Hydrogen peroxide as a central redox signaling molecule in physiological oxidative stress: Oxidative eustress^☆

3.1. Steady-state of catalase Compound I

Hydrogen peroxide is now well-known as a normal metabolite of oxygen in aerobic metabolism of cells and tissues. As mentioned, this was first shown in 1970 by the detection, in the near-infrared, of catalase Compound I in the intact perfused liver [7]. Catalase Compound I is the product of the reaction of H₂O₂ with catalase hame, denoted here in a simplified manner (Reaction (1)) [17], [18]. It can decompose upon reaction with a second H₂O₂ molecule in the catalatic reaction, releasing oxygen (Reaction (2)), or it can decompose in the peroxidatic reaction with a hydrogen donor AH₂, e.g. methanol (Reaction 3):

$\mathrm{In}\text{sum}:{\mathrm{H}}_2{\mathrm{O}}_2{+\text{AH}}_2\to A+2{\mathrm{H}}_2\mathrm{O}$

The charge-transfer band of catalase Compound I at 660 nm [17] is useful for spectrophotometry of cells or organs because of low light-scattering and practically no optical interference by other pigments in the near-infrared region of the spectrum. Unequivocal proof of the existence of H₂O₂ in normally respiring liver cells was provided by noninvasive organ spectroscopy: Catalase Compound I was shown to decrease rapidly when the oxygen concentration was lowered and, importantly, when a hydrogen donor (e.g. methanol) for the peroxidatic reaction (Reaction (iii)) of catalase was infused [7]. Using stepwise ‘titration’ with low methanol concentrations, the rate of cellular H₂O₂ production was determined to be 50 nmol H₂O₂ per gram of liver wet weight per min, which corresponds to 2–3% of total oxygen uptake [19]. That rate was increased in several metabolic conditions (Table 1) [19]. The method has also been used with isolated hepatocytes, also shown in Table 1 [20]. The steady-state concentration of H₂O₂ in the intact cell, overall, was calculated to be of the order of 10 nM [19], [21]. Early methods to quantify H₂O₂ formation included assays of the peroxidatic reaction, e.g. the use of [¹⁴C]-methanol, and others [22], [23].

Another method of quantification of extra H₂O₂ production in intact cells is illustrated in Fig. 2. Upon infusion of decanoate as a fatty-acid substrate for H₂O₂ production, there is an increase in the catalase Compound I signal at 660–640 nm, which can be calibrated against that obtained with infusion of urate, a substrate producing known amounts of H₂O₂ by urate oxidase [22], [23]. In isolated hepatocytes, a comparison of endogenous H₂O₂ production rates with the catalatic evolution of O₂ from infused H₂O₂ provided another means to analyse Compound I steady-states [24].

3.2. Genetically encoded fluorescent protein indicators of H₂O₂

The development of the HyPer probe as genetically encoded fluorescent indicator for H₂O₂ in intact cells and tissues was a landmark [25]. This new type of probe provided insight into spatiotemporal organisation [26]. Thus, it became possible to examine the dynamics of H₂O₂ in subcellular compartments [27]. HyPer is also sensitive to pH. Dissection of pH responses from responses in H₂O₂ is possible with the SypHer [28] and SypHer2 probes [29].

Probes such as roGFP2-Orp1 provide another type of tool for assaying H₂O₂, exhibiting kinetics of oxidation and reduction similar to HyPer [30], [31]. Real-time monitoring of basal H₂O₂ levels has been used with peroxiredoxin-based probes [32].

3.3. Exomarkers,’nonredox’ exogenous probes;small molecule fluorescent markers

A variety of other approaches for H₂O₂ detection is available, not to be discussed here in detail (e.g. [33], [34], [35]).

4.1. Sources: NAD(P)H oxidases, mitochondrial respiratory chain, diverse oxidases

Numerous one- or two-electron reduction reactions have been identified as sources of H₂O₂ (Fig. 3). The major enzymatic generators are the NADPH oxidases (Nox) [36], [37] and the mitochondrial respiratory chain [38], [39], [40], as well as a considerable number of oxidases. A total of 31 human cellular hydrogen peroxide generating enzymes has been compiled [41]. The one-electron reduction sources are functionally tightly linked to the superoxide dismutases, SOD1, SOD2, SOD3, with their cytosolic, mitochondrial matrix and extracellular locations, respectively. Other cell compartments also contribute to H₂O₂ production, e.g. the endoplasmic reticulum and the peroxisomes.

4.2. Sinks: catalases, peroxidases, peroxiredoxins

While the catalatic reaction of catalase (Reactions (i) and (ii)) is a dismutation reaction, regenerating one molecule of oxygen, the other sinks of H₂O₂ lead to water by reduction. Certain cysteinyl residues in peroxiredoxins or selenocysteinyl residues in glutathione peroxidases are highly reactive, the second-order rate constants for the reaction with H₂O₂ being of the order of 10⁷ M^-1s^-1 (see [42]). A number of important peroxidases utilize H₂O₂ as substrate, e.g. myeloperoxidase, eosinophil peroxidase, lactoperoxidase.

The relative contributions of catalase and glutathione peroxidase in H₂O₂ removal depend on the site of generation and the enzymatic equipment [43], [44], [45], [46]. Catalase is a predominantly peroxisomal enzyme. Therefore, it is either a localized sink for H₂O₂ produced by peroxisomal oxidases, or peroxisomes can be a sink at the cell-scale level (see comment further below). An overall assay for the rate of removal of extracellular hydrogen peroxide has been developed for cell culture experiments [47].

4.3. H₂O₂ Compartmentation: role of aquaporins as peroxiporins

Aquaporins (AQP) facilitate H₂O₂ to cross membranes [48]. Specific AQPs play a functional role in H₂O₂ translocation, for which they are called peroxiporins [49], [50].

4.4. Spatiotemporal control, H₂O₂ nanodomains

Spatial distribution of H₂O₂ in cells and tissues is not uniform. There are substantial gradients, both from extracellular to intercellular and between subcellular spaces [51], [52], [53]. It appears from a search of the available literature that H₂O₂ concentration in blood plasma is about 1–5 µM [54], which would be more than 100-fold higher than that estimated to occur within cells (see Fig. 4). Interestingly, even within subcellular organelles, there are H₂O₂ gradients·H₂O₂ in the mitochondrial cristae space originates largely from mitochondrial Complex III, whereas mitochondrial Complexes I and II contribute to mitochondrial matrix H₂O₂ [55]. In the cristae subspace “redox nanodomains” have been described, which are induced by and control calcium signaling at the ER-mitochondrial interface [56]·H₂O₂ transients sensitize calcium ion release to maintain calcium oscillations [56].

There are also circadian redox oscillations, and these are likely to comprise oscillations in H₂O₂ concentration. Peroxiredoxins exhibit pronounced circadian rhythmicity [57], [58]. There is reciprocal control of the circadian clock and the cellular redox state (see [59]).

The thioredoxin reductase-1/thioredoxin [60] and glutathione reductase/glutathione systems [61] appear to control the actions of H₂O₂ on master switches such as Nrf2/Keap1. Nuclear and cytosolic peroxiredoxin-1 differentially regulate NF-κB activities, indicating that the balance in subcellular H₂O₂ metabolism provides specificity in redox signaling [62].

For orientation on H₂O₂ concentration ranges, Fig. 4 presents physiological ranges, spanning from normal processes to adaptive ones (stress responses), denoted as “oxidative eustress”. Higher concentrations evoking inflammatory responses and others, ultimately leading to growth arrest and cell death, are denoted as “oxidative distress”. It may be mentioned that it is difficult to conceptually discern between these ranges, because inflammation, for example, certainly includes elements positive for the organism, such as phagocyte activity. So does autophagy [63], [64], mitophagy [65] and cell death in the forms of apoptosis, ferroptosis and necroptosis [66]. Assignment of a given H₂O₂ concentration to the category of eustress or of distress may vary with cell type, with the level of complexity (comparison of isolated cell/tissue and the whole organism), or with the duration of the exposure to H₂O₂. Considering the level of the whole organism, mucosal barrier tissues participate in immune defense against infection. The release of nano- to submicromolar H₂O₂ was shown to disrupt the tyrosine phosphorylation network in several bacterial pathogens as a host-initiated antivirulence strategy [67].

Estimations of kinetic parameters of H₂O₂ interactions have been performed, and models were established to improve our understanding of the complex processes in redox signaling [68], [69], [70], [71].

In Fig. 4, the gradient between extracellular and intracellular H₂O₂ concentrations is indicated, as rough orientation, to be about 100-fold, which is 10-fold higher than previous estimations [5]. Recent calculations based on results obtained with genetically encoded H₂O₂ probes came to even higher values, 200- to 500-fold [72] or 650-fold [52]; these rely on the information from the intracellular subspace where the detector probe is located. It is conceivable, of course, that there are spaces with even zero H₂O₂ concentration, which would make the gradient tend toward infinity.

5.1. Mechanism

Because of its physicochemical properties, H₂O₂ is capable of serving as messenger to carry a redox signal from the site of its generation to a target site. Among the various oxygen metabolites, H₂O₂ is considered most suitable for redox signaling [1], [73]. Redox regulation can take place via control of single enzymatic activity or at the transcriptional level. Hydrogen peroxide modulates the activity of transcription factors: in bacteria (OxyR and PerR) (see [74]), in lower eukaryotes (Yap1, Maf1, Hsf1 and Msn2/4) and in mammalian cells (AP-1, NRF2, CREB, HSF1, HIF-1, TP53, NF-κB, NOTCH, SP1 and SCREB-1) (see [1]).

The major impact on redox regulation occurs via the thiol peroxidases (see Sinks above) [75], which involves, to a large degree, reversible protein cysteine oxidation to the sulfenylated form (Fig. 3) [76], [77], [78]. According to the Second Principle of the Redox Code, “the redox proteome is organized through kinetically controlled sulfur switches linked to NAD and NADP systems” [6], and H₂O₂ plays a central role in it as an oxidant.

The linkage of redox reactions to protein phosphorylation/dephosphorylation is given by the redox sensitivity of proteine tyrosine phosphatases (PTP). Thus, oxidation and thereby inactivation of PTPs would increase steady-state protein phosphorylation. Examples are PTEN (phosphatase and tensin homolog), Cdc25 phosphatases, and PTP1B (protein tyrosine phosphatase 1B) (see [79]).

Protein phosphatase 1 (PP1) is redox-inhibited by oxidation of its metal center [80]. The intracellular labile iron pool is a determinant of H₂O₂-induced redox signaling [81], [82].

5.2. Targets

It follows that specificity and fine-tuning is exerted through control of sources and sinks. Regarding sources, the fine-control of NADPH oxidases by physical and chemical cues is pivotal (see [83]). Mitochondrial control by redox switches coordinates protein activity, localization and stability (see [84], [85]). The cysteine redox proteome has been studied thoroughly [41], [86], [87]. The median percentage oxidation of cysteine residues in the proteome is between 5% and 12%, and this can be increased to >40% by adding oxidants [86]. Regarding sinks, much interest has focused on the peroxiredoxins [88]. For example, peroxiredoxin-2 acts as highly sensitive primary H₂O₂ receptor which specifically transmits oxidative equivalents to the redox-regulated transcription factor STAT3, forming a redox relay for H₂O₂ redox signaling [89].

5.3. Processes

As many fundamental biological processes have a component of redox control, exhaustive coverage is beyond the scope of the present article. Major such processes are hypoxia, inflammation (“inflammasome“), autophagy, apoptosis, wound healing, proliferation, muscle contraction, circadian rhythm, stem cell self-renewal, tumorigenesis and aging (see above and Refs. [8], [90]).

Here, just a few recent aspects may be mentioned which illustrate the growing recognition of H₂O₂ in direct involvement: Laminar shear-stress mediates H₂O₂ formation in bovine aortic endothelial cells [91]. NADPH oxidase 4 (Nox4) has a protective vascular function, its deletion causes apoptosis [92]. Nox4 contributes considerably, by about one-third, to cellular H₂O₂ formation in vascular endothelium [92]. The redox inhibition of protein phosphatase 1, mentioned above, by Nox4 regulates elF2-alpha-mediated stress signaling [80]. Specialized cell-types such as vascular smooth muscle cells are good examples of the complex nature of the role of H₂O₂ in pathophysiology [93].

In injured zebrafish larvae, a newly generated H₂O₂ gradient mediates rapid leukocyte recruitment [94]. Cell proliferation subsequent to amputation of the tail of tadpoles involves increases in H₂O₂ concentration, as indicated by the HyPer probe (assuming that pH effects do not contribute significantly to the signal) [95]. The transients in H₂O₂ are embedded in concerted action with other signals in the process of wound responses mentioned in the Introduction above [96]·H₂O₂ controls axon pathfinding of retinal ganglion cells projecting towards the tectum in zebrafish [97].

The control of H₂O₂ gradients by peroxiporins (see above) allows for further possibilities of fine-tuning with impact on cell signaling and survival, as shown for AQP8 [98]. AQP3 is required for Nox-derived H₂O₂ signaling upon growth factor stimulation [99]. AQP3 was also shown to mediate H₂O₂ uptake to regulate downstream signaling in tumor necrosis factor-dependent activation of NF-κB [100]. The endoplasmic reticulum membrane is recognized as another important site of modulation of H₂O₂ traffic by aquaporins, notably by AQP11 [101].

6.1. Quantification and imaging: the “H₂O₂ landscape”

The detailed description of the spatial and temporal pattern of H₂O₂ concentrations in cells and tissues is a major challenge. Overall intracellular H₂O₂ production rates as given in Table 1 can serve as a first approximation. Clearly, substantial subcellular gradients exist, from practically zero H₂O₂ concentration to localized high concentrations e.g. in vesicles. The use of catalase Compound I for assaying H₂O₂ steady-states rests on a number of conditions, not elaborated here in detail; it could be argued that there is overestimation or, conversely, underestimation of total cellular rates and subcellular concentrations of H₂O₂, depending e.g. on the activity of peroxiredoxins. While it is uncontested that catalase is located in the peroxisomal matrix, other locations of catalase activity are also relevant, such as the mitochondria [104] and even the extracellular space [105]. Using D-amino acid oxidase as genetically encoded producer of H₂O₂, in conjunction with the fused HyPer probe, it becomes possible to examine intracellular H₂O₂ dynamics quantitatively [106]·H₂O₂ microdomains in receptor tyrosine kinase signaling are now amenable to imaging [107].

Mitochondrial dynamics in terms of fusion and fission is subjected to short-term regulation by redox signaling [108]. Conversely, the assembly of mitochondrial Complex I into supercomplexes determines, for example, the differential production of reactive oxygen species in neurons and astrocytes [109]. Addressing specific sites of mitochondrial H₂O₂ production may have therapeutic potential [110].

6.2. Broader context: relation to Ca²⁺

The distinction of oxidative eustress and oxidative distress may occur at a fine borderline, embedded in other fundamental control systems regulated by ion signals, notably calcium ions [111], [112], [113]. The H₂O₂ nanodomains mentioned above [56] support the emerging concept that Ca²⁺ signaling and the luminal redox state of the endoplasmic reticulum are intertwined, especially at the mitochondria-associated membranes (MAM) [114], [115]. Associated pH transients are accessible as well [29].

Modulation and propagation of H₂O₂ signals is also subject to intercellular communication. The microenvironment surrounding particular cells, notably the extracellular matrix (ECM), modifies the signal [116]. Gap junctional communication via connexin channels between cells may lead to “local oxidative stress expansion” [117].

6.3. Note

This article focused on H₂O₂ as a central redox signaling molecule. Other reactive oxygen species and reactive nitrogen species, not treated here, also have a role in redox regulation. For example, the superoxide anion radical reacts with FeS-centers, e.g. in aconitase, and by its reaction with nitric oxide peroxynitrite is formed which, in turn, modifies proteins by tyrosine nitration. S-Glutathionylation processes as well as reactions of hydrogen sulfide and its congeners are also important. This is illustrated by the secretion of S-glutathionylated proteins under oxidative stress [118].

6.4. Upshot

The concept of constitutive cellular low-level H₂O₂ steady-states is gaining increasing recognition. The localized function in modulating and maintaining key target functions by redox reactions is the essence of physiological oxidative stress, called oxidative eustress. Localized high levels of H₂O₂ occur in oxidative distress, exemplified by cell-killing in the respiratory burst of neutrophils [119] or in toxicity studies [120] involving H₂O₂ generation. A challenge for refined analysis will be the spatiotemporal and functional dissection of these two extremes in cells and tissues.