Aberrant S-nitrosylation in the TCA cycle contributes to mitochondrial dysfunction, energy compromise, and synapse loss in neurodegenerative diseases

Abstract

Neuronal synaptic activity relies heavily on mitochondrial energy production, as synaptic transmission requires substantial ATP. Accordingly, mitochondrial dysfunction represents a key underlying factor in synaptic loss that strongly correlates with cognitive decline in Alzheimer's disease and other neurocognitive disorders. Increasing evidence suggests that elevated nitro-oxidative stress impairs mitochondrial bioenergetic function, leading to synaptic degeneration. In this review, we highlight the pathophysiological roles of nitric oxide (NO)-dependent posttranslational modifications (PTMs), particularly S-nitrosylation of cysteine residues, and their impact on mitochondrial metabolism. We focus on the pathological S-nitrosylation of tricarboxylic acid cycle enzymes, particularly α-ketoglutarate dehydrogenase, as well as electron transport chain proteins. This aberrant PTM disrupts mitochondrial energy production. Additionally, we discuss the consequences of aberrant protein S-nitrosylation on mitochondrial dynamics and mitophagy, further contributing to mitochondrial dysfunction and synapse loss. Finally, we examine current strategies to ameliorate S-nitrosylation-mediated mitochondrial dysfunction in preclinical models of neurodegenerative diseases and explore future directions for developing neurotherapeutics aimed at restoring mitochondrial metabolism in the context of nitro-oxidative stress.

Introduction

Synaptic neurotransmission, which requires substantial energy consumption, enables communication within the neuronal network, thus constituting the cell-based foundation for brain function, including cognition, emotion, and memory [1,2]. To properly fulfill the great energy demand of synapses, functional mitochondria are enriched in pre- and postsynaptic regions, producing energy in the form of NADH and ATP within these cellular compartments [3]. Mitochondria play a critical role in calcium homeostasis, redox signaling pathways, epigenetics, neuroinflammation, and cell death programs, including apoptosis, necrosis, and ferroptosis [4]. Hence, physiological regulation of mitochondrial activity is essential for the maintenance of brain function.

In contrast, mitochondrial abnormalities, affecting morphology, structure, dynamics, and mitophagy, lead to bioenergetic compromise and decreased calcium buffering. Such mitochondrial dysfunction contributes to synaptic impairment and eventually neuronal cell damage and death in neurocognitive disorders (NCD). NCDs are generally characterized by a progressive decline in cognitive function and can be cortical or subcortical in character. Examples of NCDs include Alzheimer's disease (AD), Lewy body dementia (LBD), vascular dementia, and frontotemporal dementia. In these NCDs, synaptic loss, driven at least in part by mitochondrial dysfunction, represents a major pathological correlate with cognitive impairment [[5], [6], [7]]. Accordingly, elucidation of the molecular events underlying mitochondrial dysfunction will advance our understanding of the neuropathological processes leading to synaptic loss and cognitive impairment, and may point to new therapeutic targets.

Additional key hallmarks of NCDs encompass aberrant protein aggregation and proteostasis, neuroinflammation, enhanced oxidative/nitrosative stress, and excitotoxicity [[8], [9], [10]]. Among these neuropathological features, growing evidence suggests that increased production of reactive oxygen/nitrogen species (ROS/RNS), resulting in oxidative and nitrosative stress, plays a crucial role in mitochondrial impairment in neurodegenerative diseases [11,12]. Furthermore, a number of neuropathological factors, including accumulation of aggregated proteins (e.g., amyloid-β [Aβ], α-synuclein, tau, and TDP-43), environmental toxins like air pollution, excitotoxicity, and neuroinflammation, are known to stimulate ROS/RNS production in diseased brains, consistent with the notion that ROS/RNS represents important signaling molecules, contributing to mitochondrial dysfunction and synaptic deficits. In the current review article, we therefore focus on ROS/RNS-dependent signaling pathways, particularly the PTM of S-nitrosylation, mediated by nitric oxide (NO)-related species, and discuss how aberrant protein S-nitrosylation contributes to dysregulation of mitochondrial structure and function, with consequent synaptic damage and loss. Specifically, we focus on effects of aberrant protein S-nitrosylation on mitochondrial metabolism leading to synaptic impairment, and thus, contributing to cognitive decline. Strikingly given the large numbers of SNO-targets identified (as delineated below), S-nitrosylation appears to be comparable to O-phosphorylation in terms of its impact on protein activity and cellular physiology [13]. However, of the thousands of SNO-proteins discovered to date, only a few dozen have been studied in detail for their specific roles in disease progression, underscoring the need for more intensive research in this area.

Aberrant Protein S-Nitrosylation in Neurodegenerative Diseases

NO, first discovered as a vasodilating agent, has emerged as a ubiquitous signaling molecule for its role in neurotransmission, inflammation, and vascular homeostasis [9,10,14,15]. NO is synthesized in cells by three different types of NO synthases: Neuronal NOS [nNOS], inducible NOS [iNOS], and endothelial NOS [eNOS], which produce NO and L-citrulline from molecular O₂ and L-arginine, with electrons donated from NADPH [16] (Fig. 1). nNOS and eNOS activity depend on intracellular calcium signaling mediated via the Ca²⁺/calmodulin system, while iNOS is upregulated transcriptionally under various cellular stress conditions such as inflammation, and its activity is not dependent on calcium [17]. These different activation mechanisms provide temporal and spatial selectivity to NO signaling under physiological vs. pathophysiological conditions. Additionally, in vivo, nitrate converted to nitrite, in part via commensal bacteria in the gut, can be further metabolized nonenzymatically or enzymatically to NO [18,19]. Moreover, a recent study demonstrated that mitochondrial sulfite oxidase catalyzes the NOS-independent production of NO in astrocytes under hypoxic conditions [20]. In this scheme, mitochondrial sulfite oxidase uses molybdenum as a transition element cofactor, generating NO via nitrite reduction. Nitrite can be derived from dietary sources or produced endogenously, offering an additional pathway for NO production.

Fig. 1.

Formation of S-nitrosothiols in cells. (a) In cytoplasm, NOS (nNOS, eNOS, and iNOS) use substrate l-arginine, NADPH and O₂ to produce L-citrulline and NO. Transition metal ions (e.g., Cu²⁺, Fe³⁺) can produce nitrosonium ion (NO⁺)-character that supports reaction with cysteine thiolate anion to form an S-nitrosylated protein (RSNO) or low-molecular weight compound like GSNO or SNO-CoA [24]. Additionally, the thiolate anion of one protein can mount a nucleophilic attack on the nitroso nitrogen of an RSNO or LMW-SNO to effect transnitrosylation, i.e., transfer of NO-related species from one protein or compound to another. (b) Various denitrosylases maintain equilibrium levels of SNO proteins in the cell. GSNO reductase (GSNOR) and SNO-CoA reductase (SNO-CoAR) regulate levels of GSNO and CoA-SNO, thereby indirectly controlling protein S-nitrosylation. In contrast, thioredoxins (Trx) directly denitrosylate SNO-proteins. TrxR: Thioredoxin Reductase. Image created using BioRender.com.

Soon after the discovery of NO as a signaling molecule, we and our colleagues found that, in addition to the pathway activating guanylyl cyclase to produce cyclic guanosine monophosphate (cGMP), a redox-based PTM, subsequently termed protein S-nitrosylation (SNO-protein), can mediate the majority of biological activities of NO [[21], [22], [23]]. S-Nitrosylation results from the covalent reaction of NO-related species with the cysteine thiol group (more properly, thiolate anion) of the target protein. Note that NO itself does not readily react with the thiol group to generate an S-nitrosothiol modification on a protein. Instead, NO is thought to initially interact with a transition metal (such as Cu²⁺ and Fe³⁺), and the resulting NO-related species with the NO⁺-like character reacts with thiolate on the target protein [24,25]. In another possible NO auto-oxidation pathway, NO is oxidized to N₂O₃, a nitrosating agent that can react with thiolate in vitro, leading to protein S-nitrosylation; however, we favor the transition metal pathway, which is more in line with our empirical findings [26].

Similar to other PTMs, S-nitrosylation often modulates a protein's structure, function, and interaction with other proteins. The targets of S-nitrosylation encompass a wide variety of proteins including, but not limited to, transcription factors, ion channels, enzymes, G-protein-coupled receptors, and structural proteins [9,10,27]. Mechanistically, both non-enzymatic and enzymatic mechanisms can facilitate specific targeting of NO-related species to a particular cysteine residue to mediate protein S-nitrosylation [9,10,28,29]. Various cellular and chemical factors also affect protein S-nitrosylation, providing a layer of specificity to S-nitrosylation signaling. These factors include (i) location of reactants (e.g., in a close proximity to the source of NO production), (ii) amino acid motif surrounding the target cysteine (i.e., the I/L-X-C-X2-D/E, SNO motif), (iii) local environment that affects pKa, hydrophobicity, and redox potential of the target cysteine, (iv) availability of transition metal co-factors, (v) mass action (SNO-protein concentration), and (vi) protein-protein interactions among reactants [9,10,[30], [31], [32], [33]]. The enzymatic mechanism typically involves a transnitrosylation reaction in which an NO group is transferred from one protein to another or from a low molecular weight (LMW) SNO to a protein [9,10,28,29,34]. In this context, the thiolate anion of one protein mounts a nucleophilic attack on the nitroso nitrogen of the second protein, resulting in transfer of the NO group from one protein to the other [24]. In some cases, the transnitrosylating protein acts as an enzyme to catalyze transfer of NO-related species to a target protein. For example, in human AD brains and in AD model systems, we found that transfer of NO-related species from SNO-Uch-L1 to Drp1 via SNO-Cdk5, with SNO-Cdk5 apparently acting enzymatically to effect the transfer; these reactions contribute to mitochondrial fragmentation and synapse loss [28,[35], [36], [37]]. Another example of apparent enzymatic transfer of NO-related species involves SNO-CoA as a co-factor for the enzyme SNO-CoA-assisted nitrosylase (SCAN) in mediating S-nitrosylation of multiple targets, such the insulin receptor and insulin receptor substrate 1 [29].

Depending on the level of NO produced and the specific proteins that are S-nitrosylated, either neuroprotective or neurodestructive pathways can be engaged. For example, lower concentrations of NO can induce physiological S-nitrosylation, e.g., as negative feedback on the NMDA-type of glutamate receptor to prevent its overactivation, thus enhancing neuroprotection [21,23,26,38]. In contrast, excessive nitrosative/oxidative stress under pathological conditions can cause aberrant S-nitrosylation of cysteine residues located in a partial SNO motif and thus are not normally S-nitrosylated, resulting in pathologically dysfunctional activity [9].

Note that under specific conditions NO-related species can be a good leaving group, facilitating reaction of the cysteine thiol with ROS, leading to further oxidation to form sulfenylated (-SOH), sulfinylated (-SO₂H) or sulfonylated (-SO₃H) adducts [9,10,39]. Alternatively, the presence of an S-nitrosothiol can facilitate disulfide bond formation with a vicinal cysteine residue [23,26,40]. These –SO₂H, –SO₃H, and disulfide bond modifications are generally more stable cysteine modifications, with –SO₃H being considered irreversible. Such redox-based modifications may therefore provide more long-lasting effects on protein function. However, these reactions can also occur on different thiol sites that are independent of SNO modification. Additional PTMs of cysteine residues, including persulfidation, palmitoylation, glutathionylation and carbonylation, can all interfere with the formation of an S-nitrosothiol on the same thiol group and sometimes exert opposing effects on the protein activity and function [41,42]. The complex interplay between various cysteine oxidation reactions remains to be studied.

Cells typically maintain redox homeostasis through various mechanisms. For example, ascorbate, metal ions, antioxidant enzymes, glutathione, the thioredoxin (Trx) system, and molecular chaperones neutralize excessive free radicals such as NO generated during cellular and metabolic processes [[43], [44], [45]]. More specifically, denitrosylases, such as Trx [46], S-nitrosoglutathione (GSNO) reductase, and SNO-coenzyme A (SNO-CoA) reductase, remove SNO adducts from peptides or proteins [47] (Fig. 1). This action keeps the levels of SNO-proteins in check, conferring protection against nitro-oxidative stress. However, advanced age, neuroinflammation, protein aggregation, and exposure to environmental factors such as air pollution can increase the production of ROS/RNS in the brain, overwhelming cellular defense systems [12]. This redox imbalance leads to oxidative or nitrosative stress, and can activate aberrant transnitrosylation networks that contribute to neuropathological processes [35]. Future characterization of transnitrosylation networks in both physiological and pathological settings is expected to provide significant insight into the underlying mechanisms of neurodegenerative as well as systemic diseases.

Identification of SNO-proteins in vivo has been challenging, in part due to its labile nature of S-nitrosothiols in cellular contexts. Nevertheless, various research groups have developed strategies for comprehensive detection of SNO-proteins and profiling their specific sites of S-nitrosylation. The majority of these techniques involve derivatization of the SNO modification into a more stable adduct, such as biotin. This enables enrichment and subsequent detection via immunoblotting (for examining individual SNO-proteins) or mass spectrometry (MS; for identifying the S-nitroso-proteome). These SNO-protein detection methods include the biotin-switch assay, SNO site identification (SNOSID), irreversible biotinylation procedures (IBP), SNO-resin-assisted capture (SNO-RAC), Cys-biorthogonal cleavable-linker and switch technique (Cys-BOOST), organomercury-assisted enrichment, and S-nitrosothiol trapping by triaryl phosphine (SNOTRAP) [35,[48], [49], [50], [51], [52], [53], [54], [55]]. In the biotin-switch, IBP, SNO-RAC and Cys-BOOST assays, S-nitrosylated cysteine residues are selectively reduced by ascorbate, followed by derivatization of the newly formed thiol with a biotin, thiol reactive resin, or iodoacetamide-alkyne. Alternatively, organomercury-assisted enrichment and SNOTRAP directly capture SNO-cysteines in the proteome, providing more efficient and sensitive approaches for identification of both individual SNO-sites and entire transnitrosylation networks or proteins. Importantly, these techniques have enabled researchers to identify a wide range of S-nitrosylated proteins in human postmortem tissues, animal models of disease, and cell-based systems, shedding light on their role in neurodegenerative disorders such as AD and related dementias (ADRD) [[55], [56], [57], [58], [59], [60]]. In fact, many identified SNO-proteins are involved in critical cellular processes such as mitochondrial function, neurotransmission, neuroinflammation, and protein folding/misfolding or degradation; dysregulation of these processes represents a key hallmark of neurodegenerative diseases. These studies of the S-nitrosoproteome have established the pathological role of nitrosative stress in many neurological diseases.

Disruption of TCA Cycle Flux in Neurodegenerative Diseases

As one of the most energy-demanding organs in the human body, the brain utilizes glucose for its high-level energy production. Glucose metabolism is initially conducted independently of oxygen consumption through the anaerobic and cytosolic glycolytic pathway, yielding a net of two molecules each of ATP and pyruvate. For each 6C (six carbon) glucose metabolized by glycolysis, two molecules of 3C pyruvate then enter mitochondria to be processed through the tricarboxylic acid (TCA) cycle (also known as the citric acid cycle or Krebs cycle), which in conjunction with the electron transport chain (ETC) and oxidative phosphorylation (OxPhos) provides a high yield of ATP (approximately 32 more ATP). Under basal conditions, glycolytic capacity in neurons is reported to be limited [61]. Instead, glial cells are thought to take up glucose, perform glycolysis, and provide pyruvate that is converted to lactate. It has been proposed that the lactate is then shuttled to neurons, and converted back to pyruvate to enter the TCA cycle and OxPhos-dependent ATP production (termed the astrocyte-neuron lactate-shuttle hypothesis) [62]. Nonetheless, highly stimulated synaptic terminals can directly metabolize glucose, so neurons can use some glucose directly [63]. Hence, while both glycolysis and OxPhos can contribute to ATP production in neurons, mitochondrial metabolism appears to be the most critical for normal basal neuronal activity. Accordingly, a decrease in mitochondrial metabolism results in insufficient ATP production for basal neurotransmission, contributing to synaptic impairment and neuronal injury in chronic neurological conditions such as AD and Parkinson's disease (PD) [64,65].

As the initial step for mitochondria-dependent energy production, pyruvate is imported into mitochondria. The pyruvate dehydrogenase (PDH) complex catalyzes the conversion of pyruvate to acetyl-CoA, entering the TCA cycle, while the reaction also yields NADH for the ETC to produce ATP (as summarized in Fig. 2). Critically, a significant decrease in the activity of PDH complex, isocitrate dehydrogenase (IDH), and the α-ketoglutarate (αKG) dehydrogenase (αKGDH) complex occurs in sporadic AD brains and correlates with cognitive decline [66]. Additionally, the activity of the αKGDH complex is also diminished in PD brains [67]. These findings are consistent with the notion that dysfunction of TCA cycle enzymes likely contributes to dysfunctional mitochondrial metabolism in multiple neurodegenerative disorders, causing impairment in synaptic transmission because of its critical dependence on energy.

Fig. 2.

S-Nitrosylation inhibits TCA cycle flux predominantly at the αKGDH step. The TCA cycle uses acetyl-CoA from glycolysis to produce NADH, FADH₂, ATP, and CO₂. NADH molecules provide electrons to the ETC. S-Nitrosylation inhibits the activity of many enzymes in the TCA cycle, thus disrupting mitochondrial metabolism. S-Nitrosylation-dependent inhibition of TCA cycle flux predominantly occurs at the αKGDH step, with minor-to-moderate inhibition also occurring at other TCA cycle enzymes, such as Aco, IDH, and malate dehydrogenase (MDH). While some of these reactions may represent physiological regulation of TCA cycle flux [60], aberrant S-nitrosylation can inhibit flux to the point of being pathological [57].

Nitrosative/Oxidative Stress Inhibits αKGDH Complex Activity, Causing a Major Block in the TCA Cycle, Thus Contributing to Synaptic Impairment

Decreased activity of the αKGDH complex in neurodegenerative diseases

Located in the mitochondrial matrix, the αKGDH complex consists of three subunits: α-ketoglutarate dehydrogenase (αKGDH/E1), dihydrolipoyl succinyltransferase (DLST/E2), and dihydrolipoyl dehydrogenase (DLD/E3). Note that, while αKGDH/E1 and DLST/E2 are unique to the αKGDH complex, the DLD/E3 subunit is also shared by the PDH complex and the branched-chain α-keto acid dehydrogenase (BCDH) complex. Structurally, the αKGDH complex contains an inner core composed of DLST/E2 subunits, surrounded by multiple αKGDH/E1 and DLD/E3 proteins [68]. The αKGDH complex catalyzes the overall reaction of αKG ​+ ​CoA-SH ​+ ​NAD⁺ → succinyl-CoA ​+ ​CO₂ ​+ ​NADH using thiamine diphosphate (TPP), lipoic acid, and FADH₂ as cofactors. To initiate the αKGDH reaction, αKG binds to TPP on the αKGDH/E1 subunit, facilitating decarboxylation of αKG and formation of TPP-hydroxysuccinate, while releasing CO₂ [69]. The αKGDH/E1 enzyme subsequently catalyzes the transfer of a succinyl group to oxidized lipoic acid bound to specific lysine residues in the core enzyme, DLST/E2. Next, the newly introduced succinyl group on DLST/E2 reacts with the thiol group of coenzyme A (CoA) to generate succinyl-CoA (which is used by succinyl-CoA synthetase [SCS] later in the TCA cycle) as well as reduced lipoic acid (i.e., dihydrolipoic acid) on DLST/E2. After this reaction, electrons from dihydrolipoic acid on DLST/E2 are utilized to reduce the disulfide bond present at the active site of the DLD/E3 subunit, while regenerating oxidized lipoic acid on DLST/E2. The DLD/E3 subunit also contains the FAD⁺-binding domain and NAD⁺-binding domain, allowing the transfer of the reducing equivalents from the reduced sulfhydryl groups to FAD⁺ and then to NAD⁺ to produce NADH.

Evidence from our laboratory and others strongly suggests that enhanced oxidative/nitrosative stress contributes to mitochondrial dysfunction in neurodegenerative disorders at least in part via aberrant S-nitrosylation of TCA-cycle enzymes and ETC complexes, leading to defective energy production and subsequent impairment in synaptic function [12,56,57,59,70]. Concerning S-nitrosylation of TCA cycle enzymes, S-nitrosothiol formation occurs as a physiological regulator of activity for several TCA cycle proteins under the basal conditions [56,57,59,71,72]. However, our recent proteomics studies showed in postmortem brains of patients with AD, synucleinopathies (such as PD and LBD), and HIV-associated neurocognitive disorder (HAND), aberrantly increased levels S-nitrosylation on specific TCA cycle enzymes, including IDH, aconitase (Aco), and, particularly, the αKGDH complex [56,57,59] (Fig. 2). Moreover, our metabolic flux assessment using ¹³C-lactate in human induced pluripotent stem cell (hiPSC)-based models of AD and PD/LBD identified the major block in the TCA cycle at the αKGDH/SCS step, with more moderate inhibition of Aco and IDH [56,57]. Critically, these repressive effects on TCA cycle flux could be reversed by the NO synthase inhibitor, l-NAME, consistent with the notion that aberrant S-nitrosylation inhibits TCA cycle enzymes in AD and PD/LBD brains. Moreover, α-KGDH activity has been reported to be significantly decreased in human AD brains compared to healthy, age-matched controls [66,73]. With this background, our group then demonstrated that administration of a membrane permeable form of succinate, such as dimethyl succinate (DMS), can partially rescue bioenergetic compromise and synapse number in hiPSC models of AD, likely by bypassing blockade at the αKGDH step [57]. These findings support the premise that RNS/ROS-mediated inhibition of αKGDH activity contributes to cognitive impairment and other AD-related pathologies.

Considerable in vitro and cell-based evidence suggests that ROS/RNS as well as other oxidants inactivate the activity of the αKGDH complex [74], at least in part, via oxidative modifications of all three subunits [[75], [76], [77]]. For example, peroxynitrite (the reaction of NO^. plus O₂^.- to form ONOO⁻) inhibits the αKGDH complex activity via nitration of tyrosine residues in all three subunits [78]. Additionally, singlet oxygen, an oxidant generated in mitochondria, can modify Met, Trp, His, and Tyr residues of all three subunits, leading to loss of αKGDH complex activity [79]. However, whether oxidative/nitrosative modifications indeed occur in human diseased brain to downregulate αKGDH complex activity remained unclear until our recent S-nitrosoproteomics studies found that all subunits of the αKGDH complex were S-nitrosylated in human postmortem brains from patients with AD, PD/LBD, and HAND [[55], [56], [57],59].

Aberrant S-nitrosylation of the E1 subunit in neurodegenerative diseases

We recently discovered the presence of S-nitrosylated αKGDH/E1 subunit in AD, PD/LBD, and HAND brains [[55], [56], [57],59]. Moreover, consistent with other observations on diseased brains, neuronal cells differentiated from hiPSCs carrying AD-related mutations also exhibited increased SNO-αKGDH/E1 levels compared to isogenic gene-corrected control neurons [57]. In total, we identified five S-nitrosylation sites in the αKGDH/E1 subunit in human brains: Cys283, 395, 487, 566, and 802. Based on the recently solved Cryo-EM structures of the human αKGDH/E1 homodimer, most of these SNO-sites are in the α/β1 domain (ranging from amino acids 236 to 594), where the TPP ligand is bound and buried in the dimer interface [80]. It remains unresolved whether S-nitrosylation of αKGDH/E1 affects its structure to influence TPP binding, consequently inhibiting its activity.

Aberrant S-nitrosylation of the E2 subunit in neurodegenerative diseases

Our recent S-nitrosoproteomics studies revealed that human postmortem AD and PD/LBD brains contain the S-nitrosylated form of the DLST/E2 subunit [[55], [56], [57]]. For example, SNOTRAP-based proteomics found SNO-DLST/E2 in LBD brains but not in control brains [56]. We identified Cys245 of E2/DLST as a target of S-nitrosylation [55]. Cys245 resides near the N-terminal of the core domain, which contains the catalytic site of human E2/DLST [81,82].

Nitro-oxidative PTMs are present not only on cysteine residues but also sulfur atoms of the lipoyl group in DLST/E2 [55,56,83,84]. For example, glutathionylation of DLST/E2 subunits occurs on the active-site cofactor, lipoic acid [83]. Additionally, singlet oxygen-mediated oxidation of lipoic acid on DLST/E2 results in inhibition of αKGDH activity [79]. Along these lines, a recent study demonstrated that S-nitrosylation of the lipoic arm in DLST/E2 blocks its catalytic activity in macrophages [84]. Specifically, in this paradigm S-nitroso-CoA (SNO-CoA), where the –SH group of CoA reacts to form an –SNO group, acts as an endogenous transnitrosylating agent to cause a series of modifications to thiols on the lipoyl group, including S-nitrosothiol formation as well as other chemical products, such as disulfenamide and thiosulfinate. Investigations determining if SNO-CoA-dependent S-nitrosothiol formation on lipoic acid leads to inhibition of the DLST/E2 subunit in neurodegenerative disorders are eagerly awaited. Additionally, it remains to be determined if S-nitrosylation of DLST/E2 restricts transfer of the succinyl group onto CoA in the active center and/or affects core assembly of the αKGDH complex.

Aberrant S-nitrosylation of the E3 subunit in neurodegenerative diseases

The E3 component (DLD/E3) of the αKGDH complex catalyzes NADH production to supply the ETC. The human DLD/E3 protein contains 10 cysteine residues, including redox sensitive Cys80 and Cys85 (or Cys45 and Cys50 in the mature form of DLD/E3), located at the active-site center. Using our SNOTRAP-MS method, we found that S-nitrosylation of DLD/E3 at the active site is greatly increased in human postmortem AD brains compared to controls [55,57]. Intriguingly, the increase in SNO-DLD/E3 occurs in female AD brains to a greater extent than males [55]. This finding is particularly important given evidence that females are at far greater risk for AD than their male counterparts [85]. The finding of SNO-DLD/E3 in female AD brains is consistent with our hypothesis that aberrant protein S-nitrosylation may represent one of the underlying mechanisms for higher AD risk in women. Additionally, similar to findings in AD brains, AD hiPSCs-derived neurons manifest elevated levels of SNO-E3/DLD compared to isogenic, gene-corrected controls [57]. Our S-nitrosoproteomics studies also revealed that SNO-E3/DLD is present in human postmortem LBD and HAND brains [56,59], although advanced age may also be a precipitating factor facilitating generation of SNO-DLD/E3. Additionally, in full accordance with our findings that the SNO modification occurs at the active center of DLD/E3, S-nitrosylation inhibits its enzymatic activity [59]. Note that DLD/E3 is prone not only to S-nitrosylation but also to other forms of oxidative modifications such as sulfenation, carbonylation, and tyrosine nitration [[86], [87], [88], [89]].

Furthermore, a recent study found that nitroxyl (HNO) induces oxidative modifications of DLD at Cys477 and Cys484 [90]. In this context, the process of endogenous HNO generation can begin with formation of an S-nitrosothiol on the reduced lipoic acid attached to the DLST/E2 subunit. This is followed by nucleophilic attack on the S-nitroso-lipoic acid by the second thiol group within the same lipoate molecule, likely resulting in production of oxidized lipoic acid and HNO. Our group and others previously demonstrated that HNO or nitroxyl anion (i.e., specifically, singlet NO⁻) can react with critical thiols of proteins, such as the NMDA receptor, to yield a disulfide bond with a vicinal thiol, or generate a sulfinamide product [91,92]. Building on these previous findings, DLST/E2-derived HNO likely targets nearby DLD/E3 within the αKGDH complex, leading to the formation of a sulfinamide or a disulfide bridge between Cys477 and Cys484. In silico simulation suggests that these modifications impair the formation of DLD homodimers, thereby compromising DLD activity. While this HNO-dependent mechanism has been observed in the macrophage PDH complex [90], it is likely to apply to the αKGDH complex as well, since both complexes share the same DLD subunit.

Cys484 of DLD/E3 is the predominant target of S-nitrosylation according to our S-nitrosoproteomic MS studies on our hiPSC models of PD/LBD [56]. As described above, an S-nitrosothiol can facilitate disulfide bond with a nearby cysteine residue [23,26,40]. In addition, a disulfide bridge involving Cys484 is predicted to disrupt DLD activity [90]. These results suggest that NO-mediated signaling appears to suppress DLD activity via several mechanisms, with S-nitrosylation likely playing a major role, ultimately contributing to the bioenergetic failure linked to synaptic damage.

Enhanced S-Nitrosylation of Additional TCA Cycle Enzymes May Contribute to Defects in Energy Production, Contributing to Synaptic Dysfunction

Protein S-nitrosylation occurs at multiple steps in the TCA cycle

In neurons undergoing neurodegenerative processes, nitrosative stress disrupts the activity of the αKGDH complex, causing a major block at this step in the TCA cycle [57]. However, studies from our group and others have demonstrated that NO signaling also exerts inhibitory effects on other TCA cycle enzymes, including PDH, Aco, and IDH, further contributing to bioenergetic compromise [[55], [56], [57],59]. In this section, we discuss the impact of NO signaling on the PDH complex, Aco, and IDH, and their potential roles in synaptic impairment.

NO-related species inhibit the PDH complex

The PDH complex consists of three enzymatic components: pyruvate dehydrogenase (PDH/E1), dihydrolipoamide acetyltransferase (DLAT/E2), and DLD/E3. The PDH complex catalyzes the conversion of pyruvate to acetyl-CoA, which subsequently enters the TCA cycle while reducing NAD⁺ to NADH. Specifically, the PDH/E1 component forms a heterotetrametric structure, containing two catalytic α subunits and two regulatory β subunits to perform decarboxylation of pyruvate. The DLAT/E2 subunit contains covalently bound lipoate, accepting the acetyl group from PDH/E1 to produce acetyl-CoA. The DLD/E3 subunit, which is shared between the PDH and αKGDH complexes, reduces NAD⁺ to generate NADH, as described above. Notably, NO-related species target the PDH complex and impair its catalytic activity, thereby restricting carbon flow into the TCA cycle [90,93]. Mechanistically, SNO and other nitro-oxidative modifications thus exert inhibitory effects on DLD/E3 activity in neurodegenerative brains [55,57,59,93]. Concerning other components of the PDH complex, cysteine residues present in both the α and β subunits of the PDH/E1 component as well as the DLAT/E2 subunit undergo S-nitrosylation in neurodegenerative brains [[55], [56], [57],59]. As an example, Cys263 of the β subunit of PDH/E1 is S-nitrosylated not only in HAND brains but also in HAND ​+ ​Meth brains (representing patients with HAND and methamphetamine [Meth] use, resulting in further increases in oxidative and nitrosative stress). Further, Cys291 of DLAT/E2 is S-nitrosylated exclusively in HAND ​+ ​Meth brains [59]. However, the role of S-nitrosylation on the activity of PDH/E1 and DLAT/E2 remains to be fully explored.

NO-related species inhibit mitochondrial aconitase (Aco2)

Two aconitase isozymes exist in human cells: Cytosolic aconitase 1 (Aco1) and mitochondrial aconitase 2 (Aco2). Both Aco enzymes contain a [4Fe–4S] iron–sulfur cluster at their active site essential for the conversion of citrate to isocitrate. To coordinate the iron–sulfur cluster, three cysteine residues in the active site directly bind three iron ions, while the fourth iron is more loosely associated with the substrate, citrate [94]. Recently, we reported that Aco2 is S-nitrosylated in brains with AD, PD/LBD, and HAND/HAND ​+ ​Meth [[55], [56], [57],59]. Our SNOTRAP MS proteomics identified an aberrant increase in S-nitrosylation of Aco2 at Cys126, Cys385, and Cys448 in postmortem AD brains [55,57]. Among these cysteine residues, Cys385 and Cys448 are responsible for iron-sulfur cluster binding, while Cys126 is positioned adjacent to the active site. In addition to S-nitrosylation, a previous study found that Tyr151 and Tyr472, which are also located near the active site, can undergo nitration by peroxynitrite [95]. Prior studies showed that RNS/ROS inhibit mitochondrial Aco2 activity via disruption of the active site [96,97]. Consistent with these findings, we found moderate inhibition of Aco2 in hiPSC-derived cerebrocortical neurons after S-nitrosylation of Aco2 [56,57,59]. In mouse macrophage undergoing inflammatory polarization, Aco2 appears to be a primary target of iNOS-derived NO, leading to the disruption of mitochondrial metabolism [93]. In contrast, as discussed above, S-nitrosylation primarily inhibits TCA cycle flux by impairing the αKGDH complex, with moderate inhibition also occurring at the IDH and Aco2 steps in neurons in brains with neurodegenerative disorders. These findings suggest that the inhibitory effects of NO-related species on TCA cycle activity may vary depending on cell, tissue type, and disease state.

In addition to S-nitrosylation, the active site cysteine residues in Aco2 are susceptible to other types of PTMs. For example, fumarate supports transfer of a succinyl group to the thiols of the three cysteine residues that coordinate the iron-sulfur cluster, resulting in formation of S-(2-succino)-cysteine (i.e., succination) [98]. Additionally, the active site cysteine residues can undergo carbonylation as well as oxidation to sulfonic acid (-SO₃H) [95,99]. These modifications impair Aco2 activity, possibly by disrupting iron chelation. It would be important in the future to investigate whether S-nitrosylation competes with or facilitates other cysteine modifications such as succination or sulfonylation to influence TCA cycle flux in the context of AD and other neurological disorders.

NO-related species inhibit mitochondrial isocitrate dehydrogenase (IDH)

In human cells, three distinct isoforms of isocitrate dehydrogenase (IDH) are expressed: IDH2 and IDH3 are localized in the mitochondria, while IDH1 is typically found in the cytosol [100]. All three IDH isoforms catalyze the conversion between isocitrate and α-ketoglutarate. IDH1 and IDH2 are NADP⁺-dependent enzymes, whereas IDH3 is NAD⁺-dependent, producing NADH for the ETC. Structurally, IDH1 and IDH2 form a homodimer, whereas IDH3 is a heterotetramer composed of two α-subunits (IDH3A), one β-subunit (IDH3B), and one γ-subunit (IDH3G), with the α-subunits forming the catalytic site and the β- and γ-subunits acting as regulatory components [101]. IDH3 is generally regarded as the IDH enzyme involved in forward metabolic flux through the TCA cycle. Interestingly, IDH2 can catalyze both forward and reverse reactions, depending on cellular requirements.

In hiPSC-based models of AD and PD/LBD, our recent ¹³C-based metabolic flux assessments revealed modest-to-moderate inhibition of the TCA cycle at the IDH step [56,57]. Our study found that nitrosative stress is at least in part responsible for the inhibition of IDH since the inhibition was relieved by l-NAME [59]. With regard to S-nitrosylation of IDH, our S-nitrosoproteomics MS analysis identified altered levels of mitochondrial SNO-IDH proteins in human brains with neurodegenerative diseases [[55], [56], [57],59]. For example, concerning IDH3 subunits, we found S-nitrosylation of all three subunits in postmortem human brains with AD, PD/LBD, or HAND/HAND ​+ ​Meth. Specifically, our organomercury-based S-nitrosoproteomics found that S-nitrosylation of IDH3A at Cys331 is significantly elevated in AD brains. In addition, our SNOTRAP proteomics revealed that Cys185 of IDH3B is S-nitrosylated, but only in female AD brains. Regarding IDH2, our SNO-proteomics found SNO-IDH2 formation in AD, PD/LBD, and HAND brains. As an example, organomercury-mediated S-nitrosoproteome characterization identified S-nitrosylation of IDH2 at Cys402 in AD brains.

In non-CNS cells, S-nitrosylation and other oxidative modifications of mitochondrial IDH proteins also affect IDH activity. For example, similar to our findings in neurons, prior studies demonstrated that S-nitrosylation at Cys305 and 387 of IDH2, isolated from pig heart tissue, inhibits the IDH activity [102]. A recent molecular dynamics simulation study predicted that mild oxidative stress induces a disulfide bridge formation between Cys148 and 284 of IDH3G, enhancing IDH3 activity in cardiomyocytes [101]. Because S-nitrosylation can facilitate such disulfide bond formation, reaction mechanism may possibly occur in neurons but remains to be studied.

S-Nitrosylation can inhibit various ETC complexes

TCA cycle products (e.g., FADH₂ and NADH) are used as substrates by complexes I-IV of the ETC to establish a proton gradient across the inner mitochondrial membrane. The final enzyme in oxidative phosphorylation, ATP synthase (complex V), produces energy as a form of ATP using this proton gradient [103] (Fig. 3). As discussed above, inhibition of TCA enzymes by S-nitrosylation decreases the flow of metabolites into the ETC, leading to less ATP production. Additionally, the ETC itself can be impaired via S-nitrosylation, most notably at complex I. Complex I accepts a pair of electrons from NADH, while complex II, which also functions as succinate dehydrogenase (SDH) in the TCA cycle, receives electrons from FADH₂. Both complex I and complex II transfer electrons to ubiquinone (Q), which in turn passes them through complex III, cytochrome c, and then finally to complex IV, where oxygen accepts the electrons to produce water. During these electron transfers (except complex II), protons are pumped across the inner mitochondrial membrane, creating a proton gradient, which is used by ATP synthase (complex V) to generate ATP. A decrease in activity of any of these complexes leads inhibition of mitochondrial metabolism and energy production, thus contributing to neuropathological processes. For example, exposure to MPP+ (a toxic metabolite of MPTP), rotenone, or paraquat induces parkinsonian-like phenotypes, such as loss of dopaminergic neurons, at least in part via inhibition of ETC activity [104,105]. Moreover, these mitochondrial toxins trigger ROS/RNS production, further aggravating mitochondrial deficits. Along these lines, NO can suppress mitochondrial respiration through inhibition of the ETC both in an S-nitrosylation-dependent and -independent manner [106].

Fig. 3.

S-Nitrosylation inhibits the ETC. Normally, electrons from NADH flow through the ETC, pumping protons (H⁺) across the mitochondrial membrane to create a proton gradient that drives ATP formation (top). However, under conditions of extreme nitrosative stress, mitochondrial ETC complexes can be aberrantly S-nitrosylated (SNO) in various cell types, including neurons (bottom). These S-nitrosylation reactions decrease the flow of electrons and protons, diminishing ATP production. S-Nitrosylation of complex I has been studied in detail where S-nitrosylation of Cys39 during the inactive state (the D-state) prevents its re-activation (the A state) upon re-availability of NADH.

The best characterized S-nitrosylation event in the ETC involves S-nitrosylation of complex I. S-Nitrosylation inhibits complex I activity thorough Cys39 modification in the ND3 subunit. Complex I exists in two forms, a catalytically active (A) form, which is in a functional conformation where ND3 Cys39 is fully occluded; and a de-active, dormant (D) form that exposes Cys39. The reversible transition from the A-state to the D-state takes place in the absence of NADH during conditions of low oxygen or hypoxia, but complex I can reinitiate the oxidation of NADH once sufficient supply of NADH is restored. S-Nitrosylation of Cys39 blocks this reactivation process to transiently lock complex I in the D-state and prevents toxic ROS production, providing cardioprotection against acute ischemia/reperfusion [107]. A recent study showed that the irreversible alkylation of Cys39 by iodoacetamide does not inhibit its NADH-CoQ oxidoreductase activity in the A-from but permanently prevents reactivation of the D-from [108]. That study suggests that modification of Cys39 allosterically affects the activity of complex I. Furthermore, our S-nitrosoproteomics MS studies on postmortem human AD, LBD, and HAND ​+ ​Meth brains have revealed the presence of the S-nitrosylation PTM on several core subunits of complex I, including the NADH:ubiquinone oxidoreductase core subunit S1(NDUSF1) [55,56,59]. Under hypoxic conditions, it is possible that this finding is related to the presence of sulfite oxidase, which is known to promote conversion of nitrite to NO in the mitochondrial intermembrane space [20]. These studies suggest the possibility that nitrite-derived NO-related species may inhibit complex I via S-nitrosylation in various neurogenerative diseases.

The functional effect(s) of S-nitrosylation on complexes II to V in neurodegenerative diseases remains poorly understood and merits further study. For example, an age-related increase in S-nitrosylation of complex II has been observed in mouse synapses [109], but its impact on synaptic function remains uncharacterized. S-Nitrosylation at Cys501 on mitochondrial chaperone TRAP1, an SDH inhibitor, reportedly facilitates the degradation of TRAP1, thereby increasing SDH activity both in the TCA cycle and complex II [110]. Concerning complex III, S-nitrosylation of cytochrome b-c1 complex subunit 1 in complex III has been identified in mouse heart [111], although an effect on activity remains unclear. In complex IV, Cys196 and Cys200 of subunit II undergo S-nitrosylation, inhibiting complex IV activity in lung endothelial cells [112]. Additionally, in synaptosomes nanomolar concentrations of NO can interact with the heme or copper centers of complex IV, thus inhibiting activity [106,113,114]. As for complex V, the last step of oxidative phosphorylation is inhibited by S-nitrosylation of the mitochondrial F1-ATPase in the ischemic mouse heart [72]. Finally, we recently found that several proteins in complex II-V, such as succinate dehydrogenase flavoprotein subunit (SDHA) in complex II as well as cytochrome b-c1 complex subunits 1 and 2 in complex III, are S-nitrosylated in human postmortem brains with AD, LBD and HAND ​+ ​Meth [55,56,59]. Studies currently underway are examining the role of S-nitrosylation in each of these ETC complexes to determine if there is a contribution to the pathogenesis of neurodegenerative disorders. Along these lines, in a mouse model of autism spectrum disorder (ASD), a recent study has shown that every mitochondrial ETC complex (I to V) is S-nitrosylated [115], suggesting that S-nitrosylation-mediated mitochondrial dysfunction may play a role in neurodevelopmental disorders as well.

S-Nitrosylation Regulates Other Mitochondrial Functions that Affect Bioenergetics, Including Mitochondrial Dynamics and Mitophagy

Aberrant S-nitrosylation of Drp1 leads to excessive mitochondrial fragmentation, energy compromise, and consequent synapse loss

Mitochondria continuously undergo morphological alterations through the tightly regulated processes of fission and fusion, a process called mitochondrial dynamics. Regulation of mitochondrial morphology and structure are essential to ensure their distribution at synapses for adequate energy supply of synaptic activity [35,37,116,117]. In contrast, disruption in these dynamic processes can impair synaptic function and contribute to the neurodegenerative process. Studies on postmortem human AD brains have revealed altered expression of proteins involved in mitochondrial dynamics, causing excessive fission and likely contributing to mitochondrial fragmentation [118]. Excessive mitochondrial fragmentation and cristae damage lead to bioenergetic compromise, decreased ATP production, increased ROS burden, and diminished calcium buffering, all of which contribute to synaptic dysfunction and eventual neuronal cell death [37,116].

Mitochondrial dynamics are controlled by a family of GTPases, including dynamin-related protein 1 (Drp1, also known as DNM1L or DLP1) and mitochondrial fission protein 1 (Fis1), which are involved in fission, as well as the mitofusin family (Mfn1 and Mfn2) and optic atrophy 1 (OPA1), which facilitate fusion. In AD brains, Aβ and hyperphosphorylated tau interact with mitochondrial Drp1, leading to increased expression of Drp1 and likely contributing to excessive mitochondrial fission [[119], [120], [121]]. Notably, we previously demonstrated that increased nitrosative stress triggers excessive mitochondrial fragmentation via aberrant S-nitrosylation of Drp1 at Cys644, which results in pathological activation of this GTPase enzyme [37,116] (Fig. 4). Mechanistically, S-nitrosylation facilitates dimerization of Drp1, required for its GTPase activity and mitochondrial fission, which is mediated by a multimer of Drp1 composed of dimers of dimers. In cell-based models of AD, expression of non-nitrosylatable mutant Drp1 prevents mitochondrial fragmentation associated with bioenergetic compromise, thus ameliorating synaptic dysfunction. Aberrant SNO-Drp1 formation has more recently been implicated in cardiovascular disease and metabolic disease [122,123], suggesting that SNO-Drp1 may contribute to various disease conditions.

Fig. 4.

Dysregulated mitochondrial dynamics and mitophagy after aberrant protein S-nitrosylation. Under pathological conditions, aberrant S-nitrosylation of Drp1, a master regulator of mitochondrial fission, triggers its hyperactivation, leading to excessive mitochondrial fragmentation associated with bioenergetic compromise. This pathological event contributes to loss of synapses in AD/ADRD. Recent studies demonstrated that a triple transnitrosylation cascade, transferring NO from Uch-L1 to Cdk5 to Drp1, represents previously unknown redox mechanism for the formation of SNO-Drp1. Additionally, when mitochondria are pathologically depolarized, PINK1 accumulates on the outer mitochondrial membrane, promoting the recruitment of parkin to initiate mitophagy for removal of damaged mitochondria. Aberrant S-nitrosylation of PINK1 and parkin inhibits the mitophagic process, resulting in the accumulation of damaged mitochondria and injury to the cell.

Remarkably, we subsequently discovered a transnitrosylation cascade in which NO-related species transfer from ubiquitin carboxyl-terminal hydrolase L1 (Uch-L1) to cyclin-dependent kinase 5 (Cdk5), and subsequently to Drp1, thus producing SNO-Drp1. Concerning Cdk5, enhanced Cdk5 kinase activity has been reported in AD/ADRD pathogenesis [124,125]. Along these lines, S-nitrosylation of Cdk5 at Cys 83 and 157 stimulates its kinase activity [36,126]. However, we recently revealed the transnitrosylation activity of Cdk5 [35,36], which appears to function enzymatically to transfer NO from Uch-L1 to Drp1. We further demonstrated that neurons expressing non-nitrosylatable mutant Cdk5 show decreased dendritic spine damage. These findings are consistent with the notion that the transnitrosylation activity may play an even more critical role than kinase activity in synaptic loss in AD [28,36].

Of note, Uch-L1 is highly abundant in the brain, and its dysfunction has been linked to neurodegenerative disorders, such as AD and PD [[127], [128], [129]]. Our group and others have shown that SNO-Uch-L1 exhibits decreased deubiquitinating activity, but more importantly, SNO-Uch-L1 triggers a transnitrosylation cascade, leading to the formation of SNO-Cdk5 and subsequently SNO-Drp1 that contributes to synaptic impairment [35,129]. Crucially, these concerted transnitrosylation reactions appear to be both kinetically and thermodynamically favorable within cells [35,36]. For example, we found that the change in the Gibbs free energy of each transnitrosylation step is negative, approaching −20 ​kJ/mol in some cases, suggesting that these transnitrosylation reactions are highly energetically favorable. In addition, absence of Cdk5 significantly abolished the transfer of NO-related group from Uch-L1 to Drp1, consistent with the notion that (SNO-)Cdk5 acts as an enzyme to facilitate the transnitrosylation reaction. However, it remains to be conclusively established if (SNO-)Cdk5 increases the Kcat of the transnitrosylation reaction. Expression of non-nitrosylatable mutant Uch-L1, Cdk5, or Drp1 prevents synaptic loss in models of AD, including in vivo transgenic models [[35], [36], [37]], further supporting the critical role of this transnitrosylation pathway in synaptic dysfunction in AD.

Aberrant S-nitrosylation of parkin and PINK1 disrupts mitophagy

Mitophagy is a subtype of autophagy that recycles or degrades damaged/depolarized mitochondria to maintain cellular homeostasis. Accumulation of dysfunctional mitochondria due to defects in mitophagy can lead to reduced metabolic output and increased production of free radicals. Attempts to restore mitophagy using mitophagy enhancers, including both synthetic chemicals and natural compounds, have demonstrated efficacy in alleviating synaptic impairment in preclinical models of AD [130], further underscoring the pathological relevance of mitophagy dysfunction. Consistent with this notion, S-nitrosylation-induced impairments in the mitophagy processes result in adverse effects on neuronal and synaptic function [9,10,131]. Activation of phosphatase and tensin homolog (PTEN)-induced putative kinase 1 (PINK1) and parkin represents a classical step leading to mitophagy. In this pathway, damaged mitochondria exhibit decreased mitochondrial membrane potential that leads to accumulation of PINK1 on the outer mitochondrial membrane, promoting stimulation of its kinase activity. Accumulated PINK1 then recruits and activates the E3 ubiquitin ligase, parkin, and other autophagy-associated proteins such as ubiquitin. Once recruited to mitochondria, parkin ubiquitinates mitochondrial proteins, allowing the association of autophagy receptors, including p62/SQSTM1, optineurin and NDP52, at the surface of damaged mitochondria to initiate mitophagy.

We previously discovered that S-nitrosylation of PINK1 at Cys568 causes allosteric inhibition of its kinase activity. Consequently, SNO-PINK1 manifests significantly less activity in phosphorylating and recruiting parkin onto depolarized mitochondria. This SNO-mediated decrease in PINK1 activity disrupts mitophagy and enhances neuronal cell death in PD models, including hiPSC-derived A9-type dopaminergic cells, the first neurons to be affected in PD. This finding is consistent with the notion that SNO-PINK1 contributes to the early stages of disease pathogenesis [132] (Fig. 4).

The effects of S-nitrosylation of parkin are a bit more complex. S-Nitrosylation of Parkin initially promotes E3 ubiquitin ligase activity but with time results in decreased activity [[133], [134], [135]]. Hence, sustained elevation of nitrosative stress decreases parkin activity via S-nitrosylation, leading to defective mitophagy, and thus diminishing parkin's protective function against environmental stress. These findings suggest that, as PD progresses, S-nitrosylation of both PINK1 and parkin inhibits the mitophagy process, resulting in neuronal dysfunction, especially in dopaminergic neurons. Additionally, parkin-independent pathways, such as the Siah1-mediated signaling cascade, whose members are also S-nitrosylated [136], are known to regulate mitophagy; thus, future investigation will be required to examine the effects of S-nitrosylation on parkin- or PINK1-indepentent mitophagy [137]. Moreover, evidence suggests that S-nitrosylation of (macro)autophagy-related proteins, such as c-Jun N-terminal kinase 1 (JNK1), IκB kinase β (IKKβ), cathepsin B, and p62/SQSTM1, inhibits autophagic flux [138,139]. For example, our group recently demonstrated that aberrant S-nitrosylation of the autophagy receptor, p62/SQSTM1 inhibits (macro)autophagic flux, leading not only to intracellular accumulation of aggregated proteins but also cell-to-cell spread [139]. p62/SQSTM1 is involved in autophagosome formation for degradation of damaged mitochondria through mitophagy; hence, whether SNO-p62 contributes to mitophagy defects remains an important question. Similarly, we found that S-nitrosylation of cathepsin B inhibits autophagic flux in neurons, and the relationship of this reaction to mitophagy remains unexplored [140].

Our group and others have shown that both PINK1 and parkin undergo S-nitrosylation in response to environmental toxins associated with PD, thus inhibiting PINK1 and parkin activity [132,133,135]. Notably, loss-of-function mutations in the Pink1 and Parkin genes are linked to rare genetic forms of PD. These studies suggest that gene-by-environment (GxE) interactions, particularly environmental factors that increase ROS/RNS production through the inhibition of the ETC to trigger aberrant S-nitrosylation of PINK1 and parkin, can mimic the effects of these rare genetic mutations by disrupting normal protein function and thus contribute to the development of sporadic cases of PD. Clarifying whether other environmental exposures, such as air pollution and various metals, which have been linked to AD, PD, and ASD, can trigger S-nitrosylation of mitochondrial proteins and synaptic dysfunction is an important ongoing research area.

Potential Neurotherapeutic Approaches for Targeting SNO-Mediated Mitochondrial Dysfunction

Bypassing SNO-mediated αKGDH blockade in the TCA cycle offers a potential therapeutic strategy

Prevailing evidence suggests that mitochondrial dysfunction and consequent impairment in brain metabolism contribute to development of cognitive decline due to loss of synapses for the reasons just delineated. Along these lines, potential therapeutic strategies targeting mitochondrial bioenergetics to maintain or rescue brain energy homeostasis have been widely explored in both pre-clinical models and human clinical trials (reviewed in [141]). While these strategies have shown some promise for improving bioenergetics, some have adverse effects, and none have yet demonstrated clear disease modifying effects in humans. Hence, identification of new mitochondrial targets may advance the development of more effective therapeutic approaches for prevention or restoration of synaptic loss.

As discussed in this review, nitrosative/oxidative stress contributes to neuropathological processes through aberrant protein S-nitrosylation of proteins related to mitochondrial function and bioenergetics, suggesting that pathways affected by SNO-proteins may represent unique therapeutic targets. For example, our MS-based S-nitrosoproteomics as well as our metabolic flux analyses revealed that S-nitrosylation of the αKGDH complex causes significant disruption of TCA cycle flux, resulting in decreased succinate production downstream of the αKGDH/SCS steps [56,57]. Based on these findings, we hypothesized that supplementation with TCA substrate downstream of this blockade could bypass this step, thereby improving bioenergetic activity and synaptic integrity. To test this idea, we investigated the therapeutic potential of dimethyl succinate (DMS), a membrane-permeable form of succinate [142], which can be metabolized to succinate to replenish the substrate required for the subsequent step in the TCA cycle, SDH. A cell-permeable succinate derivative can also bypass complex I defects in the ETC, as SDH acts as both a TCA cycle enzyme and a component of complex II in the ETC. In line with our premise, administered in vitro, DMS partially restored mitochondrial bioenergetic function and reversed synapse loss in AD neurons derived from hiPSCs [57]. This study offers proof-of-concept evidence that pharmacologically targeting or bypassing αKGDH inhibition in the TCA cycle could serve as a potential therapeutic strategy. Such an approach could improve mitochondrial energy metabolism, thereby ameliorating synaptic function and mitigating cognitive impairment. Another recent study developed newer succinate derivatives with improved cell permeability, such as NV118 (diacetoxymethyl succinate) for treating mitochondrial dysfunction [143]. Hence, investigating the potential of these newer succinate derivatives to restore TCA cycle blockade and alleviate synaptic loss would be an important direction for future studies. However, excessive accumulation of succinate has been known to increase ROS production and cause mitochondrial damage in cardiac tissue [144]. This prior study suggests that careful optimization of succinate supplementation will be needed. Moreover, succinate and succinate derivatives are not good drug candidates for pharmacological reasons, so other types of approaches will be required.

Potential approaches to directly target S-nitrosylated mitochondrial proteins in neurodegenerative diseases

An alternative strategy involves antioxidants to prevent or reverse aberrant S-nitrosylation. However, general antioxidant therapies are often ineffective, as they typically fail to reach the precise cellular component where antioxidative activity is needed. Additionally, broad-spectrum antioxidant therapy can interfere with physiological cell signaling processes that depend on low levels of ROS/RNS [145]. Hence, selectively targeting aberrantly S-nitrosylated proteins, such as the αKGDH complex in the TCA cycle, Drp1, Uch-L1, etc., each of which contributes to mitochondrial dysfunction, would offer a more effective therapeutic approach.

While identification of specific inhibitors for mitochondria-related SNO-proteins has not yet been reported, our group and others have successfully developed strategies to discover small molecule compounds that selectively increase or inhibit S-nitrosylation of specific proteins that can be applied to mitochondria-related proteins in future research efforts. For example, through structure-based in silico screening, we identified lead chemical compounds that selectively block aberrant S-nitrosylation of DNA methyltransferase 3B (DNMT3B) without interfering with its enzymatic function [146]. DNMT3B is responsible for de novo methylation, transferring methyl groups to genomic DNA to limit transcription to RNA, but S-nitrosylation of DNMT3B appears to block this activity, thereby contributing to tumor development. To screen for SNO-DNMT3B inhibitors, we initially identified the presence of a small-molecule binding pocket near the S-nitrosylation site, confirming the druggability of this region. Next, we performed virtual screening (followed by a counter screen to exclude compounds that affected enzymatic activity) and identified promising modulatory compounds that inhibit S-nitrosylation of DNMT3B at low concentrations (IC50 ​≤ ​100 ​nM). One of our lead compounds, designated DBIC, attenuated NO-dependent carcinogenesis in both in vitro cell-based and in vivo animal models [146]. These findings are consistent with our hypothesis that selective inhibition of an aberrantly S-nitrosylated protein by small chemical compounds can serve as a potential strategy for disease modification. Moreover, our study lays the groundwork for the development of additional small compounds targeting other aberrantly S-nitrosylated proteins.

Similarly, a recent study using a virtual screening approach identified small molecules that selectively inhibit S-nitrosylation of inositol-requiring enzyme 1α (IRE1α), an endoplasmic reticulum (ER) transmembrane protein that mediates ER stress responses known as the unfolded protein response (UPR) [147]. Under conditions of ER stress, upregulation of IRE1α RNase activity induces unconventional splicing of the X-box binding protein 1 (Xbp1) mRNA, allowing expression of functional Xbp1 transcription factor. XBP1 then promotes the expression of downstream targets that maintain ER homeostasis. However, prolonged or severe activation of IRE1α can trigger cell death through JNK signaling. Our work, along with that of others, has shown that S-nitrosylation of IRE1α predominantly occurs at Cys931, with a minor modification at Cys951. This S-nitrosylation reaction attenuates the RNase activity of IRE1α [148,149]. Aberrant S-nitrosylation of IRE1α leads to impairment of the UPR and exacerbates neuronal death in cell-based models of neurodegenerative diseases [148]. The newly developed, highly-selective inhibitor of SNO-IRE1α, called 1ACTA, rescues XBP1 splicing deficits caused by nitrosative stress and demonstrates cytoprotective activity in cellular models of PD [147]. Importantly, 1ACTA spares basal Xbp1 slicing activity of IRE1α, which is essential for maintaining ER homeostasis. While further studies are required to optimize this drug's IC₅₀ and examine its neuroprotective efficacy in animal models of neurodegenerative diseases, these findings support the potential of developing selective SNO-protein inhibitors that may hold promise for mitigating the progression and severity of AD and other dementias.

It should be noted that, in contrast to the involvement of aberrant S-nitrosylation in various pathogenic processes, physiological protein S-nitrosylation plays a key role in mediating neuroprotective and synapse-protective effects under homeostatic cellular conditions. For instance, S-nitrosylation of NMDA receptors can attenuate their hyperactivity, which, if left unblocked, contributes to synaptic damage. For this reason, our group developed the drug NitroSynapsin (aka EM-036 or NitroMemantine), an aminoadamantane nitrate, that inhibits NMDA receptors only when excessively activated. The drug acts through a dual mechanism of blocking excessively-open receptor-associated ion channels and then targeting a nitro group to specific S-nitrosylation sites on the receptor, thereby prolonging and improving the efficacy of inhibition. This approach helps decrease neuronal hyperactivity and synaptic damage without triggering systemic side effects that can occur with indiscriminate NMDA receptor blockade. Critically, NitroSynapsin has shown superior efficacy to memantine (the FDA-approved NMDA receptor antagonist previously developed by our group) in ameliorating disease-relevant phenotypes in preclinical models of AD, vascular dementia, and ASD [[150], [151], [152], [153]] (reviewed in [10,[154], [155], [156]]). Collectively, these findings highlight the relevance of developing newer drugs that selectively increase or decrease SNO-proteins as a promising direction for AD therapeutic strategies.

Conclusions

A variety of pathological processes contribute to cognitive decline in neurocognitive disorders, with ROS/RNS-mediated mitochondrial dysfunction widely recognized as a key factor in this process. As highlighted in this review, aberrant protein S-nitrosylation plays a significant role in the impairment of mitochondrial function, energy compromise, and subsequent synapse loss. Hence, strategies focused on restoring mitochondrial metabolism, for instance by mitigating or bypassing SNO stress, may offer a promising avenue for the development of disease-modifying therapeutics for neurocognitive decline. Importantly, the success of these strategies will depend heavily on accurate identification and characterization of SNO-mediated signaling pathways and sites that negatively impact mitochondrial function. While several key SNO-targets, such as TCA cycle enzymes, mitochondrial dynamics proteins, and mitophagic flux regulators, have already been identified, additional studies to delineate SNO-dependent mechanisms underlying mitochondrial dysfunction will be crucial for advancing such drug development efforts.

Other considerations are important as well. For example, in addition to SNO-mediated inhibition of TCA cycle enzymes, expression of TCA cycle transcripts has been found to be consistently decreased in AD brain [157]. This suggests that both transcriptional and posttranslational changes contribute to dysregulation of the TCA cycle activity. Moreover, the TCA cycle-related metabolites, acetyl-CoA, itaconate (produced from cis-aconitate), succinate, and αKG, all have non-metabolic signaling roles affecting both physiological and pathophysiological responses, including the innate and adaptive immune systems, epigenetics, and thermogenesis [158]. For example, itaconate acts as an anti-inflammatory molecule by activating the transcription factor NRF2 via cysteine alkylation of KEAP1 as well as promoting TFEB transcription factor through direct alkylation of TFEB at cysteine 212 [159,160]. Future studies are warranted to examine if aberrant S-nitrosylation of mitochondrial targets, including TCA cycle enzymes, also alters these non-metabolic pathways to affect neurodegenerative damage.

Author contributions

All authors contributed to the conception and design of the manuscript. TN, AS, and SAL wrote the manuscript and prepared the figures. All authors participated in the editing process and provided final approval for the manuscript.

Declaration of Generative AI and AI-assisted technologies in the writing process

During the preparation of this work the author(s) used ChatGPT 4o in order to review grammar and make readability edits only. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication. This is in accordance with the editorial guidelines of Neurotherapeutics.

Declaration of competing interest

SAL discloses that he is an inventor on worldwide patents for the use of memantine and NitroSynapsin for neurodegenerative and neurodevelopmental disorders. Per Harvard University guidelines, SAL participates in a royalty-sharing agreement with his former institution Boston Children's Hospital/Harvard Medical School, which licensed the drug memantine (Namenda®) to Forest Laboratories, Inc./Actavis/Allergan/AbbVie. NitroSynapsin is licensed to the biotechnology company EuMentis Therapeutics, Inc., for which SAL is scientific founder and chair of the Scientific Advisory Board (SAB). SLA is also a member of the SAB of Point 6 Bio. Ltd., and has recently served as a consultant to Circumvent Pharmaceuticals, Inc. Further, SAL discloses that he is a named inventor on patent(s) filed by his current institution, The Scripps Research Institute, for novel NRF2 activators in the treatment of systemic and nervous system diseases via anti-inflammatory and antioxidant actions. TN serves as a consultant to Dojindo Molecular Technologies. The other author declares no competing interests.