The therapeutic opportunities and pitfalls of using iron ion chelators to treat neurodegenerative diseases Authors: Barry Halliwell Published: 2026 ABSTRACT There is considerable evidence that iron accumulation in the affected brain regions plays a role in the pathology of neurodegenerative diseases, probably because of the ability of iron ions to promote oxidative damage to important biomolecules, such as DNA, RNA, proteins and lipids (leading to lipid peroxidation). Yet recent clinical trials of iron ion‐chelating agents in human neurodegenerative diseases have given unimpressive results, in several cases making the diseases worse. In using iron chelators, it is important to remember the basics of their redox chemistry, which could explain their lack of therapeutic effects. This issue is explored here, and the potential of iron ion chelation in the treatment of neurodegeneration is critically evaluated. Abbreviations Aβ, amyloid‐beta AD, Alzheimer's disease CSF, cerebrospinal fluid DFX, desferrioxamine ESR, electron spin resonance Fe‐S, iron–sulfur FA, Friedreich's ataxia L1, the iron chelator deferiprone MCI, mild cognitive impairment MRI, magnetic resonance imaging PD, Parkinson's disease The role of iron in neurodegenerative diseases Common features found in the affected brain regions of patients with various neurodegenerative diseases include impaired mitochondrial function, increased iron deposition, accumulation of protein aggregates (e.g., plaques and tangles in Alzheimer's disease, Lewy bodies in Parkinson's disease) and increased oxidative damage (defined as the damage caused by oxygen radicals and other reactive oxygen species to important biomolecules, such as DNA, RNA, proteins and lipids, leading to lipid peroxidation), as reviewed in [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11]. For example, increased oxidative damage in Alzheimer's disease (AD) starts early in the disease process and affects DNA, RNA, lipids and proteins. Damaged proteins include enzymes essential to the clearance of abnormal proteins and to energy metabolism, thus impairing brain glucose metabolism and ATP synthesis as well as removal of damaged proteins [7, 8, 9]. There are multiple reasons as to why the brain is prone to oxidative damage, which are reviewed in [1, 2, 10, 12]. One particular reason is that several brain regions are rich in iron, which appears to be more easily mobilised in a redox‐active form (i.e., a form able to catalyse oxidative damage) during tissue injury/degeneration than it is for other tissues [2, 10], a phenomenon described as long ago as 1974 [13]. Release of redox‐active iron ions (sometimes called ‘catalytic iron ions’) can promote lipid peroxidation and oxidative damage to DNA, RNA and proteins by multiple mechanisms, including the Fenton reaction, (1) Fe2++H2O2➔FeIII+˙OH+OH− Fenton chemistry generates the highly reactive hydroxyl radical (˙OH) when the catalytic Fe2+ ions are bound to several physiological ligands (such as citrate, phosphate, ADP and ATP) or to the phosphodiester bonds of DNA and to Fe2+‐binding sites on lipids and proteins [1, 14, 15, 16]. Hydroxyl radical can attack and damage all biomolecules at very fast rates [1, 14, 15, 16]. Indeed, sequestration of iron ions in non‐redox‐active forms is a key part of human antioxidant defences in blood plasma and many tissues [1, 16, 17, 18], but this defence appears to be weaker in human brain. One reason is that cerebrospinal fluid (CSF) has very limited capacity to bind redox‐active iron and may even have such iron present all the time at low micromolar levels [19, 20]. This is unlike human plasma, which contains no redox‐active iron in healthy subjects and indeed has a substantial capacity to bind free iron ions because of the presence of high levels of transferrin and lactoferrin, which sequester free iron and render it redox‐inactive [17, 18, 21, 22]. More recently, catalytic iron and oxidative damage, especially to lipids (lipid peroxidation), have been shown to play an essential role in ferroptosis [23], a cell death pathway, which is increasingly being shown to be important in the pathology of human neurodegenerative diseases, especially Parkinson's (PD) and Alzheimer's diseases [24, 25, 26]. Increased iron deposition in the affected brain regions of patients with PD and AD was observed as far back as 1924 and 1953 respectively, and a wide range of studies has shown that this abnormal iron deposition correlates well with increased disease severity, especially in PD and AD [24, 25, 26, 27, 28]. In AD, some of this extra iron is attached to amyloid plaques [25, 29]. Indeed, iron ions can also promote aggregations of amyloid beta (Aβ), a phenomenon intimately involved in the pathology of AD [25]. Because of the above observations on the role of iron, a number of clinical trials have been conducted using iron ion‐chelating agents as putative agents to decrease the progress of neurodegeneration. Early studies examined desferrioxamine (DFX), a naturally occurring chelator produced by Streptomyces pilosus (Fig. 1), which has been widely used to treat iron overload diseases in humans, such as haemochromatosis and thalassaemia. DFX is a powerful chelator of Fe3+ ions. It inhibits iron‐ion‐dependent lipid peroxidation and the Fenton reaction (Eq. 1) by loosely chelating Fe2+ and accelerating its conversion to Fe3+, which it then binds strongly [30]. Interestingly, this generates a burst of the superoxide radical anion. (2) Fe2+–DFX+O2➔Fe2+–DFX+O2˙− However, the Fe3+ iron is then locked up into a stable redox‐inactive ferric‐DFX chelate so that it is very difficult for it to be converted back to Fe2+ to promote Fenton chemistry (Eq. 1) and other iron‐dependent oxidative damage [1, 30]. Structure of desferrioxamine B (DFX B), the most commonly used form of desferrioxamine. DFX is a linear molecule that ‘bends round’ to chelate Fe(III) with six ligands (i.e., it is a hexadentate chelator) forming a bright‐red water‐soluble complex (ferrioxamine) that is readily excreted. Commercially available as desferal® (desferrioxamine B methanesulphonate). Often called deferoxamine. Key structural features include three R—C(=O)—N(OH)—R' (hydroxamic acid) units, the N—OH (N‐hydroxy) part is crucial for iron binding. They coordinate with a single ferric iron (Fe3+) ion, creating a stable, octahedral complex (ferrioxamine). This effectively removes catalytic iron from the body, preventing iron‐mediated oxidative damage. Desferrioxamine also has direct antioxidant properties, being able to directly scavenge several reactive oxygen species [30]. It can also raise levels of H1F‐1α and lead to the activation of the cellular hypoxic response [31]. In 1991, the results of a clinical trial were published claiming that DFX administered intramuscularly in patients slowed AD progression over 2 years [32]. Unfortunately, this study was heavily criticised and has not been repeated (reviewed in [28]). Early enthusiasm for the use of DFX waned when its poor ability to cross the blood–brain barrier was realised. It also has to be administered by infusion because of poor intestinal absorption. Interest thus switched to orally active chelating agents with better brain penetrance, such as the hydroxypyridones, especially deferiprone, sometimes called L1 (Fig. 2) [27, 33]. Chemical structure of deferiprone. Deferiprone (L1) is an orally active iron‐chelating drug; its full chemical name is 3‐hydroxy‐1,2‐dimethylpyridin‐4(1H)‐one. It has methyl groups at Positions 1 and 2, and a hydroxyl group at Position 3, allowing it to bind iron in the body to treat iron overload conditions, such as thalassaemia. Its structure creates a water‐soluble complex with ferric iron (Fe3+) in a 3 : 1 ratio (deferiprone:iron) for excretion. It is a bidentate (two‐binding sites) chelator, meaning it can bind to iron ions at two points. Iron chelators as therapeutic agents—The clinical trial results The clinical trials with L1 in patients with mild cognitive impairment (MCI) or early‐stage AD have revealed that it seems to cause harm rather than conferring benefit [27, 34]. Brain iron was lowered, but there was a marked acceleration in cognitive decline compared to placebo [34]. Similarly, in PD patients, L1 accelerated motor impairment, perhaps due to lowering dopamine levels, although even when dopaminergic therapy was administered in parallel, no clinical benefit was observed [27, 35]. One human disease in which Fe2+ is very clearly implicated is Friedreich's ataxia (FA), a genetic defect (mutation on a gene on chromosome 9) that leads to decreased levels of frataxin, a protein within mitochondria that aids the assembly of iron–sulphur (Fe‐S) clusters essential for ATP production by the electron transport chain [36, 37]. This defect may also lead to increased intramitochondrial levels of catalytic iron ions, leading to aggravated oxidative damage and increased ferroptosis [36, 37]. The brain and heart are particularly affected. Yet again, a clinical trial found that L1 at high doses (40 or 60 mg/kg) aggravated symptoms of the disease [27, 38]. What went wrong? It must first be remembered that iron is essential for the function of human cells, and so prolonged administration of any powerful iron‐chelating agent to humans (other than patients suffering from iron overload diseases) will probably cause side effects, especially if it enters the brain, since iron is required for multiple pathways in the brain, including essential iron‐dependent enzymes and the process of myelination [39, 40]. The ideal would be to have a chelator that binds catalytic iron but does not remove iron from enzymes and other proteins essential for brain function, especially not from iron‐binding proteins such as transferrin and lactoferrin, which render iron redox‐inactive [21]. This seems to be the case for DFX [30]. Iron chelators have also been examined in acute situations, such as stroke or trauma‐related brain haemorrhage, in which iron ions and haem proteins (especially haemoglobin, which can degrade to release haem and iron ions [16]), are often released and promote oxidative damage [1, 2, 10, 41, 42, 43, 44]. To date, clinical trials have not shown a clear benefit of DFX for these conditions. Nevertheless, studies are continuing, and current evidence suggests that DFX does not aggravate neurological damage and may even exert neuroprotection [40, 41, 42, 43, 44]. DFX may be able to penetrate into the brain more readily because of damage to the blood–brain barrier in stroke and trauma. A second issue is that increased iron deposition does not necessarily mean that the deposited iron is redox‐active [1, 2, 10]. Protein‐bound iron (e.g., in transferrin, lactoferrin and iron‐storage proteins such as ferritin and haemosiderin) is not, unless the iron is released from these proteins, which it can be under certain circumstances, for example from transferrin at low pH or (to a limited extent) from ferritin exposed to the superoxide radical [1, 16, 17, 20, 21, 45, 46]. The same could be true for the increased iron deposits seen in neurodegenerative diseases—data establishing that they are redox‐active are actually quite sparse. Iron deposits detected by commonly used techniques such as magnetic resonance imaging (MRI), atomic absorption or Perls staining, do not necessarily represent redox‐active iron [1, 39]. Increased levels of lactoferrin have been found in PD and AD brain samples [47, 48]; any iron bound to this protein would be redox‐inactive, as mentioned earlier [17, 21, 22]. Here may lie a major problem with L1. In order for iron chelators to be therapeutically effective yet nontoxic, they should bind catalytic iron efficiently but not mobilise it from ‘safe’ (redox‐inactive) protein stores or from essential iron‐containing enzymes. Most importantly, the iron bound by the chelator must be held in a redox‐inactive form, one that cannot itself promote iron‐dependent oxidative damage. However, hydroxypyridones such as deferiprone (L1) cannot always work this way [1, 33, 49], that is L1 can sometimes mobilise iron from proteins [33] and cannot always render iron bound to it as redox‐inactive. In particular, a single molecule of hydroxypyridone occupies only two of the six coordination sites on iron ions (i.e., it is a bidentate ligand; Fig. 2) and three hydroxypyridone molecules are needed to chelate one Fe(III) completely [33, 49]. By contrast, desferrioxamine is a hexadentate ligand: one molecule occupies all six coordination sites to render the Fe3+ ion redox‐inactive (Fig. 1). 2 : 1 or 1 : 1 hydroxypyridone:iron ion complexes leave available coordination sites on the iron that can allow catalysis of free‐radical reactions, including the Fenton reaction, as discussed in [1, 49]. Thus, it is conceivable that both catalytic iron and any non‐redox‐active protein‐bound iron in the AD, PD or FA brain, once mobilised by deferiprone [33], may form deferiprone–iron complexes that aggravate oxidative damage [49]. This effect has indeed been demonstrated, using serum from thalassaemia patients. Jirasomprasert et al. [50] used the technique of electron spin resonance (ESR), which detects free radicals because of their unpaired electrons. They showed that shortly after deferiprone administration to the patients, ESR signal intensities in the serum decreased concomitantly with an increase in serum levels of deferiprone, consistent with an antioxidant effect of this chelator. Notably, enhanced ESR signal intensities were observed 300 min after dosing in patients whose serum molar ratios of deferiprone to catalytic iron were below 3. These findings suggest that the formation of incompletely coordinated iron–deferiprone complexes (i.e. chelator:iron ion ratios < 3) may facilitate free radical formation, thereby exerting a pro‐oxidant action. Jirasomprasert et al [50] recommended that to avoid adverse effects of deferiprone, a dosage regimen should be designed, according to the iron status of the patients, to maintain an adequate ratio of serum chelator‐to‐iron concentration (> 3 : 1). However, this is very difficult to control in vivo, especially in the brain. It is also possible that such pro‐oxidant effects of ‘incomplete’ L1 chelates (ratio of L1 to iron < 3 : 1) could account for the generally greater toxicity of L1 relative to desferrioxamine in the treatment of subjects with iron overload [33, 49]. This possibility should be carefully considered in any future studies using L1. Conclusions and future prospects The use of iron chelators for the long‐term treatment or prevention of neurodegenerative diseases seems problematic given the essential role of iron in the brain and body, unless a brain‐penetrant chelator that is hexadentate (binding and holding catalytic iron in a redox‐inactive form) and does not affect the essential roles of iron in enzymes, brain mitochondria, oligodendrocytes and elsewhere can be identified. However, work along those directions is proceeding, and time will tell whether the agents being tested will have clinical benefit. There is a recent resurgence of interest in the possible use of DFX given its apparent safety, for example by intranasal administration, which may increase its brain bioavailability [51, 52, 53]. Ultimately, however, its efficacy must be established through double‐blind, placebo‐controlled clinical trials. Another approach involves incorporating iron ion‐chelating ability into other brain‐penetrant molecules, such as monoamine oxidase inhibitors [54]. While this strategy is promising, its clinical utility remains to be demonstrated in controlled trials. Lactoferrin is also attracting renewed interest [47, 48] because of its ability to cross the blood–brain barrier, at least to some degree. Whether sufficient amounts can reach the brain to be neuroprotective remains to be determined. Finally, some neuroprotective brain‐penetrant nutrients are able to bind iron ions in a redox‐inactive form. Ergothioneine (ET) is one example [55]. This diet‐derived molecule, currently under clinical evaluation as a possible agent to delay the onset of dementia [55, 56, 57], is also under preclinical evaluation for the treatment of PD [58, 59]. Pilot clinical trials with ET look promising [57, 60], but large‐scale trials are needed to confirm (or refute) its effectiveness. Whether any neuroprotective actions of ET are due to its iron‐binding activity or to its several other possible actions [55, 60] would also need elucidation. Another possibility to be considered, as recently articulated, is that the emphasis placed on the role of excess iron as a neurotoxic agent has been misplaced [39]. Only time—and further evidence—will determine whether this view is correct.