Copper signalling: causes and consequences

Copper-containing enzymes perform fundamental functions by activating dioxygen (O 2 ) and therefore allowing chemical energy-transfer for aerobic metabolism. The copper-dependence of O 2 transport, metabolism and production of signalling molecules are supported by molecular systems that regulate and preserve tightly-bound static and weakly-bound dynamic cellular copper pools. Disruption of the reducing intracellular environment, characterized by glutathione shortage and ambient Cu(II) abundance drives oxidative stress and interferes with the bidirectional, copper-dependent communication between neurons and astrocytes, eventually leading to various brain disease forms. A deeper understanding of of the regulatory effects of copper on neuro-glia coupling via polyamine metabolism may reveal novel copper signalling functions and new directions for therapeutic intervention in brain disorders associated with aberrant copper metabolism.

An evaluation of the effects of copper under normal and pathological conditions depends on an accurate knowledge of copper concentrations present in vivo. In spite of this, a bewildering feature of efforts to examine the role of copper in biological processes is the limited data available on the relative distribution of copper between organs, tissues, cell types and sub-cellular compartments in mammals [ 2 , 44 , 76 – 83 ]. From a practical viewpoint, the lack of the information makes it unrealistic to determine the recommended concentration of copper in drinking water. In addition to its biological variance, the significant differences in copper levels that exist in habitats and diets may also explain difficulties in determining the impact of copper on biological systems [ 2 ]. Moreover, multiple comparisons of existing data are compromised by the use of varying techniques, characteristically atomic absorption spectroscopy (AAS), flameless atomic absorption spectroscopic technique (FAAS), inductively coupled plasma-atomic emission spectrometry (ICP-AES) (Tables 1 and 2 ) and radiotracer detection or diverse sample preparing protocols. Data obtained by FAAS on brain tissue samples taken from 38 brain regions of 7 males within 2–4 h after death showing no macroscopic signs of disease [ 77 , 78 ] disclosed significant copper concentration differences between brain areas, grey versus white matter cells, and between individuals. Brain copper concentrations were inversely correlated with age. It is worth noting that measurements of total copper levels may not necessarily reflect the biologically active metal pools [ 84 ]. Table 1 Average concentration of copper in human organs Sumino et al. [ 82 ] Margalioth et al. [ 44 ] Hamilton et al. [ 364 ] Yoo et al. [ 83 ] Lech & Sadlik [ 365 ] Haswell [ 79 ] Bárány et al. [ 76 ] FAAS AAS AAS ICP-AES FAAS AAS ICP-MS μg/g wet tissue brain 5.1 3.10 3.32 liver 9.9 7.8 5.60 3.47 kidney 2.6 1.80 2.1 1.80 2.15 stomach 1.44 1.10 intestines 2.1 1.54 lung 1.3 0.97 1.91 spleen 1.2 0.88 1.23 heart 3.3 2.40 3.26 bile 3.60 blood a 1.2 0.97 0.85 0.99 0.95 a μg/ml fluid Table 2 Average concentration of copper in different brain areas Bonilla 1984 [ 77 ] Harrison et al. [ 78 ] Ramos et al., [ 81 ] Pal et al. [ 80 ] FAAS AAS ICP-MS AAS μg/g dry tissue μg/g dry tissue μg/g dry tissue μg/g wet tissue Frontal pole 18.95 Precentral gyrus 8.68 Occipital pole 21.61 Calcarine cortex 23.07 Postcentral gyrus 18.83 Supramarginal gyrus 16.45 Uncus 16.30 Cingulate gyrus 15.14 57 Mammilay bodies 19.65 Superior colliculus 15.38 Inferior colliculus 17.92 Olfactory tract 17.66 Olfactory bulb 27.92 Optic nerve 17.79 Optic chiasm 7.06 Caudate nucleus (head) 13.49 42 61 Caudate nucleus (body) 18.46 Caudate nucleus (tail) 23.12 Putamen 14.62 44 62 Globus pallidus 12.47 35 45 Thalamus 8.75 21 Frontal lobe, white matter 5.43 22 36 Frontal lobe, gray matter 38 Occipital lobe, white matter 8.88 55 Parietal lobe, white matter 7.27 60 Temporal lobe, white matter 11.12 Red nucleus 10.41 Substantia nigra 17.42 Inferior olivary nucleus 12.00 Superior olivary nucleus 17.46 Pineal gland 17.81 Cerebellum (vermal cortex, superior half) 10.92 Cerebellum (vermal cortex, inferior half) 15.52 Hippocampus 29 70 Corpus callosum 14 Cerebellum, gray matter 47 36 2.69 Cerebellum, white matter 22 Frontal cortex 62 Superior temporal gyrus 61 Middle temporal gyrus 68 Midbrain 38 Pons 33 Medulla 35 Cortex 2.20 Striatum 2.18 Transition metals in biological tissues have been evaluated by atom absorption spectroscopy or radiotracer detection techniques, and more recently by the laser ablation inductively-coupled plasma mass spectrometry (LA-ICP-MS), secondary ion mass spectrometry, X-ray fluorescence microscopy (XFM), X-ray absorbance spectroscopy (XAS), micro particle-induced X-ray emission, and electron microscopy. Innovative imaging technologies of transitional metals were reviewed recently [ 85 – 90 ]. The recent development of recognition-based copper sensors and reaction-based copper indicators has allowed fluorescence imaging of labile copper pools [ 91 – 95 ]. Recent advances in non-destructive analytical methods will likely enable the assessment of copper dynamics over short, medium or long time scales that are relevant to signalling, metabolism and nutrition or aging. These technologies have made possible a deeper understanding of copper dynamics and distribution. Significant relationships regarding the levels of Ctr1, Atox1, ATP7A/ATP7B and copper concentrations in the human brain have been identified by the combined application of ICP-MS spectrometry, Western blot and immunohistochemistry. Copper and ATP7A levels in the substantia nigra and in the cerebellum , respectively, have been found to be significantly greater compared to other brain regions [ 96 ]. New insights into the relative distribution of copper among elements including P, S, Cl, K, Ca, Fe, Zn within the choroid plexus (CP), ventricle system, and surroundin

As outlined above, intracellular copper exists in an immense variety of static forms that involve a multiple oxidation states with favoured ligand speciations ( see below) or mixed-valent copper complexes. However, it also can change by reacting with “self” ( see Eq. 1. ) within sub-nanometer distances of multiple copper sites of vital peptides, proteins and enzymes [ 101 , 102 , 105 , 106 , 153 – 156 ]. A somewhat similar distinction can be made for a sulfur “redoxome”, a redox reaction-coupled proteomic network comprising numerous sulfur oxidation states and species and reactions with sulfur-containing peptides, proteins and enzymes, as well as the reaction of GSH with “self” yielding glutathione disulfide (GSSG) (2GSH → GSSG) [ 157 – 159 ]. Importantly, these “self”-reacting copper and sulfur “redoxomes” also interact with each other through the prominent chelation of Cu(I) by thiols of either the antioxidant copper chaperone protein Atox1 or GSH [ 63 , 142 , 160 – 162 ]. Supporting this concept, the GSH/GSSG ratio was found to be the most sensitive indicator of copper intoxication (and subsequent oxidative stress) [ 11 ]. Moreover, sulfur-doped copper clusters are relatively stable and abundant [ 163 ]. In the Cu-S-Cu unit found within active sites of copper-sulfur proteins like COX, the S − bridging a Cu 2 + component displays a short Cu-Cu distance and a small Cu-S-Cu bond angle, which are essential for the electron transport performed by COX [ 163 – 165 ]. With this in mind, toxicity of copper excess in mammalian cells is explained by obstructing the control of the “interactome” of copper-sulfur containing “redoxomes” [ 166 – 169 ]. Because of its charge density and polarizability, the oxidized cupric Cu(II) ion would tend to be found in complex with “hard” bases such as H 2 O, OH − , RNH 2 etc. (N- or O-ligands), while the “soft” acid Cu(I) does favour “soft” bases such as RS − and CN − ligands [ 46 ]. These trends in the stability of coordination complexes [ 170 , 171 ] predict that the reduced cuprous Cu(I) ion would prefer the formation of complexes with “S-ligands” such as GSH [ 172 ]. Importantly, GSH can also represent “N- or O-ligands” for Cu(II), assisting disproportionation and O 2 • − dismutase activity of Cu(I). Indeed, the complex equilibrium system GSH-Cu(I) can switch to the oxidized GSSG-Cu(II) one [ 173 ]. Taken together, these observations have been used to classify the speciation of Cu(I) with GSH as a key feature accompanying redox homeostasis [ 11 ]. Although the dissociation constant for the GSH-Cu(I) equilibrium has been predicted to be about 9 pM [ 148 ] (GSH-Cu(I)), this value has been called into question [ 174 , 175 ]. Specifically, in an experimental model of GSH-Cu(I), the formation of the tetranuclear [Cu 4 (GS) 6 ] cluster was observed as the major species within the range of pH from 5.5 to 7.5 [ 175 ]. The cluster formation equilibrium predicts that [Cu 4 (GS) 6 ] limits free Cu(I)(aq) to the sub-femtomolar concentration range in eukaryotes. These findings suggest that the affinity of GSH towards Cu(I) may be orders of magnitude higher than previously thought [ 148 ]. If valid in vivo, not only the high intracellular GSH concentration but the high-affinity formation of the [Cu 4 (GS) 6 ] cluster would also force the membrane-cytosol transfer of Cu(I) from Ctr1 to GSH. It is noteworthy, that bacteria capture copper surplus through the cytosolic protein Csp3s, which forms tetranuclear Cu(I) thiolate clusters [ 176 ] [Cu 4 (S-Cys) 5 ] − , [Cu 4 (S-Cys) 6 ] 2− , and [Cu 4 (S-Cys) 5 (O-Asn)] − . In order to avoid toxicity of cytosolic copper overload, eukaryotes gain control over excess by MTs [ 177 ], including the brain-specific MT3 (growth inhibitory factor) binding. In fact, using XRF microscopy with sub-micron resolution, Sullivan et al. [ 178 , 180 ] demonstrated the presence of Cu-rich aggregates in astrocytes of the dentate gyrus and rostral migratory stream in the rat brain. These aggregates contain Cu x S y clusters with a sulfur/Cu(I) ratio consistent with that of the Cu-MT complex. Apparently, both age-dependent [ 98 ] and overload-evoked changes [ 177 – 180 ] can be related to the copper-binding capacity of MTs. Direct and indirect effects leading to sudden and catastrophic hemolytic anemia due to the direct toxic effects of copper on red blood cells has been described in the past [ 181 – 190 ]. Nevertheless, the observation that during chronic copper poisoning in sheep there is decreased antioxidant capacity directly correlating with the level of serum copper [ 191 ] putting GSH at the centre of anti-ROS protection [ 192 ]. Underscoring the importance of this role, the level of GSH in erythrocytes is an inheritable trait [ 193 ]. Unfortunately, it is hard to obtain valid GSH and GSH/GSSG data from biological samples [ 194 ].

The concentration of copper in the cerebrospinal fluid (~ 70-80 μM) is rather high in comparison to serum (12-24 μM ) [ 215 , 216 ], raising the possibility of specific copper signalling in the brain. As outlined in previous sections, most cellular copper is strongly bound to proteins, yet the disposition of loosely bound copper can be detected by novel imaging technologies. This labile copper pool is believed to be associated with redox signalling. Labile copper has been found in the soma of cerebellar granule and cortical pyramidal neurons, in addition to the neuropil in the cerebellar and cerebral cortices, hippocampus and spinal cord [ 217 ]. The observation on the release of zinc from brain tissue during activity published in Nature in 1984 [ 218 ] provided initial evidence that transition metals could be directly involved in signalling [ 112 , 134 , 219 – 229 ]. Initial evidence suggested the potential of copper to modulate brain activity by affecting central inhibition. These include findings such as the pro-convulsant effects of a hitherto unidentified endogenous substance containing copper [ 112 , 230 ], or depolarization-induced co-release of endogenous copper with the major inhibitory neurotransmitter γ-aminobutyric acid (GABA) in different experimental models of nerve terminals in vitro (synaptosomal fraction) and ex vivo ( median eminence ) [ 225 ]. Conclusions from 67 Cu uptake and release measurements performed in hypothalamic slices by the presence of action potential blocker tetrodotoxin [ 231 , 232 ] or determination of depolarization-induced copper release from nerve endings by AAS [ 225 ] raised the concept of copper signalling in the brain. Findings, such as the N-methyl- D -aspartate (NMDA) receptor activation-induced ATP7A trafficking to the plasmamembrane in the hippocampus [ 233 , 234 ] have have provided new support for a role for copper efflux in mechanisms of excitotoxicity. During the past 30 years, there has been a renaissance of interest and an expanded view of the contributions of copper to brain function and pathophysiology, as reflected in follow-up statistics, and throughout the literature [ 18 , 42 , 63 , 69 , 235 – 248 ] (Fig. 3 . ). One may speculate about copper speciation and/or mechanisms of copper uptake and release. Apart from trafficking in complex with various neurotransmitters, carrier peptides and proteins, or as part of the protein cargo of extracellular vesicles [ 39 ] there is also the potential for copper uptake as a result of autophagy [ 29 , 249 ]. Fig. 3 Emerging themes of copper signalling and functions. Number (Left) and percentage (Right) of papers citing the first description of depolarization-induced synaptic copper release [ 225 ] in each subject category by 5-year intervals. From the time, copper signalling in brain have considerably been developed, including inhibitory and excitatory signalling, neuromodulation, neurotoxicity, Alzheimer’s and other brain disorders Zinc released from brain tissue during activity has been shown to reach concentrations in the hundred micromolar range, e.g. 300 μM [ 218 ]. In contrast, the concentration of copper in the synaptic cleft has been claimed to range from 1 to 10 μM, as determined by using the fluorescent indicator tetrakis-(4-sulfophenyl)-porphine (TSPP) in bovine chromaffin cells [ 250 ]. Notwithstanding the importance of imaging heavy metals in vivo ( see also section 2.1. “ Imaging technologies ”), the quantitative relevance of fluorescent indicators strongly depends on the affinity standards applied ( see for example [ 148 ]). Furthermore, TSPP has several drawbacks that can limit the validity of data obtained: i ) the sub-micromolar Kd value of the TSPP-copper complex may not provide accurate data at the point of/or above saturation; ii ) TSPP-copper binding is influenced by dissociation of copper from protein binding sites which necessary to validate the data an approach that is independent of protein binding, as in the case of ICP-MS technique; and finally, iii ) the weak fluorescence intensitiy of the TSPP ligand itself. Conversely, based on atomic absorption spectroscopy data on depolarization-induced release of copper from nerve endings, one can estimate an activity-dependent enhancement of copper in the synaptic cleft, in the range of 100-250 μM [ 225 ], depending on the cleft size and volume taken. Furthermore, based on the Kd (100 μM ) and saturating GABA concentration (1 mM ) for GABAa receptor binding and desensitization [ 251 – 254 ] and assuming a stoichiometry of 2 for GABA co-released with copper [ 112 ] one could also conlude that a copper concentration of 100 μM can exist transiently within the synaptic cleft. Copper can diffuse out of the synapse driven by the lower extrasynaptic concentration (1 μM ) [ 216 ]. Moreover, the extrasynaptic copper concentration has been estimated to be in the nanomolar range based on the cellular and network excitability produced by bath-applied copper i

The metal theory of AD [ 43 , 267 – 273 ] (but see the advice of Schrag et al. [ 274 ]) predicts that the disregulation of copper/zinc levels by proteins known to be involved in AD-related neurodegeneration may lead to the accumulation of amyloid fibers and oxidative stress. Indeed, by using XFM high areal concentration of copper has been detected in amyloid beta (Aβ) plaques of the hippocampal gyrus dentatus sub-region in a mouse model of AD [ 275 ]. These data corroborate previous findings on the high-affinity interaction between Cu(II) and the histidine binding motif of Aβ [ 276 ], along with the role for Aβ as a synaptic Cu(II) scavenger [ 277 ]. In addition, the experimental ‘halo’ effect in copper maps may indicate co-localization of copper with a ‘ring’ rich in lipids, observed around the Aβ plaque in AD models [ 278 ] and human AD sections [ 279 ]. This suggests a potential association between Cu-catalyzed oxidative stress and plaque formation [ 280 ]. However, the question remains as to whether changes in metal distribution are the cause or the consequence of the plaque formation and progression of AD [ 275 ] or other progressive neurodegenerative diseases. For example, the neuropathology seen in AD may also characterize individuals with Down syndrome [ 281 , 282 ], ALS or HC. By supporting a common pathway for familial and sporadic ALS, the pathological inclusions containing SOD1 fibrils may hold amyloid-like properties [ 283 ]. Abnormal copper accumulation in the striatum of HC patients has been linked to the copper binding facilitated formation of amyloid- copper transporter Ctr1, rlike bodies of the huntingtin (Htt) protein [ 284 , 285 ]. Differential effects of ATP7A and ATP7B regulating copper metabolism MURR1 domain protein 1 (COMMD1) on the formation of mutant Cu, Zn-SOD1 fibrils (increase) or parkin inclusions (decrease) as well as the Htt aggregates (unaltered), however, suggest mechanistic diversity [ 286 ].

Disease-specific autoantibodies against inwardly rectifying K + ion channel 4.1 (Kir4.1) [ 295 ], have been identified in the sera of patients suffering from the chronic inflammatory CNS disorder multiple sclerosis (MS). Feeding the copper chelator bis-cyclohexanone-oxalyldihydrazone (cuprizone, CTZ) reduces the myelin sheath and activates microglial and astroglial cells in the CNS, providing a reproducible and reversible model of pathologic processes underlying white and gray matter demyelination [ 296 – 301 ]. Expression of Kir4.1 autoantigen has been studied in the brain of CTZ-fed mice and revealed the induction of Kir4.1 protein in microvessels of the cerebral cortex [ 302 ]. The antioxidant functions of MTs [ 303 ] may have a role in MS, as suggested by the elevated level of MTs induced by CTZ in astroglia, while the oligodendroglia express low levels of MTs, which may contribute to their oxidative stress vulnerability [ 304 , 305 ]. Apart from MS modelling, several lines of evidence suggest that CTZ intoxication is an excellent paradigm to study pathology and/or therapy of epilepsy or schizophrenia as well. However, mechanistic clues claiming either copper deficiency or copper build-up (associated with hydrazide formation-dependent enzyme inhibition) remain contradictory [ 306 ].

The advent of imaging techniques that gained insight into the dynamics of labile copper pool made it possible to look beyond the molecular-level interactions in copper homeostasis and examine network-level dynamic interplays shaping copper signalling. The source-target-nphysiology (STP) scheme suggested by Chang [ 336 ] includes labile, neuronal copper pools in the Golgi compartment as a source, signal propagation via postsynaptic membrane receptor/ion channel target (the Cu(I) transporter Ctr1), and copper-dependent spontaneous activity of the neural network (physiology). Vesicular storage of canonical neurotransmitters with copper suggesting their co-release has also been described. Furthering the interactions between compartments within neurons, we conceive that cellular-level copper signalling between neurons and astrocytes, an emerging cell type of the brain, also exists and may play a fundamental part in the brain’s information processing. Several lines of evidence demonstrate memory deficits concurrent with copper deposition in the choroid plexus, astrocyte swelling (Alzheimer type II cells), astrogliosis and neuronal degeneration in the cerebral cortex, and augmented levels of copper and zinc in the hippocampus of chronically copper-intoxicated rats [ 337 ]. Furthermore, these and the other findings concerning the role for astrocytes in brain activity, dis-homeostasis and asscociated diseases [ 110 , 338 – 341 ] and brain copper and pA homeostasis in particular [ 179 , 180 , 342 , 343 ] may provide support for new astrocyte-centric directions for therapeutic intervention. It can also be depicted by the “gliocentric” alternative of the “neurocentric” STP workflow suggested by Chang [ 336 ] possibly associated with major astroglial processes and players of Glu and ammonia homeostasis underlying excitation-inhibition balance in brain [ 110 ]. Prevalent traumatic and ischaemic brain injuries are explored to validate the potential of the “gliocentric” concept of early therapeutic intervention. Now, we may add copper-dependent astroglial pA production to the list (Fig. 4 . ) [ 6 , 22 , 80 , 179 , 180 , 233 , 234 , 236 – 243 , 245 , 254 , 255 , 336 , 339 , 340 , 342 – 362 ]. GABA can be synthesized from the pA putrescine by copper-containing CuAO in astrocytes. CuAOs perform the oxidation of primary amines such as spermine, spermidine and putrescine to aldehydes and ammonia, producing H 2 O 2 as a by-product. Putrescine-derived GABA is released by the inside-out (reverse) action of glial GABA transporter subtypes. The increased GABA release and the generated tonic inhibition thereby modulate the power of gamma range oscillation in the CA1 region in vivo. The concentration of cytosolic and extracellular putrescine has been determined to be 22 nmol/g and 12 nmol/g , respectively [ 339 ]. In contrast, copper may decrease tonic inhibition via acting on delta subunit-containing extrasynaptic GABA A receptors [ 235 , 237 , 246 ], thus adding a new layer to disinhibitory mechanisms in copper-rich brain areas. Fig. 4 Copper signaling via neuro-glia coupling. Astroglia, a previously neglected cell type of the brain [ 340 ], operate a variety of copper-dependent metabolic functions [ 6 , 80 , 240 , 341 , 342 ]. For this reason, in addition to synaptic and extrasynaptic copper signalling by way of excitatory/inhibitory receptors and ionic channels [ 22 , 234 , 235 , 237 – 244 , 246 , 255 , 336 , 345 – 355 ], we place copper-dependent production of pAs in astrocytes [ 338 ] and correlated gap-junction modulation in the centre of this option. The proposed scheme conjectures activity-dependent changes of copper pools [ 179 , 180 ] and polyamines (pAs), produced by CuAOs in astrocytes. First, an enhanced gap junction communication can be achieved by pAs [ 356 – 358 ], possibly promoting activity-dependent synchronization [ 339 , 359 ]. Second, major inhibitory neurotransmitter gamma-aminobutyric acid (GABA) formed from pAs is released by astrocyte-specific GABA transporter [ 360 ]. Acting on its extrasynaptic receptor, GABA elevates tonic inhibition and enhances the fast (gamma band) neural oscillations [ 360 ]. These ways, the steady-state pA level in astrocytes determined by copper-dependent forming and consuming can be associated with neural circuit activity [ 244 , 255 , 362 ]

This work was supported by grants KMR_12-1-2012-0112 TRANSRAT, VEKOP-2.1.1-15-2016-00156 and OTKA K124558.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.