1a. Mn biology and tissue homeostasis

Although copper and magnesium can substitute for Mn as a cofactor for some enzymes, a subset of enzymes with roles in neuron and/or glial function only in the presence of Mn. These discrete Mn-binding proteins (manganoproteins) include glutamine synthetase, superoxide dismutase 2 (SOD2), arginase, pyruvate decarboxylase, and serine/threonine phosphatase [10, 17, 18].

Glutamine synthetase (GS), the most abundant manganoprotein, is predominantly expressed in astrocytes, where it converts glutamate to glutamine. Because GS contains four Mn ions per octamer [19], Mn has been proposed to regulate GS activity. In fact, insufficient manganese has been proposed to increase glutamate trafficking, glutamatergic signaling, and excitotoxicity [20]. Furthermore, it has been proposed that the increased susceptibility to seizures observed in individuals with Mn deficiency may be due in part to diminished GS levels and/or activity [21]. SOD2 is a mitochondrial enzyme that detoxifies superoxide anions through the formation of hydrogen peroxide (H₂O₂). Although the concentration of Mn within neurons is low (<10⁻⁵ M), their high mitochondrial energy demands is correlated with a propensity of increased SOD2 in neurons compared to glia [9, 14]. Furthermore, loss of SOD2 activity increases the susceptibility to mitochondrial inhibitor induced toxicity and causes oxidative stress [22].

Arginase regulates elimination of ammonia from the body by converting L-arginine, synthesized from ammonia, to L-ornithine and urea as part of the urea cycle. Moreover, in the brain, L-arginine is converted to nitric oxide by neuronal nitric oxide synthetase. Proper regulation of arginase promotes neuronal survival by impairing nitric oxide signaling [23, 24]. Pyruvate carboxylase is an essential enzyme required for glucose metabolism that interacts with Mn to generate oxaloacetate, a precursor of the tricarboxylic acid (TCA) cycle [25]. Interestingly, in the brain, pyruvate carboxylase is predominantly expressed in astrocytes [26, 27]. Protein phosphatase 1 is essential for glycogen metabolism, cell progression, regulation synthesis and release of neurotrophins, which promote neuronal survival, and synaptic membrane receptors and channels [28].

Intricate regulation of Mn absorption and tissue specific accumulation is crucial for it to properly regulate these enzymes, so understanding Mn’s essential roles and toxicity in the brain requires knowledge of its regulation in the periphery. Three major factors have been postulated to modulate plasma Mn levels. First, given that the main source of Mn is diet, tight regulation of gastrointestinal absorption of Mn is crucial. Second, following Mn absorption and a concomitant increase in plasma Mn levels, transport of Mn to target organs, including the liver, is necessary to prevent Mn-induced toxicity in the periphery. Finally, although the liver detoxifies substances, Mn must be further eliminated from the plasma via shuttling to bile [14]. The absorption of Mn by the gastrointestinal tract is highly dependent on the quantity of ingested Mn and net accumulated levels in the plasma. In vivo experiments in mice and rats have defined the range (1–3.5%) of GI absorption of Mn [29, 30]. While Mn is transported by simple diffusion in the large intestine, Mn is absorbed by active transport in the small intestine [14]. Mn excretion into bile is likely active as well because it depends on concentration gradients [31].

A plethora of plasma proteins or ligands have been implicated as specific Mn carrier proteins, including transglutaminase, beta₁-globulin, album, and transferrin [32, 33]. In fact, approximately 80% of plasma Mn is bound to beta₁-globulin [32]. Despite the demonstration that Mn preferentially binds to albumin in the plasma of both rabbits and humans, emerging evidence has provided evidence for weaker binding of Mn to albumin compared to Cd and Zn [34, 35].

Intracellular Mn²⁺ is sequestered in the mitochondria of the brain and liver via the Ca²⁺ uniporter [36, 37]. Mitochondria are the primary pool of Mn in the cell, however, nuclei have also been implied (remains debatable) to preferentially accumulate this metal [26, 38, 39]. The efflux of mitochondrial Mn²⁺ is mediated preferentially through an active but slow Na⁺-independent mechanism; a Na⁺-dependent mechanism also contributes minimally (reviewed in [10]). This slow Mn efflux has been suggested to account for the net accumulation of Mn in mitochondria. Nevertheless, the cytoplasmic Fe²⁺ exporter, ferroportin-1, has been reported to transport Mn. Interestingly, ferroportin-1 surface localization and protein expression is perturbed following Mn exposure [40].

1b. Mn transporters in the brain

Mn can cross the blood-brain barrier (BBB) and blood-cerebrospinal fluid barrier (BCB) through several carriers and in different oxidation states. Given the essential physiological functions of Mn and the neurotoxicity associated with Mn overload, Mn absorption, transport, and tissue levels are stringently regulated. Under normal physiological conditions, Mn is efficiently transported across the blood-brain barrier (BBB) in both the developing fetus and adults [14, 41]. Although Mn transport across the BBB has been actively investigated, which transporter systems are primarily responsible have not been conclusively determined. However, over the past three decades, several Mn transporter systems have been characterized, including active and facilitated diffusion modes of Mn transport [42, 43, 44]. Emerging reports have indicated that Mn can be transported via the divalent metal transporter 1 (DMT1), the transferrin (Tf) receptor (TfR) that mediates trivalent Fe uptake, the divalent metal/bicarbonate ion symporters ZIP8 and ZIP14, various calcium channels, the solute carrier-39 (SLC39) family of zinc transporters, park9/ATP13A2, the magnesium transporter hip14 and the transient receptor potential melastatin 7 (TRPM7) channels/transporters. While the tissue-specific expression of each of the aforementioned Mn transporters is yet to be determined, it is likely that optimal tissue Mn levels are maintained through the involvement of all the above and other unknown Mn transporters. In addition, other cellular processes may regulate the activity of the above transporters in response to Mn deficiency or overload. Of all the above listed polyvalent transporters, DMT1 and TfR are the most extensively documented [44]. Interestingly, a small fraction of plasma Mn is bound to transferrin (Tf), an iron-binding protein, while approximately 80% of plasma Mn is associated with albumin and beta₁-globulin[32].

The divalent metal transporter 1 (DMT1) is a member of the family of natural resistance-associated macrophage proteins (NRAMP) and crucial for the maintenance of essential metal homeostasis in the brain [45, 46, 47]. It is best known for its ability to regulate Fe homeostasis in the gastrointestinal lumen [47]. DMT1 is also known as the divalent cation transporter (DCT1) due to its ability to competently transport divalent metals including Zn²⁺, Mn²⁺, Co²⁺, Cd²⁺, Cu²⁺, Ni²⁺, Pb²⁺, and Fe²⁺ across the plasma membrane into the cytosol [45, 48]. Both alternative splice isoforms of DMT1 are predominantly localized at the plasma membrane; however, one also localizes to late endosomes and lysosomes while the other localizes to early endosomes [49]. Because only one isoform contains an iron-response element (IRE), subcellular localization depends on Fe concentration [10, 46].. The relative high affinity of DMT1 for Mn has been well established both in vivo and in vitro. Specifically, mutations in the DMT1 gene of Belgrade rats and microcytic anemia mice result in significant decreases in Mn and Fe tissue levels.[50, 51, 52] Furthermore, magnetic resonance imaging (MRI) in a recent study demonstrated the consistency of the cross-BBB transport mechanism(s) of Mn and Fe, suggesting that they are shared [53]. Finally, DMT1-mediated metal transport across rat brain endothelial cells in culture has been reported to be pH, temperature, and Fe-dependent [54, 55].

The TfR is the major cellular receptor for Tf-bound Fe, but because Tf can also bind trivalent Mn, TfR can also mediate Mn transport. Once Mn³⁺ is internalized through the endocytic pathway, it is reduced to Mn²⁺ and transported through DMT1 to the cytosol. Mn binding to Tf is time-dependent and Tf receptors are also present on the surface of cerebral capillaries [44, 56]. Additionally, the TfR is an active transporter that is pH and Fe-dependent [56]. Both in vivo and in vitro studies have reported that Mn is efficiently transported via the TfR. For example, a spontaneous mutation in a murine gene linked to the TfR, referred to as hypotransferrinemic, results in a drastic serum TfR deficiency, impairs Mn transport, and disrupts Fe deposition [57, 58]. Interestingly, autioradiography reports have indicated that the TfR is generally localized in gray matter, but not the Fe-abundant white matter tracts in humans and rodents [59, 60, 61].

The zinc- interacting protein 8 (ZIP8) and 14 (ZIP14) proteins are divalent metal/bicarbonate ion symporters known to transport Mn, Zn, and Cd under normal conditions [62, 63]. ZIP8 and ZIP14 are members of the SLC39 family of genes [63, 64] that are glycosylated and expressed on the apical surface of brain capillaries. Mn uptake through ZIP8 or Zip14 is driven by extracellular carbonate (HCO₃⁻). The expression of Zip8 and Zip14 in the brain is lower than that in the liver, duodenum, and testis [65]. In addition, voltage-gated Ca²⁺ channels, including L and P-type channels [66], such as ligand-gated Ca²⁺ channels; store-operated Ca²⁺ channels (SOCCs) [67], and the ionotropic glutamate receptor Ca²⁺ channels [68] have been implicated in Mn transport across the BBB.

Over the past decade, there has been increasing interest in the exploration and identification of novel candidate Mn transporters, including Hip14 and the product of the park9 gene. Emerging experimental data has indicated that huntingtin-interacting protein 14 and 14L (Hip14, Hip14L) mediates transport of Mn²⁺ and other divalent metals (Mg²⁺, Sr²⁺, Ni²⁺, Ca²⁺, Ba²⁺, Zn²⁺) across cell membranes [69, 70]. Hip14 is the mammalian ortholog of the ankyrin repeat protein 1 (Akr1p) that is primarily expressed in neurons of the brain. Hip14 is involved in the palmitoylation of several neuronal proteins including huntingtin (HTT) [49]. In addition, it is required for endo- and exocytosis, as well as targeting of cysteine string protein (CSP) and synaptosomal-associated protein 25 (SNAP25) to the synapse [71, 72]. Hip14 is predominantly expressed in the presynaptic terminal, Golgi apparatus, and vesicular structures localized in the axon, dendrites, and soma of neurons [73]. Biochemical studies including yeast-two-hybrid screens have demonstrated that the interaction between Hip14 and HTT is inversely correlated with the poly-Q length in the HTT protein [72].

Interestingly, Gitler and colleagues have recently reported that the park9 gene responsible for early-onset Parkinsonism also transports Mn [70]. The park9 gene encodes a putative P-type transmembrane ATPase (ATP13A2) protein. Although the exact function of park9 is unknown, it is generally thought to be a shuttle for cations, including Mn, across the cell. Biochemical studies have demonstrated that the highest and lowest park9 mRNA levels are localized within the substantia nigra and cerebellum, respectively [74].

Although Mn inhibits the choline transporter at the BBB, it has been suggested that the choline transporter may mediate Mn transport during periods of high throughput. In addition, the choline transporter has a higher affinity for Mn compared to the other metals (Cd²⁺ and Al³⁺) that it transports [75, 76, 77]. Mn transport across the choline transporter is sodium-independent, carrier-mediated, and saturable [56].

TRPM7 is ubiquitously expressed in vertebrates and functions as an active Ca²⁺ selective transporter and a serine/threonine protein kinase. Furthermore, the kinase activity is important for its metal transport function. Specifically, the transporter operates by regulating intracellular Ca²⁺ levels and Mg²⁺ homeostasis through the creation of an inward current, thus contributing to the establishment of a cellular membrane potential. The relative permeability of cations through TRPM7 has been reported to be as follows: Zn²⁺, Ni²⁺>Ba²⁺, Co²⁺> Mg²⁺> Mn²⁺> Sr²⁺> Cd²⁺> Ca²⁺. Physiological levels of Mg²⁺ and Ca²⁺ are necessary for maintaining the permeability of TRPM7 to Mn²⁺, Co²⁺, and Ni²⁺ [56].

The homomeric purinoreceptors, including P2X and P2Y, have been suggested to participate in Mn transport. These receptors are ATP-dependent and ubiquitously expressed on endothelial cells [78, 79, 80]. Purinoreceptors have a relatively lower affinity for Mn than for the other divalent metals they transport (Ca> Mg> Ba> Mn) [56]. Finally, the citrate transporter has also been implicated in Mn transport across the BBB [81].

3b. Human exposure to Mn and relationship to PD

A recent editorial [116] proposed that radiotracer imaging techniques should be used to investigate the integrity of the dopaminergic system in asymptomatic workers exposed to Mn and highlights the need to assess motor signs in addition to cognitive and behavioral symptoms. PET with [18F]FDOPA [117] in a small study with very high average blood Mn in workers and gender imbalance among groups showed that active and asymptomatic welders have presynaptic nigrostriatal dopaminergic dysfunction with different anatomical localization from that generally observed in PD, affecting the caudate more than the putamen. The welders also had significantly lower Unified Parkinson's Disease Rating Scalesubsection 3 scores than those of the control group, indicating that their occupation led to motor impairment.

3c. α-synuclein and Mn-related protein aggregation

Mn, and certain other essential and toxic metals, can directly increase fibril formation by α-synuclein. Though α-synuclein‘s function is still unclear, these fibrils form the intra-cytoplasmic inclusions (Lewy bodies and Lewy neurites) found in idiopathic Parkinson’s disease, dementia with Lewy bodies (DLB) and multiple system atrophy (MSA), disorders classified as synucleinopathies. It is known that both genetic and environmental factors affect synuclein pathology (reviewed by Eller and Williams, 2011) [40]. Thus, Mn seems to act jointly with α-synuclein in inducing neuronal cell death [119]. It has also been suggested that some metals, including Mn, can act synergistically, even at low concentrations, with certain herbicides to promote α-synuclein misfolding and aggregation [120].

Mn also increases the expression of α-synuclein in vitro [121, 122], and chronic Mn exposure leads to aggregation of α-synuclein in vivo in neurons and glial cells of - non-human primates [123]. A genetic interaction between α-synuclein and PARK9 has been reported in yeast and because PARK9, which could encode a metal cation transporter, seems to protect cells from Mn toxicity, this could provide a mechanism linking genetic and environmental causes of neurodegeneration [70]. Various mechanisms mediated by Mn could converge on α-synuclein in vivo, potentially linking Mn to Parkinson’s disease [124]. Overexpression of α-synuclein in human cells seems to facilitate Mn-induced neurotoxicity through activation of the transcription factor NF-κB, the kinase p38 MAPK, and apoptotic signalling cascades, thus possibly playing a role in dopaminergic cell death [125]. It has also been recently suggested that chronic exposure to Mn could decrease striatal dopamine turnover in transgenic mice expressing human α--synuclein [126].

4a. A role for environmental factors in HD

Over a decade after the identification of the causative HD mutation, there have been conflicting reports linking complete or incomplete penetrance of HD to triplet repeat expansion length. Environmental factors have also been suggested to contribute to the residual variation in age of onset, perhaps even more so than genetic factors [140, 141]. Moreover, Gómez-Esteban and others have implicated environmental influences that modify age of disease onset and clinical presentation in monozygotic twin studies, as twins have the same number of repeats [142, 143, 144, 145]. Such twin studies, however, fail to reveal the nature of the environmental factors involved. Furthermore, animal models of HD have provided further support for the influence of environmental factors on HD onset and progression [141, 146].

4b. Links between HTT function and metals

There are several known functional links between the HTT gene and metals. For example, wildtype HTT is important for iron metabolism and oxidative energy production, as evidenced by the decreased hemoglobin and altered iron endocytosis in Htt deficient zebrafish [147]. In fact, Fe and Cu levels are elevated in the corpus striatum of postmortem brain tissue from HD patients and animal models [148, 149]. In addition, HD postmortem brains exhibit alterations in Mn-dependent enzyme activity [3]. Furthermore, animal models have also demonstrated an increase in microglial ferritin, an intracellular iron storage protein [150]. Interestingly, Fox and colleagues have reported that wildtype Htt protein interacts with Cu and decreases its solubility [151]. Finally, the formation of inclusion bodies by CAG expansion in mutant Htt protein fragments may be associated with iron-dependent oxidative events [152]. All these studies strongly suggest that wildtype Htt is necessary for proper metal homoeostasis in the brain.

4c. A role for altered metal homeostasis and toxicity in HD neuropathology

The clinical progression of HD is associated with elevated Fe and Cu in the corpus striatum [148, 149]. Several Mn-dependent enzymes, including arginase, glutamine synthetase, pyruvate decarboxylase, and Mn superoxide dismutase 2 (SOD2) are altered in human HD postmortem brains and toxicant models of HD [3, 22, 153, 154, 155, 156]. Also, data from animal models of HD have demonstrated a significant increase in microglial ferritin (an intracellular iron storage protein) levels [150]. Interestingly, Fox et al. have recently reported that the Htt protein interacts with Cu and decreases the solubility of wildtype Htt protein [151]. However, the cellular effects of Cu or other metal ions on Htt function, proteolytic processing to generate N-terminal fragments, aggregation of fragments, and formation of mutant Htt inclusion bodies remain unknown. Finally, emerging evidence indicates that inclusion bodies formed by CAG expansion in mutant Htt protein fragments are associated with iron-dependent oxidative events, opening the possibility that other redox-reactive metal ions, such as Mn, may influence polyglutamine aggregation [152]. In essence, several studies have suggested that oxidative stress, mitochondrial dysfunction, excitotoxicity, and alterations in iron homeostasis are crucial steps in both Mn neurotoxicity and HD neuropathology. Importantly, chronic exposure to Mn in animal models leads to its significant accumulation in the striatum, providing an opportunity for Mn and HTT to interact within the neurons most vulnerable to HD pathology. Given the strong association between protein aggregation, metal ions and neurodegeneration, it is highly rational to speculate that metals may modulate HD pathophysiology.

This review was partially supported by grants from the NIH/NIEHS, RO1ES016931 (ABB) and RO1ES10563 (MA).