Iron: A Pathological Mediator of Alzheimer Disease?

Abstract

Metal-catalyzed oxidation and free radical formation are potent mediators of cellular injury to every category of macromolecule found in vulnerable neuronal populations and are thought to play an early and central role in Alzheimer disease (AD) pathogenesis. While metal-binding sites are present in proteins that accumulate in AD, metal-associated redox activity is primarily noted with nucleic acids, specifically with cytoplasmic RNA. Iron dyshomeostasis in AD is thought to arise from haem breakdown and mitochondrial turnover, and a reduction in microtubule density in vulnerable neurons increases redox-active metals, initiating a cascade of events culminating in characteristic pathologic features. Increased understanding of these early changes may be translated into more effective therapeutic modalities for AD than those currently in use.

Introduction

Trace elements, such as iron, copper, manganese and sulphur are essential for biological processes, among them cellular respiration which involves oxidation of glucose to carbon dioxide and water with the generation of energy as the high energy phosphate in ATP by transfer of electrons along the electron transport chain. Iron is crucial to this process (oxidative phosphorylation). It is a component of the haem groups of the oxygen-carrying molecules haemoglobin and myoglobin and of the electron-carrying mitochondrial protein cytochrome c. Iron is also a component of the enzyme catalase, which catalyses the decomposition of hydrogen peroxide, and peroxidase, which catalyzes oxidation of various organic substances by peroxides. Other enzymes in the electron transport chain contain non-haem iron. Electron transfer brings about metal-mediated detoxification of metabolic byproducts and, when dysregulated, can lead to toxin accumulation. Therefore, metal-dependent biological processes must be tightly regulated and in a finely-tuned homeostatic balance. When dysregulated, an excessive accumulation of metals, including iron, can result in cell death.

Degenerative diseases such as Alzheimer disease (AD) and Parkinson disease are believed to be caused by a combination of hereditary, environmental and lifestyle factors. In AD, metabolic imbalance and the resulting oxidative stress are believed to play a major role in disease pathogenesis [1, 2], accompanied by dyshomeostasis of the redox-active transition metal ions iron and copper as well as redox-inactive metal ions such as zinc.

Oxidative Stress

Oxidative stress may be defined as an imbalance between those biological processes contributing to the production of reactive oxygen species (ROS) and those responsible for their removal. In cellular respiration, during the reduction of molecular oxygen, mitochondria produce the superoxide radical O₂•, which yields H₂O₂ as a result of action of the enzyme superoxide dismutase. H₂O₂ is produced by oxidases such as monoamine oxidase. Superoxide can react with nitric oxide to yield peroxynitrite which is capable of both oxidation and nitration reactions. While O₂• and H₂O₂ are by themselves relatively non-toxic, H₂O₂, being freely permeable in tissues, can product the highly toxic radical •OH through the metal ion-catalyzed Fenton reaction which is referred to as the superoxide-driven Fenton reaction or the iron-catalyzed Haber-Weiss reaction. H₂O₂ can also be produced non-enzymatically from superoxide. Under conditions of normal cellular metabolism, superoxide favours the oxidation of Fe²⁺ to Fe³⁺ (ferrous to ferric). When the intra-cellular concentration of superoxide is higher than normal, however, the balance of the reaction is in favour of reduction of the ferric to ferrous iron (Fe³⁺ to Fe²⁺) and the elaboration of hydroxyl radicals, the principal ROS implicated in biologically-relevant oxidative stress and believed to be responsible, directly or indirectly, for most of the free-radical damage seen in AD.

In tissues subject to oxidative stress, there is ROS-mediated damage occurs in all biomolecules: nucleotides, proteins, lipids and sugars. This leads to a critical failure of cell function and ultimately to cell death. The central nervous system (CNS) is particularly vulnerable to oxidative stress both because of its high oxygen utilization as well as its relatively poor concentration of antioxidants and related enzymes. Additionally it has a high concentration of molecules susceptible to oxidative damage such as polyunsaturated lipids, proteins and nucleic acids. While oxidative balance is maintained under normal conditions and free radicals are detoxified, in the disease state, ROS species accumulate and several biomolecules are modified as a result oxidative damage. The well-established types of tissue damage seen in degenerative diseases such as AD include glycation [3], protein oxidation [4–6], lipid peroxidation [7] and nucleic acid oxidation [8–11] and its sequelae. They are all believed to result directly or indirectly from metal-catalysed superoxide radical production [12]. Many lines of evidence implicate ROS induced by redox-active metals including iron in the pathogenesis of AD [7, 13]. When metal ions become separated from specific storage and transport proteins, they can adventitiously bind to proteins, nucleic acids and circulating amino acids and remain active in ROS generation, as long as the coordination sphere of the metal ion is not saturated and supports the cycling between reduced and oxidized states of metal. Also, redox-inactive metal ions can be pathogenic because of their ability to displace redox-active metal ions from sites where the redox activity of the latter is held in check. Metals may also exert dual neurotoxic activities and contribute to neurodegeneration through their effects on protein and peptide structure such as pathological aggregation [14].

Normal Iron Metabolism in the Brain

Iron, more than any other transition metal, has been implicated in in vivo generation of ROS in tissues subjected to oxidative stress. In the brain, iron uptake is under control of the transferrin receptor in endothelial cells and cells of the choroid plexus [15] or the lactoferrin receptor on neurons and microvessels [16, 17]. A brain-specific caeruloplasmin and a transferrin whose expression are regulated by a CNS specific promoter are believed to play a role in the export of iron from neurons and non-neuronal cells [17]. In most CNS cells, iron is bound to ferritin in the redox-inactive state. Neuromelanin stores iron in the substantia nigra [17]. Also, in recent years, iron metabolism at the cellular level has been reported to be under the control of the lactotransferrin receptor, melanotransferrin, caeruloplasmin and the divalent cation transporter [14]. Disruption of the expression of these proteins in the brain can contribute to disordered iron metabolism leading to meurodegenerative diseases [18]. Overall, regulation of cellular iron metabolism is under the control of two iron regulatory proteins, IRP-1 and IRP-2 [19]. When there is iron depletion, IRPs stabilize the transferrin receptor and inhibit ferritin mRNA translation by binding to “iron responsive elements’ (IREs) within their untranslated regions. In cells replete with iron, IRP1 binding is prevented and IRP2 is degraded. IRP1, but not IRP2, is activated by extracellular H₂O₂, which provides the regulatory connection between response to oxidative stress and control of iron metabolism [14].

Neuropathology of AD

The hallmark lesions of AD pathology in the brain are extracellular amyloid plaques, known as senile plaques, and intraneuronal neurofibrillary tangles (NFT) as well as neuropil threads and selective neuronal loss. Senile plaques mainly contain the amyloid-β peptide, whereas NFT are mainly composed of paired helical filaments (PHF) containing the microtubule-associated protein, tau. Metal binding sites are present on proteins that accumulate in AD such as tau, amyloid-β and apolipoprotein E [20, 21]. Not surprisingly, zinc, iron, and copper, as well as assorted trace metals are increased in the neuropil in AD brains and often found concentrated in the core and periphery of senile plaques [3, 20]. Of significance is the fact that redox metal accumulation is not only associated with the pathological lesions, but is also found in neurons in early stages of disease development. This indicates that not only does disruption in iron homeostasis result in oxidative stress, but that, in fact, it is present before the development of pathological lesions and protein deposition in AD. Imbalance in IRP2 also suggests that iron homeostasis is impaired early in AD [22, 23]. Of importance is the observation that metal ion-mediated damage can be inhibited by pre-treatment with selective ion chelators and restored by post-chelation incubation in metal solutions [24]. Aggregates of amyloid-β can be re-solubilized by removal of metal ions with chelating agents at low concentrations (i.e., 4 μM).

Iron Deposition in Senile Plaques (Extracellular Iron)

Iron accumulation is abundant in vulnerable brain regions [13] and is seen microscopically to accumulate in senile plaques in AD [25]. Increases in oxidative stress and iron accumulation in AD brains are associated with changes in the concentration of soluble and deposited amyloid-β protein. The tyrosine residue at position 10 and three histidine residues at positions 6, 13 and 14, located at the hydrophilic N-terminal of amyloid-β are capable of binding iron. The iron bound to these sites generates H₂O₂ by the Fenton reaction described earlier and induces amyloid-β aggregation [26]. Modification of histidine residues can reduce Fe³⁺ -induced amyloid-β aggregation [26]. While it is still unclear whether the formation of senile plaques and amyloid-β aggregation are important in neurotoxicity, it appears that extracellular iron is a major source of free radicals and that the harmful effects of amyloid-β are mediated by adventitiously bound iron [27]. Furthermore, amyloid-β may even play a neuroprotective, chelating role in AD pathogenesis since amyloid-β significantly attenuates iron redox potential [27]. Senile plaques may indirectly contribute to ROS production and iron release from ferritin by activating microglia or reactive astroctyes which leads to synthesis and release of the cytokines interleukin (IL)-1, IL-6 and IL-8, which may in turn cause large amounts of ROS to be released from activated macrophages. Ferritin release results in lipid oxidation in vivo [28].

Iron in Neurofibrillary Tangles (Intracellular Iron)

Another site of iron accumulation is the other hallmark lesion of AD, namely NFT. Redox metals in NFT are known to induce oxidative stress via H₂O₂ [23]. Furthermore, oxidative stress precedes the formation of NFT since oxidative modifications accumulate in neurons lacking NFT. It remains unclear what role intracellular iron, in particular that bound to NFT, plays in AD neurodegeneration.

While tau, amyloid-β and apolipoprotein E proteins [29] each have metal binding sites, metal-associated cellular redox activity is largely dependent on nucleic acid-associated metals since nuclease pre-treatment significantly reduces this redox activity. RNA is suggested as an early target for direct metal binding and oxidative damage in AD brain and, in fact, studies show increased RNA oxidation in AD brain [9]. RNA has more non-base paired regions than DNA and oxidative damage to nucleic acids in AD is more extensive in RNA than DNA [9], supporting the idea that metals bound to RNA are major sites of redox activity. With 8-hydroxyguanosine (8OHG) as a reliable marker for nucleic acid oxidation, oxidative damage is found in the perinuclear cytoplasm of neurons. 8OHG is formed by an attack of the hydroxyl radical, cannot permeate through the plasma membrane, and must be produced within the cytoplasm in the vicinity of RNA. It would seem that metal-catalyzed redox activity, in cytoplasm, senile plaques or NFT, is dependent on adventitious binding sites [26].

Sources or Redox Iron

In addition to the increase in metal binding sites in AD brain, as compared to controls, there is an increase in the metals themselves. Possible sources of increased redox active iron include excessive haem turnover. The enzyme haem oxygenase-1 (HO-1) which catalyzes conversion of haem to iron and biliverdin and then to bilirubin was found to be increased in AD [30] suggesting haem turnover as a possible source of increased redox active iron.

Many haem-containing enzymes accumulate within mitochondria and there is a renewed interest in the potential role of this organelle as a source of redox-active metals. A 3–4 fold increase in the mitochondrial protein COX-1 and mitochondrial DNA was found in AD in vulnerable neurons [29]. Paradoxically, however, there was no increase in mitochondrial enzyme activity and morphometric ultrastructural analysis of biopsy specimens showed that, in fact, mitochondria are reduced in number in AD [26]. Ultrastructural in situ hybridization studies and immunocytochemistry showed the increased mtDNA and proteins to be localized, not to intact mitochondria, but to autophagosomes, suggesting that mitochondrial enzyme turnover in lysosomes are an additional source of increased haem turnover, which in turn induces HO-1 creating a vicious cycle. Morphometric analysis of the neuronal cytoskeleton in biopsy specimens suggests an explanation for why mitochondria in AD are targeted to lysosomes. A change in microtubule architecture has long been thought to underlie AD. Whether or not they contain NFT, there is a 50% reduction in neuronal microtubules in AD [29]. Because microtubules are essential for mitochondrial transport, a lack of microtubules would lead to either mitochondria accumulation or increased turnover in neurons. This scenario may lead to the increased metal turnover and oxidative damage prevalent in AD.

Conclusion

The role of redox active transition metals in the development of AD is a topic worthy of further investigation. Mitochondrial dysfunction, together with altered cytoskeletal architecture, is a possible sources of the increased redox active metals and initiates a series of events (i.e., nucleic acid and protein oxidation) that culminate in AD pathology. The sources and targets of oxidation need to be examined critically so that they may be translated into more effective therapeutic modalities. Currently, therapies for AD are “lesion-centered” and only provide some symptomatic relief and fail to alter disease progression or outcome. Since some of the lesions targeted by current therapy are also seen in normal aging, there has been, not surprisingly, little advance in AD treatment. Therefore, a multifaceted approach to treatment that targets early events, before the onset of pathology, is validated. Our current understanding of the role of iron in AD pathogenesis provides a legitimate alternative for new treatment options, since 1) iron is associated with oxidative stress and neurotoxicity; 2) the presence of AD lesions occurs after iron accumulation and oxidative stress; 3) iron accumulation is readily treatable; and 4) there is success in improving neuronal function with chelation therapy. Thus, redox-active iron is a valid target for the treatment of neurodegenerative disease.