Mapping Brain Metals to Evaluate Therapies for Neurodegenerative Disease

Abstract

The brain is rich in metals and has a high metabolic rate, making it acutely vulnerable to the toxic effects of endogenously produced free radicals. The abundant metals, iron and copper, transfer single electrons as they cycle between their reduced (Fe²⁺, Cu¹⁺) and oxidized (Fe³⁺, Cu²⁺) states making them powerful catalysts of reactive oxygen species (ROS) production. Even redox inert zinc, if present in excess, can trigger ROS production indirectly by altering mitochondrial function. While metal chelators seem to improve the clinical outcome of several neurodegenerative diseases, their mechanisms of action remain obscure and the effects of long‐term use are largely unknown. Most chelators are not specific to a single metal and could alter the distribution of multiple metals in the brain, leading to unexpected consequences over the long‐term. We show here how X‐ray fluorescence will be a valuable tool to examine the effect of chelators on the distribution and amount of metals in the brain.

Introduction

The brain acquires metals and exploits their unique biochemistry as part of its normal functioning [1, 2]. Metals such as iron, copper and zinc act as essential cofactors in metalloproteins required for the normal functioning of the nervous tissue (Table 1) with unique importance in myelin synthesis and neurotransmission while heavy metals such as mercury and lead are known neurotoxins. There is increasing evidence that dysregulation of iron, copper, and zinc homeostasis contributes to a vast range of neurodegenerative diseases. Here we describe how synchrotron x‐ray fluorescence (XRF) imaging can be used to quantitatively assess how the distribution and chemistry of multiple brain metals in experimental animal models are affected by chelators currently used for the treatment of neurodegenerative disease.

Table 1.

Important metalloenzymes and roles of metals.

Role	Metal	Enzyme
Neurotransmission
Dopamine synthesis	Iron	Tyrosine hydroxylase
Norepinephrine synthesis	Copper	Dopamine β hydroxylase
Neuromodulation at excitatory synapses (down‐regulates glutamate response)	Zinc	N/A (vesicular Zn2+)
Antioxidant defense
Dismutation of O2 −· to H2O2	Copper, zinc	Cu/Zn SOD
Dismutation of O2 −· to H2O2	Manganese	Mn SOD
Reduction of H2O2	Selenium	Glutathione peroxidase
Reduction of H2O2	Iron	Catalase
Protection of mtDNA against ROS	Iron	Aconitase
Iron metabolism
Registration of cellular iron status and translational regulation of expression of iron metabolism proteins	Iron	IRP1/aconitase
Ferrous iron oxidation	Copper	Ceruloplasmin
Ferrous iron oxidation	Copper	Hephaestin
Heme synthesis
Conversion of δ ALA to porphobilinogen	Zinc	δ ALA dehydratase
Insertion of ferrous iron into protoporphyrin IX with formation of heme	Iron	Ferrochelatase
Pentose phosphate shunt
Conversion of glucose‐6‐phosphate to 6‐phosphoglucono‐δ‐lactone with the formation of NADPH	Iron	Glucose‐6‐phoshate dehydrogenase
Glycolysis
Conversion of glucose to pyruvate, NADH and ATP	Magnesium, potassium	Kinases (the substrate is a Mg2+‐ATP complex; Mg2+ is essential for kinase activity)
Synthesis of acetyl‐CoA
Activation of pyruvate dehydrogenase complex	Calcium	Pyruvate dehydrogenase phosphatase (activated by Ca2+)
Citric acid cycle
Isomerization of citrate to isocitrate	Iron	Aconitase
Dehydrogenation of succinate to fumarate	Iron	Succinate dehydrogenase
Decarboxylation of isocitrate to α‐ketoglutarate	Calcium	Isocitrate dehydrogenase (activated by Ca2+)
Decarboxylation of α‐ketoglutarate to succinyl‐CoA	Calcium	α‐ketoglutarate dehydrogenase (activated by Ca2+)
Electron transport
Transport of e− to CoQ	Iron	Complex I
Transport of e− to CoQ	Iron	Complex II
Transport of e− from CoQ to cytochrome c	Iron	Complex III
Transport of e− to complex IV	Heme	Cytochrome c
Four e− reduction of O2 to H2O	Iron, copper, zinc, magnesium	Complex IV
Other
Cholesterol synthesis	Iron	HMG‐CoA reductase
Synthesis of NO	Iron	NO synthase
Synthesis of deoxyribonucleotides	Iron	Ribonucleotide reductase
Conversion of toxic sulfite to sulfate	Iron, molybdenum	Sulfite oxidase
Oxygen transport	Iron	Hemoglobin
Purine symthesis	Iron	Amido phosphoribosyl transferase
DNA binding	Zinc	NFkβ, Sp1, PARP
Formation and compaction of myelin	Zinc	MBP, MAG
Initiation and propagation of action potentials	Sodium, potassium, calcium	N/A
Long‐term potentiation	Magnesium, calcium, zinc	N/A
Second messenger	Calcium	N/A
Apoptosis	Calcium, zinc	N/A

N/A not applicable; O₂ ^−· superoxide anion; H₂O₂ hydrogen peroxide; SOD superoxide dismutase; IRP iron regulatory protein; ALA aminolevulinic acid; NADPH nicotinamide adenine dinucleotide phosphate (reduced); NADH nicotinamide adenine dinucleotide (reduced); ATP adenosine triphosphate; CoA coenzyme A; CoQ coenzyme; NO nitric oxide; NFkβ nuclear factor kβ; Sp1 specificity protein 1; PARP poly (adenosine diphosphate‐ribose) polymerase; MBP myelin basic protein; MAG myelin‐associated glycoprotein.

Mapping and Quantification of Metals in the Brain

While analytic techniques such as atomic absorption spectrometry and inductively coupled mass spectrometry have been used to quantify metals in selected brain regions [3, 4, 5, 6, 7, 8, 9, 10, 11] and some magnetic resonance imaging (MRI) sequences, such as susceptibility weighted imaging are used to image iron in vivo[12], histochemistry has long been the gold standard to localize metals in brain slices. Laser ablation inductively coupled plasma mass spectrometry (LA‐ICP‐MS) is another very powerful mapping and quantification tool that has been applied to brain (3, reviewed in 4). XRF is in some respects a ‘poor man's’ LA‐ICP‐MS because little or no sample preparation is required, no equipment needs to be purchased and the use of the synchrotron is free of charge. XRF can map multiple metals and is also nondestructive so samples can be used for other purposes following mapping.

While Perls’ and Turnbull's methods detect only nonheme iron [13, 14, 15, 16] they are useful in brain and comparable to XRF because most of the iron in the brain is nonheme iron. Detection of copper by histochemistry lacks sensitivity and specificity which has hindered studies, e.g. of Wilson's disease [17, 18, 19]. Complex techniques have been developed to distinguish protein‐bound from free zinc [20, 21]. In contrast, XRF detects total zinc which can now be quantified to allow comparison between individual brain samples. Unlike XRF which can map multiple metals simultaneously, each histochemical method employs a different chemistry, and so they cannot be combined or used sequentially on the same tissue section. XRF removes one of the barriers to elucidation of the functional interrelationships between metals by making metal co‐localization easy and quantitative.

X‐ray fluorescence spectroscopic imaging (XRF) is element specific and can simultaneously map multiple metals in all chemical forms. Our laboratory has applied rapid scanning XRF (RS‐XRF) to map the global distribution of iron, copper, and zinc in whole slices of both normal and diseased brain [22, 23, 24]. Traditional point‐to‐point X‐ray microprobe has previously been successfully employed to map multiple metals in small areas of brain tissue sections at cellular or subcellular resolutions [25, 26, 27, 28, 29], but the time required to map whole brain slices with the microprobe is prohibitive. For example, a human brain hemisphere can be mapped in 6 h at 100 μm resolution [22] but at 2 μm resolution it would take days. Therefore, RS‐XRF is ideal to get quick anatomic views of global metal distribution in slices of whole organs.

Used as a quantification tool, RS‐XRF is a cost‐effective alternative to conventional methods in which excised cubes of tissue are dissolved for bulk analytical analysis. Recently, new tools have been developed that enable the user to calibrate metal maps using XRF quantification standards to yield concentrations in μg/cm². Standard quantification foils with known amounts of deposited metal are mapped and then average fluorescence, concentration of the foil and other values (e.g. gains) are used to create a parameter file that yields a concentration scale bar for each image. While concentration per cm² is very useful for thin sections that are essentially two‐dimensional, we have developed quantification algorithms that take into account the fluorescence escape depth for each element and the x‐ray density of the brain to yield concentration in μg/g w/w (Hopp et al., unpublished). Values for iron using this method were comparable to published iron concentrations using inductively coupled plasma mass spectrometry (Table 2).

Table 2.

XRF Quantification of brain iron. (Hopp et al., in press)

Anatomic Region	XRF (counts)	Iron μg/g ww (ave foil)
Periventricular white matter	21.8	46
Subcortical white matter, precentral gyrus	39.6	84
Superior temporal gyrus	41.2	88
Insular cortex	43.0	92
Middle frontal gyrus	43.0	92
Cortex, precentral gyrus	47.3	101
Internal capsule	52.0	111
External capsule and claustrum	54.8	117
Middle temporal gyrus	57.5	122
Ventricular wall, superior to caudate nucleus	67.5	144
Superior caudate nucleus	91.5	195
Superior globus pallidus	121.7	259
Superior putamen	122.5	261
Inferior caudate nucleus	123.1	262
Anterior commissure	144.0	306
Inferior globus pallidus	157.6	335
Inferior putamen	163.5	348

Used as an imaging method, RS‐XRF shows all metals simultaneously thus surpassing all “single‐metal” histochemistry. Unlike conventional histology sections, XRF samples retain metals because they do not have to be dehydrated or embedded, and after nondestructive XRF mapping, the same sample can be processed for histology or immunohistochemistry. RS‐XRF is the only way to nondestructively and simultaneously visualize multiple metals in the same section to gauge the effect of therapeutic drugs on brain metals. Synchrotrons around the world accept proposals to access beamtime on a peer‐reviewed basis and once awarded, beamtime is free for researchers. Access can also be purchased by companies at a reasonable cost. Software to visualize and quantify images, Sam's microanalysis toolkit (SMAK), can be downloaded free of charge. Therefore XRF is a useful tool to assess the effect of drugs on preventing or reversing metal‐related brain pathology. We have also shown that it can be combined with MRI analysis of iron [30].

RS‐XRF has shown iron, copper, and zinc localized to discrete regions within the normal substantia nigra [22] and dentate nucleus of the cerebellum [23], high copper content in the normal olivary region of the medulla [24] and shown changes in metal content and distribution associated with neurodegeneration [24]. By combining RS‐XRF with X‐ray Absorption Near Edge Structure (XANES) analysis of regions of interest, we have determined the chemical form of copper in the medullary olive (data not shown).

RS‐XRF makes systematic studies of the effects of therapeutic agents on brain metals practical.

X‐ray Fluorescence Imaging and Spectroscopy Background

Rapid‐scanning X‐ray fluorescence imaging (RS‐XRF) and X‐ray absorption spectroscopy (XAS) are well‐established, element‐specific, quantitative techniques used to determine the location and chemical form of elements in a broad range of tissue samples. X‐rays, focused to a small spot or passed through a pinhole, interrogate the sample as it is raster‐scanned in the beam. For imaging, the incident X‐ray beam energy is fixed so as to excite all elements with K‐shell absorption edges below the incident energy. In our experiments 13 keV is suitable to excite K‐shell electrons of iron, copper, and zinc, but would also excite the L‐shell electrons of gold or mercury (if present) and if the detector is very close to the sample, even sulfur, calcium, and phosphorus can be mapped. After a K‐shell electron is excited and ejected into the continuum, an electron from an outer shell falls in to fill the core hole, emitting a fluorescent photon of a specific energy. These X‐ray fluorescent photons are detected and by carefully windowing on the highest energy X‐ray fluorescence emission line for each element, the location of multiple elements can be simultaneously mapped. Because XRF relies upon the physics of the atom rather than on a chemical reaction, every atom of a metal that is interrogated by the x‐ray beam emits X‐ray fluorescence regardless of its chemical form, oxidation state or physical state (bound or free). The depth from which X‐ray fluorescence, arising from a specific element, can escape from a sample is known. For example, zinc is measured from a greater depth than iron because it has a more energetic X‐ray fluorescence.

Because the X‐ray beam interrogates the sample for only milliseconds per point, radiation damage is low and the sample can be used for other analytic techniques, such as histology or MRI. After metals are mapped, points of interest can be chemically analyzed using near edge spectroscopy (XANES). Formalin can alter the chemical form of many metals, but near edge spectroscopy can still yield clues about the chemical form (e.g. heme iron vs ferritin iron).

Distribution of Metals in the Normal Adult Brain

In the adult brain, regions associated with motor functions contain two to three times more iron than non motor related regions [2, 8, 31]. Gray matter structures generally contain more iron than white matter [10, 22]. Substantia nigra, globus pallidus, putamen, caudate nucleus, red nucleus, dentate nucleus, and locus coeruleus contain the highest iron concentrations in the adult brain explaining why in movement disorders these regions are vulnerable to effects of iron imbalance [5, 8, 9, 10, 11, 12, 22, 23, 24, 31, 32].

While iron is highest in the extrapyramidal regions, copper concentrations are highest in the substantia nigra, locus coeruleus, dentate nucleus, and cerebellum [8, 10, 22, 23, 33, 34]. The basal ganglia also have a high copper content, but the difference between their copper levels and those of other brain regions is not as striking as for iron [8, 10, 22, 23, 24, 33, 34].

The white matter of the adult brain contains more zinc than the gray matter [22, 23, 24] because myelin is rich in zinc which stabilizes myelin structure. The zinc content of myelin is about 50 μM in vivo[35, 36, 37]. While most iron‐ and copper‐rich brain structures are low in zinc [7, 10, 22, 23, 24] the hippocampus and amygdala are rich in zinc because these brain regions have numerous zincergic neurons [10, 38, 39].

Thus, the brain is a major metal repository where each brain region has a unique complement of metals that is vital for their normal functioning.

Therapeutic Strategies for Combating Metal‐Associated Neurodegeneration

Iron and copper transfer single electrons as they cycle between their reduced (Fe²⁺, Cu¹⁺) and oxidized (Fe³⁺, Cu²⁺) states and this redox cycling can catalyze the production of reactive oxygen species (ROS). Reduced cations promote the one electron reduction of O₂ with formation of superoxide (O₂ ^−·) and of hydrogen peroxide (H₂O₂) with formation of hydroxyl (OH^·), the most reactive of all endogenously generated free radicals. ROS have the ability to damage carbohydrates, lipids, proteins, and DNA [40, 41].

Moreover, even redox inert metals such as zinc, if in excess, can trigger ROS production. Zn²⁺ inhibits cellular respiration and facilitates the leakage of electrons from the mitochondrial electron transport chain by inhibiting complex I and III [42]. Zinc is also able to increase production of O₂ ^−· (via activation of protein kinase C (PKC) and increase activity of NADPH oxidase) and of nitric oxide (NO) (via induction of NO synthase) [42]. NO can react with O₂ ^−· to produce peroxynitrite (ONOO⁻), a free radical that can cause oxidation and nitration of proteins, lipids, and DNA [39, 42].

Cells have developed a system of antioxidant defenses aimed at eliminating ROS. Superoxide dismutases (SOD) are metalloproteins that accelerate the dismutation of O₂ ^−· to H₂O₂ and O₂. Mammalian Cu/Zn SOD is present in the cytosol and mitochondrial intermembrane space of all cells and Mn SOD serves the same function in the mitochondrial matrix. H₂O₂ is removed by catalase or by the selenoprotein glutathione peroxidase with the concomitant oxidation of glutathione [41]. Cellular antioxidant defense is further bolstered by melatonin, flavonoids, α tocopherol, β carotene, ascorbic acid, and ubiquinol [40, 41].

Since excessive amounts of metal‐catalyzed ROS can rapidly overcome these defense systems, cells have metal chaperones and storage proteins that keep metals in a nontoxic, but bioavailable form (Table 3). The cytosolic ferritin [43, 44], mitochondrial ferritin (MtFt) [45], and mitochondrial frataxin [46, 47] have ferroxidase centers that promote the oxidation of excess ferrous iron and its storage as ferrihydrite (a biomineral formed of ferric oxides/hydroxides octahedra) [48, 49, 50]. Ferritin, frataxin, and the iron–sulfur cluster‐containing aconitase protect mitochondrial and nuclear DNA, heme, and iron–sulfur cluster proteins against ROS damage [43, 46, 51, 52].

Table 3.

Important metal storage proteins.

Cell	Protein	Role
Neuronsa	Mostly H‐chain ferritin	Ferroxidase activity (cytosol)
		DNA protection from ROS (nucleus)
	Frataxinb2	Iron chaperone, ferroxidase activity and iron storage (mitochondria)
	Metallothionein (MT) III	Sequestration of copper and zinc
Neurogliaa,c	Mostly L‐chain ferritin (microglia)	Nucleation of iron and iron storage
	H and L‐chain ferritin (oligodendrocytes)	Ferroxidase activity (cytosol) and iron storage
		DNA protection from ROS (nucleus)
	Frataxinb	Iron chaperone, ferroxidase activity and iron storage (mitochondria)
	MT I and II (astrocytes, microglia in response to injury)	Sequestration of copper and zinc

^aIn the human mitochondrial ferritin (MtFt) is expressed at high levels in erythroblasts from patients with sideroblastic anemia and testis. In mouse, MtFt expression is also detected in pancreatic islets of Langerhans, brain, spinal cord and adipocytes [146, 147].

^bIn the CNS, frataxin mRNA expression is the highest in the spinal cord, with less expression in cerebellum and very little in cerebral cortex. Further study is needed to determine the level of frataxin expression in different cells types of the nervous tissue.

^cVery little ferritin expression is seen in astrocytes.

Intracellular copper is very tightly regulated through an elegant system of chaperones that carry and insert copper ions into specific copper metalloproteins [53]. There are no known chaperones for zinc, and free zinc actually contributes to normal brain function [38, 39]. The only protein thought to function as a zinc chaperone is metallothionein (MT), a cysteine‐rich cytoplasmic protein that can sequester both copper and zinc. Three MTs are expressed in the brain, protecting it against zinc and copper toxicity [39, 54].

Both redox‐inert metal ions such as zinc and redox‐active metal ions like iron and copper can associate with proteins and exacerbate or cause, the aggregation of Aβ‐amyloid, α‐synuclein, prion protein or ataxin‐3 [2, 39, 54, 55, 56, 57, 58] triggering neurodegeneration.

Iron [59, 60], copper [54], and zinc [61, 62] are essential for myelin synthesis, stability, and maintenance by oligodendrocytes. Many enzymes involved in the production of ATP and myelin precursors (cholesterol and fatty acids) are metal dependent (Table 1). Metal dyshomeostasis leads to impaired function and degradation of metalloenzymes causing a decrease in myelin production, improper brain maturation and associated impairments [41, 54, 58, 63, 64].

Metal chelators have been proposed as potential therapeutic agents to scavenge excess brain metals by the formation a nontoxic metal‐chelator complexes and their subsequent excretion in urine. Important requirements for effective metal chelators are their ability to cross the blood brain barrier and their selectivity not only for a specific metal, but for specific chemical forms and oxidation states of that metal [65].

Indiscriminate metal chelation can induce a generalized metal deficiency with chelator‐induced inhibition of metalloenzymes, metal redistribution, interference with neurotransmitter metabolism and cytotoxicity [2, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74]. Cuprizone (bis‐cyclohexanone oxaldihydrazone) is an example of a toxic metal chelator, used to induce a reversible demyelination in rodents that mimics multiple sclerosis in humans. It is not known why oligodendrocytes are more vulnerable than other cell types to cuprizone toxicity [75].

Clioquinol, a small hydrophobic metal chelator with affinity for iron, copper, and zinc [2, 39, 58, 65], was used as an antiparasitic until 1970 when it was withdrawn because of its association with subacute myelo‐optic neuropathy (SMON) [70], but the mechanism of its toxicity remains unknown. We have recently described a case of undefined progressive spinocerebellar ataxia characterized by degeneration of olivocerebellar fibers associated with decreased copper levels in the olivary region [24]. Although there was no history of exposure to chelators, the neuropathology of this case [24] is identical to the neuropathology reported in clioquinol intoxication [74]. There is also a striking clinical resemblance between copper deficiency and clioquinol intoxication [74] and clioquinol‐induced SMON [75, 76, 77, 78, 79]. All these findings now suggest that an abnormality in copper homeostasis is responsible for clioquinol‐induced SMON. The definitive confirmation of the role of copper in SMON is essential since clioquinol is proposed for use in patients with AD and PD [2, 39, 58, 65].

The rationale for the use of clioquinol in the treatment of Alzheimer's disease (AD) arose from studies showing that copper and zinc bound to amyloid β[80, 81, 82] and that clioquinol inhibited its accumulation on amyloid [83]. Fe³⁺ and Zn²⁺ can modulate the ability of α‐secretase to cleave the amyloid protein precursor (APP) [2, 39]. Both APP and Aβ can bind Fe³⁺ and Zn²⁺ in addition to Cu²⁺ and induce precipitation and aggregation of Aβ, while Aβ‐bound Fe²⁺ and Cu¹⁺ are primarily involved in the chronic oxidative damage that characterizes AD. Metals promote both the precipitation and deposition of amyloid‐β (Aβ) and oxidative stress which is associated with the neuritic plaques [2, 38, 39, 54, 55, 57, 58]. It is not known if the metal deposition is the cause or the consequence of the oxidative stress.

Clinical trials indicate that clioquinol slows cognitive decline in AD patients [84], but its mechanism of action is controversial [83]. The beneficial effects of clioquinol may relate more to its ability to redistribute metals than remove them. White et al. have shown that clioquinol, in both its metal‐free or complexed forms, increases the synthesis of metalloproteinases that degrade amyloid β secreted by APP transfected Chinese hamster ovary cells in culture [85]. This suggests that while some of the clioquinol‐bound copper is excreted in the urine, clioquinol can also transport copper into neurons, thereby relieving neuronal copper deficiency.

XRF microprobe has the micron to submicron resolution necessary to examine metal distribution at the cellular level and so has been applied to examine metals in AD [25, 28, 86, 87, 88, 89, 90, 91]. With respect to therapeutics, RS‐XRF and XRF microprobe simultaneously map multiple metals and should be able to distinguish between metal chelators that remove metals from the brain and those that redistribute metals within the brain. When combined with X‐ray absorption spectroscopy, subtle changes in metal chemistry can be detected [49].

Regions of the brain affected in AD, the inferior parietal cortex, motor cortex, olfactory cortex and cortex of the frontal pole [92, 93, 94, 95, 96, 97, 98], accumulate iron at a rate that may exceed the capacity of ferritin to sequester it [31]. XRF is a valuable tool to localize abnormal iron or zinc. As shown in Figure 1, the metal accumulation in the AD hippocampus does not rival the normally high iron and zinc content of the substantia nigra and white matter, respectively. This suggests that inappropriate metal sequestration rather than the absolute metal amount is partially responsible for the AD pathology.

Figure 1.

Metals in the brain of a subject with AD. A. Iron map; B. Copper map; C. Zinc map; D. Overlay of iron, copper and zinc; color scales represent the normalized total Kα fluorescence counts, proportional to the total metal present, from black (lowest) to color (highest); sn substantia nigra; Hp hippocampus; axial section; scale bar 10 mm.

While iron chelators are not currently advised to treat patients with AD [99], new chelators are under development and RS‐XRF would be an excellent tool to examine their selectivity in vivo.

The most common and consistent finding in Parkinson's disease (PD) is the accumulation of iron in substantia nigra and basal ganglia (Figure 2) [8, 22, 100, 101, 102, 103, 104, 105] and this has led many researchers to seek for a potential benefit of metal chelators in PD. Ferric iron deposits have been found in dopaminergic neurons, microglia, oligodendrocytes, astrocytes, and Lewy bodies in the substantia nigra of PD patients [106]. Copper is decreased [7, 8, 22] and zinc is increased [7, 8] in the PD substantia nigra, although at least one PD case characterized by decreased zinc has been reported [22]. The caudate, putamen, and globus pallidus have also been reported to accumulate metals in PD [7, 8, 22] (Figure 2).

Figure 2.

The caudate, putamen and globus pallidus of a subject with PD (B, D, F, H) and matched control (A, C, E, G); PD shows accumulation of iron, copper and zinc when compared to control. A, B. Iron maps; C, D. Copper maps; E, F. Zinc maps; G, H. Overlay of iron, copper and zinc; color scales represent the normalized total Kα fluorescence counts, proportional to the total metal present, from black (lowest) to color (highest); c caudate; p putamen; gpe globus pallidus pars externa; gpi globus pallidus pars interna; ic internal capsule; coronal section; scale bar 10 mm.

The mechanisms of the metal changes in PD are not well understood, but the iron accumulation in the substantia nigra may trigger oxidative stress [107], followed by a protective increase in the iron binding capacity of ferritin [108] and iron chelation by neuromelanin [109, 110, 111]. The neuromelanin‐iron complexes activate microglia to release cytokines and neurotoxic compounds inducing a vicious cycle of chronic inflammation and neuronal loss in the PD substantia nigra [112, 113]. Metallothionein expression also increases in the substantia nigra [114] either as a result of the release of pro‐inflammatory cytokines [115] or in order to accommodate the dyshomeostasis of copper and zinc and scavenge free radicals [114, 116]. However, metallothioneins may not always represent a protective factor since it has been proven that they can release zinc in toxic amounts in neurons [117].

Clioquinol and desferrioxamine have been shown to reduce the iron content of the substantia nigra of PD animal models, decrease the oxidative stress markers and glutathione depletion, and prevent the iron‐induced neurotoxicity [118, 119, 120, 121, 122]. All these studies suggest a potential benefit of metal chelators in PD, but their use has not been validated in clinical trials as yet. However, it is also possible that the iron build‐up seen in PD brains is simply the consequence of neuronal loss and replacement of neurons by cells with higher iron content [2]. To date, no data are available to prove whether iron accumulation in PD brains is an early event causing the degeneration of the dopaminergic neurons or a late event caused by the neurodegenerative process.

Using animal models, RS‐XRF has great potential to improve our understanding of global metal metabolism and the interrelationships between metals. Metal maps of brains from various animal models for neurodegeneration at various stages of the disease will likely demonstrate whether metal accumulation is the cause or the consequence of the neurodegeneration. Understanding the metal dyshomeostasis and its relation to the neuronal loss and disease progression would allow for the development of nontoxic and more selective metal chelators targeted to removing those metal species that are most reactive.

Friedreich's ataxia (FRDA) is caused by deficiency of a nuclear encoded mitochondrial targeted protein named frataxin [123, 124] which can function as both an iron chaperone for biosynthesis of heme and iron sulfur clusters and as an iron storage protein [45, 46, 47, 48, 125, 126]. FRDA patients have increased iron levels in the heart [127, 128, 129] and dentate nucleus [130, 131] and it has been suggested that the severe degeneration of the cardiomyocytes in the heart and neurons in the dentate nucleus is the result of oxidative damage caused by mitochondrial iron accumulation. Deferiprone has been shown to selectively reduce the iron accumulation in dentate nuclei of FRDA patients [130] most likely because of its ability to chelate both the cytosolic and mitochondrial labile iron and redistribute it between different cellular compartments and even to different cell populations [132].

Our laboratory has used XANES to speciate the chemical form of iron in mitochondria of fibroblasts from FRDA patients. We have shown that most of the mitochondrial iron is stored as the nontoxic, but also poorly bioavailable ferrihydrite of mitochondrial ferritin (MtFt) and little ferrous redox‐active iron is present [49]. These data support the new concept that once MtFt is upregulated in FRDA (by whatever means) it competes for and sequesters the available mitochondrial iron. Since MtFt mineralizes this iron, but does not readily release it, our data support the hypothesis that the lack of iron available for heme and iron–sulfur cluster synthesis is more important than oxidative stress in FRDA pathophysiology. Like FRDA, most of the excess iron in other neurodegenerative diseases is also stored as relatively nontoxic ferrihydrite [47, 48] in ferritin [2, 48, 101] and chelation strategies designed for these diseases should take into account that at least in cells that are able to upregulate their ferritin and MtFt production, chelating the small nonferrihydrite iron pool could prevent the cells from carrying out even minimal functions such as heme synthesis.

Wilson's disease (WD) is one of the few neurodegenerative diseases where the success of the chelation therapy is proven over many years of clinical use. In WD, mutations in the gene encoding the copper transporter Atp7b result in toxic copper build‐up in hepatocytes [133]. Copper accumulation leads to copper‐mediated oxidative damage of the liver cells, hepatic necrosis, release of large amounts of free copper in the bloodstream and copper overload in all tissues, including basal ganglia [134–138].

D‐penicillamine was the first drug to be successfully employed in WD. D‐penicillamine not only chelates the tissue copper and promotes its excretion in the urine, but also detoxifies it by promoting the synthesis of metallothionein, which forms a nontoxic combination with copper [65, 139, 140]. Trientine is a less toxic alternative to D‐penicillamine also able to chelate the tissue copper and increase its urinary excretion [65, 139, 140]. Other chelating agents that can be used are ammonium tetrathiomolybdate that complexes with copper in the intestine and blood [65, 139, 141]. The advantage of tetrathiomolybdate is its specificity for copper and not other metals. Zinc sulfate has been used to prevent the absorption of copper from the intestine by inducing the expression of copper‐binding metallothioneins in enterocytes [65, 139, 142, 143]. While the benefit of these copper chelators in removing the toxic deposits of copper is beyond any doubt, the long‐term effects of copper depletion on copper‐rich brains areas, such as the olivary nucleus, have not been explored.

Excellent reviews on therapeutic chelators have been recently published [2, 65, 73, 139, 144], but many questions remain to be answered. How specific is a “specific” chelator? How does the chelation of a metal affect the distribution of other metals? Are metals chelated only in brain regions where they accumulate? RS‐XRF is an ideal, inexpensive and relatively straightforward way to compare the amount and distribution of brain metals in groups of animals (or even pathological human brain specimens) with and without chelation therapy. When combined with chemical speciation using XANES [145], RS‐XRF has the potential to become the tool of choice in mapping and speciating brain metals in situ.

Conclusion

The brain is a major metal repository [8, 9, 10, 11, 32], where large amounts of metals coexist and colocalize. The high metal content of the CNS makes it particularly susceptible to metal‐catalyzed oxidative damage, protein aggregation, neurotoxicity, and neurodegeneration [2, 53, 54].

RS‐XRF is a new and powerful tool that makes systematic studies of brain metals practical. It can be applied to any human or animal tissue and we foresee important applications in testing whether a metal chelator designed to remove a particular metal disrupts the levels of other metals. Using animal models, RS‐XRF also has great potential to improve our understanding of metal metabolism and the interrelationships between metals in the brain. With rapidly advancing beamline technology at third‐generation synchrotrons, it should be possible to map very large sections in a matter of minutes, making systematic RS‐XRF studies on serial slices of whole human brain practical.

Conflict of Interest

The authors have no conflict of interest.