The Promotion of Iron-Induced Generation of Reactive Oxygen Species in Nerve Tissue by Aluminum

Aluminum is suspected to play a role in several neurological dis orders. Reactive oxygen species (ROS) lead to oxidative stress, which is thought to be a possible mechanism for neurological damage. Inter actions between aluminum and iron, a known promoter of prooxidant events, were studied in cerebral tissues using a fluorescent probe to measure rates of generation of ROS. Al 2 (50 4 h alone failed to stimu late ROS production over a wide range of concentrations (50-1000 µ.M). The aluminum-dcferrioxamine chelate in the absence of iron could also not potentiate ROS formation. However, Al 2 (S0 1 ) 1 poten tiated FeS0 4 -induced ROS, with a maximal effect at 10 µ.M Fe and 500 µ.M AL Kaolin, a hydrated aluminum silicate, did not potentiate iron induced ROS formation. Ferritin had a minor stimulatory effect on ROS generation, but this was not potentiated by the concu'rrent pres ence of Al 2 (S0 4 )J. Transfcrrin had no effect on basal rates of ROS gen eration, but when Al 2 (50 4 h was also present, ROS production was enhanced. It is concluded that:


INTRODUCTION
The capacity of iron in a low-mol-wt form to stimulate biological lipid peroxidation is well known. This property has been reported to be enhanced within neural tissues by aluminum salts (Gutteridge ct al., 1985;Aruoma et al., 1989;Fraga ct al., 1990). The phenomenon is puzzling, in view of the inability of aluminum ions to undergo the changes in valency, characteristic of transition metals that accounts for their potential to promote excess pro-oxidant activity within cells. Neither does aluminum have a strong affinity for sulfhydryl groups. Although the intrinsic capacity of aluminum to induce the generation of free radicals has been reported, this may depend on the concurrent presence of iron (Fraga ct al., 1990;Oteiza et al., 1993).
In view of the proposed association between Ab.heimcr disease and cerebral aluminum content (Mclachlan et al., 1989), illumination of the effects of aluminum on cerebral metabolism may be important. The potentially pro-oxidant role of aluminum is especially relevant because of the reported presence of excess oxidative activity in both the aged brain, and more specifically, in those regions of the Alzheimer brain showing the most pronounced pathological changes (Smith ct al., 1991). Iron and aluminum are present at high concentrations in neurofibrillary tangles (Good et al., 1992) and may cooperate in the promotion of physiological aggregation of the {j-amyloid peptide (Shigematsu and McGcer, 1992;Mantyh et al., 1993). Such aggregation may be the result of oxidative events (Hensley et al., 1994). Lipid peroxidation rates are also elevated in autopsy samples of Alzheimer cortex (Subbarao et al., 1990), and there is other evidence of an association between this disease and oxidative stress (Pappola et al., 1992). Excess levels of iron together with aluminum have also been found in the substantia nigra of Parkinsonian patients (Hirsch et al., 1991), and aluminum accumulates within neurons in amyotrophic lateral sclerosis-Parkinsonian dementia of Guam (Perl et al., 1982).
The current study was designed both to confirm such interactions between iron and aluminum, and to understand the processes that may underlie this property. The direct assay of short-lived oxidative species was performed using a fluorescent probe, rather than measurement of more secondary indices of quantitation of damage to lipids or proteins. Issues that have been addressed include the potential of aluminum to displace iron from its intracellular binding sites, the ratio of iron to aluminum affecting such excess generation of reactive oxygen species (ROS), the molecular form of aluminum producing nonadditive pro-oxidant effects with iron, and the role of membranes in catalyzing any interaction .

Tissue Preparation
Male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) weighing 150-175 g were utilized. Rats were decapitated, the brains were excised quickly on ice, and the cerebrocortex was dissected out.

Assay for ROS Formation
ROS were assayed using 2 ', 7'-dichlorofluorescin diacetate (DCFH-DA), which is de-esterified within cells to the ionized free acid, dichlorofluorescin, DCFH. This is trapped within cells and thus accumulated (Bass et al., 1983). DCFH is capable of being oxidized to the fluorescent 2', 7'-dichlorofluorescein (DCF) by reactive oxygen. The utility of this probe in isolated subcellular cerebral systems has been documented (LeBel and Bondy, 1991). P2 or P3 suspensions were diluted in 9 vol of HEPES buffer. The diluted fractions were then incubated with 5 mM DCFH-DA (added from a stock solution of l.25 mM in ethanol) at 37°C for 15 min. After this loading with DCFH-DA, the fractions were incubated for a further 15 min in the presence of various metallic compounds. At the beginning and at the end of incubation, fluorescence was monitored on a Farrand Spectrofluorometer, with excitation wavelength at 488 nm (bandwidth 5 nm), and emission wavelength 525 nm (bandwidth 20 nm). The rate of generation was found to be linear over the incubation period (LeBel and Bondy, 1990). Autofluorescence of fractions was corrected for by the inclusion in each experiment of parallel blanks with no DCFH-DA. The correction for autofluorescence was always < 11 % of the total. ROS formation was quantitated from a DCF standard curve (0.05-1.0 mM), and results were expressed as nmol DCF formed/15 min/mg protein.

Protein Determination
Protein concentration was assayed using the method of Bradford (1976).

Statistical Analyses
Differences between groups were assessed by one-way analysis of variance followed by Fisher's least significant difference test. The acceptance level of significance was p < 0.05 using a two-tailed distribution.

RESULTS
There was a clear enhancement of FeSO,-induced rates of ROS generation within crude synaptosomes by Al 2 (S0 1 h. When the iron salt concentration was maintained at 10 µM, maximal potentiation was observed in the presence of 0.5-1 mM aluminum (Fig. 1). This amount of Al 2 (50 1 h was insufficient to lower the pH of the HEI'ES-buffered incubation medium. In the absence of added iron, Alz(S0 4 ). failed to stimulate ROS at all concentrations tested (50-1000 µM).
Tn the absence of aluminum, iron-stimulated ROS was maximal at 5 µM. However, in the presence of 0.5 mM Al 2 (S0 4 )i, the ROS-enhancing capacity of FeS0 4 was increased, and this was continued up to an iron concentration of 50 µM (Fig. 2).
There has been speculation that aluminum may enhance the potency of iron-induced lipid peroxidation by way of subtle alterations of membrane structure (Gutteridge et al., 1985;Oteiza et al., 1993). For this reason, the potential for interaction between iron and aluminum in the absence of membrane structures was examined by use of the postmicrosomal high-speed supernatant fraction of cerebral cortex. In this case,  0.52 ± 0.09 4.42 ± 0.18" 0.67 ± 0.14 0.24 ± 0.05 0.47 ± 0.09 6.79 ± 0.12" basal rates of ROS production were very low, and these were not stimulated by addition of inorganic iron or aluminum compounds, or by the presence of both ions (date not shown). Thus, the presence of a membrane fraction is needed for the potentiation described. A proposal has been made that aluminum-silicon complexes, such as zeolite or kaolin, may be the most potent accelerators of iron-based oxidative stress (Carrel et al., 1994) and that these may be related to the pathogenesis of Alzheimer disease (Evans et al., 1992). The formation of ROS within the cortical P2 fraction was inhibited by a colloidal suspension of kaolin at a concentration of 170 µg/mL. This concentration was calculated to result in a aluminum concentration of 0.5 mM. The stimulatory effect of FeS0 4 was inhibited rather than stimulated in the presence of kaolin (Table 1). Since kaolin did not inhibit basal rates of ROS generation, this result could not have been owing to the quenching of fluorescence. The ionic form of aluminum may be essential for the promotion of Data represent means from 6-10 individual determinations ± SEM. "Significantly elevated relative to basal level.
iron-induced ROS. Another issue that was addressed is whether the observed potentiation of ROS was the result of the gradual formation of an insoluble complex, perhaps of hydrated ions. The time-course of appearance of enhanced ROS generation by Fe + Al salts was determined. A freshly prepared mixture of the salts was just as active as the same mixture that had been allowed to stand for 2 d (data not shown). Centrifugation of the standard mixture of 50 µM FeS0 4 and 500 µM AIAS0 1 h did not alter the capacity of the mixture to promote excess ROS (data not shown). These data suggested that the promoting effect of aluminum on ironinduced ROS does not involve the formation of a high-mol-wt insoluble mineral complex.
Jt may be that the displacement by aluminum of iron from its biologically complexed forms can account for some of the observed interactions between these metals. To test this hypothesis, the effect of aluminum on ROS production was examined in the presence of ferritin, an iron storage protein, and transferrin, a protein involved in iron transport. Both proteins were in the iron-loaded form. Ferritin normally contains some complexed aluminum (Dedman et al., 1992), which can modulate its properties (Fleming and Joshi, 1991). Transferrin has a high affinity for, and may transport aluminum into the central nervous system (CNS) (Roskams and Connor, 1990). The concentrations of ferritin and transferrin used (13.5 µg/mL and 10 mg/mL for ferritin and transferrin, respectively) were calculated to give an iron concentration roughly equivalent to 75 µMiron. Ferritin stimulated ROS generation as has been previously described (Kukiella and Cederbaum, 1994), but this effect was not enhanced by the concurrent presence of aluminum (Table 2). Transferrin did not affect basal rates of ROS production by the crude synaptosomal fraction, but when aluminum sulfate was also present, there was a major elevation of ROS ( Table 2), suggesting that iron may have been displaced from this protein by aluminum.
A final issue that has been addressed is the possibility that chelatcs of aluminum may be able to promote ROS formation even in the absence of iron; 500 µM aluminum sulfate potentiated the ROS generating capacity of 1 mM ascorbate. The stimulation of ROS by ascorbate alone at this concentration was completely abolished by 10 µM desferrioxamine (Fig. 3). This low concentration of deferoxamine was intended to sequester iron selectively. Higher concentrations of this compound have an intrinsic ability to quench free radicals (Halliwell, 1989). In view of the high affinity of deferoxamine for iron, this was interpreted as demonstrating the requirement for the presence of trace amounts of iron for this pro-oxidant effect of ascorbate . In contrast, the promotion of ascorbate-induced ROS by aluminum was not prevented by 10 µM desferrioxamine.

DISCUSSION
The findings reported above suggest that the potentiation of ironinduced ROS generation by aluminum cannot be largely attributed to: 1. The presence of complex aluminum-silicon-based mineral surfaces on which iron catalysis of ROS synthesis can occur; 2. The intrinsic pro-oxidant properties of free aluminum salts in the absence of iron or chelators; 3. The displacement of iron from biologically complexed sites, causing its appearance in low-mol-wt forms. Although aluminum may bind to transferrin, the iron content of this protein is low (around 0.14%), and the respective affinity constants of the iron and aluminum complexes make this possibility unlikely.
The most promising remaining explanation of the synergistic interactions between these metals may then be the formation of iron-aluminum-hydroxide complexes. These ·may enhance the potency of iron by formation of a type 2 chelate, parallel to the sequestration of iron by ascorbate of EDTA. Another possibility is the destabilization of lipid bilayer structures by aluminum salts, making membranes more vulnerable to oxidative damage (Oteiza, 1994). In view of the relatively low penetrance of either metal ion across cell membranes, some of the effects described in this work may involve changes at the surface of the limiting plasma membrane of synaptosomes, which are then reflected by altered intracellular events. Aluminum appeared able to promote ascorbate-induced ROS in the absence of unchelated iron. Thus, although not possessing an intrinsic capacity to stimulate ROS production by tissues, aluminum may be able to enable the generation of ROS by chclators, in the absence of metals with multivalent potential, such as iron. Aluminum can be weakly chelated by desferrioxamine (Kruck et al., 1993), and such complexing can enhance the rate of excretion of this metal (Mclachlan et al., 1989). This chelator was only present at 1/100 of the aluminum concentration in this experiment, and results could thus not be attributed to sequestration of aluminum. However, the presence of a high concentration of aluminum may displace iron from its desferrioxamine chelate and allow iron-ascorbate interactions. This experiment therefore does not conclusively demonstrate an iron-indedendcnt aluminum-ascorbate ROS-promoting interaction.
These findings derived from an isolated system do not preclude the possibility that particulate aluminum-containing materials may gain access to the CNS, and thus have the potential to activate microglia or macrophages, which in turn could lead to deleterious oxidative events within the brain (Evans et al., 1991). Soluble aluminum may be transported to the CNS by formation of a complex with transferrin, which also chelates fe3+ (Roskams and Connor, 1990). The similarity of the ionic radii of aluminum and ferric iron (9 A) (Dean, 1979) may be relevant to the transport and release of unbound iron and aluminum in the brain.