The physiological and pathophysiological roles of copper in the nervous system

Abstract

Copper is a critical trace element in biological systems due the vast number of essential enzymes that require the metal as a cofactor, including cytochrome c oxidase, superoxide dismutase and dopamine-β-hydroxylase. Due its key role in oxidative metabolism, antioxidant defence and neurotransmitter synthesis, copper is particularly important for neuronal development and proper neuronal function. Moreover, increasing evidence suggests that copper also serves important functions in synaptic and network activity, the regulation of circadian rhythms, and arousal. However, it is important to note that because of copper’s ability to redox cycle and generate reactive species, cellular levels of the metal must be tightly regulated to meet cellular needs while avoiding copper-induced oxidative stress. Therefore, it is essential that the intricate system of copper transporters, exporters, copper chaperones and copper trafficking proteins function properly and in coordinate fashion. Indeed, disorders of copper metabolism such as Menkes disease and Wilson disease, as well as diseases linked to dysfunction of copper-requiring enzymes, such as SOD1-linked amyotrophic lateral sclerosis, demonstrate the dramatic neurological consequences of altered copper homeostasis. In this review, we explore the physiological importance of copper in the nervous system as well as pathologies related to improper copper handling.

1 |. INTRODUCTION

Copper is an essential micronutrient and redox active metal that cycles between its cupric [Cu (II)] and cuprous [Cu(I)] state. In biological systems, its ability to easily accept and donate one electron allows copper to participate in redox reactions, making it useful as a cofactor for numerous enzymes (Kaim & Rall, 1996). Indeed, copper can be found at the active site of many critical enzymes, including those essential for neurological function, such as the predominantly cytosolically localized antioxidant Cu, Zn superoxide dismutase (SOD1), the mitochondrial electron transport chain protein cytochrome c oxidase, and the secretory pathway enzymes dopamine β-hydroxylase (DBH) and peptidylglycine α-amidating monooxygenase (PAM) (Scheiber et al., 2014). Copper is also indispensable for the growth and development of eukaryotes. In mammals, the importance of copper can be appreciated by disorders of copper metabolism that lead to hypocupremia as well as those leading to copper deficiency. The consequences of hypocupremia can include anaemia, neutropenia, cardiac hypertrophy, defects in growth and development as well as neurological symptoms such as seizures, myeloneuropathy, and brain atrophy (Lazarchick, 2012; Lee et al., 2001; Nose et al., 2006; Prasad et al., 2011; Zlatic et al., 2015).

Dietary copper is absorbed through intestinal enterocytes expressing the high affinity copper transporter CTR1 on their apical membrane (Nose et al., 2010). CTR1 has been shown not only to transport copper, but to help maintain it in its reduced state via two N-terminal His-Met-Asp clusters (Kar et al., 2022). Of note, Cu (II) is reduced to Cu(I), the dominant form of copper in the cytosol, by 6-transmembrane epithelial antigen of the prostate (STEAP) proteins. STEAPs are a family of reductases, with Steap3 initially identified as an endosomal ferrireductase (Ohgami et al., 2005). A later study showed this variant, as well other members of the family, namely, Steap2, 3 and 4, are also effectively at reducing copper (Ohgami et al., 2006). Mice with intestinal epithelial cell specific knockout of Ctr1 show profound early postnatal defects in growth, hypopigmentation, skin laxity, and brittle whiskers, along with decreased activity of cuproenzymes (Nose et al., 2006)—a phenotype reminiscent of patients with Menkes disease, a genetic disease that results in hypocupremia (Ramani & Parayil Sankaran, 2023). Copper accumulation in peripheral tissues in mice with enterocyte specific loss of Ctr1 was also significantly reduced, underscoring the importance of intestinal absorption of the metal for proper distribution throughout the body (Nose et al., 2006). Indeed, once copper is absorbed by enterocytes, the ATPAse copper-transporter ATP7A translocates from the trans-Golgi network (TGN) to the basolateral membrane of the cell where it effluxes copper into the circulation (Lutsenko, Barnes, et al., 2007; Kaler, 2011). Through the circulation, copper, carried by serum proteins, is delivered primarily to the liver—the main organ for the maintenance of copper homeostasis in the body. Copper is then stored in hepatocytes, secreted in bile or incorporated into the ceruloplasmin in ATP7B-dependent manner, which contributes to the majority of copper in the plasma (Lutsenko, Barnes, et al., 2007; Lutsenko, 2021). Copper is transported to the brain by epithelial cells of the choroid plexus, which also express ATP7A (Kaler, 2011; Niciu et al., 2006). ATP7A and ATP7B are P₁-type ATPases that use energy from ATP hydrolysis to regulate intracellular copper levels by exporting copper across cell membranes. Additionally these copper transporters play a key role in incorporating copper into copper-dependent enzymes that are secreted or targeted to the plasma membrane (Linz & Lutsenko, 2007; Lutsenko, Barnes, et al., 2007). The critical importance of these copper transporters is demonstrated by the significant consequences of their dysfunction. Mutations in ATP7A prevent the transport of copper from intestinal cells into the circulation, resulting in profound hypocupremia, and are associated with the fatal, infantile-onset neurodegenerative condition Menkes disease (Kaler, 2011). ATP7B is a crucial contributor to the copper homeostatic function of the liver by facilitating the excretion of excess copper into bile by hepatocytes. Its dysfunction is linked to Wilson disease, a condition characterized by liver disease as well as neurological and psychiatric manifestations secondary to copper accumulation in the liver and brain (Bandmann et al., 2015; Członkowska et al., 2018; Kaler, 2013).

Of note, the localization of both CTR1 and ATP7A are regulated by copper levels. When extracellular levels of copper are increased, such as under conditions of high dietary copper intake, CTR1 is rapidly internalized via clathrin-mediated endocytosis and the copper transporter does not return to the plasma membrane until extracellular copper levels are decreased (Clifford et al., 2016; Molloy & Kaplan, 2009). Endocytosis of CTR1 may be facilitated by a conserved methionine-rich cluster in the transmembrane domain of the protein, allowing it to act as both a copper sensor and transporter (Guo et al., 2004). As noted previously, ATP7A’s localization from the TGN to the membrane of cells is stimulated by increased intracellular copper (Petris et al., 1996). This copper-dependent localization of CTR1 and ATP7A allows for the regulation of intracellular copper levels. This tight regulation of cellular copper levels and copper transport is highlighted by the fact that the majority of cytoplasmic copper is buffered by thiol-containing compounds, which include proteins, reduced glutathione (GSH) and cysteine, to keep the concentration of labile or ‘free’ copper in the cytoplasm in the attomolar range (Morgan et al., 2019). Indeed, given the low levels of labile copper in the cytoplasm, cuproenzymes require copper chaperones to transport copper from the plasma membrane to their intracellular locations. For example, SOD1 is metalated by the copper chaperone for superoxide dismutase (CCS) (Culotta et al., 1997; Rae et al., 1999; Wong et al., 2000), and the interaction between these two proteins is critical for both the maturation and activation of SOD1 (Banci et al., 2012; Boyd et al., 2019; Fetherolf et al., 2017). CCS has been shown to interact with CTR1 at the plasma membrane, suggesting it acquires copper directly from this copper transporter (Pope et al., 2013; Skopp et al., 2019). Similarly, ATOX1 is a copper chaperone that transfers copper and regulates the activity of the secretory pathway proteins ATP7A and ATP7B (Walker et al., 2002; Hamza et al., 2003; Strausak et al., 2003; Walker et al., 2004; Yu, Yang, et al., 2017) Like CCS, ATOX1 likely acquires copper through a direct interaction with CTR1 (Kahra et al., 2016). Although evidence suggests that, unlike the cytoplasm, the mitochondria contains a pool of labile copper (Cobine et al., 2004, 2021; Yang et al., 2005; Vest et al., 2013), mitochondrial cuproenzymes are still metalated by chaperones with the copper chaperones COX17, SCO1 and SCO2 all involved in the metalation of cytochrome c oxidase’s two copper-containing catalytic sites (Banci et al., 2008; Capaldi, 1990; Fontanesi et al., 2006; Leary et al., 2004). As the affinity of SOD1, ATP7A and cytochrome c oxidase for copper are all greater than their respective chaperones, affinity gradients facilitate the transfer of copper ions from copper chaperones to their cognate proteins (Banci et al., 2010). Of note, copper is transported across the mitochondrial inner membrane by the phosphate carrier SLC25A3 (PiC2), providing the pool necessary of this metal for both matrix storage as well as for travel back to the inner membrane for metalation reactions (Boulet et al., 2018; McCann, Quinteros, et al., 2022). Indeed, deletion of SLC25A3 results in pronounced cytochrome c oxidase defects and copper dyshomeostasis in mammalian cells (Boulet et al., 2018) (Figure 1).

FIGURE 1.

Schematic of copper handling in mammalian cells. Copper is reduced from Cu (II) to Cu(I) by Steap proteins before being transported into cells by the high affinity copper transporter CTR1. The copper chaperone CCS transports copper to the primarily cytosolically localized cuproenzyme, Cu/Zn superoxide dismutase 1 (SOD1), while the copper chaperone COX17 binds copper and delivers it to the mitochondria. The carrier protein SLC25A3 transports copper across the inner mitochondrial membrane. Here, copper is subsequently transferred to SCO1 before being delivered to cytochrome c oxidase (CCO), the final electron acceptor in the electron transport chain. SCO2 acts as a copper thiol oxidoreductase for SCO1, altering the ratio of oxidized/reduced cysteines in the protein depending on mitochondrial copper levels. ATOX1, a copper chaperone for the secretory compartment, transports copper from the plasma membrane to the trans-Golgi network (TGN), where it transfers copper to ATP7A and ATP7B, allowing these ATPAse copper transporters to metalate secretory proteins as well as translocate to the plasma membrane to mediate copper export, thereby regulating cytosolic levels of copper. Metallothioneins (MTs) and glutathione (GSH) bind copper in order keep labile levels of the metal necessarily low and avoid copper-mediated toxicity.

Given the tight regulation of cellular copper levels, the importance of copper in antioxidant defence, neurotransmitter synthesis, and energy metabolism, and the severe neurological consequences of systemic copper deficiency and overload in animal models (Hadrian & Przybyłkowski, 2021; Hwang et al., 2014; La Fontaine et al., 1999; Wang et al., 2012), as well as the growing number of human disorders with neurological phenotypes associated with defects in copper transporting proteins (Batzios et al., 2022; Członkowska et al., 2018; Jaksch et al., 2000; Kaler, 2011, 2013; Valnot et al., 2000), it is clear that copper homeostasis is critical for proper neuronal function. Moreover, copper dyshomeostasis can promptly lead to neurological disease.

2 |. PHYSIOLOGICAL ROLES OF COPPER IN THE NERVOUS SYSTEM

The brain utilizes 20% of the oxygen provided by the cardiac output and neurons are notably dependent of mitochondrial respiration for proper functioning. As such, it is not altogether surprising that nerve cells are heavily reliant on copper due to the critical role of the metal as a catalytic cofactor in cytochrome c oxidase (Complex IV), the final electron acceptor in the electron transport chain. Indeed, this critical dependence on oxidative phosphorylation make neurons particularly vulnerable in disorders associated with copper deficiencies, including the inability of mitochondria to transport copper. However, copper also serves an important role in neuronal differentiation, acts as a neuromodulator, modifies neuronal circuit excitability and is involved in neurotransmitter and neuropeptide synthesis. This section will explore the many physiological roles of copper in normal neuronal function. A later section will explore the neuropathological consequences associated with copper handling defects.

3 |. COPPER ROLES IN NEURONAL DIFFERENTIATION AND SYNAPTOGENESIS

Neurons, in addition to being highly metabolically active, carry out specific functions unique to this cell type, such as neurotransmitter and neuropeptide synthesis, as well as synaptic signalling. Given that all these functions require copper and/or cuproenzymes, it follows that these post-mitotic cells have an increased demand for the metal. Indeed, recent evidence suggests that neuronal differentiation is associated with increased copper dependence. In vitro data from differentiated PC12 cells and SH-SY5Y cells, and an in vivo study of the embryonic chick spinal cord, are all concordant in the finding that copper levels are higher in differentiated neurons as compared to non-differentiated cells (Birkaya & Aletta, 2005; Hatori et al., 2016). This increased cellular demand for copper appears to be compartment specific, as the expression of proteins localized to the secretory pathway, such as ATP7A, DBH, and PAM, is significantly increased in differentiated neurons. The expression of mitochondrial proteins such as the copper chaperone Cox17 and subunit 1 of cytochrome c oxidase are also moderately increased in differentiated neurons. In contrast, the expression of cytosolic copper-containing proteins is not different between undifferentiated and differentiated neurons, even when such proteins are directly involved in metalation (e.g., SOD1’s copper chaperone CCS) (Hatori et al., 2016). These findings are supported by observed increased copper levels in the microsomal fractions, containing secretory vesicles and the TGN, in differentiated neurons compared to levels in non-differentiated cells. The increased copper flow to the secretory pathway as well as increased expression of secretory pathway proteins in differentiated neurons suggests that Atox1, which transfers copper to ATP7A, should be more active. Intriguingly, previous research demonstrated a functional partnership between the cell’s glutathione/glutathione disulfide (GSH/GSSG) balance and Atox1’s redox state where GSH is able to reduce Atox1 and restore its ability to bind copper (Hatori et al., 2012). In both developing chick spinal cord and differentiated SH-SY5Y cells, both glutathione and Atox1 are markedly chemically reduced compared to non-differentiated neurons (Hatori et al., 2016), consistent with a model in which redox-regulation of glutathione allows Atox1 to bind copper and transfer copper to ATP7A, leading to the maturation of cuproenzymes such as PAM and DBH. Moreover, oxidizing glutathione leads to increased DBH retention suggesting that copper flow to the secretory may be regulated by the redox state of the cell (Hatori et al., 2016).

In addition to the requirement for increased copper flow to the secretory pathway for neuronal differentiation, increased intracellular copper and the proper trafficking of copper transporters have been found to be essential for neurite extension and synaptogenesis. As previously noted, Birkaya and Aletta (2005) found that differentiation of PC12 cells with nerve growth factor increased cellular copper content within three days of treatment, with intracellular copper levels rising to 14-fold when compared to control cells after 2 weeks. This increase in copper is essential for neurite outgrowth as chelating copper with tetraethylenepentamine not only diminishes intracellular copper levels, but results in a marked decrease in neurite branching and length (Birkaya & Aletta, 2005).

The copper ATPase ATP7A is an important regulator of axon extension and synaptogenesis. Indeed, multiple investigations have found that ATP7A is expressed in a developmental rather than constitutive manner in numerous neuronal subpopulations, with expression peaking early in neuronal development, prior to synaptogenesis (El Meskini et al., 2005, 2007; Niciu et al., 2006). Recent work has established that the heterogenous nuclear ribonucleoprotein hnRNPA2/B1, a ubiquitous RNA binding protein involved in RNA processing, is a negative regulator of ATP7A expression and has implicated this RNA binding protein in the developmental expression of ATP7A (McCann, Hasan, et al., 2022). Downregulation of hRNPA2/B1 leads to upregulation of ATP7A in both HeLa and SH-SY5Y neuroblastoma cells. This negative regulation of ATP7A by hRNPA2/B1 is mediated through a conserved 8-nucleotide motif in the 3′ untranslated region of ATP7A’s mRNA sequence (McCann, Hasan, et al., 2022). Of note, previous work has identified that binding of hRNAP1/B1 to this motif leads to mRNA decay (Geissler et al., 2016). Intriguingly, SH-SY5Y cells undergoing retinoic-acid (RA) and induced neuronal differentiation show a concurrent decrease in hnRNPA2/B1 levels and increase ATP7A expression. Further experiments utilizing siRNA knockdown of hnRP2A/B1 suggest that changes in expression of this RNA binding protein modulate ATP7A levels during RA-induced neuronal differentiation (McCann, Hasan, et al., 2022), underscoring both the tight regulation of copper transporter expression during neuronal development and the importance of ATP7A during neuronal differentiation.

The localization of ATP7A changes markedly during neuronal maturation. Early in neuronal differentiation, the protein is primarily localized to cell bodies, but, over time, its expression is increased in extending axons (El Meskini et al., 2005). Furthermore, following injury-induced neurogenesis in mouse olfactory receptor neurons, the localization of Atp7a follows a similar pattern, increasing its axonal expression prior to synaptogenesis (El Meskini et al., 2005), suggesting it likely has a functional role in this process. Mice with mutations in Atp7a not only show aberrant targeting of axons and dendrites, but also an accumulation of immature neurons (El Meskini et al., 2007). These findings suggest that Atp7a plays an important role in both in synaptogenesis as well as in terminal neuronal differentiation at the time of axon extension.

Although El Meskini et al. (2007) posit that this could be a copper-independent process due to the finding that Atp7a can traffic to neuronal processes of hippocampal neurons in response to N-methyl-d-aspartate (NMDA) receptor stimulation (Schlief et al., 2005), more recent evidence suggests that copper homeostasis may play an important role in ATP7A-dependent synaptogenesis. Using a Drosophila model, which allows for extensive genetic manipulation, Hartwig et al. (2021) demonstrated that copper homeostasis is essential for synaptic development and showed this process is regulated by ATP7A and the conserved oligomeric Golgi (COG) complex. The COG complex is an octamer localized to the Golgi that is involved in vesicle tethering, incoming vesicle fusion, the coordination of retrograde vesicle transport, and glycosylation (Miller & Ungar, 2012; Ungar et al., 2002). Previous work established a biochemical interaction between ATP7A and COG and demonstrated that cells deficient in the COG complex showed increased surface expression of ATP7A, along with a concomitant decrease in copper content due to increased efflux (Comstra et al., 2017). This study demonstrated a functional link between ATP7A, the COG complex, and copper homeostasis in neurons. Using genetically modified Drosophila, Hartwig et al. (2021) demonstrated perturbations in ATP7A disrupt copper homeostasis and lead to aberrant synaptic morphology in the neuromuscular junction, both of which could be rescued by alterations in COG expression. These results suggest that copper dyshomeostasis is involved in the altered synaptic morphology seen in with ATP7A mutants and that COG may be an upstream modulator of copper homeostasis in neurons. Indeed, further experiments showed that the COG complex controls the function of ATP7A and CTR1 in developing motor neurons (Hartwig et al., 2021), demonstrating the importance of Golgi integrity in neuronal function. These results not only provide insight into copper regulation in developing neurons, but also demonstrate a direct link between copper homeostasis and synaptic development. Additionally, this investigation demonstrated that ATP7A expression and copper homeostasis are critical for proper mitochondrial localization to the synapse as well as its functional integrity (Hartwig et al., 2021). Given the high metabolic demands of developing neurites and the importance of axonally localized mitochondria for neuronal development (Rangaraju et al., 2019; Son & Han, 2018; Spillane et al., 2013), these findings may offer an additional clue to the why the Golgi apparatus and ATP7A are critical in synaptogenesis.

4 |. COPPER IN THE SECRETORY PATHWAY

4.1 |. ATP7A, ATP7B and ATOX1

Several copper-binding proteins, such as ATP7A and ATP7B, needed for neurotransmitter and neuropeptide synthesis, and interestingly, vesicular copper release, reside in the secretory pathway. ATP7A and ATP7B are essential copper transporters that serve key roles in both the metalation of copper-dependent enzymes in the TGN as well in intracellular copper homeostasis by exporting excess copper across cell membranes. Although these transporters have similar functions and structure, sharing approximately 60% of the same amino sequence, several lines of evidence suggest that these proteins have distinct properties in cells that co-express both ATP7A and ATP7B (Linz & Lutsenko, 2007; Lutsenko, Barnes, et al., 2007; Lutsenko, LeShane & Shinde, 2007). In polarized cells, stimuli such as hormonal release and elevations in intracellular copper results in the trafficking of ATP7A from the TGN to the basolateral membrane whereas ATP7B moves from the TGN towards the apical membrane via endosomes (Braiterman et al., 2011; Ruturaj et al., 2024).

The apical trafficking of ATP7B in response to copper elevation is mediated by regions in the protein’s N- and C-terminal domains. A nine amino acid sequence was identified in ATP7B’s N-terminus that contain both apical targeting and Golgi retention signals (Braiterman et al., 2009) Mutational analysis has revealed that the COOH terminus of ATP7B is a critical mediator of the transporter’s apical trafficking in response to copper (Braiterman et al., 2011). In non-polarized cells, a dileucine motif in ATP7A’s COOH terminus was identified as a signal that allowed the protein to cycle between the plasma membrane and the TGN (Francis et al., 1999; Petris & Mercer, 1999). This signal was also identified in MDCK cells, an immortalized kidney cell line, as a required motif for targeting of ATP7A to the basolateral membrane. Additionally, a PDZ domain in the protein’s COOH terminus was found to help retain the protein at the basolateral membrane (Greenough et al., 2004). More recent work has suggested that clathrin and clathrin adaptor proteins, such as AP-1 and AP-2, interact with ATP7A and ATP7B to regulate their intracellular sorting (Lalioti et al., 2014; Ruturaj et al., 2024; Yi & Kaler, 2015). Although neurons do not have apical-basolateral polarity like epithelial cells, they are a highly polarized cell type with axonal and somatodendritic compartments and express both ATP7A and ATP7B. Intriguingly, recent evidence suggests that these copper transporters are differentially sorted to these neuronal compartments. In primary hippocampal neurons, Atp7b was found to be primarily restricted to the soma and dendrites and largely restricted from axons. Moreover, its somatodendritic localization was contingent upon interaction of AP-1 with the clathrin recognition signal contained in Atp7b’s C-terminal dileucine motif (Jain et al., 2015). As AP-1 has been shown to interact with the copper-dependent enzyme PAM (discussed further below) in pituitary tumour cells and pituitary tissues, and to be an important mediator of its trafficking to secretory granules (Bonnemaison et al., 2014), Jain et al. (2015), hypothesize that the somatodendritic restriction of Atp7b ensures efficient copper loading of cuproenzymes in the perikaryon. Of note, Schmidt et al. (2018), demonstrated that in primary cortical neurons Atp7b is distributed in a vesicular pattern around the nucleus and the cytosol while Atp7a was clustered at the axonal side of the nuclei suggesting distinct and non-overlapping localization of these two copper transporters under basal conditions (Schmidt et al., 2018) Intriguingly, in primary hippocampal neurons, treatment with 100 μM copper results in movement of Atp7b from the somatodendritic compartment to axons where it was colocalizes with Lamp-1, suggesting that copper treatment drives Atp7b localization to late endosomes or lysosomes for axonal trafficking (Jain et al., 2015). Previous work has shown that NMDA receptor activation in primary hippocampal cells drives trafficking of Atp7a from the Golgi to axons where it effluxes copper (Schlief et al., 2005). Therefore, it is possible that under basal conditions, Atp7a and Atp7b are restricted to the soma to metalate cuproenzymes but that these copper transporters traffic to axons to carry out their role in copper efflux when stimulated. However, more work remains to be done to further delineate the distinct localization of these copper transporting ATPases and its functional implications in neurons.

Given the negligible concentration of labile copper in the cytosol, ATP7A and ATP7B require transfer of copper by a copper chaperone that can bind the metal in the cytosolic compartment and then deliver it to the secretory pathway, where these proteins are localized. Early studies in yeast suggested that Atx1, a small cytosolically localized protein served as this putative copper chaperone as it functions upstream of the yeast ATP7A and ATP7B homolog Ccc2 in the delivery of copper from CTR1 to the secretory pathway (Huffman & O’Halloran, 2000; Lin et al., 1997; Pufahl et al., 1997; Rosenzweig et al., 1999). One important finding in the elucidation of this pathway was the that cells deficient in Atx1 fail to incorporate copper into Fet3p (Klomp et al., 1997), the yeast homolog of ceruloplasmin and a multi-copper oxidase that functions as a high affinity iron transporter (Taylor et al., 2005). More importantly, yeast lacking Ccc2 are iron deficient, a sign of Fet3p dysfunction, and this phenotype that cannot be rescued by overexpressing Atx1. Overexpressing Ccc2 in cells deficient in Atx1 however, can rescue iron deficiency, supporting a role for Atx1 in copper delivery to Ccc2 (Lin et al., 1997). Of note, in mammals, ATP7B is the enzyme responsible for the metalation of ceruloplasmin (Telianidis et al., 2013; Terada et al., 1998). Indeed, later studies confirmed that Atx1 is a metallochaperone that transfers Cu(I) to Ccc2 in a direct and reversible manner (Huffman & O’Halloran, 2000; Portnoy et al., 1999), and structural characterization of the Atx1 showed that it coordinates copper through a highly conserved metal binding site (CXXC) located near the surface of the protein (Rosenzweig et al., 1999). Structural and functional characterization of mammalian ATOX1 confirmed that there is a specific, copper dependent interaction between this metallochaperone and the metal binding domains of ATP7A and ATP7B that is dependent on the CXXC sequence motifs present in both the ATOX1 and ATP7A and ATP7B (Larin et al., 1999; Wernimont et al., 2000). Consistent with these data, cells that are deficient in Atox1 have increased intracellular copper (Hamza et al., 2001, 2003), likely reflecting a lack of copper efflux, due to the inability of Atox1 to transfer copper to ATP7A, which translocates to the plasma membrane under conditions of excess copper to transport copper out of the cell (Petris et al., 1996; Petris & Mercer, 1999). Furthermore, the C-terminus of human CTR1 has been shown to bind Atox1 with high affinity (K_d ~ 10⁻¹⁴ M) and to transfer copper to this cytosolic protein (Kahra et al., 2016), supporting a model in which ATOX1 serves as a metallochaperone for this metal and delivers it from the plasma membrane to the secretory compartment, while keeping the cytosolic concentration of labile copper necessarily low.

Investigations utilizing both yeast and mammalian systems have demonstrated the necessity of cysteine residues in the Atx1/ATOX1 metal binding domain in coordinating copper binding (Rosenzweig et al., 1999; Strausak et al., 2003; Wernimont et al., 2000). Mutagenesis experiments of yeast Atx1 demonstrated that its interaction with Ccc2 is dependent on the two cysteines present in its metal binding domain (Portnoy et al., 1999), and experiments in mammalian cells confirmed that that the two cysteines in ATOX1’s CXXC motif are required for its interaction with both ATP7A and ATP7B (Larin et al., 1999; Strausak et al., 2003). While this finding is likely a result of the structure of the protein, as the two cysteines are exposed on the surface making them accessible for docking of copper ions (Rosenzweig et al., 1999; Wernimont et al., 2000), it also highlights the potential redox sensitivity of the interaction between ATOX1 and its targets. Indeed, studies have demonstrated that Atox1 is sensitive to the cytosolic redox environment (Hatori et al., 2012, 2016; Katsuyama et al., 2021). GSH can reduce the copper-coordinating cysteines of Atox1 and endow its ability to bind copper. Alternately, when GSSG levels are high and the GSH/GSSG balance is shifted, the copper-coordinating cysteines become oxidized and Atox1 is unable to bind copper (Hatori et al., 2012). As the interaction between Atox1 and ATP7A and ATP7B is dependent on copper binding, the GSH/GSSG ratio bidirectionally modulates the flow of copper through the secretory pathway by regulating the redox state of Atox1. This coupling between Atox1 and the redox state of the cell has significant functional implications in neurons, which will be discussed below.

5 |. NEUROTRANSMITTER AND NEUROPEPTIDE SYNTHESIS

5.1 |. Dopamine-β-hydroxylase

Dopamine-β-hydroxylase (DBH) is a cuproenzyme critical for the synthesis of the neurotransmitter norepinephrine (NE) in noradrenergic neurons of the locus coeruleus and sympathetic nerve terminals. Of note, the locus coeruleus is highly enriched in copper (Naeve et al., 1999; Xiao et al., 2018). In catecholaminergic cells, DBH is responsible for the conversion of dopamine (DA) to NE, which is essential for arousal, attention, and wakefulness (Berridge & Waterhouse, 2003). Human DBH is a tetrameric glycoprotein (Wallace et al., 1973) comprised of four domains—one dopamine-β-monooxygenase N-terminal (DOMON) domain, which belongs to a class of DOMON-like domains, two copper catalytic cores and a C-terminus dimerization domain (Vendelboe et al., 2016). The catalytic cores are referred to as Cu_H or Cu_M based on whether they are coordinated by three histidines (Cu_H) or two histidines and one methionine (Cu_M). Each copper domain is thought to have a distinct function, with Cu_H transferring electrons from its co-substrate ascorbic acid, and, in the process, generating superoxide, with Cu_M hydroxylating substrates (Brenner & Klinman, 1989; Klinman, 2006). Underscoring the importance of copper for NE synthesis, DBH thus requires two bound copper for activity (Ash et al., 1984).

DBH is likely synthesized in the endoplasmic reticulum (Sabban et al., 1983), and is then trafficked to the TGN where it acquires copper from ATP7A (Xiao et al., 2018) before being packaged into secretory granules, converting their DA to NE. DBH exists in both a membrane-bound and soluble form (Sabban et al., 1983) and studies utilizing rat PC12 cells and human SH-SY5Y cells have demonstrated that once packaged into secretory granules, DBH-containing vesicles are released not only in response to stimulation but are also constitutively released into the extracellular space under resting conditions (McHugh et al., 1985; Oyarce & Fleming, 1991; Sabban et al., 1983). Importantly, extracellular DBH is enzymatically active suggesting it is exported in a copper-bound state (Schmidt et al., 2018). It is thought that this constitutive secretion of DBH also maintains the stable levels of DBH observed in the CSF, which are independent of neuronal activity. It must be noted, however, that the function of this CSF localized DBH is still unclear (O’Connor et al., 1994). Intriguingly, copper may also regulate DBH secretion as reduced copper delivery to the secretory pathway has been shown to result in increased intracellular levels of the enzyme (Hatori et al., 2016). This finding is supported by recent evidence showing that copper is necessary for the constitutive secretion of DBH in SH-SY5Y cells (Schmidt et al., 2018). Furthermore, Schmidt et al. (2018) demonstrated that neuronal ATP7B may negatively regulate the export of soluble DBH by sequestering copper in vesicles and inhibiting ATP7A from transferring this cofactor to DBH in the TGN. Thus, not only is copper necessary for noradrenergic signalling by acting as cofactor for DBH, but copper and known copper transporters affect the storage and release of this enzyme, likely to have important implications for diseases such as Menkes disease and Wilson disease, in which the function of ATP7A and ATP7B are disrupted. These observations by Schmidt et al. (2018) also further demonstrate that ATP7A and ATP7B have distinct functions in neurons.

5.2 |. Peptidylglycine α-amidating monooxygenase

PAM is an enzyme that catalyses C-terminal amidation, an important post-translational modification of many signalling peptides (Eipper et al., 1993). PAM catalyses the cleavage of glycine, resulting in a bioactive C-terminal amidated peptide and glycoxylate (Bradbury & Smyth, 1991; Kumar et al., 2016; Merkler, 1994). Underscoring the importance of this enzyme is the fact that deletion of the PAM gene results in lethality in an early larval stage in Drosophila (Jiang et al., 2000) and embryonic lethality in mice (Czyzyk et al., 2005). PAM contains two enzymatic domains—the peptidylglycine α-hydroxylating monooxygenase (PHM) domain and the peptidyl-α-hydroxyglycine α-amidating lyase (PAL) domain, which act sequentially to produce an amidated peptide. Of note, the PHM domain contains ~32% homology to DBH (Solomon et al., 2014) including a catalytic core containing Cu_H and Cu_M domains, each requiring two copper ions for maximal activity (Eipper et al., 1995; Kulathila et al., 1994; Prigge et al., 1999), which they likely acquire from ATP7A (Otoikhian et al., 2012). Additionally, like DBH, PHM requires ascorbic acid and molecular oxygen as substrates (Murthy et al., 1986), and contains Cu_H and Cu_M domains, separated by 11 Å, achieving the two-electron oxidation of the peptide (or catecholamine in the case of DNB), via two one-electron, distinct steps (Cowley et al., 2016).

In addition to copper’s role in PHM catalysis, there is evidence to suggest that this metal may also affect endocytic trafficking of PAM. Like DBH, PAM exists in membrane-bound and soluble forms. Soluble PAM is targeted to secretory granules and is released along with bioactive peptides upon stimulation. After exocytosis, membrane PAM appears on the plasma membrane, where it is rapidly internalized and is either packaged into secretory granules or undergoes degradation (Milgram et al., 1994; Steveson et al., 1999). In AtT-20 cells, a pituitary corticotrope tumour cell line, degradation of endocytosed PAM is increased by high intracellular copper levels. In contrast, cleavage of recycled PAM, its trafficking through early and late endosomes, and its subsequent return to secretory granules, which is needed for the release of the soluble enzyme, is increased in conditions of copper restriction (De et al., 2007). Importantly, De et al. (2007) demonstrated these effects were mediated through PAM’s cytosolic domain, suggesting it may act as a copper sensor, allowing the enzyme to respond dynamically to copper availability.

5.3 |. Synaptic copper

Although copper is mainly thought of as a cofactor or allosteric modulator of enzymes with its actions primarily taking place within the intracellular environment (Ge et al., 2022), there is a scattering of evidence that suggest that copper is released from presynaptic neurons and that concentrations within the synaptic cleft can rise to micromolar levels, allowing this metal to act directly, or indirectly via redox reactions, on synaptic receptors. It is important to note that due to the reducing environment of the cell, intracellular copper exists primarily as Cu(I), however when it is released into the extracellular space, it is likely oxidized to Cu (II) (Maryon et al., 2007). Early evidence for depolarization-mediated neuronal release of copper came from the rat hypothalamic explants using radiolabeled metal (Hartter & Barnea, 1988). Further studies revealed copper release from isolated synaptosomal preparations from rat cortical neurons utilizing atomic absorption measurements (Kardos et al., 1989). Indeed, depolarization with high K⁺ led to the release of copper into the extracellular space and, from these experiments, it was estimated that the putative concentration of released copper within the synaptic cleft could be as high as 200–250 μM (Kardos et al., 1989). However, it should be noted that this concentration may not take into account protein-bound copper, and indeed, later experiments using a fluorescent probe along with ICP-MS found that following depolarization the concentration of free copper released from synaptosomes was likely to be closer to 2–3 μM (Hopt et al., 2003). This concentration would be consistent with evidence that the normal extracellular copper concentration in the brain is approximately .2–1.7 μM (Tarohda et al., 2004). However, the possibility that higher concentration of this metal can exist in microdomains, as has been described for zinc (Krall et al., 2020, 2021), cannot be ruled out.

Although all these studies suggest that copper can be released from synaptic terminals upon depolarization, it has yet to be determined whether in fact, there is functional synaptically released copper in more intact preparations, as has been shown for zinc in a large number of physiological studies (Krall et al., 2021). However, treatment of primary cultured neurons with glutamate/glycine to stimulate NMDARs can also produce an increase in copper levels in the bath solution (Schlief et al., 2005). Interestingly, glutamate receptor activation leads to the rapid and reversible trafficking of Atp7a from the Golgi compartment to neuronal processes, suggesting that this P-type ATPAse may play a critical role in copper efflux (Schlief et al., 2005). Indeed, the direct role of Atp7a in copper efflux is supported by studies using primary neurons from mice with mutant Atp7a receptors. These neurons fail to efflux copper in response to glutamate/glycine treatment suggesting that proper function of Atp7a is an essential component to copper release into the extracellular space (Schlief et al., 2005). Given the rapid efflux of copper following NMDAR activation, Schlief et al. (2005) proposed that Atp7a may accumulate copper in a vesicular compartment that is readily available for exocytosis following activation of the glutamate receptor. This role for Atp7a would perhaps be akin to the function that ZnT3 (SLC30A3) plays in loading glutamatergic synaptic vesicles with zinc (Palmiter et al., 1996), although this has yet to be firmly established. Intriguingly, studies using differentiated PC12 cells demonstrated that increasing intracellular copper upregulates proteins implicated in secretory vesicle trafficking such as synaptophysin and SNAP-25, supporting a potential synaptic role for this metal (Duncan et al., 2013).

5.4 |. Potential targets of synaptic copper

Synaptically released copper may act directly on synaptic neurotransmitter receptors, alter their function via reduction/oxidation reactions, or, as mentioned earlier, influence their activity via regulation of downstream enzymatic processes, such as kinase-dependent pathways (Ge et al., 2022). In this section we highlight some the known actions of copper on potential synaptic targets (Figure 2).

FIGURE 2.

Synaptic copper. Glutamate binding to NMDA receptors stimulates ATP7A to traffic out of the trans Golgi network (TGN) to the presynaptic membrane to induce copper efflux into the synaptic cleft. Copper can then bind to post-synaptic NMDA receptors, GABA_A receptors, and AMPA receptors, subsequently modifying their activity.

5.4.1 |. Ionotropic glutamate receptors

Glutamate is a major excitatory neurotransmitter, and its receptors mediate fast excitatory synaptic transmission in the CNS. Fast synaptic action by glutamate is mediated via ionotropic receptors, which include a heterogenous family of tetrameric ligand-gated channels classified based on their structural homology and response to pharmacological agonists. NMDA receptors are comprised of an obligatory GluN1 subunit as well as two GluN2(A-D)/3(A or B) type subunits (Traynelis et al., 2010), with the majority of receptors expressing two GluN1 subunits and two GluN2 subunits (Köhr, 2006). Alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors can be comprised of homo- or heteromers of GluA1-4 subunits while kainate receptors are formed from GluK1-GluK5 subunits (Traynelis et al., 2010).

A growing body of evidence suggests that copper can modulate ionotropic glutamate receptors, with early work demonstrating that in cultured rat cortical neurons AMPA/kainate receptors are highly sensitive to inhibition by copper (IC₅₀ = 4.3 μM) and that application of micromolar concentrations of copper reduced the potency of kainate in activating the AMPA receptor, shifting the EC₅₀ of kainate from 100 μM to >300 μM. Intriguingly, application of the reducing agent dithiothreitol (DTT) blocked copper mediated inhibition of kainate-induced currents, suggesting an oxidative mechanism, rather than direct binding, may be partly responsible for copper’s inhibition of AMPA/kainate receptors. This supposition is supported by the fact that co-application of kainate with histidine, a high affinity copper chelator, did not affect recovery of receptors from copper block, suggesting that direct copper binding may not be needed to exert its inhibitory effects (Weiser & Wienrich, 1996). However, it is important to note that DTT was not washed out prior to application of copper and thus it likely bound the metal (Krężel et al., 2001), as has been shown for zinc (Paoletti et al., 1997), which would also explain a lack of copper-mediated inhibition. A more recent study utilizing primary rat hippocampal neurons found that copper’s effect on AMPA receptors is biphasic. Acute application of the metal decreased AMPA evoked currents (IC₅₀ = 22.9 μM) compared to neurons treated with AMPA alone, in addition to decreasing synaptic activity. In contrast, longer incubation with copper (3 h) produced potentiation of AMPA evoked currents, shifted the concentration-response curve for the agonist (EC₅₀ = 8.3 μM for control neurons and EC₅₀ = 2.9 μM for copper treated neurons), and increased the frequency and amplitude of miniature synaptic currents. This 3-h incubation period also increased intracellular calcium transient frequency supporting a role for increased neurotransmission. However, the increase in AMPAergic transmission disappeared after 24 h of copper treatment indicating a homeostatic response to increased levels of exogenous copper on a longer timescale (Peters et al., 2011), including increased expression of metal-binding proteins (Durnam & Palmiter, 1981). Indeed, one question that remains unanswered from this study is whether chelation or wash-out of copper after this 24-h incubation followed by acute application of the metal would still produce inhibition of AMPA currents or whether the homeostatic response prevents further response by the receptors to this stimulus. This information would be relevant to neurological disease that result in copper dyshomeostasis and increased brain copper, as lower AMPA receptor function may lead to decreased excitation and disturbance of the excitatory-inhibitory balance.

NMDA receptors are also intricately involved in the effects of synaptic copper. Not only does NMDAR stimulation allow for the release of free copper, allowing it to possibly act on post-synaptic receptors (Schlief et al., 2005), but studies have demonstrated that copper inhibits NMDAR-induced currents themselves (Trombley & Shepherd, 1996; Vlachová et al., 1996; Weiser & Wienrich, 1996), further indicating that this metal negatively modulates excitatory NMDAR-mediated neurotransmission. The estimated potency of copper in blocking NMDAR receptor currents in these studies ranged from micromolar (IC₅₀ = 15 μM-22 μM) (Trombley & Shepherd, 1996; Weiser & Wienrich, 1996) to nanomolar concentrations (IC₅₀ = 270 nM) (Vlachová et al., 1996). One potential explanation for the difference in the IC₅₀ values between those reported by Weiser and Wienrich (1996) and those reported by Vlachová et al. (1996) is the timing of their experiments. While both studies utilized primary rat cortical neurons in their investigations, Weiser and Wienrich (1996) used their cultures between 11 and 21 days in vitro while Vlachova et al. (1996) used cultures at 5–9 days in vitro. As NMDARs undergo a developmental switch, with GluN2B being the predominant subunit expressed until the second week of development, at which time GluN2A expression strongly increases (Sanz-Clemente et al., 2013; Sinor et al., 2000), it is possible that differences in NMDAR subunit identities may underlie these differences. Despite the wide range of IC₅₀ values, all studies agreed regarding the voltage-independence of this inhibition, although it must be noted that the mechanism by which copper inhibits the NMDAR remains unknown. It is important to note here that metal contaminants may be present in physiological solutions (Paoletti et al., 1997), affecting the interpretation of the results and calculation of metal affinities for a given receptor. This has led to the now widespread use of Chelex-treated water in the preparation of physiological solution when studying metal function (Anderson et al., 2015).

While the effect of copper on NMDARs may occur via direct binding, akin to the actions of zinc (Peters et al., 1987; Westbrook & Mayer, 1987), work by Marchetti et al. (2014) and Schlief et al. (2006) suggests the possibility of a redox modulatory component. Importantly, it has long been known that NMDARs are subject to redox modulation mediated by key cysteine residues on GluN1 and GluN2A subunits (Aizenman et al., 1989, 2020; Tang & Aizenman, 1993a). In contrast to Vlachova et al. (1996) which found that nanomolar concentrations of copper inhibited NMDAR-induced currents, Marchetti et al. (2014) found that in cerebellar neurons and transiently transfected HEK293 cells, low micromolar concentration of the metal potentiated the current (EC₅₀ = 4.6 μM for GluN1/GluN2B; IC₅₀ = 2 μM for GluN1/GluN2A) and increased intracellular calcium. As cerebellar neurons express GluN1/GluN2C NMDARs, as well as triheterometric GluN1/GluN2A/GluN2C NMDARs (Bhattacharya et al., 2018; Sibarov et al., 2018), differences in subunit composition could contribute to the potentiation in cerebellar neurons. This potentiation was absent when neurons were pre-treated with the reducing agent DTT, although it should be noted that DTT itself causes potentiation of NMDAR currents. This finding is consistent with previous research showing that reducing agents potentiate NMDAR currents while oxidizing agents inhibit them (Aizenman et al., 1989). Intriguingly, these experiments also demonstrated a moderate decrease in NMDAR currents when copper was applied following DTT treatment raising the possibility that copper may oxidize redox modulatory sites when they are in a reduced state, either directly (Aizenman et al., 1990; Smith et al., 1994), possibly by the generation of free radicals. However, given that copper facilitates NMDAR currents in the absence of a reducing agent, and is more likely to act as an oxidizing agent in its Cu (II) state, if copper-mediated facilitation occurs through redox modulation of the NMDAR, it is probably through an indirect, rather than direct mechanism. One possibility could involve copper-induced release of an endogenous reducing substance, such as dihydrolipoic acid, which has been found to reduce the NMDAR redox modulatory site (Tang & Aizenman, 1993b) and thus indirectly contribute to the potentiation of NMDAR currents. Clearly, additional work is required to fully clarify this process, including the use of redox-insensitive mutants of NMDA receptor subunits (Brimecombe et al., 1999; Sullivan et al., 1994).

Consistent with Trombley and Shepherd’s (1996) and Weiser and Wienrich’s (1996) investigations, concentrations above 30 μM copper inhibited NDMAR-mediated currents (IC₅₀ = 24 μM for GluN1/GluN2B, IC₅₀ = 26 μM for GluN1/GluN2A) in transiently transfected HEK293 cells and cerebellar neurons (Marchetti et al., 2014). One potential explanation for this inhibition, posited by Schlief et al. (2006), may involve S-nitrosylation of the receptor, a post-translational modification involving the attachment of a ^•NO group to the thiol side chain of cysteine. S-nitrosylation has been suggested to result in blockade of the receptor (Lei et al., 1992), consistent with inhibition of NMDAR currents by copper. Schlief et al. (2006) showed that in primary hippocampal neurons, treatment with 200 μM copper blocked NMDAR activation and resulted in endogenous nitric oxide production. Furthermore, treatment with a nitric oxide synthase inhibitor eliminated the copper induced blockade of NMDAR activation while the effect of copper was restored after treatment with an exogenous nitric oxide donor (Schlief et al., 2006). These results suggest that copper indirectly modulates NMDARs function through its effect on nitric oxide production. However, it is worth noting that the concentrations of copper used in this study are likely supraphysiological and may result in oxidative stress triggering the production of reactive oxygen and reactive nitrogen species. Furthermore, whether nitric oxide itself actually modifies redox modulatory sites on the NMDAR has been called into question (Aizenman et al., 2020; Aizenman & Potthoff, 1999; Hopper et al., 2004). Thus, the mechanism by which copper inhibits NMDAR currents, and whether it is a result of redox modulation or S-nitrosylation remains to be determined. Finally, an additional, copper-mediated inhibitory process at the NMDA receptor has been described in the literature. Cellular prion protein (PrP^C), a copper-binding protein, was shown to inhibit NMDA receptor function by decreasing the affinity of glycine, a required NMDA receptor co-agonist (Johnson & Ascher, 1987; Kleckner & Dingledine, 1988), for its binding site in a copper-dependent manner (You et al., 2012). Moreover, mutation of copper-binding sites within PrP^C_, was shown to abolish this effect (Huang et al., 2021).

Although the mechanism by which copper inhibits NMDARs has yet to be fully characterized, this process appears to have homeostatic and neuroprotective functions. When 200 μM copper, a concentration that did not cause overt neuronal toxicity, was applied to primary hippocampal neurons exposed to excitotoxic levels of glutamate/glycine, it completely abrogated cell death. Conversely, when the copper chelator bathocuproine disulfate (BCS) was applied to cultures to reduce intracellular copper prior to exposure glutamate/glycine, neuronal death was increased (Schlief et al., 2006). Importantly, all experiments were performed in the presence of the AMPA/kainate antagonist CNQX to isolate NMDAR stimulation by glutamate/glycine. These results suggest that the putatively vesicular copper released from neurons following NMDAR stimulation (Schlief et al., 2005) may act on NMDARs themselves to rein in excessive excitation, and that additional copper release can protect neurons from excitotoxic cell death. Co-application of copper with glutamate/glycine also inhibited the previously observed elevation in intracellular calcium upon NMDAR hyperactivation, strongly suggesting that copper likely protects against excitotoxic cell death by preventing downstream signalling events following receptor stimulation (Schlief et al., 2006). These findings may have implications for diseases in which neuronal copper is increased and thus, more would be available to be released from the cell upon NMDAR stimulation. In support of the homeostatic and neuroprotective nature of copper release is the finding that primary neurons derived from mice with loss of function of Atp7a, which is critical for neuronal copper release (Schlief et al., 2005), were shown to be more sensitive to excitotoxic injury induced by exposure to high K⁺ to induce synaptic glutamate release, as compared to control neurons. However, application of exogenous copper before and during high K⁺ stimulation abrogated this increased sensitivity to excitotoxic injury in Atp7a mutant neurons (Schlief et al., 2006). These findings strongly indicate a protective role for ATP7A-dependent copper release in neurons and may partly account for the neurodegeneration seen in ATP7A-linked disorders such as Menkes disease, which shows decreased function of this protein.

5.5 |. GABA_A receptors

GABA_A receptors (GABA_ARs) are ionotropic, chloridegated channels and serve as the major inhibitory receptors in the CNS. GABA_ARs are pentameric assemblies that are usually comprised of three different proteins selected from 19 different genetically encoded subunits (α1–6, β1–3, γ1–3, δ, ε, θ, π and ρ1–3). Evidence suggests that most GABA_AR are comprised of two copies of a single α subunit, two copies of a single β subunit, and one copy of another subunit such as γ or δ, although GABA_ARs can be heteromers of α and β subunits alone (Olsen & Sieghart, 2008). The subunit composition of these receptors not only affects their physiological and pharmacological properties but their localization within neurons as well. It is generally accepted that α1–3βγ subunits comprise synaptic GABA_ARs whilst those that contain α4–6 subunits as well as αβ heteromers are extrasynaptically localized (Hannan et al., 2020; Nusser et al., 1998). The localization of GABA_AR is a large determinant of the type of inhibition these receptors exert. Synaptic GABA_ARs are thought to modulate rapid phasic inhibition while extrasynaptic receptors are typically thought to influence tonic inhibition, which is important for proper regulation of the excitatory-inhibitory balance.

A large body of work has investigated the effect of zinc on GABA_ARs and it is now established that these receptors are allosterically inhibited by this metal (Krall et al., 2021). Importantly, this inhibition is subunit dependent with isoforms containing only αβ subunits being most sensitive to zinc block. Inclusion of a δ, ε or π subunit reduces the affinity of zinc for the GABA_AR but strikingly, the inclusion of a γ subunit renders these receptors almost insensitive to zinc block (Draguhn et al., 1990; Nagaya & Macdonald, 2001). This loss of zinc sensitivity is thought to be a result of a conformational change to the α and β subunits when the γ subunit is inserted, disrupting the putative zinc binding site(s) (Hosie et al., 2003). Thus, tonic GABAergic signalling is most likely to be affected by zinc binding. In contrast to zinc, there has been comparatively little research into the effect of copper on GABAergic signalling. However, early experiments utilizing rat dorsal ganglion neurons demonstrated that micromolar concentrations of copper reversibly inhibited GABA_ARs with a similar potency to zinc (IC₅₀ = 16 μM and 19 μM, respectively) (Ma & Narahashi, 1993; Narahashi et al., 1994). These concentrations are similar to what was found for zinc- and copper-mediated inhibition of GABA_A receptors in rat olfactory bulb neurons (IC₅₀ = 17 μM) (Trombley & Shepherd, 1996). Moreover, in dorsal root ganglion copper antagonized the blocking actions of zinc on GABA_ARs in a concentration-dependent manner suggesting these divalent metals share a binding site (Ma & Narahashi, 1993). These findings were supported by work in Purkinje cells which demonstrated that copper attenuates zinc induced GABA_AR block and vice versa. However, these studies found that copper inhibited GABA_ARs at nanomolar (IC₅₀ = 35 nM) rather than micromolar concentrations (Sharonova et al., 1998). These authors posited that copper and zinc bind the GABA_A receptor at distinct, but conformationally linked sites (Sharonova et al., 2000).

Further investigation into copper binding to and its inhibition of GABA_A receptors has revealed that in contrast to zinc, γ containing GABA_ARs are sensitive to inhibition by copper and that those comprised of α1β3γ2 subunits are more sensitive to block by this metal than those containing α6β3γ2 (Kim & Macdonald, 2003). These studies also revealed that an N-terminal motif in the α1 and α2 subtypes is a major determinant of high copper sensitivity and that Val¹³⁴, Arg¹³⁵ (Gln for α2) and His¹⁴¹ within this motif are the most important determinants for copper binding (Kim & Macdonald, 2003). Additionally, His²⁶⁷ within the M2 domain of the β3 subunit and His²⁷³ within the M2-M3 loop of the α6 subunit, which are major determinants of GABA_AR sensitivity to zinc (Fisher & Macdonald, 1998; Hosie et al., 2003), were found to contribute to copper inhibition (Kim & Macdonald, 2003). Together, these data suggest that zinc and copper bind to overlapping but non-identical sites on the GABA_AR.

While the finding that copper potently blocks α1β3γ2 GABA_ARs suggests that this metal inhibits fast synaptic, or phasic GABAergic signalling in contrast to zinc, which can only inhibit tonic GABAergic signalling, it is important to note extrasynaptic receptors are sensitive to copper block and, therefore, tonic GABAergic signalling as well. In cerebellar granule cells, the α6 subunit is obligatory for δ expression (Jones et al., 1997). Despite the lower affinity of copper for α6 receptors, application of micromolar concentrations of copper to α6β3δ expressing cerebellar granule cells reduces GABA evoked responses (McGee et al., 2013). Moreover, chelating copper from these cells enhances tonic conductance, further suggesting that copper binds to and inhibits extrasynaptic GABA_ARs. In addition to blocking tonic conductance in cerebellar granule cells, application of copper to medium spiny neurons from the striatum significantly inhibits their GABAergic tonic conductance (McGee et al., 2013), demonstrating the ability of copper to block tonic GABAergic signalling is not restricted to one brain region. It is important to note that like cerebellar granule cells, medium spiny neurons are also known to express extrasynaptic δ-expressing GABA_A receptors (Santhakumar et al., 2010). Given the importance of tonic GABAergic signalling in regulating neuronal excitability, it is intriguing to consider the implications of pathological conditions, such as Wilson disease, in which brain copper is elevated and may disrupt tonic GABA signalling. Finally, given that ionotropic GABA receptors, like NMDA receptors, are redox sensitive in a subunit-dependent fashion (Amato et al., 1999; Pan et al., 2000), it is entirely possible that some of the actions of the metal on this class of ionotropic receptors might be mediated through reduction and oxidation reactions. This, however, remains to be carefully evaluated, as is the case for NMDA receptors.

6 |. COPPER AND NETWORK EXCITABILITY

Given copper’s actions on both excitatory and inhibitory ligand-gated receptors, one would expect that that application of copper might affect neuronal circuit function. Indeed, using rat hippocampal slices as a neuronal network model, Maureira et al. (2015) found that broadly, copper enhances network excitability. Application of 10 nM nanomolar copper increased the spontaneous action potential (AP) firing rate in the CA1 region. More intriguingly, chelation of endogenous of copper with bathocuproine decreased spontaneous spiking activity, suggesting that endogenous copper regulates spiking activity in active hippocampal networks (Maureira et al., 2015). Additionally, consistent with previous research demonstrating that nanomolar concentraions of copper inhibit GABA_A receptors (Sharonova et al., 1998; Trombley & Shepherd, 1996), Maureira et al. (2015) found that application of 50–100 nM copper affected the excitatory/inhibitory balance in the hippocamus. Stimulation of Schaffer collaterals in the CA3 resulted in a transient silence of spiking activity in the CA1 of the hippocampus. However, application of 10 and 50 nM copper decreased the silent period (or time to first spike), consisent with a decrease in GABA_A mediated inhibition. Supporting this point, application of copper did not change basal field excitatory post-synaptic currents (fEPSCs), suggesting the decreased time to first spike was a result of decreased inhibition (Maureira et al., 2015). Based on these results, the authors posit that copper can overcome CA1 feedforward inhibition as the inhibition overcome by copper application is mediated by feedforward inhibitory neurons.

Intriguingly, Dodani et al. (2014) found that copper chelation rather than copper application increased hippocampal network excitability, albeit in cultured neurons. Confocal and two-photo calcium imaging of dissociated hippocampal cultures revealed temporally uncorrelated calcium transients. However, treatment with the copper chelator BCS resulted in an increase in the correlation between calcium transients, which reflect synchronized action potentials produced by neurons in the circuit (Mao et al., 2001). Similar results were obtained when using isolated retinal tissue (Dodani et al., 2014), which is a well-understood model for studying correlated spontaneous activity in developing circuits (Blankenship & Feller, 2010). Although BCS is a cell-impermeable chelator, it is worth noting that imaging with a fluorescent Cu(I) probe showed that BSC reduced labile copper pools in both dissociated hippocampal neurons and isolated retinal tissue, indicating that the chelator can affect intracellular copper pools (Dodani et al., 2014). Taken together, these data are consistent with a model in which disrupting copper stores via acute chelation increases network excitability as evidenced by an increase in correlated activity. Further emphasizing the importance of copper in regulating spontaneous activity, Ctr1 ^+/− mice, which have half the amount of copper in their nervous system compared to WT mice, had more spontaneous calcium transients in their retinas than those isolated from their WT littermates, paralleling previous work using a copper chelator to disrupt cellular copper (Dodani et al., 2014). Together results from this study suggest that endogenous copper alters spontaneous activity in developing neuronal circuits in a Ctr1-dependent manner.

7 |. COPPER AND CIRCADIAN RHYTHMS

The suprachiasmatic nucleus (SCN) in the hypothalamus is the principal circadian clock in the brain and is entrained to light/dark cycles via retino-hypothalamic innervation from a subset of retinal ganglion cells that release glutamate in response to light. The release of glutamate and subsequent activation of NMDARs within the SCN is the dominant cue for photic entrainment (Ding et al., 1994; Ebling, 1996). Depending on what phase the circadian clock is in, release of glutamate can also cause delays or advances in the circadian rhythm through NMDAR activation.

Intriguingly, copper may play an important role in the regulation of circadian rhythms. The copper chaperone and transporter Atox1 (Lee et al., 2013) is expressed rhythmically in the SCN, as is ATP7A (Yamada & Prosser, 2020). Additionally, electrophysiological studies in acute SCN slices from mice demonstrated that modulation of copper availability can shift circadian clock phase. Reducing copper availability via chelation with tetrathiomolybdate (TTM) or BSC at early subjective night induces a significant delay in peak neuronal activity (~3 h) (Yamada & Prosser, 2014). This phase delay is consistent with the delay induced by application of glutamate to SCN slices, suggesting that copper chelation may increase NMDAR activity, consistent with studies demonstrating copper-induced inhibition of NMDAR currents (Trombley & Shepherd, 1996; Vlachová et al., 1996; Weiser & Wienrich, 1996). Furthermore, co-application of TTM with the selective NMDAR antagonist 2-amino-5-phosphonopentanoate (AP5) inhibited the TTM-induced phase-delay supporting the supposition that this effect is NMDAR mediated. In addition to causing phase delays at early subjective night, copper chelation, like glutamate application, results in phase advances at late subjective night (Yamada & Prosser, 2014). Although previous evidence suggests that different molecular mechanisms underlie glutamate-induced phase shifts at early and late subjective night (Golombek & Rosenstein, 2010), the NDMAR appears to be involved in both copper chelation-induced phase shifts at both time points (Yamada & Prosser, 2014). As such, these findings indicate that basal copper release in the SCN may rein in excessive activation of NMDARs and serve as a homeostatic role in the regulation of circadian rhythms. Surprisingly, application of copper itself also results in similar phase shifts as copper chelation at early and late subjective night. Unlike copper chelation however, these circadian phase shifts were independent of NMDAR activation (Yamada & Prosser, 2014). Yamada and Prosser (2020) suggest that these results may best be explained by biphasic effects of copper in the SCN where too little or too much copper can phase-shift the SCN clock and activate or inhibit distinct signalling pathways.

In support of the role of copper in the regulation of circadian rhythms, a distinct splice variant of the copper transporter ATP7B (PINA) was identified in the pineal gland (Borjigin et al., 1999). This splice variant occurs via an alternative promoter and results in an open reading frame that is identical to the C-terminal half of the full ATP7B protein. PINA is expressed in pinealocytes and subset of photoreceptors in adult rats, where its expression is 100-fold greater at night than during the day. The nocturnal expression of PINA is under control of the SCN clock, which transmits its signals through innervation of the pineal gland by the superior cervical ganglion (SCG) (Borjigin et al., 1999). Of note, the SCG provides sympathetic innervation via the release of norepinephrine, which, as noted earlier, is itself copper-dependent due the requirement of copper for the conversion of dopamine to norepinephrine by DBH. Indeed, Dbh ⁻/⁻ mice have altered circadian rhythms (Swoap et al., 2004). Intriguingly, while PINA lacks the N-terminus of the ATP7B protein containing metal binding sites, when the protein is expressed in Saccharomyces cerevisiae lacking the ATP7A/ATP7B homolog ccc2, mutant S. cerevisiae present copper transport abilities, suggesting PINA is a putative copper transporter (Borjigin et al., 1999). These findings suggest that ATP7B may contain additional copper binding sites in its C-terminus and that the function of PINA in the pineal may involve copper transport.

8 |. COPPER, SLEEP, AND AROUSAL

The locus coeruleus (LC) is a brain structure located in the upper brainstem that has been shown to be important for the regulation of a wide range of behaviors including sleep, arousal, attention, learning, and memory (Aston-Jones & Bloom, 1981; Berridge & Waterhouse, 2003; Sara, 2009). Notably, it is also the brain region that has the highest concentration of copper (Que et al., 2008), which is not altogether surprising considering that the LC is the main source of central NE (Lutsenko et al., 2019), the synthesis of which requires the cuproenzyme DBH. As many behaviors linked to the LC, including sleep and arousal, are associated with NE, disruption of copper homeostasis likely also disrupts these processes.

Evidence for the importance of copper in the regulation of sleep and arousal comes from a study examining the effect of brain copper depletion in the LC of zebrafish. Zebrafish atp7a mutants were generated to investigate the effect of copper dysregulation in the LC. Intriguingly, atp7a mutant larval zebrafish were nearly identical to their wildtype siblings in regard to their gross morphology, growth, and fertility, but had significantly reduced copper levels in the brain (Xiao et al., 2018). The phenotype allowed the authors to study the behaviour of these zebrafish in the context of copper dyshomeostasis, without the confounds of growth delay and neurological abnormalities seen in animal models of Menkes disease, which also occurs due to loss of ATP7A function (Guthrie et al., 2020; Kim & Petris, 2007). Supporting the hypothesis that altered copper homeostasis in the LC would lead to alterations in sleep and arousal due to impaired NE synthesis, Atp7a mutant larvae showed differences in rest-activity cycles manifested as increased spontaneous swimming activity during the night and decreased activity during the day compared to wildtype zebrafish. Additionally, they had altered arousal behaviour as measured by the acoustic startle response. Unlike wildtype larvae which have a lower maximal escape response during the day compared to the night, indicating increased startle and decreased arousal during the night, the acoustic startle response did not differ between day and night in atp7a mutant larvae and resembled the wildtype larvae’s night startle response, suggesting overall increased arousal (Xiao et al., 2018). Importantly, larvae treated with the copper chelator BCS to reduce brain copper recapitulated the behavioural deficits observed in atp7a mutant, zebrafish confirming that these effects were a result of reduced brain copper rather than an unspecified effect of the mutation. Furthermore, larvae that had LC neurons ablated also showed similar behavioural changes to those deficient in brain copper, establishing that copper deficiency within the LC drives alterations in rest-activity cycles and arousal. Finally, increasing synaptic NE within the LC by treating mutant larvae with NE reuptake inhibitors partially rescued the rest-activity phenotype (Xiao et al., 2018). Together, these data demonstrate that copper homeostasis is important for modulation of rest-activity and arousal by the LC through a mechanism requiring metalation of DBH by ATP7A for NE synthesis. These data are supported by in vivo work in mice that utilized an Cre-inducible CRISPR/Cas9 system to knockdown (KD) dbh in LC neurons (Yamaguchi et al., 2018). Consistent with DBH’s known role in the conversion of DA to NE, mice with dbh KD had decreased NE concentrations in the LC as well as its projection areas. Functionally, this disruption resulted in increased latency of rapid eye movement (REM) and non-REM sleep-to-wake transitions in response to optogenetic stimulation of neurons within the LC (Yamaguchi et al., 2018). These results support a role for NE in arousal, consistent with the above-described work in zebrafish larvae. In addition to altered sleep-to-wake transitions, LC-specific dbh gene KD decreased the total wake time and increased the non-REM time during the dark phase of the light–dark cycle suggesting that NE release is important for the maintenance of wakefulness (Yamaguchi et al., 2018). These findings are consistent with previous work showing that spontaneous discharge of LC-NE neurons co-varied with the sleep–wake cycle in rats with the highest activity occurring during waking and virtually no activity occurring during REM sleep (Aston-Jones & Bloom, 1981). Yamaguchi et al.’s (2018) findings also support previous results from the group showing a causal relationship between LC-NE firing, sleep-to-wake transitions, and maintenance of wakefulness in mice (Carter et al., 2010) but expand them to implicate NE synthesis by DBH, and thus indirectly copper, in these processes.

9 |. PATHOPHYSIOLOGY OF COPPER IN THE NERVOUS SYSTEM

Copper is indispensable for neurological function as is homeostasis of this metal. Both excess copper and copper deficiency lead to severe neurological consequences as exemplified, respectively, by Wilson and Menkes diseases. Disorders of copper homeostasis or copper transporters can lead to a wide range of neurological symptoms including parkinsonism, sleep disturbances, seizures, muscle weakness, and ultimately result in neuronal and glial death. This section will describe important pathologies associate with copper dyshomeostasis.

10 |. WILSON DISEASE

Wilson disease (WD) is an autosomal recessive disorder of copper metabolism that leads to pathological copper accumulation, primarily in the liver and the brain. In 1993, two studies identified Atp7b, which maps to a region on chromosome 13, as the disease-causing gene (Bull et al., 1993; Tanzi et al., 1993). Since then, over 500 disease-causing mutations have been identified in ATP7B (Bandmann et al., 2015). ATP7B is highly expressed in the liver where it mediates excretion of copper into bile and delivers copper for incorporation into ceruloplasmin, which is the main copper transporting protein in the blood (Lutsenko, Barnes, et al., 2007). As the liver is the main site for metabolism of dietary copper, dysfunctional ATP7B leads to copper overload in hepatocytes, resulting in liver pathology, and excess copper, not bound to ceruloplasmin, is released into the circulation where it can pathologically accumulate in other tissues, including the brain (Członkowska et al., 2018). Indeed, brain copper in WD patients has been found to be 10–15 higher than in brains of control individuals (Poujois et al., 2017).

Although the presentation of WD patients is highly variable, patients typically present between the ages of 5–35 with clinical features that can include hepatomegaly and hepatitis, Kayser-Fleischer rings in the eyes, neurological manifestations including but not limited to dystonia, dysarthria, and parkinsonism (Lorincz, 2010; Machado et al., 2006; Svetel et al., 2001), as well as psychiatric symptoms such as depression, personality changes and psychosis (Bandmann et al., 2015; Członkowska et al., 2018). Although not typically described as a symptom of WD, a number of studies have shown sleep disturbances in WD patients (Cochen De Cock et al., 2018, 2021; Nevsimalova et al., 2011) and a meta-analysis found that 54% of patients suffered sleep disorders (Xu et al., 2020). Although the cause of sleep disturbances is not known and may be related to neurological damage, it is interesting to consider the previously described link between copper and the maintenance of circadian rhythms in this context, as copper dyshomeostasis may disrupt sleep cycles.

Given the variable clinical presentation, a diagnosis is usually made in combination with biochemical tests (Ryan et al., 2019; Salman et al., 2022). Additionally, given that treatment is most effective if given early in the disease course or before patients become overtly symptomatic (Dzieżyc et al., 2014), there is an increasing use of molecular diagnostics in first degree relatives of patients with WD to identify known homozygous or compound heterozygous mutations in such individuals (Ala et al., 2007; European Association for Study of Liver, 2012), although it should be noted that disease-causing mutations have been found to have variable penetrance (Bandmann et al., 2015). Treatment for WD consists of the use of copper chelating agents such as D-penicillamine, trientine and tetrathiomolybdate, which directly bind copper in blood and tissue and facilitate its excretion, and/or the administration of zinc salts that induce metallothioneins in intestinal cells, which, in turn, bind excess copper and thus interfere with copper absorption (Avan et al., 2022; Członkowska et al., 2018; Stremmel & Weiskirchen, 2021). It should also be noted that the zinc transporter ZIP4 is expressed in intestinal cells where it functions in zinc absorption (Cousins, 2010). Recent work has shown that human ZIP4 is able to transport Cu (II) with micromolar affinity when expressed in oocytes (Antala & Dempski, 2012). Therefore it is possible that zinc treatment leads to decreased enterocyte copper uptake through competition for uptake through ZIP4, although CTR1 is still likely to be the main copper transporter in this cell type based on knockout studies (Nose et al., 2006). While copper chelating agents are highly effective in improving hepatic function, the response rate for patients with neurological symptoms is less favourable (Stanković et al., 2023; Weiss et al., 2013), which may be reflective of irreversible brain damage. Additionally, the initiation of copper chelating agents can paradoxically worsen neurological symptoms during initial treatment (Bandmann et al., 2015; Weiss et al., 2013; Weiss & Stremmel, 2014), although the mechanism by which this occurs is still unknown.

Different brain regions show differential susceptibility to WD-induced copper toxicity. Pathological changes including astrogliosis, demyelination, and necrosis are most often reported in the basal ganglia, thalamus, cerebellum, and upper brainstem (Członkowska et al., 2018). It is of note that that many of these regions have been implicated in dystonia (Bhatia & Marsden, 1994) and patients with the neurodegenerative disease X-linked dystonia-parkinsonism show both basal ganglia and cerebellar pathology (Hanssen et al., 2018). Furthermore, both neuroimaging and neuropathological studies consistently show lesions in deep brain structures (da Costa et al., 2009; Mikol et al., 2005; Shribman et al., 2022). Thus, damage to brain regions that are vulnerable to copper-induced pathology correspond to neurological symptoms frequently experienced by patients with WD. In addition to the aforementioned lesions, patients with WD show widespread brain atrophy, which has been correlated with both neurological disease as well as increased levels of serum levels of non-ceruloplasmin bound copper (Shribman et al., 2022; Smolinski et al., 2019, 2022). In addition to neuronal pathology, a histopathological feature or WD is the presence of Opalski cells—large glial cells with a foamy cytoplasm and periodic acid-Schiff staining granules that are derived from degenerating astrocytes (Poujois et al., 2017). Two additional abnormal astrocytic cell are observed in WD patient brains—Alzheimers’s type I and type II astrocytes, both of which are misshapen and characterized as reactive astrocytes (Goldman, 2022; Meenakshi-Sundaram et al., 2008; Mikol et al., 2005; Poujois et al., 2017; Sosunov et al., 2020). Although Alzheimer’s type II astrocytes are associated with hepatic encephalopathy (Poujois et al., 2017), the cause of Alzheimer’s type I astrocytes, which are present in a number of neurodegenerative disorders, remains a mystery (Goldman, 2022), although evidence suggests that copper induces astrocyte toxicity in an oxidative stress dependent manner (Gale et al., 2023; Hu et al., 2016; Reddy et al., 2008). Unfortunately, although several animal models for WD exist (Hadrian & Przybyłkowski, 2021; Reed et al., 2018), most poorly recapitulate the neurological manifestations of disease and are better suited for investigation of hepatic pathology, making it difficult to investigate the causes of neuronal and glial cell death and injury in this disease. However, of the animals that do show neurological abnormalities, such as Atp7b ^−/− mice, evidence suggests that oxidative stress, respiratory chain defects, and copper mislocalization may all contribute to the observed neuropathology (Dong et al., 2015; Reed et al., 2018; Sauer et al., 2011).

Due to the number of identified mutations in ATP7B, there has been growing interest in finding a genotype–phenotype correlation for pathogenic variants. However, patients with the same mutation can have vastly different clinical presentations. For example, in a Japanese cohort of patients, of the two identified patients with the same c.2871delC mutation, one presented with fulminant hepatitis while the other only showed mild neurological findings at diagnosis (Okada et al., 2000). Additionally on a molecular level, even when pathogenic amino acid substitutions are in the same functional domain, the properties of these ATP7B variants vary widely in regards to stability, localization and catalytic and transport activity (Huster et al., 2012), making detailed molecular characterizations necessary to understand the consequences of each mutation on protein function. Moreover, the majority of patients with WD are compound heterozygous and recent evidence suggests that the properties of some ATP7B mutants, when expressed in vitro, differ when they are expressed alone versus when they are co-expressed with other mutants, suggesting that a functional interaction between pathogenic variants could exist (Roy et al., 2020). However, characterizing the molecular mechanisms underlying defective ATP7B function is still an important step in developing new treatment strategies—especially those that may better improve neurological outcomes. The ATP7B-H1069Q mutation is the most frequent disease-linked mutation in European and North American populations. Over 50% of patients from these populations harbour the mutation in homozygosity or compound heterozygosity (Gomes & Dedoussis, 2016). Characterization of this aberrant protein in hepatocyte-like cells derived from WD patient fibroblasts demonstrated that while a significant portion of ATP7B-H1069Q is retained in the endoplasmic reticulum where it is rapidly degraded, some protein correctly traffics to the TGN. Furthermore, exposing cells to copper increases trafficking of ATP7B-H1069Q from the TGN to endolysosomal compartments, suggesting it retains some functional properties of the wild-type protein (Parisi et al., 2018). Using proteomics, a recent study identified that this mutation promoted ATP7B interaction with heat shock protein 70 (HSP70), accelerating its degradation in the ER. Importantly, suppression of HSP70 reduced ER retention and degradation of the mutant protein and promoted transport to the Golgi and cell surface, suggesting that HSP70 could be a viable pharmacological target to increase proper localization of ATP7B-H1069Q. These results prompted a three-dimensional similarity search of FDA-approved drugs that resembled the allosteric HSP70 inhibitor used in previous experiments, identifying domperidone, an anti-emetic agent, as a possible drug that could be repurposed for WD. Indeed, domperidone increased ATP7B-H1069Q levels and promoted its export from the ER to the Golgi in HepG2 cells (Concilli et al., 2020). Although these results only show the potential utility of domperidone for the treatment of WD in vitro, this study demonstrates the importance of functional characterization of mutants, even when a clear genotype–phenotype correlation cannot be established.

11 |. ATP7A-LINKED NEURODEGENERATIVE DISORDERS

Unlike AT7B, where mutations in the gene result in a single known clinical phenotype (Wilson disease), mutations in ATP7A cause three X-linked neurodegenerative diseases—Menkes disease, occipital horn syndrome (OHS), and X-linked distal hereditary neuropathy (dHMN) (Kaler, 2011). Menkes disease is the most severe of the three and is caused by mutations in the gene that result in its loss of function. OHS is considered to be a milder version of Menkes and is typically caused by a ‘leaky’ splice junction in ATP7A, resulting in only a portion of transcripts from this allele containing the mutation, or mutations that result in a hypofunctional protein (Dagenais et al., 2001; Kaler, 2011; Kaler et al., 1994; Møller et al., 2000). dHMN is the least severe of the three conditions and has a distinct clinical phenotype from Menkes and OHS, which is thought to be due to mutations in ATP7A that impair but do not abolish correct mRNA splicing. Of note, in dHMN these mutations occur almost exclusively in the carboxyl half of the protein (Kaler, 2013; Kennerson et al., 2010).

Menkes disease and OHN reside on the same clinical spectrum with overlapping symptoms including lax skin and joints, reduced bone density, and vascular tortuosity, although Menkes disease patients typically present within the first 2–3 months of life while OHS patients can present during mid-childhood or later (Kaler, 2011, 2013; Menkes, 1988). Additionally, Menkes patients have coarse, hypopigmented hair (Ramani & Parayil Sankaran, 2023). Biochemical signs of both diseases include low serum copper and ceruloplasmin (Kaler, 2013). This copper deficiency is a result of a failure of ATP7A in enterocytes to efflux copper into the circulation, leading to enterocyte copper accumulation (Kaler, 2011; Ramani & Parayil Sankaran, 2023). However, Menkes disease and OHN diverge in the severity of their neurological manifestations. Neurological features of Menkes disease can include seizures, hypotonia, severe developmental delay, and brain atrophy, which unfortunately contribute to death by 3 years of age (Kaler, 2013; Zlatic et al., 2015). Neuropathological examination of Menkes patient brains shows widespread atrophy of both white and grey matter and neuronal loss that is most pronounced in the cerebral cortex but also affects the hippocampus, striatum, hypothalamus, and thalamus to variable degrees (Zlatic et al., 2015). In contrast, the primary neurological symptom associated with OHS is dysautonomia (Kaler, 2013). The difference in severity of neurological symptoms is thought to be a result of residual ATP7A activity in OHS patients, estimated to be approximately 20%–30% due to a mixture of properly and aberrantly spliced mRNA (Gu et al., 2001; Kaler et al., 1994; Tang et al., 2006). However, it has been reported that levels as low as 2%–5% of correctly spliced mRNA permits this milder phenotype (Møller et al., 2000). OHS patients also have a distinctive, and pathognomonic radiological feature—occipital exostoses (excess bone growths) (Kaler et al., 1994). Many symptoms of Menkes and OHS can be attributed to lack of metalation of cuproenzymes. For example, tyrosinase, which catalyses the production of melanin, is a copper-containing enzyme and lack of catalytic activity contributes to the hypopigmented hair seen in classical Menkes. Similarly, skin and joint laxity in Menkes and OHS patients can be attributed to lack of metalation of lysl oxidase—a cuproenzyme that is important for the production of extracellular matrix proteins like collagen and elastin (Ramani & Parayil Sankaran, 2023).

While it has been posited that the neurodegeneration observed in Menkes disease is a result of the lack of metalation of cytochrome c oxidase (Prasad et al., 2011; Sparaco et al., 1993; Zlatic et al., 2015), in part due to ultrastructural alterations observed in the mitochondria of neurons in Menkes patient post-mortem brain tissue (Hirano et al., 1977; Yoshimura & Kudo, 1983), this does not fully explain why OHS patients suffer from symptoms associated with loss of cuproenzyme activity but are spared the severe neurological consequences seen in Menkes—especially considering evidence that copper flux is greater in differentiating neurons than proliferating cells and that ATP7A levels are higher in differentiated neurons compared to undifferentiated cells (Hatori et al., 2016; Lutsenko, 2021). However, recent evidence from a murine model of Menkes disease supports the hypothesis that pathogenic symptoms may be a result of lack of mitochondrial copper (Guthrie et al., 2020). Atp7a-mutant mice treated with the drug elesclomol (ES)—a lipophilic molecule that binds copper with high affinity (Wu et al., 2011; Yadav et al., 2013)—in a complex with Cu (II) (CuES) survived significantly longer than vehicle-treated animals or those treated with copper histidine (CuHIS)—a hydrophilic copper complex (median age of survival for vehicle treated mice—14 days; median age of survival for CuHIS treated mice—18 days; median age of survival for CuES treated mice—203 days). These survival benefits have been recapitulated in a Caenorhabditis elegans model of Menkes disease (Yuan et al., 2022). Neurologically, CuES treated mice had significantly better neuromotor assessments and did not suffer from seizures, while, on a neuropathological level, CuES-treated animals brains showed preservation cortical and hippocampal neurons as well as a preserved Purkinje layer in the cerebellum (Guthrie et al., 2020). Importantly, mitochondrial Cox1 levels, one of the two catalytic subunits of cytochrome c oxidase that contains a copper-binding site (Capaldi, 1990; Timón-Gómez et al., 2018), showed a 14% increase, suggesting that delivery of copper to the mitochondria may underlie the improvements in Menkes pathology. This is consistent with in vitro studies showing that ES escorts copper to the mitochondria and rescues cytochrome c oxidase levels (Soma et al., 2018; Zulkifli et al., 2023). It must be noted, however, that CuES metalates cuproenzymes in other cellular compartments such as the Golgi (Guthrie et al., 2020). While these results are exciting and suggest that CuES treatment may be a viable treatment strategy for disorders of copper metabolism, whether CuES will be useful for the treatment of Menkes disease likely will rely on early detection as neocortical and cerebellar Atp7A levels peak early in development (Niciu et al., 2006), as it has been suggested that treatment of Menkes disease patients should commence within 10 days of birth due to the early onset of disease (Kaler, 2013).

X-linked distal motor neuropathy, as its name suggests, is an X-linked disorder that results in progressive atrophy and weakness of distal muscles as well as abnormal nerve conduction velocities, but unlike Menkes and OHS, patients with dHMN do not present with any CNS manifestation or other clinical abnormalities. Moreover, there does not appear to be any derangement of copper metabolism as reflected by normal serum copper and ceruloplasmin. Additionally, in contrast to Menkes and OHS, dHMN is most often an adult-onset disease. These differences in dHMN and the two other ATP7A-linked diseases reflect its milder disease course likely since mutations linked to dHMN do not cause as severe a loss of normal ATP7A function. Indeed, the first mutations linked to this disease were missense mutations (p.P1386S and p.T994I) occurring in the conserved C-terminal portion of ATP7A (Kennerson et al., 2010)—a region that does not directly involve the transporter’s known functional domains (Yu, Dolgova & Dmitriev, 2017). Additionally, in vitro characterization of these mutations revealed no change in ATP7A splicing, or a reduction in mRNA or protein levels, although both ATP7A mutants showed defects in trafficking from the TGN to the plasma membrane in response to copper supplementation in patient-derived fibroblast (Kennerson et al., 2010). However, when expressed in HEK293 and differentiated NSC-34 cells, a motor-neuron like cell line (Cashman et al., 1992), both mutant ATP7A proteins show increased plasma membrane localization under basal copper conditions compared to WT ATP7A (Yi et al., 2012). Together, these results suggest that trafficking defects are a characteristic of dHMN-linked mutations in ATP7A. It should be noted that fibroblasts derived from patients with the P1386 mutation showed an intermediate copper-retention phenotype (Kennerson et al., 2010), suggesting their ability to appropriately efflux copper is compromised by the mutation.

Although the C-terminus does not contain metal binding domains, which reside in its N-terminus, and is not known to participate in metalation of cuproenzymes (Yu, Dolgova & Dmitriev, 2017), more recent research has demonstrated that the carboxyl half of the protein is critical for anterograde and retrograde trafficking of ATP7A through its interaction with adaptor protein (AP) complexes 1 and 2 (Yi & Kaler, 2015)—which play important roles in the regulation of intracellular vesicle trafficking (Robinson, 2004). Of note, the Atp7a missense mutation P1386S partially disturbs interactions with both AP1 and AP2 and when the mutant protein is expressed in differentiated NSC-34 cells it leads to abnormal axonal localization of Atp7a. Whereas wild-type Atp7a is normally expressed in the somatodendritic compartment and traffics to axons under conditions of elevated copper, Atp7a ^P1386S is axonally localized under basal copper conditions (Yi & Kaler, 2015). The authors speculate these finding may reflect a role for AP1 in tethering Atp7a to the TGN until copper triggers its release. Intriguingly, further investigation of the ATP7A T994I missense mutation, also linked to dHMN, revealed an abnormal interaction with valosin-containing protein 97 (p97/VCP), which, like AP1 and AP2, functions in endolysosomal trafficking (Bug & Meyer, 2012). While wild-type ATP7A and ATP7A^P1386S do not strongly interact with this protein, immunoprecipitation experiments demonstrated a selective interaction between p97/VCP and ATP7A^T994I. Moreover, siRNA knockdown of p97/VCP in HEK293 cells significantly increased the amount of ATP7A^T994I retained in the TGN, directly implicating the aberrant interaction between these two proteins in the mislocalization of ATP7A^T994I at the plasma membrane (Yi et al., 2012). Thus, altered interactions between endolysosomal proteins appears to be a common mechanism underlying trafficking defects observed in dHMN-linked ATP7A mutants. Furthermore, in vitro characterization revealed the importance of the C-terminus of ATP7A in proper recycling between the TGN and the plasma membrane in response to changing intracellular copper levels. Although it is unclear why these missense mutations in ATP7A cause selective degeneration of motor neurons, it is interesting to note that a missense mutation in carboxyl half of ATP7A (p.M1311V) was recently linked to a patient with ALS and characterization of patient-derived induced pluripotent stem cell motor neurons showed trafficking defects reminiscent of those described for dHMN-linked ATP7A mutants (Bakkar et al., 2021). Therefore, motor neurons may be a cell type that is selectively vulnerable to copper dyshomeostasis.

12 |. SLC31A1 MUTATIONS

Two recent studies reported severe neurological manifestations in patients with the first-described mutations in the high affinity copper transporter CTR1, encoded by the SLC31A1 gene. In one case (Batzios et al., 2022), identical male twins homozygous for a missense CTR1 variant (p.95R > H) presented with infantile seizures, diagnosed at 17 weeks of age, and brain atrophy, a profile reminiscent of severe CNS copper deficiency, such as that seen in Menkes disease (Kaler, 2013; Prasad et al., 2011). Indeed, despite having normal serum copper and ceruloplasmin levels, both twins had low CSF copper (Batzios et al., 2022). Of note, mice heterozygous for Ctr1 knockout show no significant differences in intestinal, kidney, or liver copper levels compared to their wild-type littermates, however brain copper is reduced by ~50% suggesting proper CTR1 function is essential for transport of copper into the CNS (Kuo et al., 2001) At 9.5 months the infants were largely unresponsive while awake; treatment with CuHIS was begun, which modestly increased serum copper levels. This treatment led to some clinical improvements by 2 years of age, including behavioural responses and improved muscle tone. Unfortunately, seizures could not be alleviated, which likely reflects the presence of neurological abnormalities and brain damage sustained prior to initiation of treatment as suggested by improved clinical benefit and survival in Menkes patients who received copper replacement in the neonatal period as compared to those treated later (Kaler et al., 2008). The authors suggest that CuHIS treatment likely enhanced copper transport via residual CTR1 activity, or perhaps permeated cells via a CTR1-independent process resulting in the observed modest clinical improvement (Batzios et al., 2022). Eventually copper replacement was stopped at the request of the parents. A second study reported a missense CTR1 variant (~236 T > C) in a newborn infant presenting with severe low serum copper, respiratory distress, multifocal brain hemorrhages and seizures (Dame et al., 2023). The infant died at 1 month of age after cessation of ventilation intervention. These cases thus add CTR1 mutations to the list of copper transport-associated variants linked to severe copper deficiency and neural deficits.

13 |. AMYOTROPHIC LATERAL SCLEROSIS

Amyotrophic lateral sclerosis (ALS) is the most common adult motor neuron disease with an incidence of approximately 2 per 100,000 individuals per year (Hardiman et al., 2017; Masrori & Van Damme, 2020). ALS is characterized by selective death of upper and lower motor neurons causing progressive muscle atrophy, weakness, and spasticity. Paralysis eventually results and death is often a result of respiratory failure, typically within 2–5 years of ALS onset (Masrori & Van Damme, 2020; Rothstein, 2009). Approximately 10% of ALS is inherited (familial ALS) while close to 90% of cases are sporadic. Although over 30 mutations have been identified, mutations in four genes—C9orf72, TARDP, FUS and SOD1—comprise nearly 70% of familial ALS (fALS) cases (Hardiman et al., 2017). Of these four genes, mutations in the Cu/Zn superoxide dismutase SOD1, a copper-dependent enzyme that is important in detoxification of superoxide radicals, has been studied the longest as it was first identified as an ALS-associated mutation in the early 1990s (Rosen et al., 1993; Siddique et al., 1991). As such, the number of identified molecular mechanisms underlying SOD1-associated neuropathology/neurotoxicity including, but not limited to, protein misfolding, proteasome impairment, excitotoxicity, oxidative stress, endoplasmic reticulum stress, impaired axonal transport, axonopathy, inflammation and mitochondrial dysfunction, are vast (Ferraiuolo et al., 2011). Unfortunately, there does not yet exist effective treatment strategies for this disease. Despite initial excitement over antisense oligonucleotide (ASO) therapy for SOD1 ALS (McCampbell et al., 2018; Smith et al., 2006), a strategy that has been successfully used to treat conditions such as spinal motor atrophy (Scoles et al., 2019; Stein & Castanotto, 2017), a phase III study of an ASO targeting this mutant gene failed to improve clinical end points and was associated with adverse events (Miller et al., 2022). Thus, there is still a great need to understand underlying molecular mechanisms of this disease to develop targeted treatment strategies.

There have been several animal models developed to model SOD1 ALS, with overexpression of the mutant protein being sufficient to cause disease in mice. One common feature to both animal models of SOD1 ALS and human disease is the deposition of aggregate SOD1 protein in neuronal and glial inclusions (Philips & Rothstein, 2015;Stieber et al., 2000; Watanabe et al., 2001). One of the most widely used rodent models of SOD1 ALS is the SOD1 ^G93A model that harbours a mutant form of human SOD1 in which an alanine is substituted for glycine at codon 93 (Rosen et al., 1993). This model recapitulates important features of human ALS including the degeneration of spinal motor neurons, muscle weakness, and paralysis, as well as histological features such as pathological inclusions in neurons and glia, some of which contain mutant Sod1 (Stieber et al., 2000; Watanabe et al., 2001). Intriguingly, recent studies examining mutant SOD1 suggest that the aggregation of this protein may be due to a lack of proper metalation. The wild-type SOD1 is a homodimer with each subunit binding a one copper ion and one zinc ion. While the copper ion is necessary for catalytic activity of the enzyme, the zinc ion is thought to stabilize the structure of the SOD1 (Sirangelo & Iannuzzi, 2017). Multiple studies have demonstrated that metal deficiency increases the aggregation propensity of both wild-type and mutant SOD1 (Banci et al., 2007; Oztug Durer et al., 2009; Tiwari et al., 2009). In support of the theory that SOD1 aggregation may be linked to problems with metalation, a study utilizing NMR spectroscopy to investigate the structure of fully metalated wild-type SOD1 and the G93A SOD1 variant found this mutation resulted in local destabilization of protein structure at the site of the mutation, but more importantly selectively destabilized the metal binding region of SOD1 (Museth et al., 2009). Destabilization of the metal binding site may reduce the metal affinity or binding ability of G93A SOD1, resulting in the pathological aggregation observed in mouse models and humans with this mutation. Additionally, evidence suggests that metal deficiency destabilizes the SOD1 protein and structural distortion of apo-SOD1 results in its aberrant oligomerization (Strange et al., 2007; Teilum et al., 2009).

Based on the findings described above, it would be expected that mutations in the SOD1 copper chaperone CCS would also lead to SOD1 aggregation due to lack of copper metalation. However, surprisingly, an investigation of Ccs deficiency showed that while motor neurons in Ccs ^−/− mice showed significantly diminished incorporation of copper into wild-type and mutant Sod1, the absence of Ccs did not affect the onset and progression of motor neuron disease in fALS-linked SOD1-mutant mice. This finding suggests that motor neuron disease in Sod1 ALS is independent of Ccs-mediated copper loading (Subramaniam et al., 2002). This result is especially surprising given that finding that this SOD1 copper chaperone is necessary for the activation of the enzyme (Wong et al., 2000), and that Ccs ^−/− mice show increased vulnerability to the oxidative stressor paraquat (Subramaniam et al., 2002), which exerts its toxicity, in part, through the production of superoxide anions (Alizadeh et al., 2022). Thus, it is possible that, in vivo, protein misfolding independent of SOD1 copper metalation leads to a toxic gain-of-function independent. Supporting this point, it has been posited that the copper in zinc-deficient SOD1 is more accessible to intracellular reductants and that this increased accessibility of copper may result in redox cycling of the metal and a subsequent increase in oxidative and nitrative stress (Crow et al., 1997; Estévez et al., 1999; Trumbull & Beckman, 2009). This potential mechanism of toxicity, however, remains some-what controversial within the field.

Additionally, multiple in vitro investigations have shown that while copper is essential for the catalytic activity of SOD1, metalation by zinc, rather than copper, is crucial for the structural stability of the enzyme (Forman & Fridovich, 1973; Kayatekin et al., 2008; Potter et al., 2007). Furthermore, it has been demonstrated in several in vitro systems that zinc-deficient SOD1 as well as ALS-linked SOD1 mutants with reduced zinc-binding are more prone to aggregation and fibrillation than wild-type SOD1 (Das et al., 2023; Oztug Durer et al., 2009; Roberts et al., 2007; Sannigrahi et al., 2021). These results may explain, in part, the finding that low-dose dietary zinc supplementation extended the lifespan of SOD1 ^G39A transgenic mice (Ermilova et al., 2005), as this mutation is known to reduce Sod1’s affinity for zinc (Crow et al., 1997). Despite the possible toxic-gain of function associated with Sod1 aggregation, it should be noted that mice with neuronal specific deletion of Sod1 show increased denervated neuromuscular junctions, loss of axonal integrity, and accelerated age-related loss of skeletal muscle (Pollock et al., 2023). This suggests that loss of SOD1 function, secondary to lack of copper metalation, contributes to the ALS phenotype.

Supporting this possibility, multiple studies investigated the effect of oral administration of the metallo-complex Cu (II)-atsm [diacetylbis(4-methylthiosemicarbazonato)copper (II)], a low-molecular weight, lipophilic compound that is able to cross biological membranes, including the blood brain barrier (Dearling & Blower, 1998; Nikseresht et al., 2023). Treatment of SOD1 ^G93A mice found that this compound improved the neurological phenotype of the animals, as assessed by motor performance tests, decreased spinal cord pathology, delayed the time to symptom onset, and improved survival in a dose-dependent manner (Hilton et al., 2017; Soon et al., 2011). Transdermal administration of Cu (II)-atsm also significantly extended survival and decreased the percentage of copper-deficient mutant Sod1 in the spinal cord (Williams et al., 2016), suggesting that the improved neurological phenotype may be related to increased Sod1 function. Indeed, supplementation with Cu (II)-atsm increased the catalytic activity of mutant Sod1 in the spinal cord of transgenic animals (Hilton et al., 2017), likely contributing to the reduction in oxidative stressors markers observed in the spinal cords of SOD1 ^G93A mice (Hilton et al., 2017; Soon et al., 2011). Importantly, when Cu (II)-atsm was administered to non-transgenic animals, no increase in Sod1 activity and no change in cytochrome c activity were observed, suggesting that Cu (II)-atsm does not alter the activity of copper-replete cuproenzymes (Hilton et al., 2017). In addition to extending lifespan of SOD1^G39A mice and increasing the catalytic activity of mutant SOD1^G39A, Cu (II)-atsm has been shown to improve locomotor function and survival of SOD1^G37R mice (Hilton et al., 2018; McAllum et al., 2013; Roberts et al., 2014). Like SOD1^G93A mice treated with Cu (II)-atsm, SOD1^G37R mice that received this copper supplementation show decreased levels of copper-deficient SOD1 in their spinal cords (Hilton et al., 2018; Roberts et al., 2014). Supporting the idea that lack of copper metalation of Sod1 contributes to the ALS phenotype in these mutant animals is the finding that overexpression of human CTR1 in the SOD1 ^G37R mouse spinal cord elicited a modest improvement in Sod1 activity (Hilton et al., 2018). It should be noted that misfolded Sod1 levels do not track with disease progression in SOD1 ^G37R mice (Hilton et al., 2018), further supporting the notion that loss of Sod1 function, due to lack of copper metalation, contributes to the ALS phenotype, rather than toxic-gain-of-function secondary to Sod1 misfolding (Figure 3).

FIGURE 3.

Consequences of SOD1 dysfunction. (a) Superoxide dismutase (SOD1) is an enzyme that plays a key role in antioxidant defence by catalysing the conversion of superoxide to oxygen and hydrogen peroxide. SOD1 requires the binding of both zinc and copper for proper function. Copper is essential for catalytic activity of the enzyme, while zinc provides structural stability of SOD1. Zinc-deficient SOD1 has a much higher propensity to aggregate and form pathological aggregates while copper-deficient SOD1 is unable to metabolize superoxide leading to increased oxidative stress. Toxicity from misfolded SOD1 aggregates and oxidative stress contribute to the selective motor neuron death seen in amyotrophic lateral sclerosis (ALS). (b) Recent work has shown that when mice with SOD1 mutations receive supplementation with lipophilic compound CuII (atsm), they not only perform better on tests of motor function, such as the rotarod test, compared to sham animals, but have increased catalytic activity of mutant SOD1 and decreased copper deficient SOD1 in their spinal cords.

Intriguingly, in addition to its therapeutic effect on rodent motor neurons, — Cu (II)-atsm treatment may also improve astrocyte function in patients with SOD1 mutations (Dennys et al., 2023), which is significant as glia reprogrammed from fibroblasts of patients with SOD1 mutations have been shown to induce hyperexcitability and cell death in healthy control motor neurons (Di Giorgio et al., 2008; Ferraiuolo et al., 2016; Marchetto et al., 2008). A recent in vitro study demonstrated that when astrocytes from SOD1-linked ALS patient-derived fibroblasts were pre-treated with 1 μM Cu (II)-atsm before co-culture with healthy murine embryonic stem cell derived motor neurons, motor neuron survival was increased compared to vehicle-treated or untreated reprogrammed astrocytes. Intriguingly, however, Cu (II)-atsm does not improve astrocyte dysfunction by increasing SOD1 activity but by decreasing mitochondrial respiration, which is increased in these cells, and by increasing glycolysis (Dennys et al., 2023). These finding suggest that the beneficial effects of Cu (II)-atsm are multifaceted and involve the improvement of both neuronal and glial function in cells carrying SOD1 mutations. While these results are promising, it is necessary to acknowledge that treatments that have successfully delayed disease onset and extended lifespan in animal models of ALS, such as the antibiotic minocycline (Kriz et al., 2002; Zhu et al., 2002), have not translated to success in clinical trials (Gordon et al., 2007), and, as such, these findings should therefore be interpreted with caution.

14 |. MULTIPLE SCLEROSIS

Multiple sclerosis (MS) is a disease of the CNS affecting oligodendrocytes, the myelinating cells of the CNS, causing their cell death leading to the progressive demyelination in the brain and spinal cord. A number of animal models have been developed to study this neurodegenerative disease, including cuprizone (CPZ) treatment. The CPZ model allows the study of processes underlying primarily non-autoimmune demelyination and remyelination. This contrasts with the experimental autoimmune encephalomyelitis (EAE) model, in which mice are artificially sensitized to autoantigens. CPZ is a copper chelator, and the CPZ model of MS involves the addition of the compound to mouse chow. Despite the widespread use of this model to study MS, no consensus yet exists as to CPZ’s mechanism of action although it is thought that it may exert its toxicity by disrupting mitochondrial enzymes of the respiratory chain through copper chelation, as these require copper as a cofactor for proper function. It is posited that this disruption leads to oxidative stress and eventually results in oligodendrocyte cell death along with activation of microglia and astrocytes (Kipp et al., 2017).

Recently, however, several studies have challenged that demyelination in the CPZ is a result of copper chelation and direct oligodendrocyte death. If CPZ causes demyelination through copper chelation, then copper supplementation may offset some of its damaging effects. However, the addition of copper in stoichiometric excess of CPZ to the chow of C57Bl/6 mice did not reduce CPZ-induced demyelination (Morgan et al., 2022). Furthermore, when equivalent doses of other copper chelators including neocuproine and D-pencillamine were added to mouse chow, they failed to induce CNS demyelination, suggesting that CPZ may be acting through a distinct mechanism. Supporting this notion, the addition of D-pencillamine to the CPZ diet paradoxically prevented demyelination in a dose-dependent manner (Morgan et al., 2022), rather than exacerbating the CPZ-phenotype, as one might expect with the addition of another copper chelator. These data suggest that the binding of the CPZ to copper may give rise to a toxic species. Indeed, there is evidence that the binding of CPZ to Cu (II) gives rise to an unusually high valent Cu (II) state within a Cu³⁺(CPZ)₂⁻ complex (Messori et al., 2007; Yamamoto & Kuwata, 2009). The high oxidation state of copper could allow it to participate in reactions that lead to peroxidation of lipid-rich structures, such as myelin. Thus, evidence of CPZ-induced lipid peroxidation and subsequent oligodendrocyte death (Jhelum et al., 2020) may be a result of copper toxicity rather than copper deficiency. Furthermore, the finding that CPZ leads to ferroptosis of oligodendrocytes (Jhelum et al., 2020) is consistent with recent work demonstrating that copper can amplify sensitivity of cells to this cell death pathway (Xue et al., 2023).

Supporting the idea that copper toxicity rather than deficiency may lead to or exacerbate demyelination comes from a study that examined MS lesions in human MS patients as well as in EAE mice and CPZ mice. Colombo et al. (2021) found that in MS patients, while circulating serum copper levels were unchanged, there was a striking increase in the expression of CTR1, ATP7A, and ATP7B in GFAP-positive astrocytes during neuroinflammation. These data could indicate that in MS patients, a rise in astrocytic copper uptake from the bloodstream and subsequent redistribution to other cell types such as oligodendrocytes drives cell death and demyelination. This idea is supported by the strong inverse correlation found between CTR1 levels and myelin content in areas of active MS lesions and the direct correlation between CTR1 or ATP7B and astrogliosis, in as well as the upregulation of all copper transporters in chronic inactive lesions, with high levels of CTR1 and AT7B on GFAP-positive astrocytes in human brain samples. Consistent with what the authors observed in the human CNS, both EAE mice and CPZ mice showed intense upregulation of ATP7A, ATP7B, and CTR1 on astrocytes in the spinal cords in comparison to control mice (Colombo et al., 2021). Together these data suggest that astrocyte-mediated copper dyshomeostasis and copper toxicity is a common mechanism underlying demyelination in MS and animal models of MS.

15 |. ACQUIRED COPPER DEFICIENCY

Although Menkes disease demonstrates the devastating neurological consequences of early copper deficiency, acquired copper deficiency has become well-recognized as a cause of neurological disease (Jaiser & Winston, 2010). Risk factors for acquired copper deficiency include bariatric surgery, prolonged total parental nutrition, malabsorption syndromes such as celiac disease, and zinc overload (Btaiche et al., 2011; Doherty et al., 2011; Goodman et al., 2009; Jaiser & Winston, 2010). In the case of zinc-induced hypocupremia, excess enteral zinc induces metallothioneins in intestinal cells (Fischer et al., 1981; Hall et al., 1979; Hardyman et al., 2016). As metallothioneins have a higher affinity for copper than for zinc (Krężel & Maret, 2017, 2021; Romero-Isart & Vasák, 2002), induction of these proteins by zinc leads to reduced copper absorption into the circulation. In fact, because of zinc’s ability to interfere with copper absorption and facilitate copper excretion through the sloughing of intestinal cells containing metallothionein-bound copper, oral zinc is used as a therapeutic strategy in conditions of copper excess, such as Wilson disease (Avan et al., 2022; Stremmel & Weiskirchen, 2021) (discussed above). Patients with hypocupremia typically present with both hematological abnormalities such as anaemia and neutropenia (Lazarchick, 2012), in addition to neurological abnormalities, which often resemble subacute combined degeneration of the spinal cord, a complication of vitamin B12 deficiency involving sensory deficits, paresthesia, weakness, and ataxia (Qudsiya & De Jesus, 2021), due to the involvement of the dorsal column (Kumar et al., 2004). Although neurological symptoms due to hypocupremia can be severe and include generalized weakness, impaired sensation of position and vibration, gait ataxia, they can also be useful in distinguishing between myelodysplastic syndrome and hypocupremia. Indeed, in one case, the presence on neurological symptoms in a patient with suspected myelodysplastic syndrome, who was being evaluated for a bone marrow transplant, prompted her physicians to test her serum copper levels, which were undetectable, with symptoms resolving following copper supplementation (Kumar et al., 2005). However, it is important to identify and treat hypocupremia early as copper supplementation can resolve hematological abnormalities, neurological manifestations are not always reversible (Gabreyes et al., 2013; Rohm et al., 2019).

While many patients with hypocupremia present with hyperzincemia, the source of increased zinc levels is not always identifiable (Hedera et al., 2003; Merza et al., 2015). However, one surprising and intriguing cause of hyperzincemia leading to myeloneuropathy was the use of zinc-containing dental adhesives. Indeed, a number of reports described the development of hematological abnormalities and hyperzincemia alongside progressive myeloneuropathy that could be traced back to excessive use of dental adhesive (Carroll et al., 2017; Cathcart & Sofronescu, 2017; Doherty et al., 2011; Jamal et al., 2021; Nations et al., 2008; Tezvergil-Mutluay et al., 2010). In some cases, the neurological damage to the dorsal column of the spinal cord was severe enough to be detected on magnetic resonance imaging (Carroll et al., 2017; Gabreyes et al., 2013). The widespread recognition of hyperzincemia-mediated hypocupremia has led to the development of new, zinc-free formulations of dental adhesives; however, some popular dental adhesive brands still contain zinc. Thus, dentists should advise patients on the importance of properly fitting dentures and physicians should be aware of this association when evaluating patients with hyperzincemia and hypocupremia of unknown origin.

16 |. CONCLUSIONS

Copper is a critical micronutrient for growth, overall development, as well as brain development. Disorders of copper metabolism such as Menkes disease demonstrate the devastating consequences of early copper deficiency on neurological function (Kaler, 2011), while the neurological symptoms associated with acquired copper deficiency such as myeloneuropathy (Btaiche et al., 2011; Doherty et al., 2011; Rohm et al., 2019) highlight the continual need for copper for proper neuronal function. Indeed, the metal plays critical roles in neuronal metabolism, antioxidant defence, as well as neurotransmitter and neuropeptide synthesis (Hatori & Lutsenko, 2013; Lutsenko et al., 2019; Wen et al., 2021). Moreover, copper can act as a neuromodulator through its effects on NMDAR and GABAR activity (Marchetti et al., 2014; Narahashi et al., 1994; Peters et al., 2011; Sharonova et al., 1998; Trombley & Shepherd, 1996; Vlachová et al., 1996), has been found to alter network excitability (Dodani et al., 2014; Maureira et al., 2015), and can influence behaviour through its indispensable role in the synthesis of norepinephrine (Xiao et al., 2018; Yamada & Prosser, 2014). Although only briefly referenced in this review of the literature, the recent advent of fluorescent probes that are highly selective for copper and sensitive to oxidation states now allows for the investigation of labile Cu(I) in living cells (Chung et al., 2019; Morgan et al., 2019) and will likely lead to new and exciting discoveries regarding the physiological and pathophysiological mechanisms of copper signalling in the brain.

Abbreviations:

ALS

amyotrophic lateral sclerosis

AMPA

alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

BCS

bathocuproine disulfate

CCS

copper chaperone for superoxide dismutase

COG

complex conserved oligomeric Golgi complex

CPZ

cuprizone

Cu (II)-atsm

diacetylbis(4-methylthiosemicarbazonato)copper (II)

CuES

copper elescomol

CuHIS

copper histidine

DBH

dopamine β-hydroxylase

dHMN

X-linked distal hereditary neuropathy

DTT

dithiothreitol

EAE

experimental autoimmune encephalomyelitis

GSH

glutathione

GSSG

glutathione disulfide

LC

locus coeruleus

MS

multiple sclerosis

NE

norepinephrine

NMDA

N-methyl-d-aspartate

OHS

occipital horn syndrome

PAM

peptidylglycine α-amidating

PrP^C

cellular prion protein

RA

retinoic acid

REM

rapid eye movement

SCN

suprachiasmatic nucleus

SOD1

superoxide dismutase 1

Steap

six-transmembrane epithelial antigen of the prostate

TGN

trans-Golgi network

TTM

tetrathiomolybdate

WD

Wilson disease