Dysregulation of the glutaredoxin/S-glutathionylation redox axis in lung diseases

Abstract

Glutathione is a major redox buffer, reaching millimolar concentrations within cells and high micromolar concentrations in airways. While glutathione has been traditionally known as an antioxidant defense mechanism that protects the lung tissue from oxidative stress, glutathione more recently has become recognized for its ability to become covalently conjugated to reactive cysteines within proteins, a modification known as S-glutathionylation (or S-glutathiolation or protein mixed disulfide). S-glutathionylation has the potential to change the structure and function of the target protein, owing to its size (the addition of three amino acids) and charge (glutamic acid). S-glutathionylation also protects proteins from irreversible oxidation, allowing them to be enzymatically regenerated. Numerous enzymes have been identified to catalyze the glutathionylation/deglutathionylation reactions, including glutathione S-transferases and glutaredoxins. Although protein S-glutathionylation has been implicated in numerous biological processes, S-glutathionylated proteomes have largely remained unknown. In this paper, we focus on the pathways that regulate GSH homeostasis, S-glutathionylated proteins, and glutaredoxins, and we review methods required toward identification of glutathionylated proteomes. Finally, we present the latest findings on the role of glutathionylation/glutaredoxins in various lung diseases: idiopathic pulmonary fibrosis, asthma, and chronic obstructive pulmonary disease.

GSH METABOLISM AND UTILIZATION

Glutathione is considered a major redox buffer that can exist in the thiol-reduced (GSH) and thiol-oxidized (GSSG, glutathione disulfide) forms (219). The three amino acids that constitute GSH are either derived through dietary sources or can be synthesized de novo in cases of glycine and glutamate (259, 274) and converting methionine through the transsulfuration pathway in the case of cysteine (232). GSH is metabolized through the γ-glutamyl cycle that involves six enzymes: glutamate cysteine ligase (GCL, also known as γ-glutamylcysteine synthetase), glutathione synthase (GS), γ-glutamyltranspeptidase (GGT), γ-glutamylcyclotransferase (GGCT), 5-oxoprolinase (OPLAH), and dipeptidase (DP) (Fig. 1A) (146, 151). The synthesis of GSH from the individual amino acids requires two ATP-dependent steps and two enzymes: the formation of γ-glutamylcysteine through the enzymatic reaction of GCL followed by the synthesis of GSH from γ-glutamylcysteine and glycine through the enzymatic reaction of GS. In mammals, GCL consists of two subunits, the catalytic subunit GCL-C, which is subject to feedback inhibition by GSH, and the modulatory subunit GCL-M. GGTs break down GSH to cysteinylglycine (Cys-Gly) and transfer the γ-glutamyl group to an amino acid or a peptide. GGCT then releases glutamate from the glutamyl amino acid/peptide and converts it to 5-oxoproline. Finally, OPLAH regenerates glutamate from 5-oxoproline, whereas DP breaks down Cys-Gly into individual amino acids. In addition, a separate class of cytosolic enzymes, glutathione-specific γ-glutamylcyclotransferase (ChaC)1 and ChaC2, found in the yeast Saccharomyces cerevisiae, mouse, and human, discovered more recently, can metabolize GSH directly into 5-oxoproline and Cys-Gly (118, 129). Collectively, the existence of multiple and redundant enzyme systems points to an evolutionarily conserved need for a precise glutathione level and its precursor amino acids. GSSG can be reduced to two molecules of GSH by glutathione reductase (GR) oxidizing NADPH to NADP⁺ (247). The pentose phosphate pathway, an offshoot of the glycolysis cascade, is a major pathway that in turn reduces NADP⁺ to NADPH, thereby maintaining the glutathione redox buffer (231). The reported GSH concentrations in cells are in the range of 1–11 mM (219). The GSH/GSSG ratio in the cytosol ranges from 100:1 to >500:1 in different cell types (71), while in the endoplasmic reticulum (ER), the reported GSH/GSSG ratio ranges from 1:1 to 7:1 by use of different methods of measurements (8, 169). Lack of adequate blocking, in addition to autooxidation, may result in an overestimation of GSSG values, leading to aforementioned variations in reported GSH/GSSG ratios, and these concerns continue to plague the field (73). GSH is involved in multiple cellular reactions. GSH is consumed by glutathione peroxidases to reduce oxidized lipids (27). GSH is also involved in phase II metabolism where glutathione S-transferases (GSTs) conjugate GSH to electrophiles or other compounds in detoxification processes (281). GSH is also utilized by leukotriene C4 synthase to form leukotriene C4, the parent compound of the cysteinyl leukotrienes (132). Lastly, GSH is conjugated to proteins in a process called protein S-glutathionylation (PSSG) as a protective mechanism that prevents cysteines from overoxidation and also changes the protein structure and function, reviewed in detail below (Fig. 1B).

Fig. 1.

A: reduced glutathione (GSH) metabolism. Glutamate cysteine ligase (GCL), composed of a catalytic subunit GCLC and a regulatory unit GCLM, catalyzes the formation of γ-glutamylcysteine (1). Glutathione synthetase (GS) then completes the synthesis of GSH by adding glycine to γ-glutamylcysteine (2). GSH is broken down to cysteinylglycine by γ-glutamyl transpeptidase (GGT) (3). The γ-glutamyl group can be recycled through multiple steps where it is converted to 5-oxoproline by γ-glutamyl cyclotransferase (GGCT) (4) before it is converted back to glutamate by 5-oxoprolinase (OPLAH) (5). Cysteinylglycine is broken down into cysteine and glycine by dipeptidase (DP), where they are recycled or are shunted into other metabolic pathways (6). Alternatively, glutathione-specific γ-glutamylcyclotransferase (ChaC)1 and ChaC2 break down GSH into cysteinylglycine and 5-oxoproline (7). B: GSH utilization. GSH is conjugated to 1) oxidized lipid (LOOH) and through the catalytic reaction of glutathione peroxidase (GPXs) reduces oxidized lipid to its reduced form (LOH); 2) electrophiles and other xenobiotics through the catalytic reaction of glutathione S-transferases (GSTs) in phase II drug metabolism; 3) leukotriene A4 through the action of leukotriene C4 synthase (LTC4S), forming leukotriene C4; and 4) proteins to prevent protein cysteines from undergoing irreversible oxidation. GSH is also used in the regeneration of the sulfhydryl group to its reduced state through the catalytic activity of glutaredoxins.

PSSG: CHEMISTRY, REDOX REGULATION, DETECTION, AND PREDICTION OF PSSG SITES

One mechanism whereby GSH affects cellular physiology that has emerged within the past 10–15 years is through its covalent reaction with reactive cysteines within proteins, a process known as protein S-glutathionylation (PSSG, also known as glutathiolation or protein mixed-sulfide.) Protein S-glutathionylation represents a special type of protein disulfide, wherein the disulfide bond is formed between a protein thiol and glutathione. PSSG can occur in multiple ways. Invariably, some sort of oxidative stress precedes the formation of the disulfide bond between a protein thiol group and GSH. Proteins can be directly S-glutathionylated in vitro in the presence of compounds such as GSSG, S-nitrosoglutathione (GSNO), and glutathione disulfide S-oxide (GS(O)SG) by using supraphysiological concentrations (17, 136, 168, 208). Within cells, PSSG can be induced through the inhibition of GR to artificially increase GSSG concentration (260, 267) or by directly exposing cells to oxidants such as H₂O₂, diamide, peroxynitrite (ONOO⁻), or GSSG ethyl ester (1, 124, 260). These oxidants can lead to diverse oxidations of protein thiols including the formation of sulfenic acids or nitros(yl)ated species that can be a gateway for subsequent PSSG (72, 82, 160). More relevant to normal cellular physiology are studies demonstrating that cellular redox perturbations that occur in physiological settings also lead to S-glutathionylation of target proteins. For example, activation of NADPH oxidases (NOXs), or oxidants derived from mitochondria, have been shown to promote S-glutathionylation, and a number of PSSG targets have been identified (116, 145, 255). These more recent observations demonstrate that PSSG is not merely representative of cell stress but is instead a feature of normal cellular redox homeostasis.

Early methods to detect PSSG include the use of anti-GSH antibody and ³⁵S-labeled glutathione (37). Other avenues to detect PSSG include treatment of cells with biotinylated GSH (Bio-GSH), biotinylated GSSG (Bio-GSSG), biotinylated glutathione ethyl ester (Bio-GEE), and fluorescein isothiocyanate-labeled glutathione (FITC-GSH) (26, 133, 238). However, these modified forms of glutathione suffer the drawback of having a bulky molecule conjugated to GSH, which might reduce its accessibility to reactive protein cysteines that might be S-glutathionylated (61). Using click chemistry, another iteration of glutathione has been developed where the side chain of the glycine residue is replaced with an azide, an alkynyl, or an allyl group (61, 119, 217, 218). The subsequent S-glutathionylated proteins detected with this modified GSH are biotin labeled through click chemistry using a corresponding modified biotin such as biotin alkyne or azide-PEG3-biotin. The biotin-labeled S-glutathionylated proteins can then be enriched using streptavidin/avidin bead pulldown. These modified forms of glutathione with glycine derivatives (GSH with azido-Ala and GSH with allyl-Gly) avoid the concern of bulky molecule conjugation mentioned above. All of the aforementioned methods to detect PSSG with GSH analogs suffer from the limitation that they are added to cell culture media or are incubated with cell lysates, where they compete with endogenous glutathione and/or interfere with normal redox homeostasis.

Since the administration of modified versions of GSH cannot be used in the studies of PSSG in human tissues, alternative methods have been developed to identify protein S-glutathionylation based on specific derivatization of PSSG followed by resin-based affinity capture of the target proteins (81, 162) (Fig. 2). These resin-assisted enrichment methods take advantage of the specificity of the deglutathionylating activity of glutaredoxins (discussed below) to derivatize free thiol groups from S-glutathionylated proteins. The first step in this procedure includes cell or tissue lysis and permeabilization in the presence of a thiol-blocking agent to exclude capturing already reduced protein cysteines. Samples are thereafter incubated with GLRX to convert PSSG to proteins with newly derivatized free thiol groups. Next, these newly formed free thiol groups can be enriched directly using sulfhydryl-reactive beads such as thiopropyl sepharose beads or after incubation with modified biotins that are sulfhydryl reactive such as HPDP-biotin, which can then be enriched with streptavidin/avidin beads (70, 235). The enriched proteins are then trypsin digested and labeled with isobaric reagents such as tandem mass tag (TMT) or iTRAQ for mass spectrometry-based identification and quantification. The specific cysteine residues of a protein that are S-glutathionylated can also be identified if labeled with sulfhydryl-reactive biotins through a specific mass shift of the cysteine-containing fragment ion. In addition to these methods, our laboratory has developed an in situ method for visualizing PSSG in intact tissues. This method is applicable to paraffin-embedded sections and can reveal regional and cell-specific differences in PSSG in healthy and diseased conditions (2). Lastly, in addition to the qualitative detection of PSSG, or identification of specific PSSG targets, total PSSG can also be quantified biochemically. The quantification of total PSSG involves steps to remove free GSH/GSSG and precipitate proteins, followed by reduction to release the GSH moiety from S-glutathionylated proteins, and finally quantify GSH released (200, 256, 279). Several bioinformatics tools have been developed to predict S-glutathionylation of specific cysteine residues from a given protein that can be glutathionylated, with specificity and sensitivity ranging from 50 to 75% (180, 241, 284, 285). A web-based database, dbGSH, has also been created, where all experimentally verified S-glutathionylated proteins are manually curated (43).

Fig. 2.

Schema summarizing the steps required to identify S-glutathionylated proteins using glutaredoxin-1 (GLRX)-based derivatization and thiol affinity capture. Existing free thiols in proteins are first blocked by thiol-blocking agents (1). S-glutathionylated cysteines are thereafter derivatized by adding GLRX and cofactors (2), resulting in the emergence of new free thiols. New thiols are enriched using thiol-reactive beads (3), and enriched proteins are eluted thereafter for further analysis (4), using mass spectrometry or Western blotting as examples. We refer the reader to the text for detailed descriptions about these procedures.

GLUTATHIONYLATION/DEGLUTATHIONYLATION CHEMISTRY AND RELEVANCE TO BIOLOGICAL SIGNALING PATHWAYS

Protein S-glutathionylation, glutaredoxins, and GSH have increasingly garnered more attention in recent years as reflected by the increasing amount of publications with glutathionylation, glutaredoxin, or glutathione as key terms. Protein S-glutathionylation/deglutathionylation has been implicated in the regulation of a multitude of biological functions and diseases analogous to other reversible posttranslational modifications of amino acids. This can be attributed to the fact that the conjugation of GSH, a three-amino acid peptide, changes a protein’s tertiary structure (166, 280). The S-glutathionylation of a protein could also affect the formation of protein complexes. For example, S-glutathionylation of histone H3 induces structural changes and leads to nucleosomal instability (69, 186, 206). S-glutathionylation of cytoskeletal proteins such as actin, tubulin, and L-plastin inhibits their polymerization and leads to altered cellular functions, such as arrested mitosis and neutrophil and monocyte polarization, chemotaxis, adhesion, and phagocytosis (42, 54, 216, 255, 258). Lastly, S-glutathionylation may serve as a signal for protein degradation. For example, the caspase-dependent 14-3-3ζ degradation under low-density lipoprotein plus high-glucose-induced metabolic stress in monocytes is increased following 14-3-3ζ glutathionylation (120). S-glutathionylation of the 20S proteasome core has been shown to increase its proteolytic activity during oxidative stress, perhaps as a coping mechanism of cells under stress (227).

PSSG regulates cellular functions in multiple ways. For example, various enzymes involved in cellular bioenergetics, such as aldolase, uncoupling proteins UCP2 and UCP3, and mitochondrial complex I have been shown to be S-glutathionylated and to inhibit their respective enzymatic activities (100, 154, 187). Multiple subunits (a1 and b1) of the Na⁺-K⁺ pump have been shown to be S-glutathionylated where the addition of the GSH adduct is at least partially modulated by its associative protein, the FXYD domain-containing ion transport regulator 3 (FXYD), resulting in decreased ATPase activity (19, 144, 145). Endothelial nitric oxide synthase (eNOS), the major isoform of NOS that regulates vascular function through the production of nitric oxide (NO), is a highly studied PSSG-regulated protein (286). eNOS produces NO in its dimer form (65), and S-glutathionylation of eNOS at Cys⁶⁸⁹ and Cys⁹⁰⁸ uncouples eNOS, switching the production of NO to O₂^·−, thereby disrupting the NO signaling pathway, with implications for vascular biology (41). PSSG also indirectly regulates cellular signaling pathways through the modification of protein kinases and phosphatases, thus altering the phosphorylation/dephosphorylation of target proteins (30, 31, 52, 115, 121, 161, 245, 269). For example, glutathionylation of ribosomal S6 kinase-1 inhibits its kinase activity and leads to downstream inhibition of NOS activation due to the inability of NOS to be phosphorylated (245). Another protein kinase, Ca²⁺/calmodulin-dependent protein kinase I (CaMKI), is shown to be inhibited by S-glutathionylation at its active-site cysteine (115). On the other hand, adenosine monophosphate-activated protein kinase (AMPK) is activated by S-glutathionylation at its AMPKα catalytic subunit (52). S-glutathionylation also modulates MAPK phosphatase-1 activity through increased proteasomal degradation (121). S-glutathionylation of Signal transducer and activator of transcription 3 (STAT3) has been demonstrated to interfere with its phosphorylation, resulting in its inactivation (30, 31, 161, 269). For example, S-glutathionylation of STAT3 at Cys^328/542 inhibits STAT3 phosphorylation at Tyr⁷⁰⁵ (31) and has been correlated to increased apoptosis and decreased cell viability in alantolactone- or cynaropicrin-induced oxidative stress (30, 161).

The PSSG/glutaredoxin system also can affect other oxidative posttranslational modifications indirectly, through the control of thioredoxin (TXN) and peroxiredoxins (PRDX) redox systems, which are interconnected to control the cellular redox environment. For example, TXN has been shown to be glutathionylated at Cys⁷² and lead to the inhibition of its insulin disulfide reductase activity (37). TXN can be inactivated following oxidation of noncatalytic site cysteines, leading to the formation of a two-disulfide form of TXN that is inactive. The GLRX system allows regeneration of active TXN via a monothiol mechanism, described below (53). PRDX can be inactivated following their overoxidation, involving the formation of sulfinic (-SO₂H) and sulfonic acid (-SO₃H) residues (57). S-glutathionylation has been shown to protect PRDX from overoxidation by preventing the formation of sulfinic acid species, allowing rapid regeneration of reduced active PRDX following GLRX-mediated restoration of the SH group (192). Of special note is the importance of S-glutathionylation of PRDX6, a 1-Cys peroxiredoxin. Glutathione S-transferase pi (GSTP) has been shown to be critical in the catalysis of S-glutathionylation of PRDX6, increasing the accessibility of the otherwise buried sulfenic acid or sulfenylamide forms of PRDX6, allowing subsequent regeneration of its catalytic activity through the GLRX/GSH system (156, 202).

DISCOVERY AND BIOCHEMISTRY OF GLUTAREDOXINS

Glutaredoxin (GLRX) was first described in the 1950s–60s (175, 199). The enzyme was called “transhydrogenase”, as it was thought that hydrogen was transferred from two GSH molecules to reduce the disulfide bond. It was later demonstrated that this enzyme catalyzes deglutathionylation in a two-step reaction where a thiol-disulfide exchange between GSH and the protein disulfide bond occurs in the first step (9) (Fig. 3A). The name “thioltransferase” was thus suggested to more accurately describe this protein. Independently, an enzyme was discovered in Escherichia coli that acted as an alternate hydrogen donor to the thioredoxin system for ribonucleoside diphosphate reductase (97, 99). The author, Arne Holmgren, suggested the name “glutaredoxin”, as glutathione supplied the reducing power in the reaction. Glutaredoxins/thioltransferases were subsequently discovered in other mammals and other kingdoms (11, 89, 102, 134, 214). The name “thioltransferase” fell out of preference, and the enzyme has been known as glutaredoxin since.

Fig. 3.

Human glutaredoxin-1 (GLRX) primary sequence, 3D structure, and deglutathionylation mechanism. A: GLRX catalyzes deglutathionylation. GLRX deglutathionylates proteins and is glutathionylated in the process (1). GLRX is regenerated by consuming GSH, which is in turn regenerated through the reduction of oxidized glutathione (GSSG) by glutathione reductase (GR) with NADPH as the final electron donor (2). B: the 5 cysteines that comprise human GLRX are indicated. Note that only Cys²³ is required for deglutathionylation activity. Cys⁸, Cys⁷⁹, and Cys⁸³ have been linked to oxidative inactivation of GLRX.

GLRXs are generally known as enzymes that catalyze deglutathionylation activity (to be reviewed in a later section.) However, GLRXs have also been demonstrated to be promoting protein glutathionylation on rare occasions (17, 138, 195, 211, 230). GLRXs belong to the thioredoxin protein superfamily that catalyzes disulfide bond formation and share a distinct “thioredoxin fold” structural motif that consists of three α-helices flanking a four-stranded β-sheet (150, 158). GLRXs can be divided into two classes, the monothiol glutaredoxins such as human GLRX3 (PICOT) and GLRX5 with an active site amino acid sequence of CGFS (111, 263), and the dithiol glutaredoxins such as human GLRX1–2 with an active site sequence of C(P/S)(Y/F/D)C (152, 183). The dithiol glutaredoxins were originally classified as disulfide reductases that catalyze the disulfide reduction of certain proteins and other small molecule oxidation such as hydroxyethyl disulfide (HED) and S-sulfocysteine (Cys-SO₃) (12, 98, 141, 153, 157, 194, 262). GLRXs catalyze deglutathionylation in a ping-pong mechanism whereby the first step involves the nucleophilic displacement of the GSH moiety by the active site cysteine, followed by the rate-limiting step where the thiolate ion of the active site cysteine is regenerated by consuming one molecule of GSH (228) (Fig. 3A). GLRX’s specificity toward the γ-linkage-glutamyl moiety in both the γ-linkage-glutamyl moiety of glutathionylated proteins and free GSH is conferred by the GSH binding groove not found in other members of the thioredoxin superfamily (29, 239, 276). GLRXs contain a varying number of cysteines, ranging from two cysteines in E. coli GLRX (101), to five cysteines in human GLRX (183) (Fig. 3B). The NH₂-terminal active site cysteine of the active site motif (Cys²³ of human GLRX) is required for both disulfide bond reduction and deglutathionylation (67, 68, 270, 271). This cysteine differs from the COOH-terminal active site cysteine (Cys²⁶) in that Cys²³ is exposed to the exterior of the protein and is solvent accessible, whereas Cys²⁶ is buried and is inaccessible to solvent (117, 266). The NH₂-terminal active site cysteine has an unusually low pK_a value of ∼3.5, contributing to the formation of thiolate ion (67, 165, 271). Several factors contribute to the low pK_a, including the ion-pair interactions between the conserved positively charged residue and the thiolate ion (108, 266, 271), the interaction of the thiolate ion and the electrostatic field of the α-helix by the helix dipole (123) and the proline residue immediately after the NH₂-terminal active site cysteine (66, 110).

The functional relevance of other cysteines in deglutathionylation catalysis varies among species. In E. coli GLRX, the COOH-terminal active site cysteine contributes to the deglutathionylation reaction, as cysteine-to-serine mutation significantly reduces deglutathionylation activity (28, 215). In contrast, mammalian GLRX requires only the active site NH₂-terminal cysteine for the deglutathionylation, as the mutation of individual or all four cysteines other than active site Cys²³ does not change the deglutathionylating activity (117, 268, 270, 271). Mutation of the COOH-terminal active site cysteine to serine (Cys²⁶Ser in human GLRX) actually increases the enzymatic activity of GLRX (268, 271).

Several methods exist to measure the enzymatic activities of glutaredoxins. In the ribonucleotide reductase-coupled assay, the disulfide reduction activity is measured by the formation of [³H]dCDP from [³H]CDP, as glutaredoxin is required for GSH-dependent reduction of oxidized ribonucleotide reductase to catalyze the reduction of ribonucleotides to deoxyribonucleotides (96–99). The more commonly used assay to measure deglutathionylating activity is the NADPH-coupled assay using substrates such as HED or S-sulfocysteine, where the glutaredoxin activity is measured by the consumption of NADPH (11, 66–68, 96, 134, 165, 228). Other methods to more directly measure the deglutathionylation activity also exist, for example, peptides with glutathionylated cysteine adjacent to a tryptophan residue (189, 215), fluorescent eosin-glutathione or eosin-glutathionylated BSA (48, 111), or ³⁵S-glutathionylated proteins (114). Each assay has its limitations. The ribonucleotide reductase-coupled assay is not suitable to measure the deglutathionylation activity selectively, as the assay relies on the disulfide bond reduction of the oxidized ribonucleotide reductase and, as such, measuring general disulfide bond reduction instead of deglutathionylation. Most substrates used in the NADPH-coupled assay are small molecules that do not represent GSH bound to bulky biological protein cysteines. Furthermore, monothiol glutaredoxins do not have any measurable enzymatic activity using the HED/NADPH-coupled assay, whereas the deglutathionylation activity of monothiol glutaredoxins has been documented in vivo (246, 278). Using ³⁵S-glutathionylated proteins from whole cells, GLRX was demonstrated to preferentially deglutathionylate proteins of lower molecular weights (114). Thus, it is important to consider carefully which activity assay to utilize when investigating the biological relevance of glutaredoxin activity.

HUMAN GLUTAREDOXINS

Glutaredoxin-1

Human glutaredoxin-1 (GLRX) is primarily localized to cytosol and is encoded by 321 bases that translate to 106 amino acid residues, with the active site CPYC motif. The pK_a of the active site cysteine (Cys²³) is ∼3.5 (165), which is similar to the active site cysteine of GLRX in other mammals. The human gene encoding GLRX was mapped onto the chromosome arm 5q (178). The TATA box and the CCAAT sequence are located at 30 and 160 bp, respectively, upstream of the transcription start point (184). The promoter region contains 69 predicted transcription factor binding sites, including transcription factors implicated in apoptosis (c-Jun), cytokine production (NF-κB1, RelA), cytokine response (STAT1, STAT5), and redox sensing (AP-1) (Table 1) (60, 164, 184). Consequently, a number of growth factors and cytokines have been shown to affect GLRX mRNA expression. Notably, exposure to the profibrogenic growth factor, transforming growth factor-β1 (TGF-β1) led to decreased GLRX mRNA in mammary, tracheal and alveolar epithelial cells. Conversely, GLRX mRNA was increased in tracheal epithelial cells exposed to interferon-γ (IFN-γ) (135, 188, 209) or activation of the NF-κB pathway (4).

Table 1.

Transcription factors predicted to bind to HsGLRX promoter region

AP-1	GRα	POU2F1
AP-2α	Grβ	POU2F2
AR	HNF-1A	PPARα:RXRα
ATF-1	HNF-1B	PRα
ATF3	HNF-1C	PRβ
c-Ets-1	HNF-3α	PU.1
c-Ets-2	HNF-4α	PXR-1:RXRα
c-Jun	HOXD10	RARβ
c-Myb	HOXD9	RBP-Jκ
C/EBPα	IRF-1	RelA
C/EBPβ	IRF-2	RXRα
CTF	LEF-1	SRY
E2F-1	MEF-2A	STAT1β
EBF	NF-1	STAT4
Egr-3	NF-AT1	STAT5A
ELF-1	NF-AT1	T3Rβ1
Elk-1	NF-AT2	TCF4
ENKTF-1	NF-κB1	TCF4E
ERα	NF-Y	TFII-I
FOXP3	NFI/CTF	TFIID
GATA-1	p53	VDR
GATA-2	Pax-5	XBP-1
GR	PEA3	YY1

Prediction was done using PROMO, a web-based program for transcription factor binding site prediction (http://alggen.lsi.upc.es). GLRX, glutaredoxin-1; Hs, Homo sapiens; AP, activator protein 1; AR, androgen receptor; ATF, activating transcription factor; C/EBP, CCAAT box enhancer-binding protein; CTF, CCAAT box-binding transcription factor; E2F, E2 transcription factor; EBF, early β-cell factor; Egr, early growth response protein; ELF, E74-like factor; Elk-1, ETS like-1 protein; ENKTF, enkephalin transcription factor; ER, estrogen receptor; FOXP, forkhead box P3; GATA, GATA-binding protein; GR, glutathione reductase; Gr, glucocorticoid receptor; HNF, hepatocyte nuclear factor; HOXD, homeobox D; IRF, IFN regulatory factor; LEF, lymphoid enhancer-binding factor; MEF, mouse embryo fibroblast; NF, nuclear factor; NF-AT, nuclear factor of activated T cells; Pax, paired-box protein; PEA, polyoma enhancer activator; POU2F2, Pit-Oct-Unc family transcription factor 2F2; PR, progesterone receptor; PU.1, PU box transcription factor; PXR, pregnane X receptor; RXR, retinoid X receptor; RAR, retinoic acid receptor; RBP, retinol-binding protein; RelA, Rel-like domain-containing; SRY, sex-determining region Y; STAT, signal transducer and activator of transcription; T3R, thyroid hormone receptor; TCF4, transcription factor 4; TF, transcription factor; VDR, vitamin D receptor; XBP, X-box binding protein; YY1, yin yang 1.

Human GLRX is sensitive to oxidative stress. Incubation of GLRX with GSSG, H₂O₂, or ONOO⁻ was shown to significantly decrease its enzymatic activity (88, 196, 229). Incubation of GLRX with GSSG also resulted in a mixture of oligomers, glutathionylation at noncatalytic cysteines, and disulfide formation between noncatalytic cysteines (88, 240). Lastly, Cys⁸ is solvent exposed, and Cys⁸Ser mutation prevents its aggregation (179). These results suggest a mechanism whereby structural cysteines regulate GLRX through oxidation processes. The physiological relevance of GLRX regulation through these cysteines remains to be fully understood. As will be described below, S-glutathionylation of GLRX, at presumably noncatalytic cysteines, has been linked to its oxidative inactivation in settings of fibrotic lung disease (6).

Glutaredoxin-2

Human glutaredoxin-2 (GLRX2) has three isoforms, GLRX2a (mitochondrial) and the nuclear and cytosolic GLRX2b and GLRX2c (147, 152). GLRX2 is composed of either 164 (mitochondrial isoform; GLRX2a) or 165 amino acids (nuclear isoform; GLRX2b), with a molecular mass of 18–19 kDa. The gene encoding GLRX2 was mapped on to chromosome arm 1q and contains five exons (exons 1a, 1b, 2, 3, and 4) (147, 152). Both isoforms contain exons 2, 3, and 4 where the nuclear translocation signal is located. Alternative splicing of exon 1b gives rise to isoforms 2b and 2c (147), whereas the mitochondrial form includes an additional exon 1a, which contains the mitochondrial translocation signal (74, 152). GLRX2 contains the conserved active site motif, with an amino acid sequence of CSYC and the thioredoxin fold, and shares ~36% sequence identity with GLRX (74). GLRX2 contains four cysteines, the two cysteines at the active site, and two others (Cys²⁸ and Cys¹¹³) that form a disulfide bond (109). GLRX2 is a Fe/S cluster-binding protein. The dimeric holo GLRX2 consists of two GLRX2 monomers, a 2Fe/2S cluster and two GSH molecules and is coordinated through the active site Cys³⁷ (18, 109). GLRX2 has also been demonstrated to catalyze deglutathionylation, although its specific activity of deglutathionylation is 10-fold less than that of GLRX (152). Interestingly, holo GLRX2 is unstable in an aerobic environment or in the presence of oxidants and can regain the “lost” enzymatic activity, suggesting a redox-sensing property of GLRX in addition to its role as an additional pool of deglutathionylating enzymes in response to oxidative stress (139, 140, 265).

Glutaredoxin-3 (PICOT)

Glutaredoxin-3 (GLRX3) was originally discovered as a cytosolic, protein kinase Cθ (PKCθ)-interacting protein, and the original name is PICOT (PKC-interacting cousin of thioredoxin) (263). Unlike GLRX and GLRX2, GLRX3 is a multidomain, monothiol glutaredoxin. GLRX3 contains a thioredoxin domain but not the conserved thioredoxin active site motif. Mammalian GLRX3 has two GLRX domains, with a conserved CGFS motif. The TXN-like domain is sufficient for GLRX3-PKCθ interaction, although it is unclear whether it is sufficient for the inhibitory effect GLRX3 has exhibited toward PKCθ (263). Similar to GLRX2, GLRX3 is a Fe/S cluster-binding protein (90). Holo GLRX3 is a homodimer, with two Fe/S clusters, with the active site cysteines (Cys¹⁵⁹ and Cys²¹⁶) at each GLRX domain contributing to the binding of respective Fe/S clusters. Like GLRX2, the GLRX3 holo complex is also unstable in aerobic conditions or in the presence of oxidants such as GSNO (90). It is currently unclear whether GLRX3 exhibits any deglutathionylating activity.

Glutaredoxin-5

Glutaredoxin-5 (GLRX5) is a single-domain monothiol glutaredoxin with 157 amino acids and a molecular mass of 16.6 kDa. The active site motif has a sequence of CGFS. GLRX5 possesses deglutathionylating activity, although the activity is multiple folds less than that of GLRX2 when glutathionylated BSA is used as the substrate (111). The GLRX5 holo complex is composed of four GLRX5 monomers, two Fe/S clusters, and four GSH molecules, with each Fe/S cluster binding coordinated by the active site cysteines (Cys⁶⁷) from two GLRX5 monomers and two GSH molecules (111). GLRX5 translocates to mitochondria and plays an important role in Fe/S cluster biogenesis and cellular iron homeostasis, based on findings that ablation of GLRX5 led to a loss of cytosolic Fe and a concomitant mitochondrial Fe overload (33, 210, 273).

OTHER ENZYME SYSTEMS IMPLICATED IN S-GLUTATHIONYLATION AND DEGLUTATHIONYLATION REACTIONS

In addition to GLRX, a number of other enzymes have emerged as critical regulators of protein glutathionylation and deglutathionylation. These include members of the family of GSTs, which were originally described as enzymes that facilitate the conjugation of GSH to xenobiotics as part of phase II drug metabolism (25). More recently, GSTs have become recognized for their role in promoting PSSG, with GSTP being the prototypical S-glutathionylating enzyme (248, 281). In addition, glyoxalases (GLOs), which convert α-ketoaldehydes into D-hydroxyacids, also have been implicated in the catalysis of S-glutathionylation (59). Sulfiredoxin is an oxidoreductase that catalyzes the reduction of cysteine sulfinic acids of the active site cysteines of 2-cys PRDXs (21, 39). Sulfiredoxin has been demonstrated to catalyze deglutathionylation (32, 62, 185). Lastly, TXNs are oxidoreductases that catalyze protein disulfide reduction that share thioredoxin fold with glutaredoxins (149, 158). TXNs have also been shown to deglutathionylate GAPDH, PRDX3, the 20S proteasome, and eNOS by use of a dithiol reduction mechanism that is independent of GSH (16, 86, 226, 236). Due to space limitations, we will not further describe these other enzyme systems here.

DYSREGULATION OF THE S-GLUTATHIONYLATION/GLUTAREDOXIN REDOX AXIS: IMPLICATION FOR LUNG DISEASES

The lung, due to its biological function as the organ for gas exchange, is exposed to oxygen more than any other internal organ. It is estimated that we breathe in 10,000–20,000 liters of air daily (212). As a result, the lung experiences insults not just from oxidants derived from endogenous sources such as inflammatory cells or NADPH oxidases, mitochondria, etc., but also from exogenous sources, including cigarette smoke (discussed below), ozone, or other oxidant gases (13, 212). Oxidative stress has been linked to lung diseases for many years now, and the precise “redox mechanisms” that regulate lung function and/or pathophysiology are only slowly being teased apart owing to improved methodologies to detect oxidant-induced modifications. Similarly, the role of PSSG chemistry and glutaredoxins in the pathophysiology of lung diseases is only now emerging. In the next paragraphs we discuss relevant studies pertaining to PSSG/glutaredoxins in idiopathic pulmonary fibrosis, asthma, and chronic obstructive pulmonary disease.

IDIOPATHIC PULMONARY FIBROSIS

Idiopathic pulmonary fibrosis (IPF) is a lung disease with progressive scarring in the lung that is defined by the histopathological pattern of usual interstitial pneumonia with unknown etiology (159). The prevalence of IPF ranges from 6 to 10 per 100,000 people, with a median survival of 3–5 years after the initial diagnosis. Age is the greatest risk factor identified, with the disease being slightly predominant in males. At the cellular level, disruption of type 2 alveolar epithelial cells (AEC2s; cells that secrete surfactant to maintain alveolar structure) homeostasis has been shown to play important roles in IPF, as AEC2s in IPF lungs are found to be senescent, with genomic instability, telomere attrition, and mitochondrial dysfunction (122). A role of epithelial cells in the pathogenesis of IPF also has emerged, based on observations in patients with familial IPF-harboring mutations in the surfactant protein C (SFTPC) or ATP binding cassette member A3 (ABCA3) genes. SFTPC and ABCA3 are uniquely expressed in epithelial cells, and mutations in these genes induce endoplasmic reticulum (ER) stress and mitochondrial perturbations and culminate in pulmonary fibrosis (34, 177). Myofibroblasts, cells transdifferentiated from fibroblasts that secrete various growth factors, cytokines, and extracellular matrix during tissue repair, are also disproportionately activated, resulting in aberrant deposition of extracellular matrix proteins (122) (Fig. 5).

Fig. 5.

Summary of how an altered glutaredoxin-1/protein S-glutathionylation (GLRX/PSSG) status affects idiopathic pulmonary fibrosis (IPF), asthma, and chronic obstructive pulmonary disease (COPD). The main pathophysiological features that accompany these diseases are represented graphically. The text below summarizes what is currently known about GLRX status in these diseases. Potential PSSG targets are also indicated. Please note that many of these were derived from rodent models or cell-based studies with disease-relevant stimuli, but the majority will need to be validated in human specimens. As is described in the text, S-glutathionylated proteomes in human tissues also remain to be unraveled. Nonetheless, the proteins that were found to be S-glutathionylated are believed to have substantial implications for a number of processes and pathways that are listed and are known to control the pathogenesis of these chronic respiratory diseases. [Created with BioRender.]

Oxidative stress has been linked to the development of pulmonary fibrosis (44). However, the precise mechanisms whereby redox perturbations contribute to disease pathogenesis remain unclear, and, thus far, approaches that use the nonspecific low-molecular-weight thiol and glutathione precursor N-acetyl-l-cysteine (NAC), failed to show therapeutic benefit in patients with IPF (105). Epithelial cell death and the death receptor Fas (CD95) have been recognized as key drivers of pulmonary fibrosis. For instance, it has been demonstrated that administering FasL into mouse lungs causes epithelial cell apoptosis with an accompanying development of lung fibrosis (83, 130, 131). The amount of GSH has been documented to be significantly less in the epithelial lining fluids from IPF patients, indicating a change of redox environment and the potential role of protein S-glutathionylation (35). We have shown previously the importance of glutathionylation of Cys²⁹⁴ of Fas (Fas-SSG) for epithelial cell death (5) and that protein disulfide isomerase-3 (PDIA3) and GSTP both contributed to Fas-SSG in the ER. Fas-SSG promoted its trafficking to the surface, increased the assembly of the death-inducing signaling complex (DISC) and binding of FasL, leading to enhanced cell death (7) (Fig. 5). The interaction between GSTP and Fas was increased in lungs of IPF patients, with concomitant increased Fas-SSG, and similar findings were apparent in lungs from mice with bleomycin- or adenovirus-expressing active transforming growth factor-β1 (AdTGFβ)-induced fibrosis (163). The increase in Fas-SSG was attenuated in Gstp1^−/− mice following treatment with TLK117, a selective GSTP inhibitor, in bleomycin- and AdTGFβ-induced mouse models of pulmonary fibrosis (163). The importance of PSSG in the pathology of pulmonary fibrosis is further substantiated by observations that, overall, PSSG was increased in the lungs from mice with bleomycin- or AdTGFβ-induced lung fibrosis, and in lungs of IPF patients (6, 163). Findings that Gstp1^−/− mice or mice treated with TLK117 have attenuated collagen content and decreased overall PSSG also suggest the importance of PSSG in fibrogenesis (163). The increase in Fas-SSG was accompanied by caspase-3/8-dependent degradation of GLRX, sustaining the increases in Fas-SSG. Accordingly, epithelial cells lacking Glrx displayed more Fas-SSG and were more prone to FasL-induced cell death (5). Importantly, Glrx^−/− mice were more susceptible to bleomycin/AdTGFβ-induced fibrosis, with exacerbated collagen content in the lungs and increased overall PSSG compared with their wild-type counterparts (6). Conversely, the overexpression of Glrx specifically in epithelial cells or direct airway administration of recombinant GLRX into the airways attenuated Fas-SSG, reduced caspase-3 activity, and decreased the collagen content in mice with existing bleomycin/AdTGFβ-induced fibrosis, accompanied by increased collagenase activity (6). Although the exact mechanisms whereby the collagen degradation activity was restored/enhanced remain unknown, our data showed an uptake of recombinant GLRX by epithelial cells and cells that resemble macrophages (6). Interestingly, Glrx mRNA expression increased in monocytes stimulated with IL-6 to induce monocyte-to-macrophage differentiation in association with enhanced phagocytosis (244). Increases in PSSG have been linked to macrophage dysfunction (10). Our work has demonstrated that absence of Glrx impaired lipopolysaccharide (LPS)-induced macrophage activation and phagocytosis (3). Taken together, these observations suggest that the impact of GLRX on macrophage function may be important in regulating tissue fibrogenesis.

The increases in overall PSSG in lungs of IPF patients were accompanied by a decrease in GLRX activity. Levels of GLRX protein and mRNA in IPF lungs were slightly decreased compared with lungs from healthy subjects. Moreover, GLRX protein in lungs from IPF patients was significantly S-glutathionylated (6). The same observation was made in the mouse model of bleomycin-induced lung fibrosis. S-glutathionylation of GLRX was reduced in mice administered recombinant GLRX, in association with an overall increase in GLRX activity in these lungs. Lastly, isolation of GLRX from IPF lungs followed by dithiothreitol reduction restored its activity (6). Collectively these findings suggest that endogenous GLRX is inactivated by its own substrate, glutathione, presumably due to oxidation of cysteines outside of the thioredoxin domain, leading to runaway glutathionylation (Fig. 5).

As mentioned above, IPF is a disease associated with aging, with age being the greatest risk factor for IPF development (159). Examination of 18- to 24-mo-old mice revealed decreased GLRX activity and enhanced susceptibility to bleomycin-induced lung fibrosis compared with young adult control groups (6). Three splice variants of human GLRX (A–C; see Fig. 4 for EST) exist, and analysis of the gene expression data of all splice variants of GLRX together, in lungs from nondiseased humans, as a function of age, revealed no difference between age groups (6). However, separate analysis of the individual splice variants of GLRX revealed an increase in expression of GLRX variant A with age, whereas GLRX variant B was decreased in lung tissues as a function of age, and expression GLRX splice variant C did not change with age (Fig. 4A). In contrast to these divergent patterns of expression of GLRX with age, GLRX2, GLRX3, and GLRX5 showed weak but statistically significant trends toward decreasing mRNA expression in lungs from older age groups (Fig. 4B). Mitochondrial dysfunction has been well linked to aging and pulmonary fibrosis, where decreased expression of PTEN-induced kinase-1 leads to dysfunction in electron transport chain, reduced mitophagy, and increased apoptosis (170). Although the glutathionylated proteome of IPF lungs and epithelial cells remains unknown, mitochondrial complex I and UCP2/3 have been demonstrated to be regulated by S-glutathionylation (154, 187) (Fig. 5). Thus, the reduced expression of GLRX2 in older age groups may increase S-glutathionylation of mitochondrial complex I and/or UCP2/3 and potentially as of yet-unknown mitochondrial proteins and contribute to mitochondrial dysfunction. Examination of GLRX expression in lung tissues from patients with IPF compared with controls revealed that GLRX2 and GLRX3 mRNA increased in patients with IPF compared with the nondisease control group, in contrast to the previously detected decreases in GLRX, whereas GLRX5 mRNA decreased slightly (Fig. 4C). All together, reduced overall deglutathionylation capacity caused by reduced expression of glutaredoxins with age might lower the “threshold” level of protein S-glutathionylation, rendering older age groups with increased susceptibility to the onset of pulmonary fibrosis. Increases in GLRX2 and GLRX3 mRNA in lungs from patients with IPF may reflect compensatory, adaptive responses that are nonetheless insufficient to protect against the development and/or progression of disease.

Fig. 4.

Glutaredoxins transcript levels in lung tissue from patients with lung diseases and different age groups. A: glutaredoxin-1 (GLRX) isoform-specific expression in the lung as a function of age. Splice variant A: ENST00000379979.4; Splice variant B: ENST00000237858.6; splice variant C: ENST00000512469.2. Expression of splice variants A and B demonstrates significant correlation with age, with no significant change occurring for splice variant C (n = 25, 20–29; n = 21, 30–39; n = 60, 40–49; n = 112, 50–59; n = 97, 60–69; n = 5, 70–79). Statistical significance was calculated using Welch’s t test comparing age bins 20–39, 40–59, and 60–79. *P < 0.05; ** P < 0.01. B: GLRX2, GLRX3, and GLRX5 transcripts evaluated as a function of age in nondiseased human lungs. Spearman correlation analysis revealed a significant negative correlation between age and expression of GLRX2, GLRX3, and GLRX5, with ρ values indicated in the figure. Data represent 342 normal lung specimens, 236 derived from male donors aged 21 to 70 yr (black dots), and 106 derived from female donors aged 21 to 70 yr (red dots). C: interstitial lung disease (ILD) and chronic obstructive pulmonary disease (COPD). GLRX transcripts were assessed using publicly available microarray gene expression data from the Lung Genomics Research Consortium (LGRC; http://www.lung-genomics.org) for ILD patients with a diagnosis of idiopathic pulmonary fibrosis (I; n = 255), COPD (CO; n = 219), and nondiseased control tissues (C; n = 137) (data are available in the Gene Expression Omnibus database accession no. GSE47460). [GLRX transcripts were previously reported (6) and are reformatted here for consistency]. Asthma: GLRX transcripts were assessed from publicly available RNA-seq data from RNA isolated from nasal brushing of asthmatic patients (As; n = 53) compared with healthy subjects (C; n = 97). [The full description of RNA isolation and processing, and processing of the raw RNA-seq data are published elsewhere (181)]. Results are shown as average ± SE. Statistical significance was calculated using a Mann-Whitney test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. A and B: RNA-seq data are from the GTExPortal v.6 (A) or v.7 (B), were analyzed (accession no. phs000424.v7.p2, 06/2018) after exclusion of patients with lung diseases (58). The Genotype-Tissue Expression (GTEx) Project is supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by the National Cancer Institute, National Human Genome Research Institute, National Heart, Lung, and Blood Institute, National Institute on Drug Abuse, National Institute of Mental Health, and National Institute of Neurological Disorders and Stroke.

ASTHMA

Asthma is a complex lung disorder characterized by chronic airway inflammation and remodeling (Fig. 5). The diagnosis of asthma requires symptoms including wheeze, shortness of breath, chest tightness, cough, together with reversible airflow obstruction and/or airway hyperresponsiveness (95). Risk factors include environmental allergens such as dust mites, viral infection, air pollution, tobacco smoke, and obesity. Type 2 inflammation is found in 50% of adult with asthma, where T_H2 or innate lymphoid cells (ILCs) secrete cytokines such as IL-5 and IL-13 and is closely associated with eosinophilia (182). Separately, a subgroup of asthma patients is known to have high levels of IL17 (T_H17-high asthma) in association with neutrophilic inflammation, steroid resistance and hard-to-control severe disease (91, 106). Structurally, the airways of people with severe asthma undergo substantial remodeling, marked by increased airway smooth muscle mass, thickening of the basement membrane, and subepithelial fibrosis (95). Stiff mucus plugs, described below, are also found in airways of severely asthmatic people, contributing to chronic airflow obstruction (55). Epithelial cells play important roles in shaping the immune response (85). Epithelial cells express pattern recognition receptors that, upon ligand binding, release proinflammatory mediators that recruit, activate, and/or promote differentiation and survival of immune cells, including dendritic cells, T cells, neutrophils, and eosinophils. Epithelial cell-derived IL-25, IL-33, and thymic stromal lymphopoietin (TSLP), the “alarmins”, induce type 2 responses that, among other processes, enhance release of IL-13, thus promoting goblet cell hyperplasia and mucus hypersecretion (85, 91).

A large number of studies have been conducted to examine changes in the GSH redox status in patients with asthma by examining total glutathione, GSH, and GSSG in lung lining fluid, bronchoalveolar lavage (BAL), induced sputum, and exhaled breath condensates or systemically via the analysis of glutathione in whole blood, plasma, or erythrocytes. These studies were conducted in children as well as adults with varying degrees of asthma, in some cases after challenge with allergen or ozone or after exacerbations and used different methods to collect the clinical sample and/or to detect glutathione. It is therefore not surprising that reported values vary tremendously, with variable changes reported, resulting in an incomplete picture of how the GSH/GSSG redox couple is affected in asthma and how this relates to the endotypes of asthma, age, and/or medication use (63). Intriguingly, corticosteroids, the mainstay therapy for managing inflammation in asthmatic patients, decrease levels of glutathione (87, 191). Therefore, prior medication use, or the aforementioned concerns about the requirements to chemically protect thiols in order to prevent oxidation of GSH and other oxidation reactions, might be a factor that contributed to these variable observations.

In settings where total levels of GSH were shown to be increased in asthmatic patients, this has been interpreted as being indicative of a response to increased oxidative stress in asthmatic patients (207). Indeed, a number of enzymes critical in the synthesis and metabolism of GSH are regulated by the oxidative stress-responsive transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2, encoded by the NFE2L2 gene), which in murine models of asthma has been strongly linked to protection against disease (137, 204, 242). One study of patients with severe asthma demonstrated that, despite increases in NFE2L2 mRNA and NRF2 protein, NRF2 activation and downstream targets of NRF2, including glutathione-dependent enzymes, were not different between asthmatic and control groups, in association with decreases in levels of GSH and cysteine (64). Dysregulation of the Nrf2 pathway in severe asthma may be linked to decreases in GSH levels in those patients and may explain the aforementioned fluctuations in GSH between the different studies (63), although this scenario requires further investigation. Additional studies in mouse models of allergic airway disease have corroborated a role for glutathione redox homeostasis. Notably, mice lacking Ggt, one of the enzymes that degrades GSH, as described earlier, have increases in GSH levels in the lung lining fluid and decreases in intracellular GSH. GGT-deficient mice showed attenuated allergic airway disease in response to IL-13 (148), and similar results were reported using a GGT inhibitor (254). Furthermore, depletion of glutathione with buthioneine sulfoximine (BSO) worsened allergen-induced airway reactivity and inflammation, while conversely, NAC dampened the responses to allergens in association with modulation of glutathione pools and markers of oxidative stress (174). Glutathione S-transferases have also been strongly implicated in the pathogenesis of asthma, and polymorphic variants of a number of GSTs have been reported in patients with asthma, with variable correlations to disease severity (49, 193). Similarly, glutathione peroxidases also have been shown to be affected in patients with asthma or rodent models of disease (15, 46, 47, 63).

A role for PSSG chemistry and glutaredoxins in settings of asthma is currently emerging. Analysis of sputum samples from asthmatic patients revealed an increased level of GLRX and a decrease in PSSG compared with controls. A concomitant increase in GLRX mRNA was detected in bronchial epithelial cells isolated from asthmatic patients compared with controls (128). In contrast to those observations, RNA isolated from nasal brushing of asthmatic patients showed a decrease in mRNA expression of all GLRXs (Fig. 4C) (181), perhaps reflecting the different sites of sampling. Based on these limited observations, it remains unclear to date how GLRX status relates to asthma endotypes. Studies using mouse models of allergic airway disease show that Glrx mRNA and protein levels in lung tissues were increased in response to exposure to ovalbumin (Ova) or house dust mites (HDM), whereas in response to Ova, Glrx2 mRNA remained unchanged. Despite increases in GLRX in these settings, PSSG was also increased, showing that the increases in GLRX were not sufficient to normalize the extent of PSSG within the time frame of investigation (93, 94, 209). A kinetic assessment of GLRX, GSH, and PSSG in airways and lung tissues from mice subjected to the Ova model showed that GLRX protein increased transiently in BAL fluid 6 h after the final Ova challenge and decreased in a time-dependent manner while remaining significantly higher than baseline level 8 days post-challenge. Coinciding with this were increases in PSSG within the airway epithelium and increases in GSH levels in the BAL, collectively suggesting that the GLRX and PSSG status changes rapidly in airways in response to challenge with antigens in a sensitized host. The increases in GLRX levels in the BAL were found to positively correlate with increases in cytokines, suggestive of a link between GLRX and airway inflammation. In contrast to the acute increases in BAL, increases in GLRX in lung tissues persisted (94, 155).

To directly address a causal link between GLRX and allergic airway disease, we have evaluated mice that globally lack Glrx in the Ova and HDM models of disease. Glrx^−/− mice on a Balb/C background initially developed inflammation, mucus metaplasia, and increases in airway hyperresponsiveness in response to Ova, similar to WT littermates. However, Glrx^−/− mice resolved the disease faster compared with the respective WT groups (94), suggesting that the GLRX/PSSG axis regulates the duration/resolution of allergic airway disease. As HDM is a more relevant antigen than Ova (a large proportion of asthmatic patients is allergic to HDM), using a HDM model of airway sensitization and challenge, our laboratories demonstrated a change in patterns of inflammatory profiles and cytokine secretion in Balb/C Glrx^−/− mice in response to HDM, with a decrease of airway eosinophils and Th2 cytokines, HDM-specific IgE, an increase in airway macrophages and neutrophils, along with an increase in T_H17 cytokine IL-17, compared with respective WT groups. Re-stimulation of single-cell suspensions, derived from the lung or spleen, with HDM led to an increase in IL-17 production and a decrease in IL-5 in Glrx^−/− mice, consistent with a shift in T cell polarization (93, 94). In addition to these observations, Glrx^−/− mice displayed altered profiles of airway hyperresponsiveness, with increases in central airway resistance, yet dampened peripheral elastance in response to HDM, compared with littermates. These findings indicate that GLRX and PSSG affect T cell polarization and patterns of airway hyperresponsiveness.

The NF-κB pathway plays a central role in immune response and asthma, with downstream target genes including cytokines, chemokines, apoptosis regulators, and inflammatory mediators (282). LPS is a key activator of the NF-κB pathway via activation of the Toll-like receptor 4 (TLR4) pathway. Importantly, TLR4 activation on airway epithelial cells is also a critical step in the initiation of allergic airway disease induced by HDM (84). Work from our laboratory and that of others has demonstrated the importance of epithelial NF-κB in the pathogenesis of allergic disease models of asthma (220, 252). In the canonical pathway, activation of inhibitory κB kinase (IKK), the holoenzyme consisting of IKKα, IKKβ, and NF-κB essential modulator (NEMO), leads to phosphorylation and polyubiquitination of IκB subunit, its degradation, and the release of p50/RelA for nuclear translocation, whereas in the noncanonical pathway IKKα and p52/RelB dictate the transcriptional output, and both of these pathways are engaged in response to LPS stimulation of lung epithelial cells (253). PSSG/glutaredoxin have been shown to regulate the NF-κB pathway at multiple levels. We have demonstrated that glutathionylation of IKKβ Cys¹⁷⁹ inhibited its kinase activity, resulting in reduced RelA nuclear translocation and decreased levels of macrophage inflammatory protein 2 (MIP-2) and keratinocyte-derived chemokine (KC) in epithelial cells with LPS stimulation (208). Conversely, NF-κB activation led to transcriptional activation of GLRX, which is important in enhancing activation of NF-κB and LPS-induced proinflammatory responses in association with decreases in S-glutathionylation of IKKβ (4). We also demonstrated that GSTP constitutively interacted with the inhibitor of NF-κB, IκBα, in control cells and that after LPS stimulation this interaction was lost, whereas at later time points, when NF-κB activation decreased, GSTP showed an increased interaction with IKKβ in association with its S-glutathionylation (112) (Fig. 5). This scenario is suggestive of a multifaceted role of GSTP in the regulation of NF-κB, prevention of activation of NF-κB in the absence of stimulus through protein-protein interaction, and regulatory mechanism involving GSTP-mediated S-glutathionylation to shut down NF-κB activation. In a model of LPS-induced lung inflammation, Glrx^−/− mice had attenuated secretion of IL-1β, IL-6, granulocyte-macrophage colony-stimulating factor (GM-CSF), and TNF-α but increased secretion of KC (3). Although the composition of the inflammatory profile in Glrx^−/− mice remained the same, alveolar macrophages lacking Glrx appeared smaller than their wild-type counterparts. LPS stimulation of alveolar macrophages isolated from Glrx^−/− mice had attenuated NF-κB activation, as demonstrated by decreased levels of RelB, IκBα, and IKKα, decreased RelA nuclear translocation, and decreased secretion of TNF-α and IL-6 (3).

Given the increases in IL-17 in lungs from HDM-exposed Glrx^−/− mice, and the importance of IL-17 in the regulation of airway barrier function and remodeling, our laboratories assessed whether S-glutathionylation affected the responses of epithelial cells to IL-17 stimulation. Exposure of epithelial cells to IL-17 induced IKKα-mediated NF-κB activation, and glutathionylation of RelA and IKKα were observed. Epithelial cells lacking Glrx showed enhanced glutathionylation of RelA and IKKα, as well as decreased expression of the proinflammatory mediators KC and C-C ligand 20 (CCL20) mRNA, although, paradoxically, an increase in expression of IL-6. Absence of Glrx also led to a more robust increase in A20, a negative regulator of the NF-κB pathway, suggesting that S-glutathionylation contributed to cessation of some components of NF-κB signaling in IL-17-stimulated epithelial cells (176).

PSSG/glutaredoxins have also been implicated in the regulation of other adaptor proteins and downstream transcription factors that are relevant in allergen-induced immune responses. Notably, TNF receptor-associated factors (TRAFs) have been shown to be glutathionylated (40, 75). TRAF6 is an adaptor protein critical in transducing signals by members of the IL-1 receptor/TLR (IL-1R/TLR) family. TRAF6 was shown to be constitutively glutathionylated in various cell types and was deglutathionylated following stimulation with IL-1 (40). GLRX knockdown attenuated IL-1R1/TLR4-induced NF-kB activation, as TRAF6-SSG inhibited its polyubiquitination and failed to recruit and activate the transforming growth factor-β1 activation kinase-1 (TAK1) complex (40). In macrophages, herpes simplex virus infection led to ROS-dependent S-glutathionylation of TRAF3 (Cys⁴⁵⁵) and TRAF6 (Cys³⁹⁰), critical to the activation of NF-κB and secretion of IFN-α/β). Mutation of these cysteines, or overexpression of GLRX, decreased these herpes simplex virus-induced responses, suggesting an activating role of TRAF3/TRAF6 glutathionylation in the innate immune responses (75). Interestingly expression of a TRAF6 Cys³⁹⁰ mutant did not affect the ability of IL-1β to activate NF-kB (75), in contrast to the aforementioned inhibitory role of TRAF6 S-glutathionylation in IL-1β signaling (40), perhaps reflecting differences in cell types or sites of S-glutathionylation between these two studies. Furthermore, the adaptor protein MyD88 adaptor-like (MAL or TIRAP), important in TLR signaling, was also shown be glutathionylated, at Cys⁹¹, in response to LPS treatment of macrophages. MAL mutations (C91A and H92P), which are resistant to glutathionylation, resulted in dampened recruitment of and interaction with MyD88 and displayed decreased NF-κB activation (104), consistent with an activating role of Cys⁹¹ S-glutathionylation in proinflammatory signaling. All together, these findings point to an activating role of TRAF/MAL S-glutathionylation in TLR/IL-1-induced proinflammatory signaling (Fig. 5).

Dimethyl fumarate (DMF) is a compound with electrophilic properties that has garnered increasing attention after its FDA approval for the treatment of patients with multiple sclerosis (23). Relevant to the scope is this paper are findings demonstrating that DMF dampens HDM-induced allergic airway inflammation (107). DMF also has been anecdotally reported to reduce asthma symptoms and to improve quality of life of asthma patients (221). DMF decreased TNF-α-mediated NF-κB activation and cytokine secretion in primary human airway smooth muscle cells (221). Furthermore, in the same model, DMF reduced intracellular GSH and induced IκBα glutathionylation (IκBα-SSG), which in turn was associated with inhibition of IκBα degradation, NF-κB p65 nuclear entry and NF-κB/DNA binding, and attenuated cytokine responses induced by TNF-α (222).

As stated earlier, IL-33 is a key alarmin that is rapidly secreted from airway epithelial cells stimulated with a range of allergens or other insults. IL-33 contributes to type 2 inflammatory responses in asthma. It has been demonstrated that IL-33 secretion from airway epithelial cells occurs via a redox-based mechanism that requires the H₂O₂-producing enzyme dual oxidase-1 (DUOX) (103). A role of GLRX and PSSG in the response to IL-33 has become apparent recently in studies conducted in macrophages. Although in WT macrophages LPS led to increases in mRNA expression of IL33, this response was diminished in macrophages lacking GLRX, in association with enhanced S-glutathionylation of TRAF6, inhibiting downstream IKKβ and NF-κB activation, events required for the upregulation of IL33. Intriguingly, exposure of macrophages to recombinant IL-33 also induced IL33 mRNA expression, demonstrative of feed-forward signaling, which is also inhibited by GLRX knockdown. The importance of GLRX in IL-33 signaling was further corroborated using an intratracheal cockroach antigen model of allergic airway disease, wherein increases in GLRX protein were observed, analogous to previously described studies involving Ova or HDM. Intriguingly, GLRX induction in response to cockroach was attenuated in lungs from mice lacking IL33, demonstrating a requirement not only for GLRX in the induction of IL-33 but also a role for IL-33 signaling in the induction of GLRX in settings of allergen exposure (261), findings that collectively suggest a role for GLRX and PSSG in the regulation of type 2 responses in asthma through the regulation of the key alarmin IL-33.

Beside the aforementioned targets of S-glutathionylation important in the regulation of proinflammatory signals, asthma-relevant cytokines themselves, notably IL-1β, IL-6, and IL-17 are direct targets of S-glutathionylation (283). In particular, the activity of IL-1β was demonstrated to be inhibited following overoxidation of Cys¹⁸⁸ to sulfinic and sulfonic acids. Importantly, S-glutathionylation of Cys¹⁸⁸ protected IL-1β from oxidative inhibition and contributed to sustain the biological activity of IL-1β in vivo. Furthermore, GLRX was shown to be present and active in the extracellular space and to physiologically regulate IL-1β glutathionylation (283) (Fig. 5).

CHRONIC OBSTRUCTIVE PULMONARY DISEASE

Chronic obstructive pulmonary disease (COPD) is a progressive lung disease characterized by poorly reversible airway obstruction, in association with obstruction of the small airways (linked to mucus hypersecretion and fibrosis around the small airways), and emphysema, exemplified by a loss of alveolar integrity and airspace enlargement (14, 197) (Fig. 5). The global prevalence of COPD is estimated to be around 174 million, with ~3 million deaths attributed to it (197). COPD is currently the third leading cause of death worldwide (14). Inhalation of particulate matter and pollutants has been closely linked to the pathogenesis of COPD, with cigarette smoking the leading risk factor (14, 171, 197). However, additional factors are believed to contribute the risk of COPD development, including genetic factors. As described above in familial IPF, COPD tends to cluster in families, suggesting that there is significant heritability of COPD (14). The mechanisms that drive COPD remain incompletely understood, but the disease is associated with chronic inflammation of the peripheral airways and parenchyma, characterized by increases in diverse cell types including neutrophils, macrophages, and lymphocytes. Chronic exposure to inhaled pollutants leads to activation of pattern recognition receptors, triggering an innate immune response and activation of epithelial cells that are important in mucus hypersecretion. An adaptive immune response characterized by an increased presence of T_H1, T_H17, and CD8⁺ cytotoxic T cells, has also been described in patients with COPD. COPD-associated inflammation is frequently resistant to treatment with corticosteroids. Patients with COPD can have acute exacerbations triggered by viral or bacterial infections, and frequent exacerbations are linked to a poor prognosis. COPD airways have been shown to be colonized by Hemophilus influenzae and Streptococcus pneumoniae (14). Like IPF, emphysema is believed to be caused by accelerated aging of the lung parenchyma owing to the decreased function of sirtuins (201, 272), enhanced telomere shortening (172), increased cellular senescence (14, 251), stem cell exhaustion, and decreases in autophagy (198). Proteases derived from neutrophils, macrophages, and epithelial cells, including neutrophil elastase, matrix metalloproteinase-9 (MMP9), and MMP12 have been linked to mucus hypersecretion, extracellular matrix degradation, impaired bacterial clearance, enhanced bacterial exacerbations, and chronic inflammation (92, 171, 257). Neutrophil elastase is a major factor implicated in the pathogenesis of COPD. Administration of elastase into airways of rodents is sufficient to induce emphysema (237). α1-Antitrypsin (A1AT) is a major circulating inhibitor of neutrophil elastase (223), and importantly, humans with a genetic deficiency of A1AT are prone to the development of emphysema (50).

Oxidative stress has been considered a major driving mechanism in the development of COPD (14). This notion does not come as a surprise, as inhaled pollutants such as cigarette smoke contain a myriad of oxidants (56). Subsequent activation of epithelial cells and innate immune cells by inhaled pollutants triggers activation of NADPH oxidases, further contributing to a redox imbalance. Numerous studies have evaluated GSH in chronic smokers and COPD patients. Cigarette smoke directly reacts with GSH, leading to GSH depletion and adduct formation, notably with acrolein and crotonaldehyde (205). Although GSH levels initially decrease in response to cigarette smoke, GSH levels subsequently rebound leading to elevated levels in the lung lining fluid (76). Levels of GSH were shown to be increased in the epithelial lining fluid and BAL fluid from chronic smokers compared with nonsmokers (173), and this increase in GSH was correlated to an increased neutrophil count and increased myeloperoxidase activity (143). GSH levels were also increased in blood and sputum of smokers and COPD patients (20). These increases in GSH reflect adaptive responses, involving NRF2, to combat the disrupted redox environment (77), and, consistent with this notion, mice lacking NRF2 are more prone to the development of emphysema (203). As was stated earlier, COPD is a disease associated with aging. Of interest are findings that older individuals were shown to be deficient in the adaptive GSH response after exposure to cigarette smoke. More importantly, in a follow-up in vivo study by the same group, depletion of GSH in young mice to levels of GSH found in aged individuals led to an accelerated enlargement of airspace after exposure to cigarette smoke, along with increases in PSSG, inflammatory mediators, and elastase and MMP activities (77), strongly suggesting that defects in GSH synthesis or metabolism or increases in S-glutathionylation contribute to the development of emphysema. Indeed, it has been suggested that differences in lung glutathione metabolism that occur between rodent species may account for the variations in rodent susceptibility in elastase-induced emphysema development (24).

These and other observations have prompted a large number of clinical trials since the 1970s involving the small thiol compound and GSH precursor, NAC as potential therapeutic for COPD, reporting varying outcomes (22, 78, 250, 287, 289). However, meta-analysis of clinical studies involving NAC, and the related thiol compounds erdosteine and carbocysteine in COPD patients, illuminated a therapeutic effect of these compounds on acute exacerbations in COPD patients (38, 213, 224). It is of relevance to note that these compounds were evaluated with a putative mechanism of action as a mucolytic. Indeed, oxidation of cysteine thiol groups in mucin polymers contributes to mucus gel stiffening in the lung (277), and small thiol reductants, such as NAC, at high concentrations have the potential to reduce mucins, leading to their increased solubility. Tenacious mucus, found in plugged airways of people with asthma, has been linked to severe disease in association with formation of oxidants derived from eosinophil peroxidase, leading to persistent crosslinks (55), unlikely to be resolved with the doses of NAC, erdosteine, or carbocysteine that are clinically achievable. It remains unclear whether mucins are targets for S-glutathionylation and whether this impacts its biophysical/biological properties.

The importance of GLRX and PSSG in the pathogenesis of COPD is only now emerging. Slight decreases in GLRX protein were observed in lung lysates from COPD patients. GLRX was found to be predominantly expressed in macrophages and was decreased in patients with severe COPD compared with controls. Interestingly, the percentage of GLRX-positive macrophages directly correlated with lung function. GLRX protein was also detected in sputum supernatants, and levels were increased in COPD patients undergoing acute exacerbations, compared with nonsmoker controls (190). On the basis of these findings, it is tempting to speculate that GLRX expression in macrophages exerts a protective function and that its secretion may promote disease exacerbation. Assessment of transcript levels of GLRX mRNAs in lungs from patients with COPD showed that all GLRX mRNAs remained unchanged (Fig. 4C), pointing to a need for single-cell transcriptomics to evaluate GLRX in macrophages and other cell types from patients with COPD. In addition to the putative role of GLRX/GSSG in macrophages, GLRX/PSSG play a role in epithelial cell death in response to cigarette smoke. Exposure of lung epithelial cells to cigarette smoke extract led to a decrease in GLRX mRNA and protein levels and an increase in PSSG and cell death. In addition, cigarette smoke itself, or acrolein, an electrophile present in cigarette smoke, irreversibly oxidized GLRX, with a concomitant decrease in activity, showing that cigarette smoke affects GLRX at multiple levels (Fig. 5). Overexpression of GLRX afforded protection from cigarette-smoke-induced cell death while conversely, absence of GLRX enhanced the susceptibility of epithelial cells to cell death, in association with respective decreases and increases in S-glutathionylation (127). Paradoxically, cigarette smoke exposure in mice led to decreases in PSSG in lung tissues compared with the controls, in association with a decrease in free thiols and increases in carbonylation, suggestive of a decrease in reversible oxidation induced by cigarette smoke (126). A subsequent study by the same group reported increases in PSSG in the BAL fluid or lavaged cells (125), indicating discordant regulation of PSSG in different regions in the lung or airways after exposure to cigarette smoke. Increases in PSSG in BAL fluid induced by cigarette smoke were enhanced in Glrx^−/− mice compared with WT animals, in association with altered numbers of inflammatory cells (increases in macrophages and decreases in neutrophils, lymphocytes, and dendritic cells) and decreases in proinflammatory cytokines (125). Macrophages and tracheal epithelial cells isolated from Glrx^−/− mice exposed to cigarette smoke extract also had attenuated secretion of KC (125). In a separate study, GLRX protein expression in lung tissue was shown to be decreased in response to cigarette smoke exposure. Those authors also reported an increase in airway neutrophils and proinflammatory cytokines in Glrx^−/− mice exposed to cigarette smoke (45), in contrast to the previous study (125), possibly due to the different time of evaluation post-cigarette smoke exposure. As was described above, IKKs are key regulatory kinases in the NF-κB pathway important for proinflammatory signaling, and S-glutathionylation of both IKKα and IKKβ have been reported (176, 208). Intriguingly, in Glrx^−/− mice exposed to cigarette smoke, NF-κB activity, evaluated via phosphorylation and acetylation of RelA/p65, was enhanced compared with WT mice, in association with enhanced expression of IKKα and decreases in expression of IKKβ. Although S-glutathionylation of both IKKα and IKKβ were observed in these settings, phosphorylation of IKKα was enhanced (45). These findings point to discordant regulation of IKKs via the GLRX/PSSG axis, although additional studies will be required to elucidate the relative importance of S-glutathionylation of these kinases in regulating the proinflammatory responses of cigarette smoke (Fig. 5).

The mechanistic details whereby PSSG/GLRX regulate the pathophysiology of COPD remain unclear, and S-glutathionylation targets remain to the explored. As was mentioned above, A1AT deficiency leads to emphysema. Notably, A1AT has been shown to be prone to oxidation (249), and a number of oxidative modifications of A1AT have been shown, including S-glutathionylation (79), sulfenic acid formation (80) and nitrosylation of Cys²³² (167), and oxidation of Met^351/358 (36, 243) in association with alterations in or loss of protease inhibitor function. Consequently oxidation-resistant versions of A1AT have been created (225, 288) with the goal to provide improved therapeutic responses in patients with A1AT deficiency who receive weekly A1AT injections. A1AT can be taken up from the circulation and can protect against endothelial cell apoptosis, in association with binding to cysteine-dependent proteases, notably caspases (223). It remains unclear whether this process is affected via S-glutathionylation.

In addition to the aforementioned role of S-glutathionylation in promoting cigarette smoke-induced death of epithelial cells and affecting macrophage activation, other cell types are also affected via S-glutathionylation. Notably, PSSG and glutaredoxin regulate the biology of neutrophils, including their chemotactic migration, adhesion, and phagocytosis and bacterial clearance capacity. Glutathionylation/deglutathionylation of actin and L-plastin, a β-actin-bundling protein, regulate actin polymerization in neutrophils (54, 216). Specifically, actin and L-plastin S-glutathionylation inhibited actin polymerization and led to defects in pseudopod formation which in turn impaired migration toward chemoattractant and decreased phagocytic activity (54, 216). Glrx-deficient murine neutrophils showed impaired recruitment to sites of inflammation and reduced bactericidal capability (216). Neutrophil extracellular traps (NETs) are formed upon their degranulation and consist of granule proteins and DNA, forming net-like configurations that are critical in trapping and killing of bacteria. It has been demonstrated that NET formation was increased in patients with severe COPD in association with enhanced innate immune responses, more frequent exacerbations, and a loss of microbiota diversity (51, 264). Interestingly, due to defects in actin polymerization associated with increases in actin S-glutathionylation, neutrophils lacking Glrx also fail to degranulate or release DNA, leading to defects in NET formation. Actin Cys³⁷⁴ was identified as the target of actin glutathionylation, actin polymerization, and NET formation. Based on these findings, GLRX has been suggested to be “a master regulator of actin cytoskeleton rearrangement and DNA release” (233). Respiratory pathogens, including those that colonize the COPD airways, have been shown to evade NETs, and various mechanisms of NET evasion have emerged (234). Intriguingly expression of bifunctional peroxiredoxin-glutaredoxin, and catalase in nontypeable Hemophilus influenzae (NTHI) 86-028NP promote its resistance to oxidants, increase survival within neutrophil extracellular traps, and are determinants of bacterial persistence during chronic/recurrent NTHI infections (113). Collectively, these findings demonstrate not only the importance of PSSG/GLRX in the regulation of NET formation but also the importance of similar redox-based processes within microbes in regulating their persistence and resistance to antibiotics (Fig. 5).

Besides the regulation of the actin cytoskeleton, PSSG also regulates neutrophil migration through the glutathionylation of S100A9 (also known as MRP14) and α4-integrin (142, 275). S100A9 was shown to be glutathionylated in response to phorbol myristate acetate (PMA) or ionomycin stimulation. The S100A8/A9-SSG complex decreased neutrophil adhesion to fibronectin (142), with potential implications for neutrophil migration. Neutrophil α4-integrin plays an integral role in retention and release of neutrophils from bone marrow via binding to the ligand vascular cell adhesion molecule 1 (VCAM-1). Glutathionylation of α4-integrin increases the binding of neutrophil-associated α4-integrin to VCAM-1 on human endothelial cells. Furthermore, in an E. coli-induced acute peritonitis model, absence of GLRX dramatically elevated α4-glutathionylation and subsequently enhanced neutrophil egress from the bone marrow. Conversely, intravenous administration of recombinant GLRX inhibited α4-integrin glutathionylation and resulted in decreases of blood neutrophils (275).

SUMMARY AND FUTURE DIRECTIONS

Oxidative stress has been linked to myriad diseases, including chronic lung diseases. However, clinical trials using generic antioxidants, like NAC, as potential therapeutics have produced no or modest effects and have not resulted in a widely accepted FDA-approved treatment regimen for patients with chronic lung diseases. These failures occur in the face of a plethora of studies convincingly demonstrating the functional relevance of redox-based pathways in essentially every biological process. It is now clear that oxidants are not merely damaging molecules and in fact evoke orchestrated biological responses, in part due to targeted oxidation of protein cysteine residues, notably protein cysteine S-glutathionylation, reviewed herein. It is therefore conceivable that redox-based strategies have the potential to be powerful in the treatment of chronic lung diseases but have to be redesigned to be more specific, to prevent off-target effects, and to allow beneficial redox-based signals to occur. In this sense, protein S-glutathionylation and deglutathionylation are not dissimilar to phosphorylation and dephosphorylation reactions. Analogous to the specificity of phosphorylation and dephosphorylation reactions, general kinase or phosphatase inhibitors with off-target effects are replaced by designer drugs targeting specific phosphorylation sites in specific protein targets. Similar strategies are envisioned to combat adverse redox-based reactions. The protein S-glutathionylation/GLRX axis presents an ideal area of focus as this mechanism of redox-based signaling has emerged as a critical regulatory system that affects a number of processes relevant to chronic lung diseases (Fig. 5). A number of signaling molecules (TRAF3/6, STAT3, IKKα, IKKβ, IkBa), structural proteins (actin, L-plastin, S100A9, α4 integrin), mitochondrial proteins (UCP2 and -3, mitochondrial complex 1), the death receptor Fas, the protease inhibitor A1AT, and cytokines (IL-1β, IL-6, IL-17) are targets for S-glutathionylation and therefore have the potential to affect cell death, migration, mitochondrial function, release of alarmins, innate and adaptive immune responses, epithelial and immune cell activation, and protease activities. The aforementioned processes accompany IPF, severe asthma, and COPD. Therefore, additional studies to corroborate the exact PSSG targets in these diseases will be critical. This review focused on studies performed predominantly in mouse models and cell lines exposed to disease-relevant stimuli and stresses. The extent to which findings obtained from such studies bear relevance to humans with these disorders will necessitate translational studies to validate these PSSG targets. Indeed, S-glutathionylated proteomes remain to be discovered in specimens from patients, and the targets highlighted herein for the most part remain to be validated in human tissues. Recent advances in detection methods and mass spectrometry quantification techniques, as well as bioinformatic tools for predictions of glutathionylation sites, provide improved opportunities to identify glutathionylated proteins. In addition, the development of single-cell technologies such as single-cell RNA-seq and proteomics could aid in deciphering the roles that glutathione and glutaredoxins play in cell and tissue homeostasis. Most studies described herein have used mouse models that globally lack the Glrx gene. A recently developed conditional Glrx-deficient mouse model (unpublished observations) will enable the elucidation of the role of GLRX and indirectly, PSSG, within specific cell types, and these approaches will be informative toward an improved understanding of PSSG/GLRX in chronic lung diseases. Although GLRX2, GLRX3, and GLRX5 expressions are modified in lungs from patients with IPF and nasal brush obtained cells from asthmatics, and their expressions decrease in lung with aging, a setting where COPD and IPF predominate, GLRX2, 3 and 5 remain unstudied in these diseases to date. Therefore, investigation of these GLRXs, along with other redox enzymes described briefly, will be critical to gain an improved understanding about the processes that govern S-glutathionylation. Similarly, an improved understanding about the fate of GSH itself (for example its transport into the airways, degradation, formation of GSNO and GSSG), in settings of chronic diseases also will be critical, as GSH status governs PSSG chemistry and GLRX activity. Despite these current limitations, findings of diminished activity of GLRX in lungs from patients with IPF, coupled to the reversal of fibrosis in diverse mouse models following direct administration of GLRX into the lungs (6), offer an exciting platform for elucidation of mechanisms of action whereby GLRX exerts its protective function and offer an exciting opportunity for development of new therapeutics. Targeting redox-based chemistry in chronic lung disease may, in fact, be a treasure trove for drug development.

DISCLOSURES

Y. M. W. J-Heininger and V. Anathy hold patents: United States Patent No. 8,679,811: “Treatments Involving Glutaredoxins and Similar Agents” and United States Patent No. 8,877,447: “Detection of Glutathionylated Proteins”. Y. M. W. J-Heininger and V. Anathy are scientific founders of Celdara Medical LLC and have received consulting fees from Celdara Medical LLC. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

GRANTS

This work was supported by the National Institutes of Health Grant Nos. R35HL135828 (to Y. M. W. Janssen-Heininger); R01HL137268, R01HL085646, R01HL138708, and R21AG055325 (to A. van der Vliet); R01HL122383, R01HL141364, and RO1HL136917 (to V. Anathy); F31 HL144051 (to E. A. Elko); and HL076122 (to A. M. Manuel).

AUTHOR CONTRIBUTIONS

S.B.C., E.A.E., and D.J.S. prepared figures; S.B.C. drafted manuscript; S.B.C., E.A.E., R.A., A.M.M., C.v.d.W., J.E.D., J.v.d.V., D.J.S., V.A., C.G.I., Y.-W.L., A.v.d.V., and Y.M.W.J.-H. edited and revised manuscript; S.B.C., E.A.E., R.A., A.M.M., C.v.d.W., J.E.D., J.v.d.V., D.J.S., V.A., C.G.I., Y.-W.L., A.v.d.V., and Y.M.W.J.-H. approved final version of manuscript.