Endoplasmic Reticulum-Mitochondria Contacts Modulate Reactive Oxygen Species-Mediated Signaling and Oxidative Stress in Brain Disorders: The Key Role of Sigma-1 Receptor

Lipid metabolism

Biochemical studies show that the synthesis of triacylglycerol, phosphatidylcholine (PC), and phosphatidylethanolamine (PE) requires the activity of enzymes present at both ER and mitochondria (85, 141). These findings demonstrate the relevance of MAMs in lipid metabolism and transfer between both organelles. The MAMs involvement in lipid synthesis and exchange results from the accumulation of GM1-ganglioside enzymes in the glycosphingolipid-enriched microdomain of MAMs (151) and the presence of lipid rafts domains associated with Sig-1R receptors and erlin-1 and erlin-2 (27).

The MAMs are enriched in phospholipid-synthesizing enzymes, such as FACL4, which is involved in the ligation of fatty acids to coenzyme A (CoA) and in triacylglycerol synthesis; PSS1/2, which catalyzes phosphatidylserine (PS) synthesis; PEMT2, which is involved in PC synthesis; and ACAT1/SOAT1, which catalyzes the production of cholesterol esters, which are incorporated into lipid droplets (161, 171). PS is synthesized in the ER by PSS1 and PSS2 and translocated to the inner mitochondrial membrane (IMM) where it is decarboxylated by phosphatidylserine decarboxylase into PE, which is rapidly exported to other organelles, such as the ER, where it is converted to PC by PEMT2 (161, 171) (Fig. 2).

Enzymes involved in cholesterol biosynthesis are also accommodated at MAMs, namely ACAT1 that catalyzes the intracellular esterification of free cholesterol, allowing a dynamic equilibrium between membrane-bound and cytoplasm-stored cholesterol in resting state (84), and cholesterol-binding caveolin-1 (CAV1) that promotes cholesterol transfer into mitochondria (150). Cholesterol was found to negatively regulate MAMs-mitochondria association. Using an in vitro membrane association assay, it has been observed that depletion of cholesterol in isolated MAMs with methyl-β-cyclodextrin significantly increased the association between MAMs and mitochondrial fractions.

These results suggest the existence of two different mechanisms regulating MAMs-mitochondria association: one that promotes the membrane association via specific membrane-tethering proteins, such as Mfn2, and another that restricts the association wherein cholesterol plays a pivotal role (Fig. 2).

Calcium homeostasis

The ER, mitochondrial, and cytosolic Ca²⁺ homeostasis are regulated at MAMs through SERCA transport of Ca²⁺ from the cytoplasm into the ER lumen, IP₃R-mediated ER Ca²⁺ release, and mitochondria uptake by VDAC at the OMM (84, 161) (Fig. 2). The MAMs form a Ca²⁺ signaling hub regulating ER chaperone-assisted folding of de novo synthesized proteins and activity of mitochondria-resident Ca²⁺-dependent dehydrogenases that are required for mitochondrial bioenergetics, as well as proteins involved in cell death such as BAX and BCL-2 (135, 142).

Therefore, fine-tuned Ca²⁺ transfer between the ER and the mitochondria is essential to determine cell fate. The MAMs form a microdomain for Ca²⁺ transfer, and physical proximity of ER and mitochondria allows a direct and selective transmission of Ca²⁺ signals (135, 161). Agonists of plasma membrane Ca²⁺ permeable ionotropic receptors activate phospholipase C that catalyzes the formation of IP₃, which binds to IP₃R and stimulates ER Ca²⁺ release (135). Due to the proximity to mitochondria, Ca²⁺ crosses the freely permeable OMM through VDAC channels and the impermeable IMM by the low-affinity channel of the mitochondrial Ca²⁺ uniporter (MCU), finally accumulating in the mitochondrial matrix (135, 161).

The MAM chaperone GRP75 mediates the interaction between the IP₃R and VDAC, acting as a bridge between these membrane proteins and therefore regulating Ca²⁺ transfer to the mitochondrial matrix via MCU (135). Usually, mitochondria take up Ca²⁺ released from ER in a very short time frame after its release; however, the amplitude of ER-associated Ca²⁺ puffs decreases in the presence of MCU inhibitors. Low or very high cytosolic Ca²⁺ concentrations reduce the IP₃R opening and consequently diminish ER Ca²⁺ release (141), which might negatively impact mitochondrial function.

Sig-1R binds to IP₃R, preventing the ubiquitin-mediated degradation of IP₃R-associated Ca²⁺ channels, increasing Ca²⁺ transfer from the ER to the mitochondria. By doing this, Sig-1R stimulates ATP production due to activation of Ca²⁺-dependent metabolic enzymes in the Krebs cycle, and consequently, preserves cell viability (141, 161). On the other hand, SERCA overexpression leads to ER Ca²⁺ overload, which stimulates ER Ca²⁺ release, namely through the IP₃Rs. As a consequence, ER-mitochondria Ca²⁺ transfer is exacerbated, promoting mitochondrial Ca²⁺ overload and dysfunction (105, 161).

Several proteins recruited to MAMs are involved in Ca²⁺ modulation, namely AKT/PKB and cytochrome c. In response to survival signals, the anti-apoptotic protein AKT/PKB translocates to MAMs and inactivates IP₃R reducing ER Ca²⁺ release. Cytochrome c, which is released from mitochondria by apoptotic stimuli, can bind IP₃Rs at MAMs, increasing ER Ca²⁺ release and subsequent mitochondrial Ca²⁺ overload, as well as cytochrome c release, thus enhancing pro-apoptotic signaling (135). Ca²⁺ is taken up into ER through Ca²⁺ pumps of the SERCA family, after release from mitochondria through the mitochondrial Na⁺/Ca⁺ exchanger (NCLX). This closes the cycle of intracellular Ca²⁺ signaling between the ER and the mitochondria, which determines the correct functioning of these organelles (141).

In addition to the regulation of Ca²⁺ and lipid homeostasis, MAMs are also involved in other physiologic events, namely the modulation of energy metabolism, ER stress response, autophagy, apoptosis, and innate immunity (Fig. 3).

OPEN IN VIEWER

FIG. 3.

MAMs regulate mitochondrial metabolism, UPR, inflammation, and cell death. The uptake of ER-derived Ca²⁺ by mitochondria is required for normal mitochondrial metabolism (ATP production) and is dependent on different channels: IP₃Rs allow Ca²⁺ release from the ER; Ca²⁺ entry into mitochondria occurs through VDAC and MCU, which are, respectively, located at the outer and inner mitochondrial membranes; SERCA transports cytosolic Ca²⁺ to the ER by using energy from ATP hydrolysis. Sig-1R is a MAM's resident chaperone that facilitates IP₃R-mediated release of Ca²⁺ from the ER. Several components of the PERK, IRE1α, and ATF6 branches of the ER UPR, which mediate cellular responses to ER stress, are located at MAMs. The GRP78, which is an ER-localized chaperone involved in ER stress response involved in protein folding, also buffers Ca²⁺ in the ER lumen. Mitochondrial Ca²⁺ overload commits cells to death by mPTP-mediated apoptosis, leading to depolarization of the mitochondrial membrane and, consequently, induction of cytochrome c release into the cytosol. The Fis1-Bap31 complex, which results from the interaction between mitochondrial Fis1 and Bap31 localized at ER, provides a platform for apoptosis by recruiting procaspase-8. In case of mitochondrial dysfunction, DAMPs such as ROS are released from mitochondria, leading to NLRP3 inflammasome activation and, consequently, to inflammation-mediated cell death. Fis1, fission protein Fission 1; ROS, reactive oxygen species; DAMP, danger-associated molecular pattern.

Energy metabolism

Mitochondria play a central role in cell metabolism since they are responsible for ATP synthesis, which is the main cellular energetic source. Mitochondrial uptake of Ca²⁺ released by the ER is required to activate the tricarboxylic acid cycle (TCA), which provides reducing equivalents required for ATP production at the electron transport chain (ETC) (31). On the other hand, the ATPase activity of the ER resident GRP78 (32) depends on ATP produced by mitochondria (79). Disruption of ER-mitochondria tethers, such as VAPB-PTP1P51, perturbs Ca²⁺ flux between these organelles and, consequently, impairs mitochondrial ATP production (130).

IMM and OMM are strongly enriched in lipids. Although mitochondria can synthesize lipids such as cardiolipin (CL), the structural and functional maintenance of these organelles is highly dependent on lipids imported from the ER (108). CL is synthesized in mitochondria from ER-derived phosphatidic acid (PA) (87). CL deficiency inhibits the ETC complex IV, supporting the fact that CL plays a key role in mitochondrial metabolism (115). CL modulates OPA1 functions and therefore regulates mitochondrial morphology and quality control (13), which is fundamental to adjust ATP synthesis to metabolic needs (55). PC, the most abundant mitochondrial phospholipid, is only synthesized in the ER and then exported to the mitochondria (77) and decreased mitochondrial PC/PE molecular ratio is associated with decreased oxygen consumption and ATP depletion (166).

The earlier findings indicate that lipid transfer between the ER and the mitochondria is crucial to maintain normal cell functioning, namely bioenergetic metabolism. Therefore, the impairment of ER-mitochondria contacts negatively affects mitochondrial metabolism and cell survival.

ER stress response

One of the major ER functions is the folding of proteins synthesized de novo. Several insults, such as nutrient deprivation, hypoxia, loss of Ca²⁺ homeostasis, as well as accumulation of mutant proteins, can compromise the efficiency of protein folding, leading to the accumulation of unfolded/misfolded proteins in its lumen, a state known as ER stress that affects cell homeostasis. In response to ER stress, the adaptive UPR signaling pathway, which is mediated by activation of the ER transmembrane protein sensors PERK, IRE1α, and activating transcription factor 6 (ATF6), is induced to reestablish ER homeostasis and promote cell survival (72).

On ER stress, PERK phosphorylates eukaryotic translation initiation factor 2 subunit-α (eif2-α), attenuating global protein synthesis (64), which limits the loading of the ER lumen with misfolded proteins. Phosphorylated eif2-α paradoxically increases translation of the ATF4 (102) that upregulates the genes involved in redox homeostasis, amino acid metabolism, protein synthesis and degradation by macroautophagy or ER-associated protein degradation (ERAD), and also apoptosis (11, 62, 65).

IRE1α oligomerizes and autophosphorylates on ER stress to promote its RNase activity (195). IRE1α cuts off a small 26-nucleotide intron from the mRNA encoding the transcription factor X-box-binding protein 1 (XBP1) (190), increasing the levels of a spliced active form of XBP1 (XBP1s), which, in turn, upregulates the genes involved in ER protein translocation, folding, and secretion, as well as degradation of misfolded proteins (95). ATF6 is translocated to the Golgi apparatus when the ER is under stress; it is cleaved by Site 1 and Site 2 proteases (S1P and S2P) into an active ATF fragment (ATF6p50) that migrates to the nucleus and stimulates the transcription of UPR target genes involved in protein folding and degradation (185, 188).

Numerous evidence demonstrate that the ER stress response is intimately associated with ER-mitochondria contacts. Activation of the PERK-ATF4 arm of UPR was found to prevent ER stress-related events through parkin upregulation, which, in turn, avoids mitochondrial fragmentation and ATP depletion (22). Accordingly, stress-induced parkin overexpression has been associated with enhanced ER-mitochondria coupling and mitochondrial Ca²⁺ uptake to preserve energetic metabolism (29).

Moreover, Verfaillie et al. demonstrated that PERK is enriched at MAMs and that its ablation disturbs ER morphology and Ca²⁺ signaling, as well as diminishes ER-mitochondria contact sites (172). Remodeling of these contacts on ER stress has been shown to be mediated by protein kinase A (PKA) that enhances Ca²⁺ transfer toward mitochondria to avoid cell dysfunction and death. These protective effects were abolished in the presence of the MAMs-located CAV1, and cells become unresponsive to ER stress and more sensitive to Ca²⁺-mediated cell death (23).

Calnexin is a chaperone located at MAMs that interacts with SERCA regulating ER-mitochondria Ca²⁺ exchange. Under short-term ER stress, it is preferentially located at the ER and shifts its function from Ca²⁺ signaling regulation toward modulation of protein folding and quality control in the ER lumen (104). During the early adaptive phases of ER stress, there is a temporary increase in ATP production in an attempt to preserve cell survival (174). Increased mitochondrial energetic metabolism results from enhanced ER-mitochondria coupling and Ca²⁺ transfer (27).

Autophagy

Macroautophagy is a catabolic mechanism that is involved in several physiological processes, namely quality control strategies to eliminate dysfunctional organelles and intracellular misfolded proteins (121). Macroautophagy is initiated by the isolation of specialized membranes from compartments of the ER enriched in phosphatidylinositol 3-phosphate (PI₃P), called omegasomes, to form autophagosomes (10).

Several findings support the hypothesis that autophagosomes are formed at ER-mitochondria contact sites. For example, the pre-autophagosome/autophagosome marker ATG14 and the autophagosome-formation marker ATG5 are recruited to the ER-mitochondria platform on starvation. Moreover, the disruption of coupling between organelles avoids the formation of ATG14 puncta, preventing autophagosome formation (61). The MAMs have also been proposed as platforms for autophagy based on their role in modulating lipid composition of ER-mitochondria contacts. For instance, PE upregulation at MAMs improves autophagic flux (144).

In addition, the molecular interaction between the components of MAM's lipid-rafts microdomains, such as the ganglioside GD3, and core-initiator autophagic proteins, including AMBRA1 and WIPI1, is required to initiate autophagy (50). The stearoyl-CoA desaturase 1 (SCD1), an enzyme enriched in MAMs that regulates membrane content of fatty acids, promotes autophagy and its inhibition depletes membrane phospholipids, resulting in the impairment of autophagosome formation and autophagic cargo degradation (83).

During mitophagy, which is the autophagic process that selectively eliminates damaged mitochondria, the ER-mitochondria tether Mfn2 is rapidly degraded and ER-mitochondria contacts are destroyed. This process involves Mfn2 phospho-ubiquitination by the PINK1/parkin system, which is a signal to mark damaged mitochondria and initiate their lysosomal degradation. Moreover, ER-mitochondria contacts themselves are negative mitophagy regulators, through the Mfn2 antagonist effect (116).

Apoptosis

The activation of the intrinsic apoptotic pathway when cells are exposed to prolonged stress implicates the communication between the ER and the mitochondria. Ca²⁺ released from the ER IP₃R- or RyR-associated Ca²⁺ channels located at MAMs increases cytosolic Ca²⁺ levels in the vicinity of the mitochondria. Excessive Ca²⁺ entry through VDAC and MCU leads to mitochondrial Ca²⁺ overload that is noxious for cells, leading to permanent opening of the high conductance mitochondrial permeability transition pore (mPTP), which is considered as a non-return point of apoptosis.

The permanent pore opening is involved in cytochrome c release and dissipation of the mitochondrial membrane potential, with concomitant cell death associated with the release into the cytosol of mitochondrial proteins such as cytochrome c, apoptosis-inducible factor, and Smac/DIABLO proteins. Once released, these proteins trigger both caspase-dependent and caspase-independent apoptosis (30).

Some BCL-2 family proteins are located at MAMs and modulate the amount of Ca²⁺ transfer between the ER and the mitochondria (20). For instance, BCL-2 inhibits pro-apoptotic signals by reducing Ca²⁺ release through its interaction with the IP₃R at MAMs (146). In addition, the interaction of BCL-xL with the IP₃R promotes a pro-survival Ca²⁺ transfer to mitochondria (182). On the other hand, during severe stress conditions, the IP₃R-interacting protein IRBIT inhibits the anti-apoptotic BCL2l10 and promotes ER-mitochondria contact and Ca²⁺ transfer, which, in turn, induces mitochondrial Ca²⁺ overload, culminating in apoptosis (19).

Apoptotic cell death is induced under chronic or severe ER stress conditions on UPR activation (62). The role of ER-mitochondria coupling in apoptosis is further supported by data showing that the ER stress sensor PERK, which is mainly localized at MAMs, is required for ER stress-induced apoptosis triggered by ROS that promote the transfer of Ca²⁺ from the ER to the mitochondria (172). Under stress conditions, Mfn2-deficient cells, which present alterations in ER-mitochondria contact sites, were shown to present reduced apoptosis markers and PERK deficiency rescued these cells from apoptosis triggered by ER stress (124). Further, IRE1α ubiquitination at MAMs by the mitochondrial ubiquitin ligase (MITOL/MARCH5) inhibits apoptosis under ER stress conditions (164).

Innate immunity

The MAMs play a critical role in innate immunity, providing a molecular platform for NLRP3 inflammasome assembly (120). Inflammasomes are multiprotein complexes that are activated in the presence of pathogen-associated molecular patterns and damage-associated molecular patterns (DAMPs) (60). They are composed of a sensor protein, an adaptor protein called apoptosis-associated speck-like (ASC) protein containing a caspase activation and recruitment domain (CARD) and pro-caspase 1. Once assembled, caspase 1 is activated by auto-cleavage from pro-caspase 1 and induces the maturation and secretion of the pro-inflammatory cytokines interleukin-1β/18 (IL-1β/18) (60).

During NLRP3 inflammasome formation, de novo synthesized NLRP3 and ASC are translocated to MAMs, resulting in the apposition of ASC on the mitochondria to NLRP3 on the ER (196). At MAMs, inflammasome senses DAMPs that emanate from damaged mitochondria such as ROS, DNA, and CL externalization, membrane depolarization, and Ca²⁺ deregulation, among others (120). Cellular processes that are modulated by ER-mitochondria contacts at MAMs such as Ca²⁺ and lipid transfer play a crucial role in NLRP3 activation. In the presence of NLRP3 agonists, blocking mitochondrial Ca²⁺ uptake by silencing VDAC isoforms 1 and 2 prevents NLRP3 inflammasome activation, as well as mitochondrial ROS production (196). Moreover, cholesterol trafficking to the ER is required for NLRP3 inflammasome activation (92).

ROS generation

The ROS are biological molecules with at least one oxygen atom and with higher reactivity than molecular oxygen (O₂). These molecules include non-radical and free radical (with at least one free electron) species such as hydrogen peroxide (H₂O₂) and the superoxide anion radical (O₂ ^.−), respectively, which are key redox signaling agents (154).

Apart from transmembrane NADPH oxidases (NOXs) and the mitochondrial ETC, which are the major endogenous sources of H₂O₂ and O2^−•, H₂O₂ is also produced by other oxidases present in the ER, peroxisomes and plasma membrane, as well as by superoxide dismutases (SOD) (154). Mitochondria produces O₂ ^−• as a by-product of the inefficient transfer of electrons by the ETC during oxidative phosphorylation (OXPHOS).

This ROS is rapidly converted to H₂O₂ by the copper-zinc SOD (SOD1) in the intermembrane space and by the manganese-dependent SOD (SOD2) in the mitochondrial matrix space and then reduced to water by catalase (CAT) or glutathione peroxidase (74, 153). H₂O₂ can react with metal ions, such as Fe²⁺, via the Fenton reaction, to generate highly reactive hydroxyl radicals (88). In the ER, H₂O₂ is produced during the protein folding process mediated by the PDI and Ero1α chaperones (194).

Although the oxidative environment is conducive to protein folding, excessive H₂O₂ accumulation can deregulate the ER redox balance, leading to the accumulation of misfolded proteins causing ER stress (194). Similarly, the flavoprotein quiescin sulfhydryl oxidase catalyzes disulfide formation by coupling disulfide oxidation to the reduction of O₂ generating H₂O₂ (90). NOX4 located at the ER membrane also produces H₂O₂, which may be involved in apoptotic cell death activation after prolonged ER stress (191). To maintain the ER redox balance, some oxidoreductases such as peroxiredoxin 4 (PRDX4), glutathione peroxidase 7/8 (GPx7/8), and ascorbate peroxidase scavenge H₂O₂. In addition, reduced glutathione (GSH) also contributes to the removal of excessive H₂O₂ to avoid oxidative stress (6).

The ER-mitochondria interface hosts dynamic H₂O₂ nanodomains, which originate from the mitochondrial cristae and are induced by and influence Ca²⁺ signals. These structures are compressed on Ca²⁺ uptake to the matrix that stimulates the influx of K⁺ through Ca²⁺-activated mitochondrial K⁺ channels, causing rapid osmotic changes in the matrix, compensated by H₂O influx (21). Sig-1R, which is a MAMs-resident chaperone that modulates Ca²⁺ exchange between the ER and the mitochondria by interacting with IP₃Rs, is also involved in ROS generation.

In physiological conditions, Sig-1R agonists were found to generate moderate oxidative stress in mouse brain mitochondria stimulating ROS production at ETC complex I level. Under pathological conditions, Sig-1R activity may contribute toward restoring mitochondrial function (54). Ero1α, a key controller of oxidative folding and ER redox homeostasis, also localizes at MAMs and regulates Ca²⁺ fluxes. Downregulation of Ero1α inhibits mitochondrial Ca²⁺ fluxes and modifies the activity of MCUs. On the other hand, overexpression of redox active Ero1α increases passive Ca²⁺ efflux from the ER, lowering ER Ca²⁺ concentration and fluxes to mitochondria in response to IP₃R agonists (4).

Ero1α activity results in H₂O₂ formation in the ER (197). During oxidative protein folding, PDI chaperones (e.g., ERp44) oxidatively fold proteins before Ero1α oxidase transfers electrons from reduced PDI to the terminal acceptor, which is usually O₂ that is reduced to H₂O₂ (197) (Fig. 4). The interplay between Ero1α and ERp44, which is an ER luminal chaperone also present in the MAMs fraction, is responsible for the regulation of ER Ca²⁺ release via IP₃R. Ero1α interacts with IP₃R1 and oxidizes it, potentiating the Ca²⁺ release during ER stress (4).

OPEN IN VIEWER

FIG. 4.

Redox regulation at MAMs. Due to the close proximity between the ER and the mitochondria, ROS generated in the mitochondria under physiological conditions can diffuse to the ER and modulate the redox status. The ROS can also influence ER-mitochondria Ca²⁺ signaling via oxidizing Cys thiol of Ca²⁺ channels/pumps/exchangers present at both organelles. Several MAMs-resident proteins such as PDI, ERO1α, Sig-1R, and p66Shc are involved in ROS production in the ER or in the mitochondria. In the ER, during the oxidative protein folding process, Ero1α oxidase transfers electrons from reduced PDI chaperones to O₂ that is subsequently reduced to H₂O₂. Under physiological conditions, Sig-1R activity stimulates ROS production at the ETC complex I level, and it reduces their toxic levels in the presence of pathological stimuli. Under oxidative stress conditions, p66Shc is activated and can translocate to the mitochondria and/or to the MAMs fraction, where it contributes to ROS formation. ERO1α, endoplasmic reticulum oxidoreductase 1 alpha.

Thioredoxin-related transmembrane protein 1 (TMX1), another member of the PDI family of ER proteins that catalyze protein folding and thiol-disulfide interchange reactions, also regulates Ca²⁺ flux between the ER and the mitochondria. At MAMs, TMX1 inhibits SERCA2b, increasing cytoplasmic Ca²⁺ levels that activate Ca²⁺ import in mitochondria to stimulate OXPHOS (59). TMX downregulation alters mitochondrial morphology, repositions mitochondria close to the plasma membrane, enhances bioenergetics, and increases mitochondrial- and NOX-derived ROS generation (193).

GRP75 is essential to maintain the physical contact between the ER and the mitochondria by promoting the interaction between IP₃R and VDAC1, and it also plays a major role in redox homeostasis regulation. GRP75 silencing was demonstrated to attenuate both cytosolic and mitochondrial Ca²⁺ overload in conditions of oxidative stress, block ROS formation, and preserve mitochondrial respiration (76). The canonical ER stress sensor PERK also serves as a structural tether at the ER-mitochondria interface that regulates inter-organelles crosstalk during ROS-induced cell death. It was observed that weak ER-mitochondria contacts are present in PERK^−/− cells, preventing the propagation of ROS signaling to the apoptotic machinery (172).

Another MAMs-resident protein, the p66Shc, is activated by oxidative stress and translocates to the mitochondria where it participates in ROS formation. p66Shc protein is a growth factor adaptor protein belonging to the ShcA family involved in Ras signaling under physiologic conditions (136). Under oxidative stress, p66Shc can be phosphorylated at Ser36 by JNK (89), protein kinase (PK)Cβ (184), or ERK (78), leading to a cascade of events including isomerization of Ser36-phosphorylated p66Shc by the prolyl isomerase Pin1, dephosphorylation of Ser36-P-p66Shc by phosphatase A2, and translocation of p66Shc to the mitochondria and/or MAMs fraction, where it participates in ROS production (52) (Fig. 4).

A positive correlation between increased p66Shc phosphorylation at Ser36 and oxidative damage of DNA and proteins was found in fibroblasts expressing mutations in complex I subunits. PKCβ inhibition decreased oxidative stress levels and restored complex I activity under these conditions (184, 187). Accordingly, knockdown of p66Shc alleviated ROS levels and apoptosis in an in vitro oxidative stress model (63), further supporting a role of p66Shc in ROS production. Sirt1-regulated acetylation modulates the capacity of p66Shc to induce ROS formation. Lysine acetylation of p66Shc facilitates its phosphorylation on Ser36 and translocation to the mitochondria where it promotes H₂O₂ production (91).

Alzheimer's disease

A growing body of evidence has revealed the impact of altered communication at MAMs in the pathophysiology of neurodegenerative disorders (175) (Fig. 5). Some of them are associated with increased ER-mitochondria contacts (7, 8, 140, 165), whereas in others a decrease of these structures has been described (43).

OPEN IN VIEWER

FIG. 5.

Ca²⁺ transfer at MAMs in physiologic versus pathologic conditions. Under physiological conditions, a constitutive Ca²⁺ transfer from the ER to the mitochondria, which is regulated by ER-mitochondria contact sites at MAMs, is essential to preserve cellular bioenergetics, namely ATP production and mitochondrial membrane potential. Under pathological conditions, the ER-mitochondria communication at MAMs is altered, affecting the homeostasis of both organelles, which can lead to cell death due to mitochondrial dysfunction. Neurodegenerative diseases are closely related to alterations in ER-mitochondria contacts, whereas some are associated with increased and others with decreased distance between the ER and the mitochondria. As a consequence, exchange of biomolecules and ions between the organelles is impaired, leading to several changes including ROS generation and mitochondrial membrane depolarization, which ultimately results in the permanent opening of the mPTP.

Area-Gomez et al. proposed the “MAM hypothesis” for AD based on evidence of mitochondria toxicity that arises from enhanced ER-mitochondria contacts caused by increased levels of C99, a product of amyloid-β precursor protein (APP) cleavage by β-secretase (8). Accordingly, ER-mitochondria communication was shown to be increased in fibroblasts from both familial and sporadic AD patients (7) and several MAM proteins were found to be overexpressed in human post mortem AD brain samples (68).

Interestingly, the activation of MAMs has been associated with the presence of the ɛ4 variant of apolipoprotein E (ApoE4), the major risk factor for AD (165). Del Prete et al. reported increased ER-mitochondria contact sites in SH-SY-5Y cells overexpressing the APP Swedish mutation (APPswe) (140). Using in vitro and in vivo AD models, these authors also found that both APP and its proteolytic products, namely the Aβ peptide, are located at MAMs where they modulate the ER and mitochondria functions, namely the activities of the E3 ubiquitin-protein ligase and complex III or cytochrome b-c1 complex, respectively (140). These data highlight the role of APP processing in AD-associated MAMs deregulation.

However, it is not clear how α- and β-secretases may function in the ER, mitochondria, and MAMs compartments in intact cells. Recent studies by Fernandes et al. that observed a decrease of close ER-mitochondria contacts together with activation of the ER UPR, reduced ER-mitochondria Ca²⁺ transfer, and impaired mitochondrial function in a mouse neuroblastoma cell line (N2A) overexpressing human APPswe (43) further support MAMs alterations in AD.

Using the MAMs fraction isolated from the cerebral cortex of the APP/PS1 mouse model of AD, Völgyi et al. detected an increase in the levels of markers of protein synthesis, ERAD, oxidative stress response, and evidence of decreased mitochondrial protein transport and ATP production. These changes precede the onset of dementia-like symptoms in the APP/PS1 model that is characterized by cerebral accumulation of the Aβ peptide, suggesting their importance in AD progression (176).

Mutant presenilin-2 (PS2) that causes familial AD and also induces Aβ deposition was demonstrated to decrease mitochondrial ATP production by reducing ER Ca²⁺ content and thus ER-mitochondria Ca²⁺ transfer, which impairs the ETC (147). The ATP depletion directly affects cell homeostasis and could explain changes in protein degradation, synaptic function, neuritic growth, neurotransmission, and long-term potentiation involved in memory formation, which are closely associated with AD (170).

There is strong evidence that p66Shc is implicated in numerous oxidative stress-associated pathologies, such as AD and other neurodegenerative disorders (159). The expression and activation of p66Shc renders central nervous system (CNS) cells more sensitive to the toxicity of Aβ by promoting mitochondrial OXPHOS and ROS production (101). In brain mitochondria isolated from p66Shc^−/− aged mice, redox imbalance was demonstrated to be attenuated and both ROS generation and clearance were decreased (137). Moreover, p66Shc ablation in PS/APP transgenic mice rescued Aβ-induced cognitive deficits accompanied by improvement of brain mitochondrial respiration, reversal of mitochondrial complex I dysfunction, restoration of ATP production, and reduction of ROS levels (40).

These results support a role for p66Shc in Aβ-related mitochondrial dysfunction and oxidative damage in vivo, and they suggest that p66Shc silencing may be a promising novel therapeutic strategy against AD.

Parkinson's disease

The MAMs also play a crucial role in the pathogenesis of PD mediating PD-related events, including proteostasis loss, namely compromised autophagy (56). Besides that, several proteins implicated in PD are involved in ER-mitochondria signaling. For instance, α-synuclein regulates mitochondrial Ca²⁺ homeostasis by stimulating ER-mitochondria contacts (28). Recently, Basso et al. demonstrated that parkin plays a key role as regulator of ER-mitochondria communication by showing decreased inter-organelle interaction in parkin-deficient cells and in human fibroblasts expressing the familial PD-related mutant parkin (14).

In addition, it was demonstrated that MFN2 mutants that are not ubiquitinated by parkin do not promote ER-mitochondria interactions. These findings were strengthened by using an in vivo drosophila PD model in which the reestablishment of ER-mitochondria tethering using a synthetic linker was able to rescue PD-related motor deficits (14). Mitochondrial impairment and altered ER-mitochondria contacts are closely associated with other familial PD-related genes such as PINK1 and DJ-1 (134). Decreased ER-mitochondria contacts and impaired mitochondrial motility in neurites have been observed in PINK1 knockdown cells (134).

DJ-1 physically interacts with and is an essential component of the IP₃R3-Grp75-VDAC1 complexes at MAMs. Depletion of DJ-1 levels, decreased IP₃R3-DJ-1 interaction, as well as compromised ER-mitochondria association were found in the substantia nigra of sporadic PD patients (98).

Amyotrophic lateral sclerosis

Several strands of evidence support the role of MAMs disruption in ALS pathophysiology. Stoica et al. demonstrated that the ALS-related pathologic TDP-43 protein disrupts ER-mitochondria contacts, specifically targeting the interaction between the ER-resident protein VAPB and the mitochondrial protein PTPIP51, which is associated with impaired Ca²⁺ homeostasis (156). Loss-of-function mutations in Sig-1R are also strongly implicated in the familial forms of this disease (2, 180) and are closely associated with decreased ER-mitochondria contacts (16, 58). The ALS-linked Sig-1R variants are not able to bind IP₃Rs and thus, to control Ca²⁺ homeostasis at MAMs, leading to calpain activation, mitochondrial dysfunction, and neurodegeneration.

Further, MAMs collapse was observed in motor neurons isolated from Sig-1R-deficient mice (180). SOD1 mutations are one of the most common causes of inherited ALS, and they are also linked with MAMs disruption and Sig-1R aggregation (180), which highlights MAMs dysfunction as a common mechanism underlying both Sig-1R- and SOD1-linked ALS forms.

Huntington's disease

Cherubini et al. were pioneers in implicating ER-mitochondria miscommunication in HD. These authors reported reduced ER-mitochondria contact sites in striatal neurons from R6/1 HD mice (36). Moreover, they demonstrated that HD-related defects on mitochondrial dynamics, caused by increased Drp1-mediated mitochondrial fragmentation, disrupt ER-mitochondria communication, affecting Ca²⁺ homeostasis and leading to excessive ROS production (36). Further, it was reported that a significant depletion of ER-mitochondria tethering proteins, namely IP3R, Grp75, and Mfn2, occurs in the knock-in mutant HdhQ7/Q111 and R6/1 HD mice models (36).

Finally, a significant decrease of MAMs proteins in the putamen of HD patients compared with controls was observed. These findings suggest that the disruption of ER-mitochondria contacts, caused by either upregulated mitochondria fission or loss of MAM-resident proteins, is a relevant mechanism implicated in striatal dysfunction and neurodegeneration in HD (36).

Regulation of mitochondrial and ER functions by Sig-1R

In physiological conditions, Sig-1R forms a complex with the ER-resident chaperone GRP78 at MAMs. When activated by agonists or under ER stress conditions, Sig-1R dissociates from GRP78 and binds to IP₃Rs. The chaperone activity of free Sig-1Rs attenuates aggregation and degradation of IP₃R, enhancing Ca²⁺ uptake by the mitochondria (66). As a result, redox reactions and ATP production are facilitated due to the regulation of Ca²⁺-dependent enzymes in the TCA cycle and of cell death pathways that can be initiated due to mitochondrial Ca²⁺ overload, namely, necrosis and apoptosis (143).

Under stress conditions, Sig-1R also stabilizes IRE-1α at MAMs and enhances cell survival by prolonging the activation of the IRE1-XBP1 signaling pathway. Accordingly, Sig-1R is upregulated by ER stress to preserve cell homeostasis (122) and is translocated to the plasma membrane to interact with and regulate many different membrane proteins, including receptors, ion channels, and kinases (158). Sig-1R also translocates from MAMs to the nuclear membrane on ER stress, interacting with chromatin remodeling factors that, in turn, regulate the activity of transcription factors and thereby gene expression (169) (Fig. 6).

Role of Sig-1R in oxidative stress

Sig-1R promotes cell survival by protecting against oxidative stress in several tissues, including the brain tissue (54), through activation of the antioxidant response element (ARE) and expression of genes encoding the antioxidant enzymes NAD(P)H quinone oxidoreductase 1 (NQO1) and SOD1 (132). Sig-1R knockout (KO) mice showed elevated endogenous ROS levels and glutathione depletion in Müller cells compared with wild-type mice. In addition, protein levels of the antioxidant enzymes SOD1, CAT, NQO1, and GPX1 were found to be decreased in Sig-1R KO Müller cells.

The Sig-1R agonist dehydroepiandrosterone (DHEA) and its sulfated analog (DHEAS) exhibited a neuroprotective role against H₂O₂-induced toxicity in mouse neuroblastoma cells. Moreover, the antioxidant enzymes peroxiredoxin 3 (Prx3) and SOD2 have been shown to be reduced in the hippocampi of mice deficient of zinc finger protein 179 (Znf179) that was identified as a Sig-1R downstream effector (157). Inhibition of Sig-1R by RNAi or antagonists significantly blocked the increase of Ferritin heavy chain 1 (FTH1) and transferrin receptor protein 1 (TFR1) levels induced by erastin and sorafenib, indicating that Sig-1R prevents ROS accumulation by negatively regulating iron metabolism (12).

Metal ions such as iron, copper, and zinc are important in several metabolic processes; therefore, their homeostasis is tightly regulated, particularly in the brain. Dysregulation of redox-active metal ions may cause uncontrolled free radical generation and contribute to the pathophysiology of neurodegenerative diseases (170).

Alzheimer's disease

Sig-1R has been associated with AD pathophysiology. In fact, Sig-1R density is significantly lower in AD patients' brains (68, 119, 167) and genetic studies demonstrated an association between SIGMAR1 polymorphisms and the risk for developing AD (42).

Sig-1R knockdown causes degeneration in mouse hippocampal neurons (68). Moreover, impaired neurogenesis in the dentate gyrus of adult hippocampus due to the dysfunction of glutamate NMDA receptors has been observed in Sig-1R KO mice (152). Accordingly, the protective effect of Sig-1R agonists in in vitro and in vivo models of AD has been reported. The Sig-1R agonists PRE-084, SA4503, ANAVEX1-41, and ANAVEX2-73 prevented learning and memory impairments in mice injected with Aβ_25–35 (111, 112, 173).

Similarly, anti-amnesic and neuroprotective effects induced by the Sig-1R agonist (-)-methyl(1S,2R)-2-{[1-adamantyl(methyl)amino]methyl}-1-phenylcyclopropane-carboxylate have been observed in Aβ-treated rats (5). Improved cognitive function, as well as decreased synapse loss, BACE1 and GSK3β activity, p25/CDK5, neuroinflammation, soluble and insoluble Aβ levels, and plaque and tau pathologies have been reported in 3xTg-AD mice treated with AF710B, a muscarinic and Sig-1R agonist (45). Further, 2-73-(tetrahydro-N,N-dimethyl-2,2-diphenyl-3-furanmethanamine hydrochloride) blocked the accumulation of Aβ_1–42 and C99 in the hippocampus of Aβ-injected mice, a sporadic model of AD (94).

ANAVEX2-73 and PRE-084 prevented dysfunction of the mitochondrial ETC and subsequent oxidative stress and apoptosis in Aβ_25-35-injected mice (93). It has been observed that PRE-084, ANAVEX1-41, ANAVEX3-71, and DHEA induce mild oxidative stress in mitochondria and activate complex I in physiological conditions but protect against pathological oxidative stress, reducing Aβ_1-42-induced ROS formation (54). These results highlight the role of Sig-1R in modulating mitochondrial function and ROS generation in both physiological and pathological conditions.

Although the studies cited earlier support the neuroprotective effect of Sig-1R activation, there are studies showing contradictory results. For example, Sig-1R deficiency was shown to reduce Aβ_25-35-induced hippocampal neuronal cell death and cognitive deficits in mice (189).

Parkinson's disease

Sig-1R has been reported to be decreased in the brain of PD patients (123, 167). In a 6-hydroxydopamine (6-OHDA)-induced PD mouse model, pharmacological modulation of Sig-1R with PRE-084 improved motor deficits, which was accompanied by an increase in dopaminergic fiber density and dopamine levels, neurotrophic factors (BDNF and GDNF), and survival signaling mediators (ERK1/2 and Akt) (48). In another study, afobazole and PRE-084 have been demonstrated to normalize motor function and prevent dopamine depletion in the striatum of the 6-OHDA-induced parkinsonism model (177).

Moreover, it has been observed that physiologic concentrations of dopamine induce apoptosis in Sig-1R knockdown CHO cells by an NF-κB/BCL-2-dependent mechanism (123). These results highlight the potential protective effect of Sig-1R in PD. Clinical benefits have also been reported, namely the attenuation of levodopa-induced dyskinesia in PD patients treated with Sig-1R agonists (47, 133). On the contrary, another study demonstrated that Sig-1R deficiency and pharmacological inactivation reduces 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced motor deficits and loss of dopaminergic neurons, suggesting that Sig-1R deficiency is neuroprotective (75).

Huntington's disease

Sig-1R levels are reduced in the PC6.3 cell line, a neuronal HD model overexpressing the mutant huntingtin protein (80). In this in vitro HD model, PRE084 upregulates antioxidants and inhibits NF-κB signaling, decreasing oxidative stress (80). SIGMAR1 siRNA significantly promoted the cytoplasmatic and nuclear accumulation of aggregates in HeLa cells transfected with N-terminal mutant huntingtin, whereas SIGMAR1 overexpression enhanced the degradation of intranuclear inclusions (118).

Pridopidine was originally described as a dopamine stabilizer. However, its mode of action includes Sig-1R activation, which is required for its effect on BDNF signaling pathways (51). Pridopidine stabilizes striatal medium spiny neurons spines, normalizes ER Ca²⁺ content and release in aged corticostriatal co-cultures from YAC128 transgenic HD mice, and the deletion of neuronal Sig-1R prevented spine protection (125, 148).

Amyotrophic lateral sclerosis

Sig-1R is enriched in motor neurons in the CNS (113). Further, Sig-1R accumulates in postsynaptic densities associated with cholinergic synapses and is enriched at ER structures of alpha motor neurons. A mutation in the gene encoding Sig-1R causes FTLD-ALS and juvenile ALS (2, 103, 180) and its protein levels are reduced in lumbar ALS patient spinal cord. Pharmacological Sig-1R activation induced the clearance of mutant protein aggregates in mice-cultured motor neurons (139). In mutant SOD1-overexpressing mice, Sig-1R deficiency accelerated the onset of ALS manifestations and induced MAMs disruption, calpain activation, and mitochondrial dysfunction (180).

Further, SIGMAR1 knock-out in an ALS mice model potentiated motor performance impairment and decreased longevity (114). Recently, Couly et al. provided evidence that mutant Sig-1R overexpression recapitulates ALS pathology in vivo and that Sig-1R rescues locomotor activity and ATP levels in flies overexpressing TDP-43, the key ALS protein (37).

The benefits of Sig-1R activation in ALS were obvious. Administration of the Sig-1R agonist SA4503 to an ALS mouse model prevented motor neuron degeneration and symptom progression, together with extended survival. Moreover, SA4503 reduced SOD1G93A mutation and serum free-induced cell death in NSC34 cells (128). It has also been reported that SA4503 can normalize bradykinin-sensitive intracellular Ca²⁺ stores and avoid cytosolic Ca²⁺ overload in spinal neurons isolated from embryonic mice overexpressing human SOD1 with the ALS-causing mutation G93A (162). In this mouse model, improved spinal motor neuron function, locomotor performance, preservation of neuromuscular junctions and of motor neurons in the lumbar spinal cord have been described on treatment with the agonist PRE-084 (107).

Further, pridopidine attenuated deficits in Sig-1R axonal transport, disruption of neuromuscular junction, and cell death in SOD1G93A motor neurons and in neuromuscular co-cultures (81). Recently, Couly et al. provided evidence that mutant Sig-1R recapitulates ALS pathology in vivo whereas Sig-1R rescues locomotor activity and ATP levels of flies overexpressing TDP-43 (37).

All the studies cited earlier support the involvement of Sig-1R in the pathophysiology of neurodegenerative diseases, and the effects of its modulation pinpoint Sig-1R as a potential therapeutic target for these disorders.