost 50% of patients in North America and is associated with a severe form of FALS, although SOD1 activity is only slightly reduced (49). This mutation affects a residue involved in the formation of contacts at the dimer interface, as several other FALS-typical mutations. Furthermore, different mutations affecting the active site result in different effects in terms of the severity of the disease. For instance, mutation H48Q confers to the protein scarce dismutating activity, but was described for patients severely affected, with a rapid course (8 mo), at difference with H46R (49).

These results primed the hypothesis that neurodegeneration in FALS is not related to the loss of function by SOD1, but to a gain of additional noxious properties by the enzyme.

3.1.2. Additional Functions of SOD1 in FALS

In addition to redox cycling with O2- substrate (Table 4), copper in the active site of SOD1 can also react with the peroxide product (Table 4). Whereas cycle A will remove O2- from any potentially harmful interaction with metals or other radicals to generate more potent oxidants like OH- (see Table 1), cycle B becomes a generator of OH- and envisages a paradoxical pro-oxidant activity of SOD1. However, cycle B can be effective only under a certain set of circumstances, because it is much less kinetically favored than cycle A, unless in the presence of large excesses of H2O2 or in the presence of enzyme variants with abnormal affinity for H2O2. The latter case may be the result of either distorted copper coordination or altered geometry of the active-site channel. In fact, markers of oxidative stress such as peroxidated lipids, oxidized proteins, and DNA are elevated in several experimental systems, ranging from both sporadic and FALS patients to transgenic mice (53-55).

Some of the FALS-typical enzyme variants have been purified and some of them indeed showed an increased propensity for the reaction with H2O2 (56-58). At least two of the ALS-linked mutant SOD1 (A4V and G93A) can increase the rate of formation of OH- in vitro, as demonstrated through the electron spin resonance (ESR) technique, by the use of spin traps (56-58). Those spin-trapping experiments were criticized by some authors (59) and the relevance of such reactions in vivo has also been questioned. For instance, conflicting evidence for increased OH- formation was obtained in transgenic FALS mice (54,55,60). Furthermore, Marklund et al. (61) did not find increased peroxidase activity with another FALS mutant, D90A.

Imperfectly folded mutant SOD1 may catalyze the nitration of tyrosines through use of peroxynitrite as a substrate (Table 4). The group of Beckman has found that, in agreement with previous reports in vitro (62), mutant FALS-SODls have decreased affinity for zinc in vivo and this might enhance catalysis of tyrosine nitration by peroxynitrite (63,64). Indeed, the presence of elevated free-nitrotyrosine levels has been reported in transgenic mice expressing SOD1 mutants (60).

These findings could explain how SOD1 enzymes act as antioxidants and pro-oxidants at the same time. In fact, antioxidant activities (i.e., superoxide dismutation) and pro-oxidant activities (i.e., peroxidase and nitration reactions) utilize similar redox mechanisms. Wild-type SOD1 seems to achieve its specificity by restricting access of potential substrates other than O2-: structural perturbations in FALS-SOD1 mutants may allow access of alternative substrates to the copper center, leading to more diverse catalytic redox reactions. Both hypotheses, of FALS-SOD1 peroxidative activity and tyrosine nitration, point to a different, increased availability of active-site copper to the substrate as a prerequisite for the gain of function of SOD1s typical of FALS patients. Enhanced peroxidase activity of FALS-SOD1 is copper dependent, in that it can be blocked by stoichiometric copper-chelating agents as diethyldithiocarbamate and D-penicillamine in vitro, whereas the peroxidase activity of the wild-type enzyme is unaffected unless copper is removed from the active site by an excess of chelators (56,65). Furthermore, the metal-binding properties of the FALS mutant proteins and their redox behavior are altered. It has been proposed that alteration of the properties of the zinc site will alter the metal-binding affinity of the copper site as well, possibly inducing a loosening of the protein structure (62). This hypothesis has been recently supported by the finding that zinc-deficient SOD1 (both wild-type and FALS mutants) acquire toxic properties (64).

Direct measurements of purified recombinant FALS-SOD1 have demonstrated a decrease in copper binding for yeast G85R mutant (66) and human H46R (50) but not for other mutants, such as yeast G93A (66). In H46R, a critical residue for maintenance of the geometry of the active site is substituted. This mutant is structurally stable but lacks significant enzyme activity and has impaired capability of binding catalytic copper (50). Recombinant mutated protein was expressed in the Escherichia coli strain QC779, which is defective in bacterial SODs. We showed that this protein contains about 5% copper; the copper ESR signal for H46R increases in intensity upon titration with copper chloride, but the spectrum displays an axial geometry, at variance with the tetraedrally distorted square planar geometry of the wild-type enzyme, thus suggesting that arginine in the 46 position does not correctly coordinate the copper. Furthermore, reconstituted protein displays only poor dismutating activity. In a recent article, it has been confirmed that this particular FALS mutant does not bind copper at the copper-specific site; rather, copper ions either compete for zinc at the active site or bind to a surface solvent-exposed residue near the dimer interface, individuated as Cys111 (67). Based on several evidences obtained in their yeast model system, Corson et al. (68), by analyzing several FALS-SOD1 mutants overexpressed in Sacchoromyces cerevisiae, have suggested that aberrant copper-mediated chemistry catalysed by a less tightly folded mutant enzyme might be responsible for SOD1-linked FALS phenotype.

3.1.3. Mutant SOD1 Involvement in FALS-Associated Neurodegeneration

The molecular mechanisms hypothesized to underlie SOD1 toxicity in FALS are under study in our laboratory in models of neural cells. We devised a model system made of several human neuro-blastoma SH-SY5Y cell lines where the ratio of expression of mutant SOD1 to endogenous SOD1 approximated 1:1, a situation resembling that of heterozigous patients (69). Mutant SOD1s were chosen as to have the opposite situation in terms of residual dismutating activity and relevance to the

Fig. 2. D-Penicillamine protects SH-SY5Y neuroblastoma cells transfected with H46R FALS mutant from paraquat-induced toxicity. (A) SOD1 activity measured spectrophotometrically; (B) cell survival evaluated by trypan blue dye exclusion and expressed as percentage of corresponding untreated cell line. WT: cells transfected with wild-type SOD1. For details, see ref. 70.

Fig. 2. D-Penicillamine protects SH-SY5Y neuroblastoma cells transfected with H46R FALS mutant from paraquat-induced toxicity. (A) SOD1 activity measured spectrophotometrically; (B) cell survival evaluated by trypan blue dye exclusion and expressed as percentage of corresponding untreated cell line. WT: cells transfected with wild-type SOD1. For details, see ref. 70.

severity of the disease (i.e., the H46R or the G93A variants). In those studies, mutations have been introduced in the cDNA-encoding human SOD1 and several monoclonal cell lines have been established constitutively expressing either the wild-type or the mutant enzymes. Cells overexpressing the G93A mutant, in addition to a significant increase in SOD1 protein and activity, also showed a loss of mitochondrial membrane potential, an increased sensitivity to valinomycin, and a parallel increase in cytosolic calcium concentration. Therefore, mitochondrial damage and calcium levels may represent early factors in the pathogenesis of FALS.

When SH-SY5Y cells have been transfected with plasmids directing constitutive expression of H46R mutant, they showed increased levels of immunoreactive SOD1. However, they have much lower dismutase activity than wild-type transfected or control neuroblastoma cells. This led us to the conclusion that expression of this mutant enzyme affects the activity of the endogenous enzyme, through formation of unstable or partially inactive heterodimers (70). In fact, misfolded SOD1 have been reported in spinal cords from transgenic mice expressing several FALS mutants in the presence or in the absence of overexpression of wild-type human SOD1 (71). Overall data indicate that FALS-SOD1 is made less stable by mutations and is capable of forming precipitates independently of the coexpression of wild-type protein. However, this seems not to be a general rule for FALS-SOD1s, as many other mutants do not aggregate (72) and the possibility that SOD1 aggregation is the consequence of a dysfunction in some other cell component is still open.

In our transfected cell system, we have also observed that expression of mutant H46R induces a selective increase in sensitivity to paraquat-induced cell death. This cannot arise only from reduced

Fig. 3. SOD1 protein and activity in SH-SY5Y neuroblastoma cells transfected with WT or G93A FALS mutant enzyme. (A) Protein content was assessed by Western blotting using a polyclonal antibody. (B) SOD1 activity was measured by a polarographic method. For details, see ref. 80.

SOD1 activity, which still remains about 75% of that of control cells (Fig. 2A), but may be likely ascribed to aberrant metal chemistry of this mutant enzyme. In fact, cells could be protected by treatment with the copper chelator D-penicillamine (Fig. 2B). Our data are in line with a previous report (73) in a different system, where cell death of adenovirus-infected PC12 cells expressing FALS-SOD1s was partially inhibited by administration of a copper chelator. Survival of transgenic mice expressing a different FALS mutation (G93A) was extended by administration of D-penicillamine (74).

The actual mechanism of motor neuron loss (whether necrosis or apoptosis) in ALS is still actively debated. Although activation of proapoptotic Bax protein and inactivation of antiapoptotic Bcl-2 has been reported in sporadic ALS patients (75), some authors have not been able to detect apoptotic cells in FALS-SOD1 transgenic mice (76). However, mutations associated with FALS convert SOD1 from an antiapoptotic gene to a proapoptotic gene in yeast and in conditionally immortalized mammalian neural cells (77). Furthermore, caspase-1 is activated in neural cells and tissue with SOD1-FALS (78) and inhibition of caspase-1 and caspase-3 delays progression of disease and extends survival in FALS-SOD1 transgenic mice (79). Furthermore, NO-dependent apoptosis seems to be increased by FALS-SOD1 in motor neurons (64).

In order to extend the mechanistic understanding of SOD1 action in FALS, we challenged our model of neuroblastoma cells transfected with the G93A mutant with NO (80). Cells were treated with several NO donors, including nitroso-glutathione (GSNO), which represents the physiological intracellular dispenser of NO (81). The G93A transfected cells showed an increased amount of both SOD1 protein and activity, comparable to that of wild-type SOD1 transfected cells (Fig. 3). However, they

Fig. 4. Pro-oxidant status of SH-SY5Y cells transfected with FALS mutant G93A. (A) Intracellular ROS were detected by incubation with 2'-7'-dichlorofluorescein diacetate (DCF), followed by FACScan analysis. Tert-butyl-hydroperoxide (t-butOOH) was used as a positive control. (B) Bcl-2 immunoreactive protein was measured by Western blotting using a monoclonal antibody. For details, see ref. 80.

Fig. 4. Pro-oxidant status of SH-SY5Y cells transfected with FALS mutant G93A. (A) Intracellular ROS were detected by incubation with 2'-7'-dichlorofluorescein diacetate (DCF), followed by FACScan analysis. Tert-butyl-hydroperoxide (t-butOOH) was used as a positive control. (B) Bcl-2 immunoreactive protein was measured by Western blotting using a monoclonal antibody. For details, see ref. 80.

had an increased flux of ROS (Fig. 4A), as if a pro-oxidant status lurked within them. In addition, they also showed a lower basal level of the antiapoptotic, antioxidant protein Bcl-2 (Fig. 4B). When exposed to NO donors, they showed an increased susceptibility to undergo apoptosis, with respect to control SH-SY5Y cells, whereas wild-type SOD1 transfected cells were spared. NO-mediated apoptosis in these models was associated with a canonical sequence of events, including Bcl-2 downregulation, increased expression, and phosphorylation of the tumor suppressor protein p53 and of the product of one its target genes, the cell cycle inhibitor p21 (80). Furthermore, cytochrome-c was released from the mitochondria (Fig. 5A), thus leading to activation of the cysteine protease caspase-3 (Fig. 5B). Interestingly, pretreatment of cells with the copper chelator D-penicillamine protected cells transfected with the G93A SOD1 mutant from NO-induced apoptosis (Fig. 6), thus conferring them with behavior typical of the wild-type SOD1 transfected cells. In other words, it seems that G93A cells behave as the wild-type ones when the copper is properly constrained. These results confirm the role of aberrant copper chemistry both in oxygen radical formation and in NO susceptibility, as we already showed for the H46R mutant (70). In conclusion, high levels of SOD1

Fig. 5. Molecular markers of apoptosis in SH-SY5Y cells upon GSNO treatment. (A) Immunoreactive cytochromes in cell cytosol was detected by Western blotting by a monoclonal antibody; (B) caspase-3 activity was measured fluorometrically. For details, see ref. 80.

Untreated +GSNO +GSNO

Fig. 6. Increased susceptibility of G93A cells to GSNO is abolished by D-penicillamine. Apoptotic cells were detected by FACScan analysis, upon propidium iodide staining. For details, see ref. 80.

Untreated +GSNO +GSNO

Fig. 6. Increased susceptibility of G93A cells to GSNO is abolished by D-penicillamine. Apoptotic cells were detected by FACScan analysis, upon propidium iodide staining. For details, see ref. 80.

activity protect cells from apoptosis, unless other nasty side activities are present; in that case, they may become a damage mediator and be proapoptotic.

In conclusion, neurodegeneration in FALS appears to be a consequence of aberrant copper chemistry deriving from improper handling of the metal by one of its physiological carriers, even though its concentration limits are normal. In such circumstances, copper can become available in the cell to harmful reactions that generate oxidative stress. This is actually the most likely mechanism of copper toxicity in other late-onset neurodegenerative diseases.

3.2. Alzheimer's Disease

The argument of improper copper binding as possible mediator of brain injury is relevant to Alzheimer's disease (AD), the most common cause of senile dementia with both familial and sporadic forms. It is a progressive neurodegenerative disorder characterized by neuronal cell loss or dysfunction. The neocortex and the hippocampus are both severely affected, but neurodegeneration is selective (82): thus, learning, behavior, and memory functions are impaired. Neurochemical hallmarks of this disease include the presence of proteinaceous deposits in neurons (neurofibrillary tangles) as well as in the extracellular space (cerebrovascular and neuritic plaques) (83). The principal component of these deposits is a peptide of 39-43 amino acids (about 4 kDa), called amyloid p protein (Ap), which has an important role in neuronal dysfunction because it is toxic to neurons (84). This peptide derives from the processing of a full-length, transmembrane protein, called amyloid precursor protein (APP), displaying several different isoforms, as a result of alternative splicing of a single gene. It consists of a large N-terminal extracellular region, a transmembrane domain, and a small cytoplasmic tail. All of the family components show binding sites for zinc, iron, and copper. Zinc binds to the extracellular domain and may regulate protein folding; furthermore, it has been recently suggested that it may favor precipitation of Ap, converting it into a less harmful form than the soluble one (85). Copper also binds to the extracellular region to a cysteine-rich domain by two conserved histidines (86). The physiological role of APP is not clear: however, owing to its membrane location and its capability of copper binding, a possible involvement in neuronal copper transport has been suggested (87). When bound copper(II) is reduced to copper(I), an intramolecular electron transfer may occur, leading to the formation of a disulfide bridge (88). Copper(I) may in turn be oxidized back by H2O2 with the production of OH- and random fragmentation of APP. Copper(II) may favor the aggregation process of Ap (89), especially under conditions representing physiological acidosis resembling those present in the inflammated brain parenchyma (90). Conversely, zinc may exert a preventive role of free-radical formation by quenching abnormal Ap-mediated redox activity (85). It should also be kept in mind that zinc, iron, and copper are more concentrated than normal in the neuropil of Alzheimer patients and are further concentrated in the core and periphery of plaques (91) and that this increase may be the result of metals binding to Ap. This is a very important point because Ap is present as a soluble protein in normal fluid and tissues and tends to form aggregates also in healthy aged individuals, although the aggregates forming in Alzheimer's brain are much less soluble. A plausible hypothesis would be that exposure to excessive amounts of these metals or predisposition to their selective accumulation is a risk factor for the onset of the disease.

The Ap interaction with redox active copper is a significant source of oxidative stress in AD (92,93). There is wide consensus indicating increased oxidative stress in the brain of Alzheimer patients, represented by increased protein oxidation (94,95), increased lipid peroxidation (96), and damage to both nuclear and mitochondrial DNA (97,98). Mitochondrial dysfunction is implicated in AD and it has even been proposed as a causative genetic factor in the pathogenesis of familial forms (82). There is also evidence that Ap impairs the activity of mitochondrial respiratory chain and reports evidenced decrease of the activity of cytochrome-c oxidase (99-101), although this may be a secondary event (102). Cortical, cerebellar, and hippocampal neuronal cultures derived from knockout APP mice were less susceptible than wild-type cells to physiological concentrations of copper, in terms of decreased levels of lipid peroxidation (103) and copper(II) markedly potentiates the neurotoxicity exhibited by Ap in cell cultures (104).

3.3. Parkinson's Disease

Recently, some evidence for the involvement of copper in the mechanism underlying the onset of Parkinson's disease (PD) has also emerged. This disease is characterized by resting tremor, rigidity, bradykinesia or slowness, and postural instability. Hallmarks of this disease include degeneration of dopaminergic neurons in the substantia nigra, with intracytoplasmic inclusions known as Lewy bodies

(105). Degeneration also involves neurons in the locus ceruleus, nucleo basalis, hypothalamus, and cerebral cortex. Established etiology of this disease involves environmental factors (exposure to pesticides, herbicides, industrial chemicals) and genetic factors (approx 5-10% of PD patients have a familial form of parkinsonism with an autosomal-dominant pattern of inheritance). Familial PD was linked to the q21-23 region of chromosome 4 (106); mutations in the gene coding for the protein a-synuclein, leading to replacements of the alanine residue in position 30 or 53 with threonine or proline, respectively, were identified (107,108). Although the physiological role of a-synuclein is still largely unknown, accumulation of this small protein seems to be a crucial event in the development of PD, because it is an abundant component of Lewy bodies in all of the cases of the disease. Furthermore, mutation of this protein in the familial cases might favor the accumulation process.

The pathogenesis of PD may involve oxidative stress, because oxidized marker molecules have been detected in the brains of PD patients (105). Oxidative insult may derive from several sources. Among those, increased turnover of dopamine, the oxidation of which leads to the formation of H2O2 (12) or derangements in mitochondrial complex I activity (109). The relevance of oxidative stress for this disease is also strengthened by the finding that in the PD brain, the antioxidant glutathione decreases (110) and that oxidized proteins (111), DNA (112), and lipids (113) are present in PD brain. Furthermore, excitotoxic damage mediated by NO and peroxynitrite seems to be relevant to cell damage, because, for instance, NO-dependent nitration of tyrosine residues on cellular proteins has been found in PD (114). The mechanism of cell death in PD seems to involve apoptosis, repeatedly detected in postmortem PD brains (105).

A direct involvement of copper in the pathogenesis and/or progression of PD has been envisaged. Defects of cytochrome-c oxidase in nigral neurones of PD patients have been found (115), as well as a significantly lower ceruloplasmin concentration and ceruloplasmin activity and Cu,Zn superoxide dismutase activity in blood (116). More interestingly, it has been demonstrated that copper(II) induces a-synuclein to form self-oligomers and that the acidic C-terminus of the protein is essential for the copper effect (117). Interestingly enough, fragments of a-synuclein represent the non-Ap component of senile plaques in AD, thus reinforcing the role of copper in AD progression. Therefore, abnormal copper homeostasis can be considered a risk factor also for the development of PD.

0 0

Post a comment