Search tips
Search criteria 


Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. 2008 August; 28(16): 5106–5119.
Published online 2008 June 16. doi:  10.1128/MCB.00060-08
PMCID: PMC2519710

Reactive Oxygen Species Regulate Hypoxia-Inducible Factor 1α Differentially in Cancer and Ischemia[down-pointing small open triangle]


In exercise, as well as cancer and ischemia, hypoxia-inducible factor 1 (HIF1) transcriptionally activates hundreds of genes vital for cell homeostasis and angiogenesis. While potentially beneficial in ischemia, upregulation of the HIF1 transcription factor has been linked to inflammation, poor prognosis in many cancers, and decreased susceptibility of tumors to radiotherapy and chemotherapy. Considering HIF1's function, HIF1α protein and its hydroxylation cofactors look increasingly attractive as therapeutic targets. Independently, antioxidants have shown promise in lowering the risk of some cancers and improving neurological and cardiac function following ischemia. The mechanism of how different antioxidants and reactive oxygen species influence HIF1α expression has drawn interest and intense debate. Here we present an experimentally based computational model of HIF1α protein degradation that represents how reactive oxygen species and antioxidants likely affect the HIF1 pathway differentially in cancer and ischemia. We use the model to demonstrate effects on HIF1α expression from combined doses of five potential therapeutically targeted compounds (iron, ascorbate, hydrogen peroxide, 2-oxoglutarate, and succinate) influenced by cellular oxidation-reduction and involved in HIF1α hydroxylation. Results justify the hypothesis that reactive oxygen species work by two opposite ways on the HIF1 system. We also show how tumor cells and cells under ischemic conditions would differentially respond to reactive oxygen species via changes to HIF1α expression over the course of hours to days, dependent on extracellular hydrogen peroxide levels and largely independent of initial intracellular levels, during hypoxia.

The transcription factor hypoxia-inducible factor (HIF) plays a critical role in the mammalian response to oxygen (O2) levels. HIF1, the first characterized member of the HIF family, transcriptionally activates hundreds of genes associated with angiogenesis in cancer, exercise, and ischemia, as well as energy metabolism, nutrient transport, cell cycle, and cell migration (85, 98).

HIF1α and HIF1β make up the HIF1 heterodimer. The β-subunit is constitutively expressed in cells. Expression of the α-subunit may be induced by a number of pathways, and its degradation is highly sensitive to O2 levels. Called a master switch for hypoxic gene expression (76, 85), intracellular HIF1α in normoxia is experimentally undetectable; during hypoxia, it rapidly accumulates in the cell nucleus and triggers gene expression. Molecular players involved in this process have come to light over the past 6 years; research has begun to define roles for prolyl hydroxylases, iron, ascorbate, hydrogen peroxide, 2-oxoglutarate, succinate, and von Hippel-Lindau protein in the HIF1 pathway.

Concomitantly, the study of reactive oxygen species (ROS) and the interest in antioxidants as potential dietary supplements for prevention of cancer, cardiac dysfunction, and neurodegeneration has grown rapidly. Ongoing debate surrounds the role of these compounds in hypoxic responses and the utility in pursuing them as preventative therapeutics. Some studies have shown increased ROS expression in hypoxia (10, 40), while others show a decrease (33, 96). Increased HIF1α expression has been found to contribute to mitochondrial activity (1), and specifically ROS formation, during hypoxia (26, 40, 81). However, other studies have demonstrated a decrease in HIF1α with increasing ROS (22, 99). Finally, some studies have shown no effects of H2O2 (87) or mitochondrial ROS in general (96). Related observations seem nearly as conflicting. Under hypoxic conditions, mitochondrial complex III may produce ROS, and the presence of high ROS concentrations generated from the mitochondria has been shown to stabilize HIF1α (8, 9, 20, 26). On the other hand, ROS may be produced in the cytosol, derived from NADPH oxidases (17, 33), and ROS may play a larger role in HIF1α expression during normoxia than hypoxia (43).

There are several hypotheses as to how ROS interact with the HIF1 pathway and alter HIF1α expression (recent related reviews include references 41 and 75). One possibility is that hydrogen peroxide oxidizes ferrous iron (Fe2+) to its ferric form (Fe3+), prohibiting the necessary binding of ferrous iron to the HIF1α hydroxylation enzymes, prolyl hydroxylases (PHDs) (71). Another change could be in the recruitment of ascorbate as a free radical scavenger, preventing ascorbate from reducing ferric iron and/or preventing ascorbate from binding directly to the PHDs. If ROS increased rather than decreased free Fe2+, as suggested by some experiments, HIF1α hydroxylation would instead increase (56). Additionally, 2-oxoglutarate (2OG) and succinate (SC) are also compounds involved in HIF1α hydroxylation whose concentrations could be altered by free radicals and mitochondrial dysfunction (38, 56, 71). A fourth mechanism by which ROS could influence the HIF1 pathway is through changing the availability of oxygen to bind directly to the PHDs or changing PHD phosphorylation.

To address these alternate mechanisms and analyze possible competing factors involved in pro- and antioxidant therapy in cancer and ischemia, we developed a computational model describing the in vivo system and used it to observe dynamics currently inaccessible at the molecular level in vivo. Experimentally, ROS have been shown to affect the HIF1 pathway through changes in H2O2, Fe2+, Asc, 2OG, or SC levels (61, 71), and mechanisms involving these compounds were the focus of this study.

The model consists of kinetic equations mapping the molecular steps in HIF1α degradation in normoxia, HIF1α synthesis in chronic hypoxia, and effects of the enzyme and cofactors involved in the HIF hydroxylation pathway. Kinetic values were estimated from in vitro studies, and results were validated by comparison to a series of independent experimental data. The input is cellular oxygen level, and the output is HIF1α levels in the nucleus in relation to necessary intermediate reactions, including reactions with prolyl hydroxlylase, iron, 2-oxoglutarate, ascorbate, succinate, and von Hippel-Lindau (VHL) ligase. The model was expanded to represent two possible mechanisms of how ROS interact in the HIF1 pathway: (i) at high concentrations, ROS induce HIF1α by decreasing the activity of prolyl hydroxylases, and ROS effects can be silenced by antioxidants; (ii) in some cells with damaged mitochondria, the opposite effect (ROS decreasing HIF1 activity) is possible through increased iron and 2-oxoglutarate, cofactors in HIF1α degradation.

Using this model, we demonstrate how ascorbate, iron, hydrogen peroxide, 2-oxoglutarate, and succinate would alter HIF1α expression in two representative hypoxic microenvironments: cancer cells and cells during ischemia. We show how these compounds affect adaptation to chronic hypoxia, taking into account possible changes in succinate and reactive oxygen species levels associated with increased anaerobic metabolism and oxidative phosphorylation, such as those found in cancer. Results offer insight into the pro- and antioxidant effects of five compounds present in the HIF1 pathway and how they differ in a tumor microenviroment compared to ischemia. The model demonstrates temporal-specific molecular mechanisms that could be harnessed for use in cancer prevention, recovery from ischemic injury, and repression of angiogenesis and inflammatory signaling regulated by HIF1α expression.


Formulation of the computational model.

A model of oxygen sensing by HIF1α was introduced and validated elsewhere (77). Here, we built an HIF1 computational model to incorporate potential mechanisms of ROS and antioxidants reacting within the HIF1α pathway. We represented ROS through two distinct means: (i) changes in the cofactors involved in HIF1α hydroxylation and (ii) the addition of hydrogen peroxide. The first mechanism is through direct increases or decreases in the concentration (availability) of certain cofactors of the hydroxylation pathway, i.e., by altering [Fe2+], [Asc], and [2OG] levels, independent of H2O2 concentrations. The second mechanism focuses on H2O2 as a representative ROS and affects HIF1α hydroxylation through Fe2+ and a Fenton reaction.

The complete model, based on an extensive analysis of experimental data, includes the hydroxylation of HIF1α by PHDs and the ubiquitination of hydroxylated HIF1α by VHL. Table Table11 lists the compounds relevant to the current model. Equation 1 describes the overall scheme of HIF1α degradation. Equations 2 and 3 depict the oxidation of iron by reaction with hydrogen peroxide and the reduction of iron by ascorbate, respectively. The complete model includes HIF1α hydroxylation, independent reactions of iron and ascorbate, hydrogen peroxide production, succinate accumulation and product inhibition, PHD2 synthesis and HIF1α synthesis in chronic hypoxia, and the binding of HIF1α to VHL.

Model variables and their abbreviations, as used in the paper

The scheme of the overall biochemical reaction of HIF1α hydroxylation, with succinate product inhibition is as follows:

An external file that holds a picture, illustration, etc.
Object name is zmb0160876360009.jpg

Iron oxidation by reaction with hydrogen peroxide is depicted as follows:

equation M1

Iron reduction by ascorbate (2, 101) is depicted as follows:

equation M2

Enzyme-substrate binding kinetics are used to describe the hydroxylation reactions. Governing equations are determined from mass balances for the substrate and the intermediate enzyme-substrate complexes. A combination of enzyme-substrate saturation assumptions was used for the binding of iron, ascorbate, 2-oxoglutarate, and oxygen to PHD2, PHD2 hydroxylation of HIF1α, and VHL-mediated ubiquitination. In the hydroxylation reaction of PHDs with HIF1α, we represented the binding of PHD2 with the substrates iron, 2-oxoglutarate, and oxygen sequentially; the redox reactions for ascorbate and iron are included as separate equations (77). These hydroxylation steps and their output were validated against experiments previously (77, 78). Model inputs are initial compound concentrations, including cellular O2 levels (Table (Table2).2). The output is HIF1α levels in the cell cytoplasm. The described kinetic model includes all previously studied reactions (77, 78).

Parameters for the production and degradation of H2O2 in different microenvironmentsa

Three feedback loops, HIF1 autocrine upregulation, HIF1 induction of PHD2, and succinate product inhibition, were determined to govern HIF1 levels and regulate the response to chronic hypoxia (Fig. (Fig.1)1) (78). Production terms were included for the synthesis of HIF1α and PHD2 at 3 and 4 h of hypoxia, respectively. The PHD2/HIF1α synthesis ratio is 0.01:0.05 in all cases presented in this paper; this ratio was chosen by estimating that between 4 and 8 hours there is a sixfold increase in HIF1α compared to initial conditions. C1, a constant in the synthesis terms for PHD2 and HIF1α, was set at 0.1 (78).

FIG. 1.
Schematic of the HIF1α system during chronic hypoxia. Three feedback loops govern HIF1α hydroxylation: HIF1α synthesis (dashed line), PHD2 synthesis (light gray line), and succinate production inhibition (dark gray line). Succinate ...

The effects of succinate were represented by one of two mechanisms, as previously described (78). Briefly, product inhibition by succinate was included by modifying the backward kinetic rates for the PHD2 complex binding to unhydroxylated HIF1α. This inhibition could result from 2-oxoglutarate being converted into a succinic acid salt or succinate and therefore not allowing 2OG to be available to the PHD2 forward hydroxylation reaction. As an alternate possibility, succinate accumulation could trigger a change in PHD2 activity or HIF1α levels independent of product inhibition, through a yet-unknown signaling mechanism. A hypothesis is that succinate accumulation and related tricarboxylic acid (TCA) cycle changes correlate to a redistribution of intracellular oxygen or changes in oxygen demand. There is not yet experimental evidence to rule for or against this hypothesis about succinate's role; there is potentially related work on nitric oxide's involvement in intracellular oxygen redistribution (44) and structural evidence linking ROS and succinate dehydrogenase (45, 104) that leave this as a possibility to explore. One way of modeling this potential change is by altering the oxygen available for hydroxylation as a function of succinate production. In the figures shown below, we refer to the two mechanisms as “production inhibition” and “SC signals O2 depletion.”

In the present study, the model was expanded to include the production, transport, and metabolism of hydrogen peroxide, a representative ROS that is overproduced by mitochondria malfunction and during ischemic stress (Fig. (Fig.1).1). Here we only present equations not included in our previous work.

Hydrogen peroxide.

Hydrogen peroxide is a reactive oxygen species that is also involved in the oxidation of Fe2+, a cofactor in HIF1α hydroxylation (equation 3). Changes in the concentration of hydrogen peroxide were incorporated into the model by a mass balance to represent H2O2 production (either release from the cell mitochondria during hypoxia or NADPH oxidase-dependent H2O2 production), H2O2 degradation by glutathione peroxidase and catalase, and influx from the surrounding extracellular environment. H2O2 degradation and production were represented in all cell types through additional kinetic terms. The kinetic equation for hydrogen peroxide production and degradation is as follows:

equation M3

where p is a constant production term, estimated from experimental data in different microenvironments. This includes the oxidation of Fe2+ by H2O2. H2O2 degradation is dependent on concentrations of catalase and glutathione peroxidase, GPx. Catalase is a mitochondrial enzyme that degrades H2O2 to water in a catalytic reaction (and in a peroxidatic reaction, it degrades H2O2 in the presence of a hydrogen donor; here we consider only the catalytic reaction) (25). Glutathione peroxidase is a second enzyme that controls intracellular H2O2 levels, also by consuming H2O2 to produce water; unlike catalase, it is predominantly active in the cytosol. The kinetic rates kcatalase, kGPx, and kFe3 are apparent first-order rates for a given initial H2O2 concentration. Table Table22 provides the initial concentrations for H2O2 under different conditions, as well as estimates for the kinetic term p and pseudo-first-order rate constants for the terms kcatalase × [free catalase] and kGPx× [GPx]. kFe3 was estimated as 1.1 × 10−2 μM−1 min−1 (12, 62). Table S1 in the supplemental material shows relative catalase and GPx activities for a number of cancer cell types, compared to noncancerous tissue. These values lend additional weight to the assumption that there is a threefold difference in catalase activity between tumor and noncancerous conditions (59).

To complete the representation of H2O2 intracellular concentration, possible diffusion of H2O2 into and out of the cell needs to be addressed. The H2O2 mass balance including transport is as follows:

equation M4

where [H2O2]extracellular is the extracellular concentration of hydrogen peroxide. The last term of equation 5 includes the change in [H2O2] from the extracellular environment into the cell (units of μM·min−1), {(pmembrane·A)/V}·[H2O2]extracellular and {(pmembrane·A)/V}·[H2O2], the outward H2O2 flux from the cell. Equal permeability is assumed for flux into the cell and flux out of the cell. From experimental analysis using T cells suspended in vitro, an estimate of the permeability coefficient for H2O2 across cell membranes, pmembrane, is 2 × 10−4 cm·s−1. If the cells are represented as spheres with area A of ~627 μm2, radius of 7.5 μm, and volume V, then pmembrane·A/V ≈ 0.8 s−1 or 48 min−1 (5).

In representing nonischemic, noncancerous cells in vivo, where intracellular and extracellular H2O2 are at low levels, transport into and out of the cell is neglected, and the mass balance includes only production and metabolism (25), as in equation 4. For other conditions, including transport, the mass balance is given by equation 5. H2O2 diffuses rapidly, and for the current model, H2O2 concentration is represented as uniform throughout the cell.

In vivo concentrations of H2O2.

While in vivo intracellular H2O2 concentrations in most cell types and in humans are not yet precisely known, estimates from measurements in bacteria and rat livers indicate levels on the order of 0.2 μM. In Escherichia coli, intracellular H2O2 concentrations were measured as 0.13 to 0.25 μM by one study (37) and estimated to be lower than 20 nM in the absence of exogenous sources in another study (82). Levels for mammalian cells range from 10−3 to 10−1 μM, depending on H2O2 production rates (25). Maximal H2O2 concentrations used for signaling in mammalian cells have been estimated as 0.5 to 0.7 μM (88). Variability among steady-state intracellular H2O2 in a single cell type is anticipated to be 50% or less (under different conditions of superoxide dismutase expression, relative steady-state intracellular H2O2 varied less than 20% in hamster lung fibroblasts [93]). However, it also should be noted that H2O2 concentrations can rise as high as 100 μM in phagocytes (72), and viable, transient levels of H2O2 in brain cells of >200 μM have been proposed (7).

The effects of added extracellular H2O2 have been studied in vitro by many research groups. As one example, extracellular H2O2 concentrations of 0.01 to 0.25 mM were shown to induce changes in characteristics of the potassium current in endothelial cells (higher concentrations of up to 0.5 mM had different effects on the potassium channel potential) (21). A transport model of H2O2 and experiments on H2O2 consumption indicate that following a change in H2O2, an equilibrium is reached within seconds between intracellular and extracellular H2O2 levels, with a 7- to 10-fold higher concentration in extracellular space, dependent on membrane permeability (5).

For the purposes of this study, initial in vivo intracellular concentrations of H2O2 were estimated from experiments, for normal physiology, as well as cancer and ischemic microenvironments (Table (Table2).2). For tumor cells, noncancerous (normal) cells, and noncancerous cells in ischemia, the default initial concentrations for H2O2 were 0.2 μM, 0.02 μM, and 0 μM, respectively. In the model, extracellular H2O2 is estimated as 1.4 μM, approximately sevenfold greater than the initial intracellular level for tumor and ischemic cells, and 14 μM during reperfusion (Table (Table3)3) (5, 88). Experimentally determined relative values for H2O2 kinetic parameters are provided in Table Table5,5, below, for specific tumor types.

Duration of hypoxia, hydrogen peroxide production, and estimated extracellular hydrogen peroxide in tumors and during ischemiaa
Experiments showing the up- or downregulation of HIF1α under different conditions explored in the modela

Numerical solution.

The system of nonlinear differential equations was solved using Mathworks Matlab software. The ode23s solver, based on a modified Rosenbrock formula, was used to find a solution for the series of 19 differential equations. For the time integration, the solver used adjustable time steps with a default absolute error tolerance in the solution of 10−6 μM.


Tumor microenvironment.

The tumor microenvironment is predominantly associated with hypoxia, high rates of glycolysis, and high levels of reactive oxygen species. Furthermore, select enzymes that degrade ROS (e.g., catalase for H2O2) appear in lower levels in tumor-bearing mammals than in healthy ones (19, 25). Results of the model show effects of ROS on the HIF1α pathway in tumors through two proposed mechanisms: (i) direct changes in concentrations of the hydroxylation cofactors Fe2+, Asc, and 2OG, or (ii) introduction of elevated levels of H2O2 and changes in H2O2 metabolism, production, and transport.

ROS in tumors represented through Fe2+, Asc, and 2OG.

Experiments suggest that ROS interact with the HIF1α pathway by altering the availability of Fe2+ and Asc, hydroxylation cofactors (22, 34, 38, 54). The model showed effects of high ROS levels on a hypothetical microenvironment by changes in Fe2+ and Asc levels (Fig. (Fig.2A).2A). Where [Fe2+]0 and [Asc]0 were upregulated, the peak amount of HIF1α occurred within the first hour, as in transient hypoxia (see Fig. S1A in the supplemental material). Without ROS and baseline levels of Fe2+ and Asc, the model predicted a maximum in unhydroxylated HIF1α concentration near 4.5 h, whereas when ROS were represented by a decrease in [Fe2+]0 and [Asc]0, the time to maximum HIF1α expression shifted by several hours, and HIF1α remained elevated over several days (Fig. (Fig.2A).2A). Tenfold decreases in [Fe2+]0 and [Asc]0 led to an approximately fivefold increase in maximum HIF1α expression (Fig. (Fig.2A2A).

FIG. 2.
HIF1α levels during chronic hypoxia are predicted for cells in the presence or (hypothetical) absence of ROS, where ROS is represented by changes in Fe2+ and Asc or in Fe2+ and 2OG concentrations. (A) HIF1α levels versus ...

Other experiments indicate that Fe2+ and 2-oxoglutarate are released by damaged mitochondria. 2-Oxoglutarate, like succinate, is a product of the citric acid cycle. In hypoxia, it has been hypothesized that nitric oxide causes mitochondrial dysfunction and thereby 2OG and Fe2+ upregulation mediated by the reactive oxygen species peroxynitrite (56). Increased intracellular 2OG and Fe2+ then contribute to increased hydroxylation of HIF1α by the PHDs. The model showed effects of high ROS levels on a hypothetical microenvironment by changes in Fe2+ and 2OG levels (Fig. (Fig.2B2B).

ROS in tumors with succinate effects.

Certain tumors, such as pheochromocytomas and paragangliomas, are associated with mitochondrial mutations and deficiencies in the enzyme succinate dehydrogenase (SDH). Succinate and 2-oxoglutarate are also intermediate products in the citric acid (TCA) cycle. In the hydroxylation reaction, PHD2 simultaneously splits oxygen to hydroxylate HIF1α and oxidizes and decarboxylates 2-oxoglutarate to succinate (83). Downregulation of SDH, an enzyme that degrades succinate to fumarate, leads to intracellular succinate accumulation, HIF1α stabilization, and HIF activation (83). Through in vitro and in vivo experimental observations, succinate was hypothesized to act as a product inhibitor of the PHD hydroxylation reaction of HIF1α (55, 74, 78, 83). Deciphering the relationship between succinate and ROS has been the subject of numerous studies, as well. Because of the structure of SDH redox centers, mutations in SDH have been predicted to result in ROS formation (104). Studies using both Caenorhabditis elegans (48, 86) and tumors with SDH mutations (49) have shown increased ROS. However, other studies have provided evidence that ROS is not necessary for succinate accumulation to produce a pseudo-hypoxic effect and stabilization of HIF1α (64, 74, 84). Shedding light on both possibilities occurring under different conditions, recent research showed a link between mutations in particular subunits of mitochondrial complex II (succinate-ubiquinone oxidoreductase B [SdhB]), ROS production, and HIF1α activation, and no ROS connection with other complex II mutations (SdhA) (42). Another relevant study showed that SDH mutations could specifically contribute to tumor formation via both ROS production and a proliferative response associated with succinate accumulation (91). The computational model's predicted effect of ROS on tumors with SDH deficiency is shown in Fig. Fig.2C;2C; Table Table44 shows the initial concentrations used to represent conditions found in vivo (78). ROS, represented by a 10-fold decrease in initial Fe2+ and 2OG concentrations, shifted the time to the peak accumulation of HIF1α by several hours (Fig. (Fig.2B).2B). The time to peak HIF1α accumulation was shifted up to ~48 h, when the model was used to represent in vitro conditions of succinate product inhibition (see Fig. S1B and C in the supplemental material; in these figures, [SC]0 is 500 μM, the maximal concentration used to represent SDH deficiency in published in vitro experiments) (83). ROS, represented by a 10-fold decrease in initial Fe2+ and Asc, also shifted the time to peak accumulation of HIF1α (see Fig. S1D in the supplemental material).

Initial concentrations for O2, Fe2+, Asc, SC, and 2OG for conditions represented in the model

ROS in tumors represented through H2O2.

The effect of ROS on HIF1α could be solely related to one ROS species, H2O2, its initial concentration, and its production in tumor cells (61). Measurements of H2O2 in vivo are difficult to obtain, and estimates for a base value of 0.2 μM in tumors were obtained from experiments, as described above. Fibroblasts that became tumorigenic with the expression of Nox1, the catalytic subunit of an NADPH oxidase, expressed H2O2 at a level 10-fold that of normal fibroblasts (6), in agreement with a number of studies that had established the relationship between increased H2O2 and tumorigenicity (24, 30, 90). Results from the model show the effect of an increase in initial [H2O2]0, H2O2 production, and H2O2 metabolism predicted for a tumor cell (Fig. (Fig.33 and and4)4) and? a tumor cell with SDH deficiency represented by two distinct mechanisms (Fig. (Fig.4;4; see also Fig. S2 in the supplemental material). Additionally, the effects of extracellular H2O2 on HIF1α are predicted and compared to available in vitro experimental data (see Fig. S4 in the supplemental material).

FIG. 3.
Model results showing the effects of H2O2 on the cellular hypoxic response. HIF1α expression is predicted from 0 to 72 h for noncancerous (normal) cells, tumor cells, and cells exposed to ischemia. [HIF1α]0 = 1 μM. Both ...
FIG. 4.
Model results showing the effects of H2O2 on the cellular hypoxic response via HIF1α expression from 0 to 72 h in tumor cells and tumor cells with an SDH deficiency. [HIF1α]0 = 1 μM. Tumor cell conditions correspond to ...

Ischemia microenvironment.

The ischemic microenvironment is associated with hypoxia, and high levels of reactive oxygen species are found both following the ischemic insult and during reperfusion.

ROS in ischemia represented through H2O2.

The mechanism of ROS involvement in HIF1α expression during ischemia could be solely related to H2O2 initial concentrations and production. If this is the case, the model predicts a temporal expression of HIF1α in cells exposed to ischemia, similar to that of tumor cells (Fig. (Fig.3).3). This expression changes at the onset of reperfusion, which induces a larger amount of H2O2 production and alters H2O2 transport (Fig. (Fig.5;5; see also Fig. S4 in the supplemental material). If H2O2 extracellular levels remain elevated, the resulting increase in HIF1α lasts, uninhibited by any mechanism currently modeled, whether the intracellular conditions are approximated as hypoxia or normoxia (see Fig. S4 in the supplemental material).

FIG. 5.
Model results showing the effects of H2O2 on the cellular hypoxic response via HIF1α expression from 0 to 12 h in ischemic cells and ischemic cells exposed to ROS generated by reperfusion. Conditions correspond to those detailed in Table ...

ROS in ischemia represented through H2O2, Fe2+, and Asc or 2OG.

Using the ischemia microenvironment, a third and a fourth mechanism for how ROS affect the HIF1α pathway were tested, with increases in H2O2, Asc, and Fe2+ or increases in H2O2, 2-oxoglutarate, and Fe2+ simultaneously. The model results for changes in Fe2+ and Asc or 2OG are shown, in conjunction with increased production of H2O2 (Fig. (Fig.6).6). In comparison to the Fe2+ and Asc mechanism of ROS effects (a 10-fold increase in Fe2+ and Asc) (Fig. (Fig.6A),6A), an increase in 2OG accelerates the production of HIF1α and elevates its expression much higher by 8 h of hypoxia (Fig. (Fig.6B6B).

FIG. 6.
Model results showing the effects of H2O2 on the cellular hypoxic response via HIF1α expression from 0 to 12 h in ischemic cells with reperfusion. (A) ROS is represented by both Fe2+ and Asc increases and by H2O2 production (dashed line) ...


The variable effects of ROS on HIF1α can be attributed to three main factors: (i) the degree of hypoxia, (ii) the form and intracellular location of ROS produced, and (iii) the molecular microenvironment of the cell. The third was the focus of this model. Before discussing the relevance of the microenvironment, it is worthwhile to briefly describe the roles of the first two factors, as they relate to the presented results.

ROS, O2 levels, and ROS species.

ROS production requires oxygen, so it is not surprising that ROS are expressed in different concentrations during anoxia, hypoxia, and normoxia. However, a direct correlation between ROS levels and O2 availability remains elusive. Low O2 limits formation of superoxide and its by-products (33), while hypoxia also has been shown to increase ROS, possibly through release by the mitochondria electron transport chain (41). Equally intriguing, ROS appear to play distinctly different roles in anoxia, hypoxia, and normoxia, with respect to the HIF1 system. In anoxia, the electron transport chain in mitochondria may serve as oxygen sensor and regulator of HIF1α expression (41); as oxygen consumption by mitochondria becomes limited by O2 availability only when O2 falls below ~0.1%, mitochondria would not likely be an effective detector of moderate hypoxia. In hypoxia, some experiments suggest that at low or intermediate ROS concentrations induced by a superoxide generator, HIF1α is downregulated, as PHD activity is upregulated (22, 54). In related experiments, during normoxia, high ROS levels increased HIF1α, by blocking PHD hydroxylation. Additional studies have also indicated that in long-term hypoxia (12 h), the effect of ROS is to signal HIF1α degradation and downregulate HIF1α expression (27). In contrast, other experiments have shown in hypoxia that ROS upregulated HIF1α (6, 40, 75).

One explanation for the conflicting effects of ROS on HIF1α may be that HIF1α expression is dependent on particular reactive oxygen species, and not others (20); experiments differ in how they have measured ROS and in what cell types. ROS is produced by several means, which will affect ROS location and signaling. In normal cells, mitochondrial complex I and III and cytosolic monoamine and NADPH oxidases produce H2O2. In cancer cells, ROS, and H2O2 in particular, are additionally overproduced by mitochondrial respiration, while in ischemia, cells may be susceptible to ROS from both mitochondrial dysfunction and the extracellular environment, including effects of reperfusion injury. In both cases, inflammatory cells may also release ROS locally and affect intracellular levels in tumors and ischemic cells. Of all reactive oxygen species, H2O2 seems to be the one with the noticeable effect on the HIF1 system (20, 34, 61). However, the reactive oxygen species peroxynitrite may serve as an oxygen donor during HIF1α hydroxylation, as well as mediate Fe2+ and 2OG release from the mitochondria (56, 89), and effects of superoxide on HIF1α have been shown in renal carcinoma (46) and renal medullary interstitial cells (103).

Predictions: ROS and the cellular microenvironment.

The molecular environment, the focus of this work, distinguishes why the effects of ROS differ between tumor cells and ischemic cells. In tumor cells, the duration of hypoxia is generally longer, and the cell adapts to anaerobic metabolism and likely relatively stable levels of ROS production. HIF1α levels peak, as in normal cells, before 12 h (Fig. (Fig.4),4), and do not become elevated again until several days (see Fig. S3A in the supplemental material). In contrast, for the ischemic cell, the duration of ischemia is often hours or less, and cell is starved of oxygen but unable to adapt as readily as a cancer cell; additionally, the levels of ROS rapidly increase following the infarct or occlusion, and with perfusion, and then remain elevated for days, eventually decreasing (Fig. (Fig.5;5; see also Fig. S4A and B in the supplemental material).

ROS mechanisms.

We represented the effects of ROS on the HIF1α through two different mechanisms: (i) altering the relative concentrations of cofactors in PHD2 hydroxylation of HIF1α and (ii) H2O2 production, transport, and metabolism. Both mechanisms involve the availability of free ferrous iron (Fe2+), a cofactor in the hydroxylation reaction, via a mechanism that has been shown in experiments (34). They differ in that the first mechanism was modeled by changing the availability of Fe2+ and Asc (Fig. (Fig.2A2A and and6A)6A) or Fe2+ and 2OG (Fig. 2B and C and and6B),6B), while H2O2 only altered the availability of free ferrous iron in the model. Changes in H2O2 concentration were highly dependent on the microenvironment (Tables (Tables22 and and33).

The results from the model lend credence to the hypothesis that ROS interacts with the HIF1 system through several different mechanisms, and they may help explain conflicting experiments on HIF1α expression and ROS (Table (Table5).5). If ROS work by both increasing Fe2+ and Asc or 2OG (Fig. (Fig.22 and and6)6) and increasing H2O2 (thereby decreasing free Fe2+) (Fig. (Fig.3,3, ,4,4, and and5),5), depending on the hypoxic conditions, then there could be a dual effect of ROS so that in some cases it upregulates HIF1α and in others it downregulates it. Moreover, in the first mechanism, Fe2+, Asc, and 2OG are quickly depleted in ischemic-reperfusion conditions, and the relatively lower HIF1α level is a transient effect (Fig. (Fig.6).6). Beyond 5 to 12 h (with a 10-fold increase in 2OG and Asc, respectively), HIF1α levels reach the same level or higher, as they do using H2O2 alone to represent ROS (Fig. (Fig.6;6; see also Fig. S4B and C in the supplemental material). Depending on the duration of hypoxia in experiments, the model then predicts the same ROS conditions would yield either a relative upregulation or downregulation of HIF1α. Model results using H2O2 as the representative ROS are in agreement with experimental studies showing H2O2 depletes ascorbate (70) (data not shown). The effect of an infinite source of Asc and free Fe2+ has not yet been measured experimentally or modeled with respect to the HIF1α system and ROS; this would be an interesting test of altering hypoxic response through nutritional supplementation.

ROS in cancer and in ischemia. (i) H2O2 in cancer.

Restricting the effect of the ROS in the model to changes in H2O2 levels offers insight into how differences in cancer and ischemic microenvironments determine distinct cell fates. The effects of H2O2 on the HIF1α concentrations are dependent on the dose of H2O2, levels of hypoxia, and cell type. Seemingly contradictory, high levels of H2O2 are associated not only with the progression of cancer to metastasis but also the susceptibility of cancer cells to cell death (61). An existing hypothesis to explain this observation is that all cells, cancer cells included, have an upper threshold for H2O2 concentration, above which apoptosis occurs and below which cells proliferate at a rate proportional to the amount of H2O2 (36, 59) (Fig. (Fig.7).7). The model predicts cancer cells reach a lower steady-state level of H2O2 than cells exposed to ischemia and reperfusion (Fig. (Fig.8).8). Assuming the threshold hypothesis is true, the model implies cancer cells are more apt to survive in their hypoxic microenvironment well beyond the survival time of cells exposed to ischemia-reperfusion, at the same duration of hypoxia.

FIG. 7.
Potential effects of the threshold hypothesis. Thresholds are represented by the horizontal dashed-dotted lines. (A) If intracellular levels of H2O2 above ~1 μM induced apoptosis (4), cells exposed to ischemia-reperfusion injury would ...
FIG. 8.
Circuit representations of ROS effects on the HIF1α pathway. (A) Coherent feed-forward loop type 4 in hypoxia, where ROS blocks PHD2 expression, leading to upregulation of HIF1α; another upregulation of HIF1α by ROS may come as ...

Additionally, HIF1α expression is increasing as H2O2 increases or remains elevated below the threshold (Fig. (Fig.7),7), making the cancer cells more virile and less susceptible to radiation and chemotherapy (50). Increased HIF1α expression triggers the production of angiogenic factors in both cancer and ischemic cells, but the ischemic cells have less of a chance to survive their high intracellular ROS levels, while angiogenesis fuels the cancer cells' proliferation. If ROS works through an increased Fe2+, Asc, or 2OG mechanism too, the time to peak HIF1α levels is delayed (Fig. (Fig.22 and and6;6; see also Fig. S1 in the supplemental material), potentially delaying the onset of angiogenesis. Furthermore, there may be a threshold for maximum HIF1α levels like that hypothesized for H2O2 levels, above which apoptosis occurs, in which case, again the cells exposed to ischemia-reperfusion would be more susceptible than cancer cells (Fig. (Fig.7B).7B). In support of the model's predic-tion of extracellular H2O2 effects on HIF1 in cancer cells, results were compared to experimental data looking at in vitro conditions where cancer cells were supplemented with H2O2 and insulin. While no direct comparison can be made, as the conditions are not the same and the experiment assessed HIF1 activity through hypoxic gene activation, the trend of increasing HIF1 activity and then saturation at high extracellular H2O2 agree (see Fig. S4 in the supplemental material).

(ii) H2O2, ischemia, and reperfusion.

Along with differentiating the effects of ROS in cancer cells and cells exposed to ischemia, the model results shed light on two additional experimental observations: concentration-dependent effects of H2O2 and the extent of reperfusion injury. The model described H2O2 concentration as a function of production, transport, and metabolism. The intracellular concentration of H2O2 (high or low) rapidly reaches an equilibrium level dependent on extracellular concentrations. This is a function of its quick transport across the cell membrane (5, 16), combined with the driving force of intracellular metabolism (see Fig. S2B and S3D in the supplemental material). As a consequence of the rapid equilibrium, initial concentrations of H2O2 are of minimal relevancy, and the model predicts that sustained intermediate levels of intracellular H2O2 have the most significant effect on increasing HIF1α expression in tumors and ischemia and altering the hypoxic response long term (Fig. (Fig.7A).7A). While there may be many explanations other then sustained, elevated intracellular H2O2 levels, this conclusion also offers an interesting perspective on experimental evidence showing that intermediate H2O2 levels produce more DNA lesions than higher or lower H2O2 concentrations (69) and supports experiments showing some cells can fully recover from transient very high levels of H2O2 (7).

The model predicts reperfusion produces a far greater increase in H2O2-dependent HIF1α expression than ischemia alone (Fig. (Fig.5),5), even while H2O2 production decreases to normal levels with reperfusion after 3.5 h (Table (Table2).2). This is true even if reperfusion is considered to provide enough oxygen for cells to reach normoxia (see Fig. S3A and B in the supplemental material). Anticipated high extracellular levels of H2O2 during reperfusion drive this elevated H2O2-dependent HIF1α (Table (Table33).

The model then highlights how regulating hydrogen peroxide temporally is essential in treatments for ischemia-reperfusion injury. In vivo, high oxidant stress following ischemia may drive angiogenesis, allowing recovery of normoxia through the growth of vessels, while on the other hand too high levels of oxidants damage tissues. Modeling is key to correctly pinpointing the effective time frame for therapeutics, and as the computational modeling advances with new experimental and clinical measurements, antioxidant and proangiogenic treatments could be tailored to individuals.

Metabolism of ROS: catalase, GPx, and variability.

In experiments using human hepatoma cells, catalase overexpression did not show an appreciably different effect on HIF1 expression or transcriptional activity (87). The opposite has been shown in other cell types (20). The authors of the first study suggested that H2O2 played little or no role as a signaling molecule in the hypoxic response. As another possibility, which could explain discrepancies between studies, H2O2 may play a role in HIF1 activation only or predominantly in hypoxia, but not anoxia (81). Our model suggests a third explanation. An increase in intracellular catalase concentration does not have a strong effect on the hypoxic response via direct HIF1 activation, while extracellular H2O2 and catalase levels do. Lending support to the model's finding that extracellular H2O2 greatly determines the cellular hypoxic response, recent studies have indicated catalase added to cell medium has a strong effect on cell-cell communication in microglia via NO signaling (47). Furthermore, another study showed that an adenovirus containing catalase affected the activity of apoptosis signal-regulating kinase 1, another signaling molecule involved in hypoxia and ischemia, whose activity is dependent on H2O2 and indirectly related to HIF1 (57).

The model offers plausible mechanisms to explain several phenomena associated with ROS effects on the HIF1 system; however, it is worthwhile to mention limitations of the current model. In vivo intracellular H2O2 concentrations depend on a number of parameters that vary by cell type. These parameters include the concentration and activities of the peroxisome enzyme catalase and glutathione peroxidase, predominant in the cell cytosol; both enzymes degrade H2O2 into water. The concentrations and activities of these enzymes were approximated as constants in the model (Table (Table2),2), and further studies would represent how known changes in their activities affect intracellular H2O2 concentrations and the HIF1 system. For example, the concentration of catalase is known to vary with cell type, ranging from 4 ng/106 cells in lymphocytes to 850 ng/106 cells in macrophages (80). The variability among endothelial cells is expected to be less; however, it has not been established experimentally to our knowledge (in bacteria, a variability in catalase activity of 10- to 20-fold has been assessed [79]). Studies measuring tissue-level catalase activity lend weight to the model's assumption of constant catalase activity during ischemia and reperfusion (39). However, the concentration and activity of intracellular catalase, as well as the scavenging enzyme GPx, are likely specific to the cell and tissue type, as well as age (67) (see Fig. S1 in the supplemental material). H2O2 membrane transport has been modeled here as unlimited diffusion (65), while facilitated transport by aquaporins may also occur (15). Additionally, extracellular H2O2 is estimated as a constant for each condition considered in the model. When more observations become available on the temporal changes in extracellular H2O2 during cancer and ischemia, it would be interesting to include effects of changing local extracellular H2O2.

The model predicts that in tumors, the metabolism of H2O2 drives increased transport from the extracellular microenvironment, while H2O2 production by a single cell is a fraction of what is being metabolized (see Fig. S3D in the supplemental material). The model approximated uniform metabolism and production throughout the cell, while production occurs in subcellular units or areas of the cytoplasm. As another consideration, the transport of H2O2 into and out of the cell and intracellular membranes (i.e., peroxisomes) affects the H2O2 gradients present in the cell. How the response to an isolated, high concentration of ROS differs from a cell's response to uniform intracellular ROS elevation merits exploration.

Another characteristic of the model suggests an avenue for experimental pursuit. The current model has no specific feedback on ROS effects. Model predictions for HIF1α accumulation in chronic conditions correlate well with a range of experiments showing a peak in HIF1α before 12 h (Fig. (Fig.3;3; see also Fig. S2A in the supplemental material). However, without a way to negatively regulate ROS, for hypoxic and ischemic durations beyond 48 h, the model predicts elevated HIF1α levels that have not been seen in many experiments (see Fig. S2A and S3B and Table S2 in the supplemental material). Potential sources of adaptation to ROS or downregulation of ROS effects on HIF1α could be changes in PHD2 activity or O2 redistribution.

A systems view of HIF1α signaling.

As HIF1 is a transcription factor, its protein regulation is interesting in the greater context of systems biology. From experiments and computational modeling thus far, in chronic hypoxia HIF1α has two positive autoregulatory feedback mechanisms, HIF1α upregulating itself and the hydroxylation product succinate downregulating PHD2 to upregulate HIF1α, and one negative one, HIF1α-dependent upregulation of PHD2 (78). Furthermore, ROS regulates HIF1α expression; this provides another form of autocrine regulation by a hypoxic cell.

What does a hypoxic cell gain from these multiple regulatory pathways, both negatively and positively? Negative autoregulation alone speeds the response time for a closed regulatory circuit and leads to saturation in the protein. If HIF1α only regulated PHD2, and not itself, this would be the expected result; on the other hand, if HIF1α regulated HIF1α alone and not PHD2, the positive feedback would lead to a slower response time and could lead to indefinitely increasing HIF1α conditions. From experiments, HIF1α auto-upregulation occurs an hour or more prior to HIF1α-dependent PHD2 upregulation. One might expect from this a characteristic peak found in HIF1α expression, with a chronic decay rate of HIF1α dependent on the synthesis ratio of PHD2:HIF1α (78). The staggered response coupling both positive and negative feedback then potentially optimizes both the degree of HIF1α upregulation and the speed of its regulation during periods of chronic hypoxia.

HIF1α protein regulation during hypoxic durations of less than 3 hours is modulated without feedback. This provides a response within minutes (HIF1α accumulation in hypoxia and HIF1α oxygen-dependent degradation). There is no need to keep the response on, once oxygen is restored. It may also be that some genes are upregulated by HIF1 specifically during long-term hypoxia, and the concentration or duration of the transient HIF1 response is insufficient to trigger their activation.

To summarize and weigh current hypotheses on ROS at a systems level, a brief network motif analysis of possible ROS mechanisms in regards to HIF1 and hypoxia is presented (Fig. (Fig.8).8). Figure Figure88 provides circuit representations of hypotheses of ROS and HIF1α interactions. A coherent feed-forward loop refers to a circuit where the indirect path and the direct path yield the same response; an incoherent loop refers to one where the two paths cause opposite effects (3). Of the three proposed mechanisms of ROS, the incoherent feed-forward type 1 (Fig. (Fig.8B),8B), is one of the most prevalent circuits in biological systems, albeit at the gene level (3). In this case ROS may increase HIF1α through limiting O2 availability, ferrous iron, and Asc or decrease HIF1α by increasing 2OG and Fe2+ or donating O2 in hypoxia, all cofactors in PHD2 hydroxylation. If ROS mechanisms can be described by this circuit, pulse-like dynamics in HIF1α expression may be possible; ROS would begin to upregulate HIF1α independently of PHD2 and then PHD2 would be upregulated by ROS, leading to HIF1α downregulation (3). Other possible mechanisms of ROS interactions are shown, as well as the effects of succinate and hypoxia globally on HIF1α. At the transcriptional level, these configurations are rarer than the incoherent feed-forward type 1, and they are characterized by limited function.

If the prevalence of specific motifs at the transcriptional level can be extrapolated to molecular species and proteins, the presented circuit analysis helps support the hypothesis from the molecular model that ROS acts by both up- and downregulating HIF1, rather than one or the other. The allure of representing the system in circuit diagram form is both enhanced and tempered by the known biological complexity of the HIF1 system. In graphical form, hypothesized positive and negative controls on HIF1 become simplified; adding components, e.g., metabolic pathways, becomes relatively easy. However, even while highlighting characteristics of simple circuits and network motifs, the interactions, delay time, protein versus gene response, and multiple connections can alter the excepted benefits or characteristics of a particular motif.

The presented model demonstrates several molecular mechanisms for how ROS signaling can affect the HIF1 pathway through Fe2+, Asc, 2OG, SC, and H2O2. We showed how tumor cells and cells exposed to ischemia would differentially respond to ROS via changes to HIF1α expression over the course of hours to days. Model results also show that in hypoxia (both in cancer and ischemic microenvironments), H2O2 intracellular levels rapidly reach equilibrium with extracellular levels, largely independent of initial intracellular levels. H2O2 transport and metabolism, more than cellular production, dictate HIF1α levels. Antioxidants (e.g., Asc) can alter the amount of ROS that is metabolized intracellularly and restore free Fe2+ levels, but unless the extracellular ROS levels change too, this effect is transient. Applied to therapeutic manipulation of hypoxic response, model results imply that antioxidants would need to be applied judiciously at the correct intervals (a sustained, moderate level) to have a noticeable effect on HIF1α levels either in the cancer or ischemic-reperfusion microenvironment. The optimal concentration would be dictated by the hypoxic microenvironment, and the model suggests noticeably higher doses would be needed to avoid reperfusion injury than those needed to prevent cancer cell proliferation or reduce ischemic damage alone.

Supplementary Material

[Supplemental material]


This work was supported by NIH 1F32HL085016-01 (A.Q.) and NIH HL079653 and NIH HL087351 (A.S.P.).

We thank J. Pouyssegur for useful discussions.


[down-pointing small open triangle]Published ahead of print on 16 June 2008.

Supplemental material for this article may be found at


1. Agani, F. H., P. Pichiule, J. C. Chavez, and J. C. LaManna. 2000. The role of mitochondria in the regulation of hypoxia-inducible factor 1 expression during hypoxia. J. Biol. Chem. 27535863-35867. [PubMed]
2. Al-Ayash, A. I., and M. T. Wilson. 1979. The mechanism of reduction of single-site redox proteins by ascorbic acid. Biochem. J. 177641-648. [PubMed]
3. Alon, U. 2007. An introduction to systems biology.Chapman & Hall, Boca Raton, FL.
4. Antunes, F., and E. Cadenas. 2001. Cellular titration of apoptosis with steady state concentrations of H2O2: submicromolar levels of H2O2 induce apoptosis through Fenton chemistry independent of the cellular thiol state. Free Radic. Biol. Med. 301008-1018. [PubMed]
5. Antunes, F., and E. Cadenas. 2000. Estimation of H2O2 gradients across biomembranes. FEBS Lett. 475121-126. [PubMed]
6. Arnold, R. S., J. Shi, E. Murad, A. M. Whalen, C. Q. Sun, R. Polavarapu, S. Parthasarathy, J. A. Petros, and J. D. Lambeth. 2001. Hydrogen peroxide mediates the cell growth and transformation caused by the mitogenic oxidase Nox1. Proc. Natl. Acad. Sci. USA 985550-5555. [PubMed]
7. Avshalumov, M. V., L. Bao, C. P. J., and M. E. Rice. 2007. H2O2 signaling in the nigrostriatal dopamine pathway via ATP-sensitive potassium channels: issues and answers. Antioxid. Redox Signal. 9219-231. [PubMed]
8. Bell, E. L., and N. S. Chandel. 2007. Mitochondrial oxygen sensing: regulation of hypoxia-inducible factor by mitochondrial generated reactive oxygen species. Essays Biochem. 4317-27. [PubMed]
9. Bell, E. L., T. A. Klimova, J. Eisenbart, C. T. Moraes, M. P. Murphy, G. R. Budinger, and N. S. Chandel. 2007. The Qo site of the mitochondrial complex III is required for the transduction of hypoxic signaling via reactive oxygen species production. J. Cell Biol. 1771029-1036. [PMC free article] [PubMed]
10. Bell, E. L., T. A. Klimova, J. Eisenbart, P. T. Schumacker, and N. S. Chandel. 2007. Mitochondrial reactive oxygen species trigger hypoxia-inducible factor-dependent extension of the replicative life span during hypoxia. Mol. Cell. Biol. 275737-5745. [PMC free article] [PubMed]
11. Bellisola, G., M. Casaril, G. B. Gabrielli, M. Caraffi, and R. Corrocher. 1987. Catalase activity in human hepatocellular carcinoma (HCC). Clin. Biochem. 20415-417. [PubMed]
12. Beltran, F. J., M. Gonzalez, F. J. Rivas, and P. Alvarez. 1998. Fenton reagent advanced oxidation of polynuclear aromatic hydrocarbons in water. Water Air Soil Pollut. 105685-700.
13. Bergeron, M., A. Y. Yu, K. E. Solway, G. L. Semenza, and F. R. Sharp. 1999. Induction of hypoxia-inducible factor-1 (HIF-1) and its target genes following focal ischaemia in rat brain. Eur. J. Neurosci. 114159-4170. [PubMed]
14. Bernaudin, M., A. S. Nedelec, D. Divoux, E. T. MacKenzie, E. Petit, and P. Schumann-Bard. 2002. Normobaric hypoxia induces tolerance to focal permanent cerebral ischemia in association with an increased expression of hypoxia-inducible factor-1 and its target genes, erythropoietin and VEGF, in the adult mouse brain. J. Cereb. Blood Flow Metab. 22393-403. [PubMed]
15. Bienert, G. P., A. L. Moller, K. A. Kristiansen, A. Schulz, I. M. Moller, J. K. Schjoerring, and T. P. Jahn. 2007. Specific aquaporins facilitate the diffusion of hydrogen peroxide across membranes. J. Biol. Chem. 2821183-1192. [PubMed]
16. Bienert, G. P., J. K. Schjoerring, and T. P. Jahn. 2006. Membrane transport of hydrogen peroxide. Biochim. Biophys. Acta 1758994-1003. [PubMed]
17. Biswas, S., M. K. Gupta, D. Chattopadhyay, and C. K. Mukhopadhyay. 2007. Insulin-induced activation of hypoxia-inducible factor-1 requires generation of reactive oxygen species by NADPH oxidase. Am. J. Physiol. Heart Circ. Physiol. 292H758-H566. [PubMed]
18. Boveris, A., N. Oshino, and B. Chance. 1972. The cellular production of hydrogen peroxide. Biochem. J. 128617-630. [PubMed]
19. Bozzi, A., I. Mavelli, B. Mondovi, R. Strom, and G. Rotilio. 1981. Differential cytotoxicity of daunomycin in tumour cells is related to glutathione-dependent hydrogen peroxide metabolism. Biochem. J. 194369-372. [PubMed]
20. Brunelle, J. K., E. L. Bell, N. M. Quesada, K. Vercauteren, V. Tiranti, M. Zeviani, R. C. Scarpulla, and N. S. Chandel. 2005. Oxygen sensing requires mitochondrial ROS but not oxidative phosphorylation. Cell. Metab. 1409-414. [PubMed]
21. Bychkov, R., K. Pieper, C. Ried, M. Milosheva, E. Bychkov, F. C. Luft, and H. Haller. 1999. Hydrogen peroxide, potassium currents, and membrane potential in human endothelial cells. Circulation 991719-1725. [PubMed]
22. Callapina, M., J. Zhou, T. Schmid, R. Kohl, and B. Brune. 2005. NO restores HIF-1α hydroxylation during hypoxia: role of reactive oxygen species. Free Radic. Biol. Med. 39925-936. [PubMed]
23. Canbolat, O., J. Fandrey, and W. Jelkmann. 1998. Effects of modulators of the production and degradation of hydrogen peroxide on erythropoietin synthesis. Respir. Physiol. 114175-183. [PubMed]
24. Cerutti, P. A. 1985. Prooxidant states and tumor promotion. Science 227375-381. [PubMed]
25. Chance, B., H. Sies, and A. Boveris. 1979. Hydroperoxide metabolism in mammalian organs. Physiol. Rev. 59527-605. [PubMed]
26. Chandel, N. S., D. S. McClintock, C. E. Feliciano, T. M. Wood, J. A. Melendez, A. M. Rodriguez, and P. T. Schumacker. 2000. Reactive oxygen species generated at mitochondrial complex III stabilize hypoxia-inducible factor-1alpha during hypoxia: a mechanism of O2 sensing. J. Biol. Chem. 27525130-25138. [PubMed]
27. Chang, T. C., C. J. Huang, K. Tam, S. F. Chen, K. T. Tan, M. S. Tsai, T. N. Lin, and S. K. Shyue. 2005. Stabilization of hypoxia-inducible factor-1α by prostacyclin under prolonged hypoxia via reducing reactive oxygen species level in endothelial cells. J. Biol. Chem. 28036567-36574. [PubMed]
28. Chavez, J. C., F. Agani, P. Pichiule, and J. C. LaManna. 2000. Expression of hypoxia-inducible factor-1α in the brain of rats during chronic hypoxia. J. Appl. Physiol. 891937-1942. [PubMed]
29. Chavez, J. C., and J. C. LaManna. 2002. Activation of hypoxia-inducible factor-1 in the rat cerebral cortex after transient global ischemia: potential role of insulin-like growth factor-1. J. Neurosci. 228922-8931. [PubMed]
30. Chung-man Ho, J., S. Zheng, S. A. Comhair, C. Farver, and S. C. Erzurum. 2001. Differential expression of manganese superoxide dismutase and catalase in lung cancer. Cancer Res. 618578-8585. [PubMed]
31. Deudero, J. J., C. Caramelo, M. C. Castellanos, F. Neria, R. Fernandez-Sanchez, O. Calabia, S. Penate, and F. R. Gonzalez-Pacheco. 2008. Induction of hypoxia-inducible factor 1alpha gene expression by vascular endothelial growth factor. Role of a superoxide-mediated mechanism. J. Biol. Chem. 28311435-11444. [PubMed]
32. Dirnagl, U., U. Lindauer, A. Them, S. Schreiber, H. W. Pfister, U. Koedel, R. Reszka, D. Freyer, and A. Villringer. 1995. Global cerebral ischemia in the rat: online monitoring of oxygen free radical production using chemiluminescence in vivo. J. Cereb. Blood Flow Metab. 15929-940. [PubMed]
33. Fandrey, J., S. Frede, and W. Jelkmann. 1994. Role of hydrogen peroxide in hypoxia-induced erythropoietin production. Biochem. J. 303507-510. [PubMed]
34. Gerald, D., E. Berra, Y. M. Frapart, D. A. Chan, A. J. Giaccia, D. Mansuy, J. Pouyssegur, M. Yaniv, and F. Mechta-Grigoriou. 2004. JunD reduces tumor angiogenesis by protecting cells from oxidative stress. Cell 118781-794. [PubMed]
35. Ginouves, A., K. Ilc, N. Macias, J. Pouyssegur, and E. Berra. 2008. PHDs overactivation during chronic hypoxia “desensitizes” HIFα and protects cells from necrosis. Proc. Natl. Acad. Sci. USA 1054745-4750. [PubMed]
36. Giorgio, M., M. Trinei, E. Migliaccio, and P. G. Pelicci. 2007. Hydrogen peroxide: a metabolic by-product or a common mediator of ageing signals? Nat. Rev. Mol. Cell Biol. 8722-728. [PubMed]
37. Gonzalez-Flecha, B., and B. Demple. 1997. Homeostatic regulation of intracellular hydrogen peroxide concentration in aerobically growing Escherichia coli. J. Bacteriol. 179382-388. [PMC free article] [PubMed]
38. Gottlieb, E., and I. P. Tomlinson. 2005. Mitochondrial tumour suppressors: a genetic and biochemical update. Nat. Rev. Cancer 5857-866. [PubMed]
39. Grisham, M. B., L. A. Hernandez, and D. N. Granger. 1986. Xanthine oxidase and neutrophil infiltration in intestinal ischemia. Am. J. Physiol. 251G567-G574. [PubMed]
40. Guzy, R. D., B. Hoyos, E. Robin, H. Chen, L. Liu, K. D. Mansfield, M. C. Simon, U. Hammerling, and P. T. Schumacker. 2005. Mitochondrial complex III is required for hypoxia-induced ROS production and cellular oxygen sensing. Cell. Metab. 1401-408. [PubMed]
41. Guzy, R. D., and P. T. Schumacker. 2006. Oxygen sensing by mitochondria at complex III: the paradox of increased reactive oxygen species during hypoxia. Exp. Physiol. 91807-819. [PubMed]
42. Guzy, R. D., B. Sharma, E. Bell, N. S. Chandel, and P. T. Schumacker. 2008. Loss of the SdhB, but not the SdhA, subunit of complex II triggers reactive oxygen species-dependent hypoxia-inducible factor activation and tumorigenesis. Mol. Cell. Biol. 28718-731. [PMC free article] [PubMed]
43. Haddad, J. J., and S. C. Land. 2001. A non-hypoxic, ROS-sensitive pathway mediates TNF-alpha-dependent regulation of HIF-1α. FEBS Lett. 505269-274. [PubMed]
44. Hagen, T., C. T. Taylor, F. Lam, and S. Moncada. 2003. Redistribution of intracellular oxygen in hypoxia by nitric oxide: effect on HIF1α. Science 3021975-1978. [PubMed]
45. Hederstedt, L. 2003. Structural biology. Complex II is complex too. Science 299671-672. [PubMed]
46. Herr, B., J. Zhou, S. Drose, and B. Brune. 2007. The interaction of superoxide with nitric oxide destabilizes hypoxia-inducible factor-1α. Cell Mol. Life Sci. 643295-3305. [PubMed]
47. Holmquist, L., G. Stuchbury, M. Steele, and G. Munch. 2007. Hydrogen peroxide is a true first messenger. J. Neural Transm. Suppl. 200739-41. [PubMed]
48. Ishii, N., M. Fujii, P. S. Hartman, M. Tsuda, K. Yasuda, N. Senoo- Matsuda, S. Yanase, D. Ayusawa, and K. Suzuki. 1998. A mutation in succinate dehydrogenase cytochrome b causes oxidative stress and ageing in nematodes. Nature 394694-697. [PubMed]
49. Ishii, T., K. Yasuda, A. Akatsuka, O. Hino, P. S. Hartman, and N. Ishii. 2005. A mutation in the SDHC gene of complex II increases oxidative stress, resulting in apoptosis and tumorigenesis. Cancer Res. 65203-209. [PubMed]
50. Jensen, R. L. 2006. Hypoxia in the tumorigenesis of gliomas and as a potential target for therapeutic measures. Neurosurg. Focus 20E24. [PubMed]
51. Jiang, Y., J. Wu, R. F. Keep, Y. Hua, J. T. Hoff, and G. Xi. 2002. Hypoxia-inducible factor-1α accumulation in the brain after experimental intracerebral hemorrhage. J. Cereb. Blood Flow Metab. 22689-696. [PubMed]
52. Jung, S. N., W. K. Yang, J. Kim, H. S. Kim, E. J. Kim, H. Yun, H. Park, S. S. Kim, W. Choe, I. Kang, and J. Ha. 2008. Reactive oxygen species stabilize hypoxia inducible factor-1 alpha protein and stimulate transcriptional activity via AMP-activated protein kinase in DU145 human prostate cancer cells. Carcinogenesis 29713-721. [PubMed]
53. King, A., M. A. Selak, and E. Gottlieb. 2006. Succinate dehydrogenase and fumarate hydratase: linking mitochondrial dysfunction and cancer. Oncogene 254675-4682. [PubMed]
54. Kohl, R., J. Zhou, and B. Brune. 2006. Reactive oxygen species attenuate nitric-oxide-mediated hypoxia-inducible factor-1α stabilization. Free Radic. Biol. Med. 401430-1442. [PubMed]
55. Koivunen, P., M. Hirsila, A. M. Remes, I. E. Hassinen, K. I. Kivirikko, and J. Myllyharju. 2007. Inhibition of hypoxia-inducible factor (HIF) hydroxylases by citric acid cycle intermediates: possible links between cell metabolism and stabilization of HIF. J. Biol. Chem. 2824524-4532. [PubMed]
56. Kozhukhar, A. V., I. M. Yasinska, and V. V. Sumbayev. 2006. Nitric oxide inhibits HIF-1α protein accumulation under hypoxic conditions: implication of 2-oxoglutarate and iron. Biochimie 88411-418. [PubMed]
57. Kwon, S. J., J. J. Song, and Y. J. Lee. 2005. Signal pathway of hypoxia-inducible factor-1α phosphorylation and its interaction with von Hippel-Lindau tumor suppressor protein during ischemia in MiaPaCa-2 pancreatic cancer cells. Clin. Cancer Res. 117607-7613. [PubMed]
58. Lacy, F., M. T. Kailasam, D. T. O'Connor, G. W. Schmid-Schonbein, and R. J. Parmer. 2000. Plasma hydrogen peroxide production in human essential hypertension: role of heredity, gender, and ethnicity. Hypertension 36878-884. [PubMed]
59. Laurent, A., C. Nicco, C. Chereau, C. Goulvestre, J. Alexandre, A. Alves, E. Levy, F. Goldwasser, Y. Panis, O. Soubrane, B. Weill, and F. Batteux. 2005. Controlling tumor growth by modulating endogenous production of reactive oxygen species. Cancer Res. 65948-956. [PubMed]
60. Leonardi, F., L. Attorri, R. D. Benedetto, A. D. Biase, M. Sanchez, F. P. Tregno, M. Nardini, and S. Salvati. 2007. Docosahexaenoic acid supplementation induces dose and time dependent oxidative changes in C6 glioma cells. Free Radic. Res. 41748-756. [PubMed]
61. Lopez-Lazaro, M. 2007. Dual role of hydrogen peroxide in cancer: possible relevance to cancer chemoprevention and therapy. Cancer Lett. 2521-8. [PubMed]
62. Lovstad, R. A. 2003. A kinetic study on iron stimulation of the xanthine oxidase dependent oxidation of ascorbate. Biometals 16435-439. [PubMed]
63. Lu, H., C. L. Dalgard, A. Mohyeldin, T. McFate, A. S. Tait, and A. Verma. 2005. Reversible inactivation of HIF-1 prolyl hydroxylases allows cell metabolism to control basal HIF-1. J. Biol. Chem. 28041928-41939. [PubMed]
64. MacKenzie, E. D., M. A. Selak, D. A. Tennant, L. J. Payne, S. Crosby, C. M. Frederiksen, D. G. Watson, and E. Gottlieb. 2007. Cell-permeating alpha-ketoglutarate derivatives alleviate pseudohypoxia in succinate dehydrogenase-deficient cells. Mol. Cell. Biol. 273282-3289. [PMC free article] [PubMed]
65. Makino, N., K. Sasaki, K. Hashida, and Y. Sakakura. 2004. A metabolic model describing the H2O2 elimination by mammalian cells including H2O2 permeation through cytoplasmic and peroxisomal membranes: comparison with experimental data. Biochim. Biophys. Acta 1673149-159. [PubMed]
66. Mansfield, K. D., R. D. Guzy, Y. Pan, R. M. Young, T. P. Cash, P. T. Schumacker, and M. C. Simon. 2005. Mitochondrial dysfunction resulting from loss of cytochrome c impairs cellular oxygen sensing and hypoxic HIF-alpha activation. Cell. Metab. 1393-399. [PMC free article] [PubMed]
67. Meng, Q., Y. T. Wong, J. Chen, and R. Ruan. 2007. Age-related changes in mitochondrial function and antioxidative enzyme activity in Fischer 344 rats. Mech. Ageing Dev. 128286-292. [PubMed]
68. Moeller, B. J., Y. Cao, C. Y. Li, and M. W. Dewhirst. 2004. Radiation activates HIF-1 to regulate vascular radiosensitivity in tumors: role of reoxygenation, free radicals, and stress granules. Cancer Cell 5429-441. [PubMed]
69. Nakamura, J., E. R. Purvis, and J. A. Swenberg. 2003. Micromolar concentrations of hydrogen peroxide induce oxidative DNA lesions more efficiently than millimolar concentrations in mammalian cells. Nucleic Acids Res. 311790-1795. [PMC free article] [PubMed]
70. Page, E. L., D. A. Chan, A. J. Giaccia, M. Levine, and D. E. Richard. 2008. Hypoxia-inducible factor-1α stabilization in nonhypoxic conditions: role of oxidation and intracellular ascorbate depletion. Mol. Biol. Cell 1986-94. [PMC free article] [PubMed]
71. Pan, Y., K. D. Mansfield, C. C. Bertozzi, V. Rudenko, D. A. Chan, A. J. Giaccia, and M. C. Simon. 2007. Multiple factors affecting cellular redox status and energy metabolism modulate hypoxia-inducible factor prolyl hydroxylase activity in vivo and in vitro. Mol. Cell. Biol. 27912-925. [PMC free article] [PubMed]
72. Park, S., X. You, and J. A. Imlay. 2005. Substantial DNA damage from submicromolar intracellular hydrogen peroxide detected in Hpx- mutants of Escherichia coli. Proc. Natl. Acad. Sci. USA 1029317-9322. [PubMed]
73. Peters, O., T. Back, U. Lindauer, C. Busch, D. Megow, J. Dreier, and U. Dirnagl. 1998. Increased formation of reactive oxygen species after permanent and reversible middle cerebral artery occlusion in the rat. J. Cereb. Blood Flow Metab. 18196-205. [PubMed]
74. Pollard, P. J., J. J. Briere, N. A. Alam, J. Barwell, E. Barclay, N. C. Wortham, T. Hunt, M. Mitchell, S. Olpin, S. J. Moat, I. P. Hargreaves, S. J. Heales, Y. L. Chung, J. R. Griffiths, A. Dalgleish, J. A. McGrath, M. J. Gleeson, S. V. Hodgson, R. Poulsom, P. Rustin, and I. P. Tomlinson. 2005. Accumulation of Krebs cycle intermediates and over-expression of HIF1α in tumours which result from germline FH and SDH mutations. Hum. Mol. Genet. 142231-2239. [PubMed]
75. Pouyssegur, J., and F. Mechta-Grigoriou. 2006. Redox regulation of the hypoxia-inducible factor. Biol. Chem. 3871337-1346. [PubMed]
76. Powell, F. L. 2003. Functional genomics and the comparative physiology of hypoxia. Annu. Rev. Physiol. 65203-230. [PubMed]
77. Qutub, A. A., and A. S. Popel. 2006. A computational model of intracellular oxygen sensing by hypoxia-inducible factor HIF1 alpha. J. Cell Sci. 1193467-3480. [PMC free article] [PubMed]
78. Qutub, A. A., and A. S. Popel. 2007. Three autocrine feedback loops determine HIF1 alpha expression in chronic hypoxia. Biochim. Biophys. Acta 17731511-1525. [PMC free article] [PubMed]
79. Schlegel, H. G. 1977. Aeration without air: oxygen supply by hydrogen peroxide. Biotechnol. Bioeng. 19413-424. [PubMed]
80. Schraufstatter, I., P. A. Hyslop, J. H. Jackson, and C. G. Cochrane. 1988. Oxidant-induced DNA damage of target cells. J. Clin. Investig. 821040-1050. [PMC free article] [PubMed]
81. Schroedl, C., D. S. McClintock, G. R. Budinger, and N. S. Chandel. 2002. Hypoxic but not anoxic stabilization of HIF-1alpha requires mitochondrial reactive oxygen species. Am. J. Physiol. Lung Cell Mol. Physiol. 283L922-L931. [PubMed]
82. Seaver, L. C., and J. A. Imlay. 2001. Hydrogen peroxide fluxes and compartmentalization inside growing Escherichia coli. J. Bacteriol. 1837182-7189. [PMC free article] [PubMed]
83. Selak, M. A., S. M. Armour, E. D. MacKenzie, H. Boulahbel, D. G. Watson, K. D. Mansfield, Y. Pan, M. C. Simon, C. B. Thompson, and E. Gottlieb. 2005. Succinate links TCA cycle dysfunction to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer Cell 777-85. [PubMed]
84. Selak, M. A., R. V. Duran, and E. Gottlieb. 2006. Redox stress is not essential for the pseudo-hypoxic phenotype of succinate dehydrogenase deficient cells. Biochim. Biophys. Acta 1757567-572. [PubMed]
85. Semenza, G. L. 2004. Hydroxylation of HIF-1: oxygen sensing at the molecular level. Physiology (Bethesda) 19176-182. [PubMed]
86. Senoo-Matsuda, N., K. Yasuda, M. Tsuda, T. Ohkubo, S. Yoshimura, H. Nakazawa, P. S. Hartman, and N. Ishii. 2001. A defect in the cytochrome b large subunit in complex II causes both superoxide anion overproduction and abnormal energy metabolism in Caenorhabditis elegans. J. Biol. Chem. 27641553-41558. [PubMed]
87. Srinivas, V., I. Leshchinsky, N. Sang, M. P. King, A. Minchenko, and J. Caro. 2001. Oxygen sensing and HIF-1 activation does not require an active mitochondrial respiratory chain electron-transfer pathway. J. Biol. Chem. 27621995-21998. [PubMed]
88. Stone, J. R., and S. Yang. 2006. Hydrogen peroxide: a signaling messenger. Antioxid. Redox Signal. 8243-270. [PubMed]
89. Sumbayev, V. V., and I. M. Yasinska. 2006. Peroxynitrite as an alternative donor of oxygen in HIF-1α proline hydroxylation under low oxygen availability. Free Radic. Res. 40631-635. [PubMed]
90. Szatrowski, T. P., and C. F. Nathan. 1991. Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res. 51794-798. [PubMed]
91. Szeto, S. S., S. N. Reinke, B. D. Sykes, and B. D. Lemire. 2007. Ubiquinone-binding site mutations in the Saccharomyces cerevisiae succinate dehydrogenase generate superoxide and lead to the accumulation of succinate. J. Biol. Chem. 28227518-27526. [PubMed]
92. Tanriverdi, T., H. Hanimoglu, T. Kacira, G. Z. Sanus, R. Kemerdere, P. Atukeren, K. Gumustas, B. Canbaz, and M. Y. Kaynar. 2007. Glutathione peroxidase, glutathione reductase and protein oxidation in patients with glioblastoma multiforme and transitional meningioma. J. Cancer Res. Clin. Oncol. 133627-633. [PubMed]
93. Teixeira, H. D., R. I. Schumacher, and R. Meneghini. 1998. Lower intracellular hydrogen peroxide levels in cells overexpressing CuZn-superoxide dismutase. Proc. Natl. Acad. Sci. USA 957872-7875. [PubMed]
94. Tuttle, S. W., A. Maity, P. R. Oprysko, A. V. Kachur, I. S. Ayene, J. E. Biaglow, and C. J. Koch. 2007. Detection of reactive oxygen species via endogenous oxidative pentose phosphate cycle activity in response to oxygen concentration: implications for the mechanism of HIF-1α stabilization under moderate hypoxia. J. Biol. Chem. 28236790-36796. [PubMed]
95. Vangeison, G., D. Carr, H. J. Federoff, and D. A. Rempe. 2008. The good, the bad, and the cell type-specific roles of hypoxia inducible factor-1 alpha in neurons and astrocytes. J. Neurosci. 281988-1993. [PubMed]
96. Vaux, E. C., E. Metzen, K. M. Yeates, and P. J. Ratcliffe. 2001. Regulation of hypoxia-inducible factor is preserved in the absence of a functioning mitochondrial respiratory chain. Blood 98296-302. [PubMed]
97. Vordermark, D., and J. M. Brown. 2003. Evaluation of hypoxia-inducible factor-1α (HIF-1α) as an intrinsic marker of tumor hypoxia in U87 MG human glioblastoma: in vitro and xenograft studies. Int. J. Radiat. Oncol. Biol. Phys. 561184-1193. [PubMed]
98. Wang, G. L., B. H. Jiang, E. A. Rue, and G. L. Semenza. 1995. Hypoxia-inducible factor 1 is a basic-helix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc. Natl. Acad. Sci. USA 925510-5514. [PubMed]
99. Wartenberg, M., F. C. Ling, M. Muschen, F. Klein, H. Acker, M. Gassmann, K. Petrat, V. Putz, J. Hescheler, and H. Sauer. 2003. Regulation of the multidrug resistance transporter P-glycoprotein in multicellular tumor spheroids by hypoxia-inducible factor (HIF-1) and reactive oxygen species. FASEB J. 17503-505. [PubMed]
100. Wellman, T. L., J. Jenkins, P. L. Penar, B. Tranmer, R. Zahr, and K. M. Lounsbury. 2004. Nitric oxide and reactive oxygen species exert opposing effects on the stability of hypoxia-inducible factor-1α (HIF-1α) in explants of human pial arteries. FASEB J. 18379-381. [PubMed]
101. Williams, N., and J. Yandell. 1982. Outer-sphere electron transfer of ascorbate anions. Aust. J. Chem. 351133-1144.
102. Xia, C., Q. Meng, L. Z. Liu, Y. Rojanasakul, X. R. Wang, and B. H. Jiang. 2007. Reactive oxygen species regulate angiogenesis and tumor growth through vascular endothelial growth factor. Cancer Res. 6710823-10830. [PubMed]
103. Yang, Z. Z., A. Y. Zhang, F. X. Yi, P. L. Li, and A. P. Zou. 2003. Redox regulation of HIF-1alpha levels and HO-1 expression in renal medullary interstitial cells. Am. J. Physiol. Renal Physiol. 284F1207-F1215. [PubMed]
104. Yankovskaya, V., R. Horsefield, S. Tornroth, C. Luna-Chavez, H. Miyoshi, C. Leger, B. Byrne, G. Cecchini, and S. Iwata. 2003. Architecture of succinate dehydrogenase and reactive oxygen species generation. Science 299700-704. [PubMed]
105. Zhong, W., T. Yan, R. Lim, and L. W. Oberley. 1999. Expression of superoxide dismutases, catalase, and glutathione peroxidase in glioma cells. Free Radic. Biol. Med. 271334-1345. [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)