Reactive Oxygen Species in Metabolic and Inflammatory Signaling

Abstract

Reactive oxygen species (ROS) are well known for their role in mediating both physiological and pathophysiological signal transduction. Enzymes and subcellular compartments that typically produce ROS are associated with metabolic regulation, and diseases associated with metabolic dysfunction may be influenced by changes in redox balance. In this review, we summarize the current literature surrounding ROS and their role in metabolic and inflammatory regulation, focusing on ROS signal transduction and its relationship to disease progression. In particular, we examine ROS production in compartments such as the cytoplasm, mitochondria, peroxisome and endoplasmic reticulum, and discuss how ROS influence metabolic processes such as proteasome function, autophagy and general inflammatory signaling. We also summarize and highlight the role of ROS in the regulation metabolic/inflammatory diseases including atherosclerosis, diabetes and stroke. To successfully develop improved therapies that target oxidative signaling, it is vital to understand the balance ROS signaling plays in both physiology and pathophysiology, and how manipulation of this balance and the identity of the ROS may influence cellular and tissue homeostasis. An increased understanding of specific sources of ROS production and an appreciation for how ROS influence cellular metabolism may help guide us in this effort to treat cardiovascular diseases.

Introduction

Reactive oxygen species (ROS) regulate cellular homeostasis and act as prime modulators of cellular dysfunction contributing to disease pathophysiology. ROS are byproducts of numerous enzymatic reactions in various cell compartments, including the cytoplasm, cell membrane, endoplasmic reticulum (ER), mitochondria and peroxisome, as part of basal metabolic function. They are also generated specifically by enzymes such as NADPH oxidases and serve a signaling function in the cell. Depending upon the source of ROS, cell type and tissue environment, ROS signaling may participate in normal physiological processes or contribute to a maladaptive response that leads to metabolic dysfunction and inflammatory signaling. Diseases associated with elevated inflammatory signaling and metabolic dysfunction such as atherosclerosis, diabetes and stroke are associated with an altered redox balance.^1–3 Understanding the role of ROS signaling in the regulation of metabolic activity, inflammatory activation and diseases associated with metabolic dysfunction is important in our pursuit of novel therapies to treat these diseases. This review highlights the role of ROS signaling in basic metabolic processes and inflammatory signaling, and focuses on how this regulation contributes to disease development.

1. Sources of ROS and their role in metabolic function

I. Cytoplasmic

Cytoplasmic ROS (cytoROS) production is a cornerstone of cellular signaling and disease pathophysiology. One of the most well-known sources of cytoROS is the NADPH oxidase (NOX) family of enzymes. NOX2 (or gp91^phox) is well-characterized for its role in phagocytic function;⁴ however, it and three homologs, NOX1, NOX4, and NOX5, are expressed throughout the cardiovascular system.⁵ NOX proteins produce O₂⁻ through NADPH electron exchange, and NOX-dependent ROS production influence many metabolic processes and disease states.⁶ Briefly, endothelial cell (EC) NOX-dependent ROS production drives hypoxia inducible factor 1α (HIF1α)-mediated glucose transporter 1 (GLUT1) expression, hexokinase activity and resultant glycolysis in response to low oxygen tension as part of the angiogenic response.⁷ In response to inflammatory stimuli, neutrophil phosphofructokinase 2 colocalizes with NOX2, inducing its activation. NADP⁺ produced as a byproduct of NOX2 O₂⁻ generation is used to facilitate an elevated glycolytic rate. A locally increasing NADPH concentration as a by-product of increased glycolysis is hypothesized to further enhance NOX2 activity; however, this has yet to be proven.⁸

In addition to NOX-dependent ROS production, the nitric oxide synthases (endothelial, neuronal and inducible) are sources of cytoplasmic ROS. Endothelial nitric oxide synthase (eNOS) produces O₂⁻ through its oxygenase domain in a Ca²⁺/calmodulin-dependent reaction in the absence of tetrahydrobiopterin (BH₄).⁹ Dysregulated or uncoupled eNOS is a hallmark of cardiovascular and metabolic diseases.¹⁰ The oxygenase domain is vital for O₂⁻ production from neuronal and inducible NOS (nNOS and iNOS, respectively), and O₂⁻ generation is dependent upon reduced L-arginine.^{11, 12} NOS uncoupling in various disease states results in a dysregulated NO response whereby synthesized NO combines with O₂⁻ to produce peroxynitrite. Peroxynitrite enhances pentose phosphate pathway (PPP) activity through stimulation of glucose-6-phosphate dehydrogenase (G6PD), leading to elevated NADPH.¹³ A similar mechanism has been noted in response to H₂O₂, a downstream metabolite of O₂⁻.¹⁴ Thus, in the transient setting, peroxynitrite, and even H₂O₂, may promote a protective response through PPP-dependent increases in reducing equivalents. However, chronic stimulation of the PPP by ROS may instigate a toxic feedback loop whereby increased NADPH production results in excessive O₂⁻ through NOX.¹⁵ In support of both hypotheses, G6PD deficiency and overexpression are protective against oxidative stress^16,17 However, further investigation into how ROS and metabolic diseases affect NADPH shuttling is needed.

CytoROS also induce AMP-activated protein kinase (AMPK) activity.¹⁸ AMPK is a central regulator of cellular metabolism implicated in multiple metabolic functions including glycolysis, lipid metabolism, mitochondrial function, cell growth and autophagy (discussed below).¹⁹ Hypoxia can regulate AMPK activity through an ROS-dependent ER Ca²⁺/Stromal interaction molecule 1/Ca²⁺/calmodulin-dependent protein kinase kinase beta pathway in addition to indirect regulation via a change in the AMP/ATP ratio.²⁰ H₂O₂ can directly modulate AMPK activity and downstream metabolic function through oxidation and S-glutathionylation of the α- and β-subunits of AMPK by targeting cysteines 299 and 304.²¹ ROS-mediated impairment of the mitochondrial respiratory chain can also increase the AMP/ATP ratio which can activate AMPK,²² and peroxynitrite induces AMPK activation in bovine aortic ECs through a c-SRC(Tyr⁴¹⁶)/phosphoinositide 3-kinase (PI-3K)/phosphoinositide-dependent kinase-1 (Ser²⁴¹) pathway. Peroxynitrite-dependent AMPK phosphorylation at Thr¹⁷² leads to phosphorylation of acetyl CoA carboxylase at Ser⁷⁹, resulting in inactivation and increased fatty acid oxidation.²³

Low oxygen tension in ECs increases glycolysis as part of the angiogenic response through NOX-mediated, ROS-induced HIF1α signaling.⁷ However, increased ROS signaling in some disease states inactivates the glycolysis rate-limiting enzyme pyruvate kinase M2, and diverts glycolytic substrates into the PPP pathway to generate reducing equivalents needed for ROS detoxification, thereby acting as a protective mechanism.²⁴ In total, cytoROS signaling plays a critical role in the PPP and glycolytic pathways (Figure 1). Below we will outline the contributions of various cytoROS-producing entities in the development of metabolic disorders.

Figure 1. Cytosolic ROS production and regulation of cytosolic metabolic pathways.

Cytosolic ROS are formed most notably through NOX activity and influence metabolic processes including glycolysis and downstream oxidative phosphorylation, pentose phosphate pathway activity and autophagy. Please refer to the abbreviation table for full names of listed proteins.

II. Mitochondria

Mitochondria are central regulators of aerobic energy production. Proper respiratory chain function requires a delicate balance between pro-oxidant and anti-oxidant systems. Importantly, mitochondrial respiration relies on electron transfer and a proton gradient to drive ATP production. ROS are a natural byproduct of this process; however, inflammatory and metabolic diseases are associated with perturbed mitoROS production.^{25, 26} MitoROS are generated by numerous mechanisms including Complexes I-III (Figure 2). Complex I serves as an entry point for electrons from NADH into the respiratory chain. O₂⁻ is produced from the interaction of O₂ with reduced FMN when the matrix NADH/NAD+ ratio is high, resulting in O₂⁻ release into the matrix.²⁷ In addition, mitoROS is produced through complex I via reverse electron transfer (RET). RET occurs in a 2-step process involving (1) reduced coenzyme Q (CoQ), and (2) a change in proton motive force that drives electrons back into complex I. Complex III is also a source of mitoROS. Under normal conditions, electrons flow from the CoQ pool to cytochrome C. Although complex III produces very low levels of O₂⁻, in the presence of Antimycin A, the Q_i site is inhibited, which promotes O₂⁻ production from the Q_o site due to the interaction between O₂ and a ubisemiquinone bound to the Q_o site.²⁸ O₂⁻ generated from complex III is mainly released into the intermembrane space, but upon dismutation, H₂O₂ may diffuse into the matrix.²⁹

Figure 2. Mitochondrial ROS production.

MitoROS are produced as a normal byproduct of mitochondrial respiration and metabolic enzymatic activity. Under settings of increased ROS generation as a result of dysregulated enzymatic activity and cellular stress, mitoROS can influence metabolic pathways including the Krebs cycle, fatty acid synthesis, ATP generation, glycolysis and mitophagy. Cyto=cytoplasm, IMS=intermembrane space, LPP=lipid peroxidation product. Please refer to the abbreviation table for full names of listed proteins.

While complex I and III are well known for their contribution to mitoROS production, complex II may also serve as an ROS producing complex.³⁰ Complex II oxidizes succinate to fumarate as part of the Krebs cycle and acts as the site of ubiquinone reduction in the electron transport chain (ETC). ROS are produced from complex II when both complex III and complex I are inhibited. Complex II ROS production is believed to proceed through a forward mechanism involving electrons from succinate, or in a reverse mechanism where electrons are provided through a reduced ubiquinone pool. Both mechanisms result in ROS production from the complex II flavin site.³⁰

Other than complex-derived mitoROS, enzymes involved in metabolic reactions produce mitoROS (reviewed in ^{31, 32}). Of importance, the α-ketoglutarate dehydrogenase (KGDHC) and pyruvate dehydrogenase (PDC) complexes produce ROS as a result of both forward electron transfer and RET.^{33, 34} KGDHC, PDC and branched-chain α-keto acid dehydrogenase complex, a dehydrogenase complex that catalyzes the oxidative decarboxylation of α-ketoacids, are thought to produce significantly more ROS than complex I.³⁵ Recent evidence indicates PDC and KGDHC-derived ROS may be regulated through protein-S-glutathionylation. Interestingly, S-glutathionylation during forward electron transfer may attenuate ROS production, whereas S-glutathionylation during RET may increase ROS from PDC and KGDHC³⁶ In addition to generating ROS, KGDHC is also an early target of oxidative stress. Similarly, aconitase, an enzyme responsible for the isomerization of citrate to isocitrate, is responsive to ROS.^{37, 38} In conditions of low oxidative stress, when aconitase activity and Krebs cycle substrate production is diminished, α-ketoglutarate levels and NADH production can be maintained through glutamate via glutaminolysis; however, in the presence of high oxidative stress, KGDHC is inhibited, reducing NADH production and mitochondrial respiratory capacity.³⁸ RET-induced mitoROS can inactivate aconitase and pyruvate dehydrogenase kinase 2 (negative regulator of PDC) through reversible oxidation of cysteine 45 and 392 when the NADH/NAD⁺ ratio is elevated. This regulation is hypothesized to promote acetyl-CoA production via PDC from carbohydrates while simultaneously inhibiting β-oxidation, resulting in cytoplasmic export of citrate and stimulation of fatty acid synthesis.³⁹ Reduced aconitase activity is associated with aging and diseases associated with metabolic dysfunction.⁴⁰

Recent evidence indicates that a subset of NOX4 is localized to the mitochondria and is a regulator of mitoROS generation. NOX4 expression is increased in kidney cortex from diabetic rats and facilitates glucose-induced mitoROS production. Likewise, NOX4 induces cysteine oxidation, and resultant decreased activity, of aconitase and citrate synthase in cardiac myocytes,⁴¹ and influences mitochondrial morphology and complex I activity.⁴² Additionally, ATP production from normal mitochondrial respiration negatively regulates mitochondrial NOX4-induced ROS production through direct interaction between ATP and the NOX4 Walker A binding motif,⁴³ which suggests mitochondrial NOX4 may play a dynamic role in metabolic function.

Apart from ROS producing pathways, impairment of mitoROS scavenging pathways results in ROS accumulation leading to organelle and cell dysfunction. Within the mitochondria, the predominant ROS buffering systems include the glutaredoxin (Grx), glutathione and thioredoxin (Trx) systems.⁴⁴ Dismutation of O₂⁻ into H₂O₂ occurs through the superoxide dismutase (SOD) family of proteins.⁴⁵ In the matrix, dismutation primarily proceeds through SOD2 (MnSOD), whereas in the intermembrane space, dismutation is carried out by SOD1 (Cu, Zn-SOD). H₂O₂ decomposition into O₂ and H₂O then occurs via the GSH redox system, which includes glutathione reductases, peroxidases (GPX) and peroxiredoxins (Prdx).⁴⁶

The Trx and Grx systems also play a prominent role in mitochondrial ROS buffering. The mitochondrial Trx system involves thioredoxin-2 (Trx2), thioredoxin reductase-2 (TrxR2) and members of the Prdx family of proteins. Importantly, the antioxidant effect of Trx2 activity results from reduction of other oxidized proteins, mainly the Prdxs. Oxidization of Trx2 is reversible and is remedied through TrxR2 using NADPH as an electron donor. ⁴⁷ Homozygous knockout (KO) of Trx2 is embryonically lethal, and heterozygous mice, while viable, show decreased mitochondrial respiratory function and increased mitoROS production,.⁴⁸ Likewise, cardiac-specific TrxR2 KO mice exhibit cardiac structural changes, dysregulation of autophagy, decreased oxygen consumption and a change in their metabolic profile.⁴⁹

Mitochondrial Grx family members include Grx2 and Grx5. Grx2 regulates O₂⁻ production from complex 1 by catalyzing glutathionylation of two thiol groups, and Grx5 regulates iron/sulfur enzymes.^50–52 However, information regarding how the GRX and Trx systems regulate cardiovascular function is limited. Cumulatively, mitochondria maintain a delicate balance between oxidant and anti-oxidant systems, and dysregulation can result in organelle dysfunction resulting in metabolic stress.

MitoROS can also influence HIF1α stabilization and cell proliferation.^{53, 54} Hypoxia-induced HIF1α promotes a metabolic shift favoring anaerobic glycolysis and reduced mitochondrial respiration by upregulating glucose transporters and glycolytic enzymes while also inhibiting PDC through activation of pyruvate dehydrogenase kinase 1.^{55, 56} Reduced PDC activity causes decreased Krebs cycle and ETC flux, which may in turn attenuate mitoROS production in hypoxic conditions. Likewise, HIF1α induces mitophagy resulting in decreased mitochondrial mass, O₂ consumption and resultant ROS generation. Thus, in hypoxic conditions, mitoROS favor the stabilization of HIF1α, which induces a metabolic shift towards glycolysis, leading to reduced mitochondrial activity and mitoROS production.^{57, 58} In normoxic conditions, inhibition of ETC function using rotenone (complex I inhibitor) or TTFA (complex II inhibitor) induces cell death through induction of mitoROS-mediated autophagy.⁵⁹ AMPK, a regulator of uncoordinated 51-like kinase 1 (ULK1) activity and downstream autophagy, can be regulated by mitoROS, and mitoROS contribute to autophagy through AMPK/ULK1-dependent signaling (discussed below).^{60, 61} Similarly, AMPK is responsible for the mitoROS/HIF1α-dependent extension of life span observed in Caenorhabditis elegans (C. elegans).⁶² Slightly elevated levels of mitoROS in mice are also associated with extension of lifespan.⁶³ Interestingly, mice with heterozygote deletion of mouse clock-1, a protein involved in CoQ biosynthesis, exhibit increased longevity, increased inflammatory cytokine production and macrophage activation that is dependent upon elevated mitoROS. These mice also show enhanced resistance to infection, which may be due to the observed increase in macrophage phagocytic activity.⁶⁴

Mitochondrial fission and fusion are closely associated with mitochondrial function and can influence, and be influenced by, mitoROS production. (For a comprehensive review of mitochondrial fission and fusion, please refer to references.^{65, 66} With regard to mitochondrial fission, the small cytoplasmic GTPase dynamin-related protein-1 (Drp1) is implicated in ROS signaling. Drp1 activity regulates mitoROS production and downstream mitochondrial functional changes in a variety of environments,^67–69 and oxidative stress influences Drp1 mitochondrial translocation and resultant fission.⁷⁰ Mitochondrial fission also precedes mitophagy, which can be regulated by oxidative stress and is thought to be a negative regulator of mitoROS signaling through selective degradation of dysfunctional mitochondria.^{71, 72} Drp1 undergoes numerous post-translational modifications including S-nitrosylation and phosphorylation, and data suggest that ROS signaling contributes to serine 616 phosphorylation and activation of Drp1 GTPase activity.^73–75 However, the contribution of ROS signaling to the regulation of other fission proteins is currently unknown and is an area ripe for future investigation.

Similar to Drp1 and mitochondrial fission, ROS signaling may regulate mitochondrial fusion. The inner mitochondrial membrane GTPase optic atrophy protein-1 (OPA1) is regulated by reactive oxygen species modulator 1 (ROMO1). In response to ROS, ROMO1 is inactivated leading to OPA1 cleavage, cristae remodeling and mitochondrial fission.⁷⁶ Likewise, deletion of OPA1 induces morphological irregularities, respiratory defects and ROS generation.⁷⁷ In addition to OPA1, mitofusin 1 and 2 activity can regulate and be regulated by ROS.^78–80 The precise mechanism and contribution of fission/fusion-regulated ROS signaling to metabolic function remains to be explored; however, the contribution of ROS and mitochondrial morphology regulation in the setting of metabolic diseases will be explored below.

In summary, the mitochondria are dynamic players in metabolic regulation and signaling. MitoROS are produced as part of normal mitochondrial function, but various cellular stresses augment ROS levels either through increased oxidant production or decreased antioxidant activity. As discussed, mitoROS can regulate cellular metabolic function (illustrated in Figure 2), and in section 2, we will discuss the contribution of mitoROS to metabolic and cardiovascular diseases.

III. Peroxisome

Peroxisomes, like mitochondria, are vital organelles in aerobic metabolism that regulate key processes such as α- and β-oxidation, glyoxylate metabolism, amino acid catabolism, the pentose phosphate pathway, ketogenesis, polyamine oxidation and isoprenoid and cholesterol metabolism.⁸¹ Peroxisomes are also a significant source of ROS. In particular, peroxisomes produce H₂O₂ due to an abundance of O₂-consuming oxidases, which include acyl-CoA oxidases (ACOX), D-amino acid oxidase, D-aspartate oxidase, polyamine oxidase, xanthine oxidase (also produces O₂⁻), L-α-hydroxyacid oxidase and L-pipecolic oxidase.^{82, 83} Unlike the mitochondria, peroxisomal electron transfer does not lead to ATP generation. Instead, free electrons are transferred to H₂O to form H₂O_2.⁸³ In addition, peroxisomes can produce nitric oxide through NOS.⁸⁴ However, similar to mitochondria, peroxisomes contain numerous oxidant scavenging enzymes including GPX, catalase, Prdx1 and 5, peroxisomal membrane associated protein 20, SOD1 and SOD2.⁸² For an in-depth review of peroxisomal enzymes, please refer to⁸⁵.

Recently, a new role for peroxROS in mTOR complex (mTORC) 1 activity was defined. In response to elevated peroxisomal β-oxidation, peroxROS activate tuberous sclerosis proteins 1 and 2 (TSC1 and TSC2), which are bound to the peroxisomal assembly proteins peroxin 19 and 5 (Pex19 and Pex5), respectively. ROS induce TSC2-mediated GTPase activity of Rheb, leading to mTORC1 inhibition and autophagy induction.⁸⁶ Pex5 binds ataxia-telangiectasia mutated (ATM) and localizes it to the peroxisomal membrane. In response to peroxROS, ATM activation (Ser¹⁹⁸¹) induces AMPK and TSC2 activity and downstream mTORC1 inhibition, ULK1 activation and pexophagy. As part of this mechanism, ATM activation also induces the phosphorylation of Pex5 (Ser¹⁴¹), triggering ubiquination at lysine 209 and subsequent binding of p62, which is required for peroxisomal targeting to autophagosomes (Figure 3).⁸⁷ In addition to pexophagy regulation, peroxROS can also disrupt the mitochondrial redox balance and promote mitochondrial fission/fragmentation⁸⁸ and mitochondrial-mediated cell death.⁸⁹

Figure 3. Peroxisomal ROS and metabolism.

PeroxROS are produced as byproducts of enzymatic reactions within β-oxidation, polyamine synthesis, D-amino acid deamination and hypoxanthine oxidation, and have been found to be key regulators of pexophagy. Full names of abbreviations are listed in the accompanying table.

IV. ROS and the Endoplasmic Reticulum

The ER has a well-established role in metabolic and cardiovascular diseases^{90, 91} due to its roles in Ca²⁺ handling, protein synthesis/folding and regulation of the secretory pathway. Protein-folding is highly sensitive to ER redox status and dysregulation of disulfide bond formation in response to ER stress increases luminal oxidative stress leading to a decline in ER function.⁹² One of the most well understood routes for disfulfide bond introduction into folded proteins is the protein disulfide-isomerase (PDI) and ER oxidoreductin 1 (ERO1) pathway. PDI introduces disulfide bonds through thiol oxidation in folding substrates leaving PDI in a reduced state. Reduced PDI is re-oxidized through ERO1, which transfers acquired electrons through a flavin adenine dinucleotide cofactor to molecular oxygen, forming H₂O₂. ER H₂O₂ can further be used by Prdx4 to re-oxidize PDI, thereby increasing the efficiency of ERO1-mediated disulfide bond transfer.^{93, 94} Overexpression of a human hyperactive mutant of ERO1 induces severe oxidative stress and induction of the unfolded protein response (UPR), an ER stress response involved in the pathogenesis of metabolic and cardiovascular diseases, highlighting the sensitivity of the ER to changes in redox balance.⁹⁵ Furthermore, UPR activation can induce ERO1 activation leading to increased oxidative stress and sustained UPR signaling, and administration of antioxidants can attenuate the UPR and improve downstream protein secretion.⁹⁶ In addition to the PDI/ERO1 pathway, ROS is produced through the membrane associated monooxygenase system via cytochrome p450, cytochrome b₅ reductase and through ER-localized NOX4 (Figure 4).^97–99

Figure 4. Endoplasmic reticulum and ROS.

The ER is highly sensitive to redox status, and altered ROS signaling can influence protein folding, Ca²⁺ release and mitochondrial respiration. Please refer to the abbreviation table for full names of depicted proteins.

As part of this redox balance, the ER contains an antioxidant system that is vital to proper ER function: peroxiredoxins (Prdx4, mentioned above) and glutathione peroxidases (GPX7 and GPX8). GPX7/8 contain a KDEL ER localization sequence, making them ER-specific ROS scavengers. GPX7/8 both interact with ERO1α, one of two isoforms of ERO1, at cysteine²⁰⁸/cysteine⁴¹. Reduced GPX7/8 scavenges ERO1-derived H₂O₂, leaving GPX7/8 in an oxidized form.¹⁰⁰ Oxidized GPX7, through its cysteine 86 residue, can bind to the cysteine^41/420 residue of glucose-regulated protein, 78 kDa (GRP78). This binding promotes GRP78 oxidation and enhanced protein refolding chaperone activity. Silencing of GPX7 induces oxidative stress, accumulation of misfolded proteins and induction of the UPR.¹⁰¹ PDI oxidation is regulated by GPX7/8 as well.¹⁰⁰

The ER and mitochondria sit in close proximity to one another, and changes in the ER redox balance influence mitochondrial function. In response to ER stress, activating transcription factor 4 and C/EBP homologous protein (CHOP) induce ERO1α. In mitochondrial-associated ER membranes (MAMs), ERO1α oxidizes the type 1 inositol 1,4,5-trisphosphate receptor (IP₃R) inducing mitochondrial Ca²⁺ uptake and mitoROS production through mitochondrial respiration.^{102, 103} ERO1α-mediated cytoplasmic Ca²⁺ efflux through the ER IP₃R is also hypothesized to contribute to NOX2-dependent ROS stimulation via CAMKII.¹⁰⁰ SHC-transforming protein 1 isoform p66 (p66^shc) is a regulator of cellular oxidative stress¹⁰⁴ that translocates to MAMs in response to cellular stress and produces mitoROS through cytochrome C oxidation,¹⁰⁵ and is capable of inducing ER stress¹⁰⁶ and inhibiting mTOR-dependent anabolic metabolism (Figure 4).¹⁰⁷

In summary, maintenance of the ER redox balance is critical to proper ER function, and alterations in ER ROS producing and scavenging pathways provoke ER stress and contribute to metabolic dysfunction.

V. ROS and Inflammation

Inflammation and metabolism are intricately intertwined, considering that numerous metabolic and cardiovascular disorders exhibit chronic low-grade inflammation.¹⁰⁸ Canonical NF-κB signaling is associated with insulin resistance, obesity and atherosclerosis,^108–110 and circulating dietary factors such as fatty acids and glucose can trigger inflammatory signaling. It has also been suggested that NF-κB may regulate metabolic reprogramming favoring aerobic glycolysis.¹¹¹ The influence of ROS on NF-κB signaling may depend upon the cellular location of oxidation (cytoplasmic vs nuclear).¹¹² In general, ROS are known to activate NF-κB in response to inflammatory agonists.¹¹³ NF-κB nuclear translocation occurs in response to H₂O₂¹¹⁴ through a mechanism involving IκBα tyrosine phosphorylation (Tyr⁴²), phosphorylation of the serine/threonine PEST domain with subsequent degradation via calpain, and p65 phosphorylation (ser⁵²⁹).^{115, 116} ROS-induced NF-κB is inhibited by SOD2 overexpression,¹¹⁷ and the NOX family of proteins also influence,¹¹⁸ and are influenced by, NF-κB activity.^{119, 120}

Specific inflammatory agonists utilize ROS as part of their signaling cascades. IL-1β induces active endosomal IL-1R complex assembly that involves MyD88 and NOX2 ROS-induced tumor necrosis factor (TNF) receptor associated factor (TRAF) 6 endosomal recruitment.¹²¹ NOX4 activity is required for lipopolysaccharide (LPS)-induced NF-κB activation,¹²² and TNFα-induced NF-κB activation increases antioxidant expression leading to decreased TNFα-induced apoptotic signaling through a ROS/JNK pathway.¹²³

Flow-induced activation of NF-κB is regulated by ROS signaling as well. Flow-mediated EC dysfunction and monocyte adhesion is dependent upon NOX-derived O₂⁻ regulation of NF-κB inducing kinase (NIK) and IKK signaling leading to NF-κB activation.^124–126 As part of this response, bone morphogenic protein 4 and p21-activated kinase act as upstream regulators of flow-induced ROS generation and downstream NF-κB signaling, contributing to oscillatory shear stress-induced vascular dysfunction and atherosclerotic lesion development.^{126, 127}

An abundance of evidence points to a role for mitoROS in regulating inflammatory signaling. In response to inflammatory stimuli, proinflammatory cytokines are synthesized and released from cells. Importantly, mitoROS contribute to LPS-induced cytokine release²⁵, thrombin induced NF-κB activation via IP₃R Ca²⁺ signaling,¹²⁸ and lysophosphatidylcholine-induced AP-1 activity and downstream endothelial activation.¹²⁹ Likewise, mitochondrial H₂O₂ production contributes to endothelial NF-κB activation in aged rat arteries,¹³⁰ and inhibition of mitoROS through ETC inhibition abrogates hypoxia-induced endothelial NF-κB activation and IL-6 secretion.¹³¹ Evidence also indicates that mitoROS may be a downstream result of NF-κB activation as well.¹³²

RET-induced mitoROS are involved in metabolic changes associated with macrophage activation during inflammation. Macrophage proinflammatory signaling is supported in part by metabolic repurposing that favors glycolysis-derived ATP production and RET-induced mitoROS generation through mitochondrial hyperpolarization and succinate oxidation.¹³³ RET-derived mitoROS promote HIF1α stabilization,¹³⁴ leading to regulation of glycolytic capacity and IL-1β mRNA and protein expression.^135,133 In response to LPS, immune cells secrete IL-1β through the inflammasome. MitoROS signaling is a primary regulator of NLR Family Pyrin Domain Containing 3 (NLRP3) inflammasome activation. NLRP3 inflammasome activation consists of two phases: (1) a priming phase where agonists such as LPS stimulate NF-κB-mediated NLRP3, IL-1β and interleukin-18 (IL-18) transcription, and (2) an activation phase where a multi-protein complex consisting of NLRP3, Apoptosis-associated speck-like protein containing a CARD, and caspase 1 is assembled.¹³⁶ The fully assembled NLRP3 inflammasome, along with activation via potassium efflux,¹³⁷ lysosomal destabilization and mitoROS,^{138, 139} regulates the maturation of IL-1β and IL-18.¹⁴⁰ Thus, in LPS-stimulated macrophages, RET-induced mitoROS production, as a result of a metabolic shift towards glycolysis, regulates 1L-1β transcription (inflammasome priming), but may also regulate the maturation and secretion of IL-1β (inflammasome activation). However, alternative pathways of IL-1β processing independent of inflammasome activity also exist.¹⁴¹

Thioredoxin-interacting protein (TxNIP), a negative regulator of Trx, is an ROS-regulated proapoptotic factor that mediates mitoROS and NOX4 activity and influences glucose-induced inflammasome activation in two ways.^{142, 143} First, TxNIP-dependent inhibition of Trx induces ROS which further exacerbate inflammasome activity and inflammatory cytokine signaling. Second, during inflammasome activation, TxNIP dissociates from Trx and interacts with NLRP3, which is required for proper inflammasome activity in response to glucose stimulation.¹⁴² In addition, IRE1α, an ER stress protein, increases inflammasome activity by regulating TxNIP mRNA stability.^{144, 145}

Apart from ROS-dependent inflammasome activation, ROS are critical to macrophage phagocytic activity. Importantly, metabolic and cardiovascular diseases associated with chronic inflammation display impairments in macrophage phagocytic/efferocytic activity that are correlated with changes in ROS signaling. NOX2-dependent oxidative signaling is required for sufficient phagocytic function and pathogen/apoptotic cell degradation,¹⁴⁶ as shown by increased lung inflammation and abdominal aortic aneurysm progression in myeloid NOX2 KO mice.^{147, 148}

Macrophage NOX-derived ROS induce microtubule-associated protein 1A/1B-light chain 3 (LC3) translocation to the phagosome, which is required for lysosomal fusion and phagosomal clearance in LC3-associated phagocytosis (LAP).¹⁴⁹ Toll-like receptor signaling promotes mitochondrial recruitment to phagosomes where augmented mitoROS kill phagocytosed bacteria and enhance NOX-dependent ROS production.¹⁵⁰ As part of LAP, Drp1-mediated mitochondrial fission increases cytosolic Ca²⁺ through inhibition of the mitochondrial Ca²⁺ uniporter and mitoROS production, which are required for efficient phagosomal sealing and LAP-mediated apoptotic cell degradation.¹⁵¹ Drp1-dependent mitochondrial fission and resultant ROS production also contribute to NF-κB activation in T cells,¹⁵² and changes in mitochondrial dynamics regulate T cell metabolic reprogramming.¹⁵³

Mer tyrosine kinase (MerTK) is an essential membrane protein in macrophages that participates in efferocytosis. In response to inflammatory stimuli, proteolytic cleavage and inhibition of MerTK activity impair efferocytosis and downstream resolution of inflammation.¹⁵⁴ In response to LPS, TLR4-TIR-domain-containing adapter-inducing interferon-β signaling induces MerTK shedding through a NOX2/PKCα/ p38 MAPK and a Disintegrin and metalloproteinase domain-containing protein 17 (ADAM17) pathway, and macrophage-specific ADAM17 deletion protects against MerTK shedding in vivo.¹⁵⁵ MitoROS and AMPK also play a significant role in macrophage efferocytosis. Apoptotic cell-released lysophosphatidylcholine diminishes mitochondrial membrane potential and ATP production coupled with concomitant mitoROS generation and downstream AMPK activation. AMPK facilitates both metabolic reprograming towards glycolysis and tubulin synthesis that are needed for macrophage chemokinesis and efficient efferocytosis.¹⁵⁶ Macrophages experience defective efferocytosis in response to TNFα as well, potentially through phospholipase A2/arachidonic acid-dependent ROS and Rho activity.¹⁵⁷

Overall, inflammation is an underlying component to many diseases including those that exhibit metabolic distress. ROS act as central regulators of inflammatory signaling, particularly with respect to NF-κB activation and inflammasome signaling (Figure 5). The following sections will highlight the contribution of ROS to specific cellular functions and their role in regulating metabolic dysfunction in various diseases.

Figure 5. Inflammation and ROS.

Various inflammation-inducing stimuli including TNFα, LPS, thrombin and oscillatory shear stress influence ROS production through sources including NOX and the mitochondria. Elevated ROS production as a result of inflammatory signaling can mediate canonical NF-κB activation and downstream inflammatory gene induction, proteasome activity, antioxidant gene transcription, inflammasome activation and cytokine secretion. Full names for abbreviations are listed in the accompanying table.

VI. Secondary Products of ROS

Secondary products of ROS also mediate inflammatory signaling in a variety of cellular environments. For example, lipid peroxidation products, such as 4-hydroxynonenal, and oxidized phospholipids (OxPLs) regulate NF-κB activation and inflammatory signaling through numerous pathways,^158–160 and can propagate ROS signaling; however, whether secondary ROS product signaling is protective or instigates a pathophysiological response may depend on ROS concentration, cell-type and cellular stress.¹⁵⁹ In regards to pathophysiology, lipid peroxidation and OxPLs have been implicated in atherogenesis.^161,162 Mechanistically, both products can regulate NF-κB activity and chemokine release resulting in arterial wall inflammation and immune cell recruitment.^163–165 However, incubation of LPS-treated mice or macrophages with OxPLs suppresses NF-κB activation and TNFα secretion, suggesting that OxPLs may be protective in some conditions.¹⁶⁶ It should also be noted that in a single agonist condition, OxPLs may influence macrophage polarization and contribute to the macrophage pro-inflammatory response.¹⁶⁷

Secondary products of ROS can also regulate metabolic function in various tissues and cellular environments. In response to OxPLs, EC glycolytic and proliferative capacity are increased and are dependent upon NRF2 signaling.¹⁶⁸ Similarly, oxidation of cholesterol may promote a protective response in macrophages by increasing autophagy and promoting effective efferocytosis.¹⁶⁹ Lipid peroxidation can increase autophagy in rat VSMCs through an ER stress/JNK-dependent mechanism.¹⁷⁰ In the mitochondria, lipid peroxidation affects membrane fluidity and is capable of modulating ETC complex activity, Krebs cycle enzymes, proteostasis and mitochondrial membrane potential (Figure 2).¹⁷¹ Overall, while secondary products are known to influence various enzymes and contribute to diseased states such as atherosclerosis and cardiac diseases, specific signal transduction pathways have yet to be fully elucidated. Nevertheless, secondary products of ROS signaling have a clear role in metabolic and inflammatory signaling.

VII. ROS and Autophagy

Autophagy is a highly conserved catabolic process in which cytoplasmic macromolecules and organelles are delivered to lysosomes for degradation.¹⁷² Constitutive basal levels of autophagy support metabolic homeostasis by promoting a balance between protein synthesis and degradation, as well as organelle biogenesis and degradation; however, dysregulated autophagic flux contributes to metabolic and inflammatory diseases.^172,173. For an extensive review on the regulation of autophagic machinery, please refer to¹⁷⁴.

Oxidative stress regulates autophagic flux through its influence on autophagic gene transcription, protein activity and organellular degradation.¹⁷⁵ ROS induce macroautophagy (commonly referred to as autophagy), and selective degradation of oxidized proteins through chaperone-mediated autophagy (CMA) is important for cellular viability during periods of oxidative stress.¹⁷⁶. Oxidative modification of CMA substrates increases their susceptibility to degradation, and CMA is enhanced through LAMP2a upregulation in cells challenged with ROS.¹⁷⁶ Little is known about the effect of ROS on microautophagy and further work is needed to distinguish the roles of macroautophagy, microautophagy, and CMA in response to oxidative stress.

In addition to the role of ROS in autophagic flux, ROS are required for autophagy induced by starvation,¹⁷⁷ dopamine,¹⁷⁸ sodium selenite,¹⁷⁹ mitochondrial electron transport chain inhibitors TTFA and rotenone,¹⁸⁰ TNFα¹⁸¹ and LPS.¹⁸² The source and identity of ROS that mediate these effects remain unclear. Sources of ROS located at or near the plasma membrane, such as NOXs, are prime candidates to transduce signals from external stimuli into the cell. However, this hypothesis has only been verified in macrophages through LAP. ¹⁴⁹

As mentioned, ROS can influence AMPK activation. In response to ROS, cytoplasmic ATM promotes liver kinase B1 (LKB1)-dependent AMPK activation and downstream autophagy through the regulation of TSC2 activity and mTOR inhibition.^183,184 Furthermore, overexpression of catalase blocks H₂O₂-induced autophagy,^177,182 and starvation-induced H₂O₂ production regulates cysteine oxidation of Atg4, a negative regulator of autophagy that delipidates LC3-II to LC3-1 and inhibits autophagosome formation.¹⁷⁷ Starvation methods including the removal of glucose, L-glutamine, pyruvate and selenium or amino acids can also increase O₂⁻ levels.¹⁸⁵ Interestingly, amino acid starvation alone increases intracellular H₂O₂ levels, and overexpression of SOD2 in serum-starved, amino acid-starved and H₂O₂-treated cells attenuates ROS production and inhibits autophagy.¹⁸⁵ Mitochondrial antioxidants, such as SS-31 and SOD2, can block autophagy induced by stressors including mitochondrial inhibitors,¹⁸⁰ sodium selenite,¹⁷⁹ and immobilization.¹⁸⁶ Given the critical role of ROS in the regulation of autophagy, further work is needed to clarify the specific mechanisms by which ROS influence autophagic machinery.

ROS are also implicated in regulating the selective degradation of mitochondria through mitophagy. MitoROS can trigger mitochondrial permeability transition pore opening and a burst of mitoROS, which has been described as ROS-induced ROS release.^{187, 188} Thus, ROS produced by damaged mitochondria may act as a self-removal signal,¹⁷⁵ and while this hypothesis needs further verification, the benefits of selectively recycling mitochondria with oxidized proteins and damaged DNA as a secondary defense against oxidative stress are clear and have been demonstrated in yeast.^72,189 This is supported by recent work suggesting that ROS are required for Phosphatase and tensin homolog-induced kinase 1/Parkin-dependent mitophagy.¹⁹⁰

VIII. ROS and the Proteasome

The proteasome is a highly organized multimeric complex responsible for the selective hydrolysis of cytoplasmic, nuclear and ER proteins¹⁹¹ which supports metabolic homeostasis by maintaining a stable pool of free amino acids for protein synthesis.¹⁹² The proteasome also promotes cellular viability in response to oxidative stress.^{175, 193} Amino acids are continuously subjected to oxidative modification as a consequence of aerobic respiration, and most of these modifications are irreversible and irreparable. Studies using antisense oligonucleotides against critical proteasome subunits ¹⁹⁴ or proteasome inhibitors ¹⁹⁵ have implicated the proteasome as the primary pathway for degrading oxidized proteins. Thus, the proteasome serves as an important secondary defense against oxidants.

The central catalytic component of the proteasome is the barrel-shaped 20S core particle (CP) composed of four heptameric rings consisting of two outer “gate-keeping” α rings and two identical inner β rings that contain catalytically active/proteolytic β subunits.^191,196 The 20S CP may be flanked by regulatory subunits through α subunit interaction¹⁹³ which modulate 20S CP activity and substrate specificity.¹⁹⁷ The most common of these subunits is the 19S regulatory particle (RP),¹⁹⁷ which along with the 20S CP, comprises the 26S proteasome. The 26S proteasome is a key component of the ubiquitin proteasome pathway (UPP) and regulates the degradation of ubiquitinated proteins.^197,198 The 19S RP recognizes polyubiquitinated substrates and hydrolyzes ATP to unfold and translocate target proteins into the 20S CP.¹⁹⁷ The 20S CP may also associate with other regulatory proteins, such as PA28αβ, PA28γ, and PA200, for ubiquitin/ATP-independent protein degradation.^198,199

Mildly oxidized proteins selectively and rapidly undergo proteolysis, and strong evidence suggests the 20S proteasome is largely responsible for this degradation.^193,195,200 H₂O₂ impairs ubiquitin conjugation and reduces ubiquitin/ATP-dependent proteolysis;²⁰¹ however, disruption of the ubiquitin system does not affect oxidized protein degradation,¹⁹⁵ suggesting that the 20S ubiquitin/ATP-independent proteasome is responsible for protein degradation during oxidative stress. Concordant with these findings, in vitro studies using purified 26S show that the 26S proteasome, even in the presence of ATP and a functional ubiquitination system, does not degrade oxidized proteins.¹⁹³ However, both the 26S and 20S proteasome are sensitive to oxidative stress.²⁰² In fact, in the presence of the activator PA28αβ, exogenous ROS enhances 20S protease activity²⁰² while H₂O₂ disrupts 26S proteasome complex integrity and reversibly dislocates the 20S CP from the 19S RP.²⁰⁰. This dissociation is accompanied by a loss of 26S proteasome activity and enhances cell survival following H₂O₂-induced oxidative stress.²⁰⁰ Interestingly, dissociation of the 26S proteasome increases the fraction of 20S proteasome, which may account for an increase in cell survival.²⁰⁰ Consistent with this notion is the observation that yeast deficient in 26S assembly are more resistant to oxidative stress than their wild type counterparts.²⁰³

Some evidence, however, supports a role for the 26S proteasome in degrading oxidized proteins. Ubiquitin carboxyl-terminal hydrolase 14 (USP14), a 26S proteasome-associated deubiquitinating enzyme, decreases ubiquitin-protein conjugate degradation by disassembling polyubiquitin chains.²⁰⁴ Inhibition of USP14 enhances cell survival and reduces the accumulation of oxidized proteins in cells challenged by menadione,²⁰⁴ suggesting that ubiquitination and the 26S proteasome are important for degrading oxidized proteins. Moreover, the expression of Ubiquitin conjugating enzyme (UBC) 4, an E2 enzyme, promotes the degradation of glutathionylated proteins in lens fiber cells, and this degradation is blocked by a dominant negative form of ubiquitin and proteasome inhibitors.²⁰⁵ Several studies also suggest that the 26S proteasome is critical for cellular viability during recovery from oxidative stress. Cells treated with H₂O₂ exhibit a transient increase in proteolytic activity and in ubiquitin conjugation post-treatment.^{201, 206} While oxidative stress negatively regulates 26S activity, 26S proteasome activity is almost completely restored 24 hours post-treatment.²⁰² Finally, it has been observed that certain oxidized proteins are preferentially ubiquitinated.²⁰⁷ Thus, there is evidence to support a role for both the 20S and 26S proteasome in degrading oxidized proteins. However, further work may help to distinguish the different roles of the 20S and 26S proteasome in dealing with specific oxidized substrates.

The proteasome plays an important role in mitigating the effects of ROS, and oxidative modification of proteasome subunits modulates proteasome activity. Carboxylation of Regulatory Particle Triphosphate (RPT) 3, an ATPase subunit in the 19S RP, impairs ATPase activity and decreases 26S proteasome activity.²⁰⁸. Additionally, both carbonylation and 4-hydroxy-2-nonenal (HNE) modification of two α-subunits in the 20S proteasome impair ubiquitin/ATP-independent proteolysis.²⁰⁹ Interestingly, S-glutathionylation directly modulates specific proteolytic activity of the 20S proteasome.²¹⁰ Low doses of GSH or GSSG enhance the chymotrypsin-like activity of purified proteasome, while higher doses inhibit this activity.²¹⁰ The effect of S-glutathionylation seems specific to the chymotrypsin-like activity, as GSH and GSSG do not affect the trypsin-like activity of purified proteasomes.²¹⁰

Other post-translational modifications also modulate proteasome activity in response to oxidative stress. Poly-ADP ribosylation of the 20S proteasome by the redox-sensitive enzyme poly-ADP ribose polymerase enhances the chymotrypsin-like activity of the 20S proteasome in K562 leukemia cells.²¹¹ In addition, apoptosis signal-regulating kinase 1 (Ask1) is activated by oxidative stress,²¹² and Ask1 phosphorylation of RPT5, a 19S subunit, negatively regulates 26S proteasome activity.²¹³ Significantly, Ask1 is required for ROS-induced inhibition of 26S proteasome activity.²¹⁴ Thus, Ask1-dependent proteasome phosphorylation may act as a critical regulatory mechanism for proteasome activity during oxidative stress.

Following oxidative stress, E1 activating enzyme expression and ubiquitin conjugation are increased in bovine lens epithelial cells.²⁰⁶ Increased expression appears to be due to increased translation of E1 mRNA.²⁰⁶ Consistent with the 20S proteasome playing a critical role in degrading oxidized proteins, H₂O₂ treatment increases α3, α4, β1, and β2 20S proteasome subunit expression without affecting the 19S RPS subunit S4.²¹⁵ Interestingly, upregulation of immunoproteasome subunits has also been reported.^{215, 216} Low-molecular-mass protein (LMP) 2, LMP7, LMP10 are β subunits that may replace constitutive β subunits in the 20S proteasome upon interferon-γ stimulation,²¹⁷ and increased expression of these subunits has been reported in neural cells in response to exogenous ROS.²¹⁶ The transcription factors that regulate immunoproteasome and standard proteasome subunit expression following oxidative stress require further study.

2. Cardiovascular and Metabolic Diseases

I. Atherosclerosis

NOX

Atherosclerosis is mediated in large part through subendothelial lipoprotein retention, endothelial dysfunction, vascular remodeling and a heightened inflammatory response, and is an underlying cause of heart disease and stroke.^218,219 Of importance, metabolic disease risk factors typically observed in patients with obesity and diabetes (changes in cholesterol, elevated triglycerides, insulin resistance) increase the risk for atherosclerosis.^220,221 Although the link between oxidized LDL and atherosclerosis was discovered nearly 30 years ago,²²² the NOX enzyme family was only defined in the mid-1990s. Some of the first studies performed exploring NOX-induced ROS concluded that NOX-derived ROS have little influence on lesion development in ApoE^−/− mice on normal chow diets. Neither p47^phox−/− (NOX1 and NOX2 activator) nor NOX2 (gp91^phox−/y) mice crossed with ApoE^−/− mice show changes in aortic sinus lesion area, although NOX2^−/yApoE^−/− mice exhibit reduced plasma triglyceride and cholesterol levels.^223,224 However, these findings appear to be aortic sinus-specific, as further analysis of p47^phox−/− ApoE^−/− and NOX2^−/y ApoE^−/− mice (on chow and high-fat/western-type diets) revealed a reduction in descending aortic lesion development.^225–227

NOX1^−/y ApoE^−/− mice are protected from both atherogenic diet-induced and diabetes-induced (atherogenic diet + streptozatocin, STZ) atherosclerosis,^228–230 associated with decreased ROS production, inflammatory signaling and aortic macrophage infiltration. NOX1 likely exerts its effects in part by reducing VSMC proliferation and migration,²³¹ and its phosphorylation on threonine 429 (Th4⁴²⁹) may be a contributing factor.²³² However, the role of NOX1 in the initiation and progression of atherosclerosis remains controversial because recent findings in NOX1-deficient ApoE^−/− mice indicate NOX1 may be protective against hyperlipidemia and plaque instability in response to a western-type diet.²³³

Similar to NOX1, the role of NOX2 in lesion etiology remains controversial. As mentioned, NOX2^−/y ApoE^−/− mice fed an atherogenic diet show reductions in the development and progression of atherosclerosis.²²⁶ Likewise, treatment of high fat diet fed ApoE^−/− mice with a NOX2-specific inhibitor reduces aortic lesion area,²³⁴ and while there are compelling data regarding a role for global NOX2-induced ROS in plaque development, the tissue-specific role of NOX2 remains ill-defined. In ApoE^−/− mice, EC-specific overexpression of NOX2 increases inflammatory signaling and macrophage infiltration in early lesions, but does not influence diet-induced lesion progression.²³⁵ Bone marrow transplantation between p47^{phox −/−} ApoE^−/− mice and p47^+/+ ApoE^−/− mice shows that suppression of either bone marrow cell or vascular wall O₂⁻ production attenuates western-type diet induced atherosclerosis.²³⁶ Moreover, suppression of NOX2 in bone marrow cells is associated with reduced plasma oxLDL, suggesting NOX2-derived ROS may regulate oxLDL production. oxLDL is scavenged and internalized by the lectin-like oxidized low density lipoprotein receptor 1 (LOX-1) and contributes to atherogenesis through ROS-dependent mechanisms involving NF-κB activation,²³⁷ inhibition of AKT/eNOS signaling,²³⁸ macrophage proinflammatory cytokine production and cell death,^239,240 and VSMC apoptosis.²⁴¹ The importance of NOX2-derived ROS within macrophages is also seen in its role in pathogen degradation after phagocytosis through a process called macropinocytosis. ^{242, 243}

Given the diverse roles of both NOX1 and NOX2 in vascular disease, it is not surprising that NOX4 can also be both a harbinger of lesion development and a protector. As a protector, NOX4-derived ROS appear to be critical in maintaining vessel homeostasis in mouse models of atherosclerosis, because NOX4^−/− Ldlr^−/− and NOX4^−/− ApoE^−/− mice experience endothelial dysfunction and increased plaque burden as well as increased plaque formation in partial ligation plus high fat diet models.^244,245 However, in diabetes-induced atherosclerosis, global NOX4 deletion may be protective or augment diabetes-induced lesion development, depending on the time frame.^{227, 229} NOX4 deletion in a 10-week STZ model of diabetes reduces plaque burden associated with T cell activation and infiltration,²²⁷ but in a 20-week model of STZ-induced diabetes, NOX4 deletion augments VSMC collagen deposition and proliferation, which may exacerbate lesion progression.²⁴⁶ The contribution of NOX4 in diabetic lesion development may also differ between early and advanced lesions, as NOX4 deletion has a minimal effect on early diabetes-induced plaque progression in the aortic arch,²²⁹ but is critical to suppressing the inflammatory response in advanced lesions.²³⁰

In tissue-specific NOX4 KO models, a more specific role for NOX4 signaling in atherosclerosis emerges. Expression of an EC-specific NOX4 mutation (human P437H dominant negative mutation) in ApoE^−/− mice exacerbates lesion progression in response to STZ;²⁴⁷ whereas EC overexpression of wild-type NOX4 attenuates high-fat diet-induced lesion development.²⁴⁸ As mentioned, NOX4 is upregulated in advanced lesions, as opposed to the upregulation of NOX1 and NOX2 in the early phases of plaque development.^225,249,250 While these studies indicate that this upregulation may be protective, other studies have shown NOX4 upregulation may promote lesion progression. Importantly, upregulated NOX4 in advanced plaques correlates with VSMC dysfunction and plaque instability,²⁴⁹ and smooth muscle-specific NOX4 deletion protects against western-type diet-induced atherosclerosis.²⁵¹ Aged ApoE^−/− mice fed a western-type diet and aged human carotid VSMCs exhibit increased mitochondrial NOX4 induction with parallel increases in mitoROS production, and inhibition of NOX4 increases aged VSMC mitochondrial complex I and II activity, decreases mitoROS and attenuates VCAM-1 induction.²⁵⁰ ER-derived NOX4 signaling is also implicated in VSMC apoptosis in response to 7-ketocholesterol, a major cholesterol oxidation product found in human plaques. 7-ketocholesterol upregulates IRE1 activity leading to JNK/AP-1-dependent NOX4 gene induction. Increased ROS production through NOX4 induces ER stress proteins GRP78 and CHOP and the apoptosis regulator Bcl-2-associated X protein (Bax).²⁵² A role for NOX2 in CHOP induction has also been observed,²⁵³ and in vivo silencing of CHOP attenuates lesion area and plaque necrosis in hyperlipidemic mice.^254,255

The PPP pathway can serve as both a biosynthetic pathway for nucleotides and a factory for NADPH production resulting in increased reducing capacity or NOX activity and O₂⁻ production. G6PD catalyzes the first step of the PPP and generates NADPH as a by-product. With regard to atherosclerosis, deficiency in G6PD abrogates O₂⁻ production and resultant inflammatory signaling and lesion growth in ApoE^−/− mice fed a western-type diet.²⁵⁶ However, G6PD overexpression in ECs reduces TNFα-induced ROS production and increases eNOS activity, suggesting a protective role for this protein.²⁵⁷ Future studies should aim to clarify the circumstances by which G6PD contributes to pro- and anti-oxidant signaling, as well as the role of other PPP enzymes, in NOX activity and atherogenesis.²⁵⁷

As noted above, a plethora of evidence indicates a role for NOX-induced ROS in atherogenesis, although these molecules play both a positive and negative role in lesion progression. It is becoming increasingly clear that no model of atherosclerosis is the same. Changes in diet (composition and timing), genetic background, and age all influence NOX signaling. Tissue-specific knockouts have helped to create a clearer picture of the role of NOX in various tissues as it pertains to atherosclerosis, and continued investigation will determine the suitability of NOX enzymes as potential therapeutic targets.

Mitochondrial ROS

MitoROS signaling plays a major role in atherosclerosis and associated vascular complications, and induction of mitoROS is correlated with human plaque development.⁵³ In general, scavenging of mitoROS attenuates atherosclerosis development and complications associated with atherosclerosis.^250,258 Genetic inhibition of mitoROS may also have beneficial effects on lesion development. Suppression of mitoROS signaling in macrophages through the overexpression of mito-targeted catalase reduces lesion area, inflammatory signaling and immune cell infiltration into the aortic root of Ldlr^−/− mice fed a western-type diet. In cultured mitoCatalase-overexpressing macrophages, LPS and oxLDL-induced mitoROS promote IKKβ phosphorylation (serine 177) and downstream p65 phosphorylation (serine 536), and inhibition of this pathway blocks macrophage MCP-1 induction.²⁵⁹ Furthermore, bone marrow transfer from mitoCatalase-expressing mice to aged WT Ldlr^−/− mice diminishes neutrophil extracellular traps and lesion area, but is unable to perturb increased plasma cholesterol levels.²⁶⁰

As discussed previously, a dynamic interplay exists between mitoROS production and mitochondrial morphology. With regard to atherosclerosis and metabolic dysfunction, high glucose can induce mitoROS through Drp1-dependent mitochondrial fission in human ECs, and inhibition of fission with the pharmacological inhibitor mitochondrial division inhibitor-1 (Mdivi-1) attenuates diabetes-induced aortic oxidative stress, cell adhesion molecule expression and aortic root lesion development.²⁶¹ However, there appears to be a cell type-specific role for mitochondrial fission in disease progression, as myeloid Drp1 silencing in Ldlr^−/− mice fed a western-type augments plaque necrosis due to impaired efferocytosis.¹⁵¹

Uncoupling protein 2 (UCP2) is a mitochondrial inner membrane protein and is reported to regulate ROS generation through feedback inhibition involving ROS-induced uncoupling and proton leak.²⁶² UCP2 expression is increased in C57BL6 mice in response to an atherogenic diet, and deletion of UCP2 increases oxidative stress, endothelial dysfunction, VCAM-1 expression, macrophage infiltration and atherogenesis.^263,264 Likewise, xanthine oxidase is found in various cellular compartments including the mitochondria, and pharmacological inhibition via febuxostat or tungsten reduces chemokine expression, endothelial dysfunction and lesion development.^{265, 266}

Altogether, mitoROS play a significant role is lesion etiology. Given their unique position in regulating cellular energy metabolism, it is no surprise that alteration in mitoROS production influences mitochondrial function and cellular homeostasis and is a possible cause of lesion progression. There are numerous ROS producing complexes within the mitochondria, and teasing out their role in atherosclerosis is of paramount importance.

ROS Scavenging

Similar to inhibition of ROS producing enzymes, increasing ROS scavenging appears to have a beneficial effect on lesion development. Mice overexpressing catalase and catalase + SOD1,²⁶⁷ Prdx4,²⁶⁸ and Trx2 (EC-specific)²⁶⁹ all exhibit decreased lesion development. In contrast, deletion of ROS scavenging systems exacerbates lesion progression, as seen in mice with deletion of NF-E2 related factor 2,²⁷⁰ GPX1,^271,272, SOD2,²⁷³ and Prdx1 and 2.^274,275

The paraoxanase (PON) family of proteins may also serve an anti-oxidant role in atherosclerosis through their ability to hydrolyze lipid peroxides.²⁷⁶ Overexpression of the PON transgenic cluster (PONs 1,2 and 3 together) promotes plaque stability via increased collagen synthesis, decreased necrotic core area and reduced oxLDL and inflammatory markers. PON1 overexpression may also reduce monocyte-to-macrophage differentiation and promote macrophage resistance against oxLDL-induced foam cell formation²⁷⁷ and induction of the LPS + IFNγ inflammatory phenotype.^278,279 In contrast, deletion of PON1 increases vascular oxidative stress and leukocyte adhesion,²⁸⁰ and deletion of PON2 exacerbates lesion development.^281,282 Recent evidence suggests that PON2 may localize to the mitochondrial inner membrane where it is bound to CoQ and contributes to ROS scavenging and maintenance of complex I and III activity.²⁸¹

Cumulatively, ROS signaling has been shown to be an underlying causative factor that contributes to the complex etiology of atherosclerosis in animal models (Supplementary Table I). The studies presented highlight a unique role for ROS in the regulation of cellular responses to various atherogenic stimuli, and tissue-specific studies have begun to elucidate the role ROS signaling might play in various cell types and their contribution to lesion progression. However, even in the presence of an abundance of data indicating a causative role for ROS in atherosclerosis, mechanistic details regarding metabolic regulation and its role in ROS signaling are lacking.

II. Diabetes/Obesity

Obesity creates an increased risk for the development of cardiovascular and metabolic diseases including atherosclerosis, diabetes mellitus and hypertension.²⁸³ Obesity is associated with insulin resistance and hyperglycemia, which contribute to the development of type II diabetes mellitus, and clinical investigation has yielded novel insight into how diabetes-related complications contribute to systemic vascular and metabolic dysfunction. Clinical and animal research consistently point to oxidative stress as an underlying factor that mediates obesity- and diabetes-related health ailments.^284,285 Importantly, glucose metabolism can produce ROS via sorbitol metabolism, hexosamine metabolism, α-ketoaldehyde production, PKC activation, glycation and oxidative phosphorylation.²⁸⁶

Vascular System

Hyperglycemia is a consequence of diabetes and insulin resistance and plays a significant role in the accompanying vascular complications (atherogenesis, endothelial dysfunction, inflammation, altered vascular tone). Early studies highlighted the ability of glucose and free fatty acids to stimulate ROS production in vascular cells including VSMCs, ECs and human leukocytes, and showed that high glucose can impair PPP activity and generation of reducing equivalents.^287–290 Since then, investigation has focused on unique pathways by which ROS may regulate vascular complications in diabetes and obesity.

ROS can regulate vascular tone in response to hyperglycemia via regulation of PKCβ and inhibition of eNOS activity.²⁹¹ Hyperglycemia-induced ROS production activates the 26S proteasome leading to ubiquitination and degradation of the BH₄ synthesis rate limiting enzyme guanosine 5′-triphosphate cyclohydrolase I (GTPCH). BH₄ deficiency dysregulates eNOS activity which impairs endothelial-dependent relaxation, and treatment of STZ-injected mice with TEMPOL restores GTPCH, BH₄ and attenuates endothelial dysfunction.²⁹² Furthermore, PKCβII induces endothelial dysfunction through p66^shc-mediated mitoROS production.^291,293 Hyperglycemia increases p66^shc acetylation (Lys⁸¹), which precedes PKCβII-induced phosphorylation (Ser³⁶) of p66^shc, and inhibition of p66^shc reduces mitoROS, aortic lipid peroxidation and restores eNOS activity and endothelial-dependent relaxation in diabetic mice.^{291, 293} PKCβ also regulates vascular tone through impairment in the large conductance Ca²⁺-activated K⁺ channel. In response to high glucose, the β1 subunit of the Ca²⁺-activated K⁺ channel is degraded through PKCβ-stimulated NOX1 and NOX4 activation, inhibition of AKT and activation of the forkhead box protein O3a/F-box only degradation pathway.²⁹⁴

As with atherosclerotic lesion development, NOX-derived ROS play an important role in vascular homeostasis in diabetes and obesity. Inhibition of NOX activity attenuates diabetes-induced impairment of endothelial-dependent relaxation,²⁹⁵ and aged NOX2^−/y mice fed a high fat diet are protected from obesity, dyslipidemia, insulin resistance and endothelial dysfunction.²⁹⁶ Similar results have been observed in arteries from diabetic NOX1^−/y mice,²⁹⁷ and mice treated with siRNA targeting p22^phox,²⁹⁸ as well as NOX2^−/y mice with EC-specific overexpression of a human dominant negative (DN) variant of the insulin receptor that causes EC-specific insulin resistance and reduced vascular relaxation.²⁹⁹ Proinflammatory stimuli such as TNFα can induce NOX activity in diabetic mice resulting in impaired vasorelaxation,³⁰⁰ and NOX-derived ROS downregulate cGMP-dependent protein kinase I, which normally promotes vascular relaxation.³⁰¹ Furthermore, high glucose induces NOX4 upregulation via a PKCζ and NF-κB-dependent pathway.³⁰² Insulin like growth factor 1 (IGF-1) enhances high glucose-induced NOX4/p22^phox complex formation and activation of VSMCs, and induces NOX4 Tyr491 phosphorylation leading to NOX4 binding to the SH2 domain of growth factor receptor-bound protein 2 (Grb2). NOX4/Grb2 association is needed for NOX4 localization to the SHPS-1 plasma membrane scaffold in VSMCs and STZ-injected mouse vasculature. Disruption of the NOX4/Grb2 association inhibits Src oxidation in vivo as well as VSMC proliferation.³⁰³

ROS produced by NOX also reduce vascular endothelial growth factor (VEGF) expression and post-ischemic neovascularization in diabetic mice. NOX2^−/y mice show increased bone marrow mononuclear cell (BM-MNC) to EC differentiation and injection of BM-MNCs from NOX2^−/y mice into diabetic mice enhances neovascularization.³⁰⁴ Furthermore, endothelial progenitor cells from diabetic patients show a reduced reendothelialization capacity that can be normalized through inhibition of p47^phox.³⁰⁵ Together, these data suggest a major role for ROS-mediated signaling in both facilitating dysregulated vascular tone and impairment of angiogenesis/neovascularization in diabetic ischemic tissue.

One consequence of hyperglycemia is the generation of advanced glycation end products (AGEs), covalent adducts formed between glucose and plasma proteins. AGEs contribute to diabetic-related maladies including neuropathy, retinopathy and cardiomyopathy.³⁰⁶ Given its direct contact with blood, the endothelium is a prime target for AGE-mediated signaling. Importantly, aortic ECs express the receptor for AGEs (RAGE), which is increased in response to high glucose via mitoROS production,³⁰⁷ and incubation of ECs with diabetic red blood cells expressing AGEs increases VCAM-1 induction and tissue factor production through RAGE-induced NOX activation.³⁰⁸ AGE stimulation may also dysregulate eNOS activity via increased NOX and mitoROS within the coronary endothelium of diabetic patients.³⁰⁹ Silencing of NOX2 in macrophages prevents AGE-induced tissue factor expression as well.³⁰⁸ NOX silencing is also beneficial in reducing AGE-induced apoptosis and NF-κB signaling in VSMCs.^{308, 310} Further, high glucose-induced mitoROS production increases PKC activation, hexosamine pathway activation and AGE formation through DNA strand break-mediated poly(ADP-ribose) polymerase activation and GAPDH poly(ADP-ribosyl)ation.³¹¹

As mentioned, mitoROS play a significant role in modulating vascular function in diabetes and obesity through both AGE signaling and modulating vascular tone, and potentially through changes in mitochondrial morphology. Isolated coronary ECs from diabetic mice exhibit increased mitochondrial fission that can be attenuated with in vivo TEMPOL delivery.³¹² High glucose-induced mitochondrial fission induces mitoROS production,^312,313 which may be a result of increased pyruvate uptake following increased fission,³¹⁴ as well as an instigating step in apoptosis.³¹³ Silencing of Drp1 or Fis1 inhibits high glucose-induced mitoROS and restores eNOS activity and cGMP production, suggesting a role for mitochondrial fission-induced ROS in regulating vascular tone.³¹⁵ Likewise, high glucose-induced VSMC proliferation can be inhibited by attenuating mitochondrial fission-dependent NOX activity.³¹⁶ Low blood glucose also commonly occurs in diabetic patients who lack tight glycemic control, and low glucose enhances mitoROS production through Drp1-dependent mitochondrial fission. Suppression of mitochondrial fission in response to low glucose inhibits mitoROS production, increases NO bioavailability and restores endothelial-dependent vascular relaxation.³¹⁷ Moreover, NOX-induced Drp1-driven mitochondrial fission contributes to inflammasome activation by palmitate and elevated free fatty acids, which are often observed in diabetes.³¹⁸

Much evidence supports an ROS contribution to the regulation of vascular tone and inflammatory signaling in diabetes and obesity (Supplementary Table II). The interaction of ROS and the metabolic changes that occur in obesity and diabetes is less clear, and is a subject for further investigation.

III. Stroke

Ischemic stroke is a leading cause of death and long term disability in the United States,³¹⁹ and patients who exhibit metabolic risk factors including diabetes mellitus, obesity and dyslipidemia are at a greater risk of experiencing stroke-related events.³²⁰ Within minutes of hypoxia and glucose deprivation, a complex cascade of molecular events ensues, involving depolarization of neurons, increased Ca²⁺ influx, ATP depletion and release of the excitatory neurotransmitter glutamate.³²¹ Activation of glutamate receptors leads to a further increase in intracellular Ca²⁺, activation of NOS and NOX signaling, mitochondrial dysfunction and neuronal death. Although hypoxia and glucose deprivation play a major role in the neurodegeneration induced by stroke, a role for ROS is clear. Indeed, several clinical studies have shown a correlation between elevated oxidative stress and brain ischemia,^322,323 and decreasing oxidative stress may be protective against stroke-induced complications.^324,325.

MitoROS are involved in the pathophysiology of cerebral ischemia as well as reperfusion injury.³²⁶ Importantly, depletion of SOD2 increases mortality, stroke volume, brain edema and cytochrome C-mediated neuronal apoptosis following transient and permanent middle cerebral artery occlusion (MCAO).^327,328 In response to stroke, mitoROS arise from various sources including, complex I and IV, monoamine oxidase, p66^shc, mitochondrial BKCa and mK_ATP channels, cytochrome b5 reductase and dihydroorotate dehydrogenase.^329–333 In patients who experience acute ischemic stroke, p66^shc gene expression is increased in peripheral blood monocytes, and post-ischemic knockdown of p66^shc in mice undergoing transient MCAO improves survival and functional outcomes and reduces stroke lesion volume.³³⁴ p66^shc knockdown attenuates transient MCAO (tMCAO)-induced blood-brain barrier (BBB) permeability, which is associated with a reduction in NOX ROS-induced claudin-5 degradation.³³⁵ NOX signaling represents an important source of ROS in the pathophysiology of stroke. Several studies have correlated NOX2 and NOX4 depletion with decreased infarct volume, oxidative stress, BBB permeability and neutrophil infiltration after MCAO.^{324,336–339} Similar results have been reported with non-specific Nox2 inhibitors, apocynin and diphenylene iodonium,^337,340 and injection of glutathione into rats subjected to tMCAO reduces infarct volume and increases cell survival signaling.³⁴¹ Further, NOX2 deletion prevents tMCAO-induced IL-1β, TNFα, and CC-chemokine ligand 2/3 upregulation,³⁴² as well as glutamate toxicity through suppression of complexin II/soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) interaction.³⁴³ However, a recent study hypothesizes that deletion of NOX2 only delays infarct progression while simultaneously increasing angiogenesis, but is unable to prevent neuronal loss.³⁴⁴ Similarly, NOX2 deletion may have a protective effect only in models of transient occlusion, as NOX2 deletion has no effect on permanent MCAO-induced neurological dysfunction.³⁴⁵

A role for tissue-specific NOX4 signaling in cerebral ischemia has also recently been elucidated. Overexpression of NOX4 in pericytes increases BBB permeability following permanent MCAO in mice, which coincides with increased MMP-9 activity and phosphorylation of NFκB.³⁴⁶ Similarly, in response to tMCAO, EC NOX4^−/− mice have increased BBB stability associated with less autophagy-related stress, but show no reduction in neuronal cell death. In contrast, neuronal NOX4^−/− mice display reduced neuronal cell death but no change in BBB stability.³³⁹ Separate from NOX2 and NOX4, there have been conflicting reports regarding the role of NOX1 in stroke pathophysiology. NOX1 depletion decreases infarct volume, improves neurological outcome and reduces cerebral edema in mice subjected to tMCAO; however, these effects may be independent of NOX1-ROS signaling.³⁴⁷ Adenoviral-mediated knockdown of NOX1 in rats undergoing tMCAO promotes increased functional recovery, which is associated with reduced infarct size and neuronal cell death.³⁴⁸ However, two different groups have shown that NOX1 depletion does not affect infarct volume, brain edema or neurological score following tMCAO.^324,349

As mentioned, diabetes is a risk factor for stroke,³⁵⁰ and elevated blood glucose at stroke onset predisposes individuals to more severe functional outcomes and increased mortality.³⁵¹ Mechanistically, glucose injection into mice at the time of reperfusion after tMCAO increases NOX-dependent ROS production and neuronal cell death compared to normoglycemic controls, and p47^phox−/− mice are protected from this response.³⁵² Furthermore, systemic glucose is regulated by the ventromedial nucleus of the hypothalamus (VMH), and glucose sensing in the VMH is partly regulated by Drp1-dependent mitoROS signaling.³⁵³ Glucose load induces Drp1 activity through UCP2 in VMH, and UCP2-mediated mitochondrial fission reduces ROS, a mechanism that seemingly argues against the hypothesis that mitochondrial fission induces ROS, although increased Drp1-dependent mitophagy in response to permanent brain ischemia may play a role in mitigating ROS production.³⁵⁴ Depletion of UCP2 increases infarct volume in mice after tMCAO³⁵⁵ and alters whole body glucose utilization and insulin sensitivity. Re-expression of UCP2 in the VMH rectifies changes in glucose utilization and insulin sensitivity,³⁵⁶ and UCP2 overexpression in mice subjected to tMCAO reduces brain damage and neurological dysfunction.³⁵⁷

Oxygen sensing and glucose metabolism are central to regulating brain homeostasis and contribute to ischemic pathophysiology. Prolyl hydroxylase domain proteins (PHDs) regulate hypoxic signaling, most notably through regulation of HIF1α activity. PHD1^−/− mice are protected against the deleterious consequences of permanent brain ischemia.³⁵⁸ PHD1 deficiency reduces neuronal glycolysis and glucose consumption while concomitantly increasing glutamine oxidation, thereby maintaining mitochondrial respiration and energy homeostasis. This is vital for protection against ischemic-related injury as glucose is diverted towards the PPP pathway, thus increasing oxidative PPP flux and reducing equivalents needed for ROS scavenging in response to an ischemic insult.³⁵⁸ A role for oxidant scavenging has also been noted in ischemic pathophysiology. Prdx2 overexpression reduces infarct size and neurological deficits in mice through attenuation of DNA damage and PARP1/p53 pro-death signaling.³⁵⁹ GPX3 deficiency promotes vascular dysfunction and platelet-dependent arterial thrombosis and increases cerebral infarct size in mice with permanent MCAO.³⁶⁰ Conversely, overexpression of GPX1 attenuates tMCAO-induced edema, microglial activation, neutrophil infiltration and neuronal cell death compared to WT mice.³⁶¹ Similar results have been observed in SOD1 mice.³⁶²

Cumulatively, although evidence supports a major role for ROS signaling in stroke pathophysiology, few studies have delved into cell-type specific roles of ROS. Given the dynamic interplay between ECs, pericytes, VSMCs, astrocytes and immune cells in the development and progression of stroke, additional experiments are warranted. In particular, with the unique metabolic alteration that occurs during stroke, more emphasis should be placed on how ROS signaling affects nutrient shuttling and metabolism in the ischemic tissue.

3. Antioxidant Trials

Animal studies have unequivocally suggested that ROS are a viable target for therapeutic intervention in the treatment of cardiovascular disease. However, clinical trials have yielded less than favorable results,^363,364 and although these results are negative, they have helped to identify why current approaches fail and how to target future interventions. Most antioxidant trials were conducted with various vitamins (A, C and E), and although vitamins have antioxidant properties, their effects may be too broad and concentrations may never reach the levels needed to attenuate ROS production, nor be as effective in ROS scavenging by tissue antioxidant defenses. Concerns have arisen regarding efficacy of synthetic compared to natural antioxidant formulations, as well as mode of delivery (e.g., pill vs diet). Furthermore, disease progression takes time, especially in the case of atherosclerosis and heart failure. To date, most studies have only looked at a small window of time (~5 years) during more advanced stages of disease, which may not be long enough to uncover positive effects. It has also become increasingly clear that ROS are vital to maintaining physiological function, and thus general scavenging of ROS may compromise cellular homeostasis and augment disease progression in some cases. This is especially important as we consider the location and source of ROS, the specific roles of various types of ROS and how changes in concentration may affect cellular signaling. For instance, H₂O₂ has been found to induce both vasoconstriction and vasorelaxation in blood vessels treated with KCL and phenylephrine, respectively.³⁶⁵ Likewise, suppression of endothelial H₂O₂ in mice results in hypotension and prevents the upregulation of eNOS in response to exercise.³⁶⁶ On a similar note, O₂⁻, while largely known for its role in arterial vasoconstriction and mediating EC and VSMC dysfunction, has also been found to regulate both vasoconstriction and vasodilation in a concentration-dependent manner in the cerebral vasculature,³⁶⁷ and is critical to platelet activation and immune cell phagocytic function/recruitment.^368,369 ROS signaling in the setting of hypoxia has also been speculated to both instigate and inhibit angiogenesis, which may be dependent upon the local tissue environment and disease setting.^305,369 Finally, in many instances, antioxidants scavenge free radicals but do not affect H₂O₂, which as described throughout this review, is responsible for many of the pathophysiological effects on the vasculature.

While numerous studies highlight the physiological and pathophysiological differences in ROS signal transduction and their effect on cellular function, the takeaway remains the same. Instead of broad-spectrum ROS scavenging, it may be more effective to consider therapies that target specific sources of ROS, or limiting the production of specific ROS so as not to disrupt basic physiological function. We also need to consider how specific ROS influence different diseases. For example, in some cases, inhibition of O₂⁻ may not be as effective as scavenging H₂O₂ and vice-versa. However, inhibiting a specific source of ROS such as NOX may be beneficial. To this end, GKT137831, a promising NOX1/4 inhibitor, is currently in phase II clinical trials for diabetic kidney disease.³⁷⁰ Human studies will be needed to test if new therapies targeted to specific sources of ROS, such as mito-tempol, will be beneficial in the treatment of cardiovascular disease.

4. Future Investigation/Concluding Remarks

One of the challenges in dissecting the role of ROS in cardiovascular pathology is that ROS are produced not only as natural by-products of metabolic reactions in various cellular compartments, but they also serve as signaling molecules that regulate specific biochemical pathways in normal cell function and survival. Dysregulation of ROS signaling, or excess non-specific production of ROS, can influence disease pathophysiology. As highlighted in this review, ROS are particularly important in cellular metabolism and inflammatory signaling. Thus, it is not surprising that ROS play a significant role in diseases associated with metabolic dysregulation and inflammation.

There are myriad areas where further investigation is warranted. Importantly, we have seen much emphasis on cytoplasmic and mitochondrial ROS signaling; however, there is a dearth of information regarding the contribution of peroxisomal and endoplasmic reticular ROS to cellular homeostasis. New organelle-targeted probes and antioxidants will help to tease out the contribution of organellular/compartmental-specific ROS and their influence on critical cellular processes including aerobic/anaerobic respiration, β-oxidation, fatty acid synthesis, protein translation and post-translational modifications. Such tools will also provide insight into how ROS-induced inflammatory signaling may affect and be affected by this regulation, given the relationship between inflammatory and metabolic signaling. In addition, with the advent of cre-lox technology and the development of tissue-specific and inducible knockout mouse models, research should continue to delineate the tissue-specific role of ROS signaling in metabolic and cardiovascular diseases. Cardiovascular and metabolic disease etiology is complex, and understanding tissue-specific redox signaling will be important in our effort to develop new and novel therapies to treat disease. Metabolic dysregulation is a primary driver of cellular dysfunction and disease progression, and understanding the contribution and influence of ROS on metabolic processes poses an exciting area for scientific discovery.

Abbreviation

ACOX

Acyl-CoA oxidase

ADAM17

A disintegrin and metalloproteinase domain-containing protein 17

AGE

Advanced glycation end product

AMPK

AMP-activated protein kinase

AP-1

Activator protein 1

ASC

Apoptosis-associated speck-like protein containing a CARD

ASK1

Apoptosis signal-regulating kinase 1

ATM

ataxia-telangiectasia mutated

BBB

Blood-brain barrier

BH4

tetrahydrobiopterin

BM-MNC

Bone marrow mononuclear cell

BMP4

Bone morphogenic protein 4

CaMKKβ

Ca²⁺/calmodulin-dependent protein kinase kinase Beta

Cat

Catalase

CCCP

Carbonyl cyanide 3-chlorophenylhydrazone

CHOP

C/EBP homologous protein

CMA

Chaperone-mediated autophagy

CoQ

Coenzyme Q

CP

Core particle

CytoROS

Cytoplasmic ROS

DN

Dominant negative

Drp1

Dynamin-related protein 1

EC

Endothelial cell

eNOS

Endothelial nitric oxide synthase

ER

Endoplasmic Reticulum

ERO1

ER oxidoreductin 1

ETC

Electron transport chain

G6PD

Glucose-6-phosphate dehydrogenase

GlUT1

Glucose transporter 1

GPX

Glutathione peroxidase

Grb2

Growth factor receptor-bound protein 2

GRP78

Glucose-regulate protien 78

Grx

glutaredoxin

GTPCH

Guanosine 5′-triphosphate cyclohydrolase I

HIF1α

Hypoxia inducible factor 1α

IGF-1

Insulin like growth factor 1

IKK

IκB kinase

IL-18

Interleukin 18

IL-1β

Interleukin 1 Beta

iNOS

Inducible NOS

IP3R

Inositol 1,4,5-trisphosphate receptor

IκBα

Nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, alpha

JNK

c-Jun N-terminal kinase

KGDHC

α-ketoglutarate dehydrogenase

KO

Knockout

LAP

LC3-associated phagocytosis

LC3

Microtubule-associated protein 1A/1B-light chain 3

LKB1

Liver kinase B1

LMP

Low-molecular-mass protein

LOX-1

Lectin-type oxidized LDL receptor 1

LPP

Lipid peroxidation product

LPS

Lipopolysaccharide

MCU

Mitochondrial calcium uniporter

MerTK

Mer Tyrosine Kinase

MICU1

mitochondrial calcium uptake 1

Mitochondrial-associated ER membrane

MAM

MitoROS

Mitochondrial ROS

mTORC

Mammalian target of rapamycin complex

NAC

N-acetyl cysteine (CC

NF-κB

Nuclear factor kappa-light-chain-enhancer of activated B cells

NIK

NF-κB inducing kinase

NLRP3

NLR family pyrin domain containing 3

nNOS

Neuronal NOS

NOX

NADPH oxidase

ONOO

Peroxynitrite

OPA1

Optic atrophy protein 1

OxPL

Oxidized Phospholipid

P66SHC

SHC-transforming protein 1 isoform p66

PDC

pyruvate dehydrogenase

PDHK

Pyruvate dehydrogenase kinase

PDI

Protein disulfide-isomerase

Pex

Peroxin

PHD

Prolyl hydroxylase domain protein

PINK1

(PTEN)-induced kinase 1

PON

Paraoxanase

PPP

Pentose phosphate pathway

Prdx

peroxiredoxin

RAGE

Receptor for AGE

RET

Reverse electron transfer

ROMO1

ROS modulator 1

ROS

Reactive oxygen species

RP

Regulatory particle

RPT

Regulatory particle triphosphate

SNARE

Soluble N-ethylmaleimide-sensitive factor attachment protein

SOD

Superoxide dismutase

STZ

Streptozatocin

tMCAO

Transient middle cerebral artery occlusion

TNFα

Tumor necrosis factor alpha

TRAF

TNF Receptor associated factor

Trx

Thioredoxin

TrxR

Thioredoxin reductase

TSC

Tuberous sclerosis protein

TxNIP

Thioredoxin-interacting protein

UBC

Ubiquitin conjugating enzyme

UCP2

Uncoupling protein 2

ULK1

uncoordinated 51-like kinase 1

UPP

Ubiquitin proteasome pathway

UPR

Unfolded protein response

USP14

Ubiquitin carboxyl-terminal hydrolase 14

VCAM-1

Vascular cell adhesion molecule 1

VDAC1

Voltage dependent anion channel 1

VEGF

Vascular endothelial growth factor

VMH

Ventromedial nucleus of the hypothalamus

VSMC

Vascular smooth muscle cell