THE EMERGING ROLE OF THE THIOREDOXIN SYSTEM IN ANGIOGENESIS (original) (raw)

. Author manuscript; available in PMC: 2011 Nov 1.

Published in final edited form as: Arterioscler Thromb Vasc Biol. 2010 Aug 26;30(11):2089–2098. doi: 10.1161/ATVBAHA.110.209643

Abstract

Notwithstanding a multitude of studies, the mechanisms of angiogenesis remain incompletely understood. Increasing evidence suggests that cellular redox homeostasis is an important regulator of angiogenesis. The thioredoxin (TRX) system functions as an endogenous antioxidant that can exert influence over endothelial cell (EC) function via modulation of cellular redox status. It has become apparent that the cytosolic thioredoxin-1 (TRX1) isoform participates in both canonical and novel angiogenic signaling pathways, and may represent an avenue for therapeutic exploitation. Recent studies have further identified a role for the mitochondrial isoform thioredoxin-2 (TRX2) in ischemia-induced angiogenesis. Thioredoxin interacting protein (TXNIP) is the endogenous inhibitor of TRX redox activity that has been implicated in growth factor-mediated angiogenesis. As TXNIP is strongly induced by glucose, this molecule could be of consequence to disordered angiogenesis manifest in diabetes mellitus. This review will focus on data implicating the TRX system in EC homeostasis and angiogenesis.

Keywords: Thioredoxin, Thioredoxin Interacting Protein, Angiogenesis, Ischemia, Endothelial cell

Introduction

Angiogenesis is a tightly controlled process originally observed during embryogenesis as new blood vessels arise from the primitive vascular plexus. Postnatal angiogenesis is now a recognized adaptive response to ischemia and hypoxia, in which the thioredoxin (TRX) system is increasingly implicated.1, 2 Angiogenesis may also be dysregulated - a key feature of tumor growth and other diseases, including atherosclerosis and the vascular complications arising from diabetes mellitus. Despite great promise, the ability to regulate angiogenesis therapeutically remains a largely unrealized goal.

The endothelial cell (EC) is central to angiogenesis. Notably, the endothelial cell senses low oxygen tensions, in part due to the inhibition of prolyl hydroxylase. Under normoxic conditions this enzyme hydroxylates proline groups in the oxygen degradation domain of the transcription factor hypoxia inducible factor-1α (HIF1α), leading to its degradation. As hypoxia inhibits prolyl hydroxylase, HIF1α is stabilized, and can translocate to the nucleus to induce the transcription of angiogenic genes, such as vascular endothelial growth factor (VEGF).3, 4

Cellular redox homeostasis is also an important regulator of angiogenesis. In ECs, the NADPH oxidases generate reactive oxygen species (ROS), which at low levels are important signaling moieties but at high levels have deleterious consequences.5–7 More recent work directly implicates the thioredoxin (TRX) system and its endogenous inhibitor, thioredoxin interacting protein (TXNIP) in EC homeostasis and key angiogenic processes including EC migration, proliferation and survival.8–11 Herein, we review the TRX system with a particular focus on the emerging evidence of its critical role in the modulation of angiogenesis.

The Thioredoxin System

In conjunction with glutathione and glutaredoxin, the TRX system maintains the reducing environment of the cell and detoxifies ROS. The antioxidant activity of the TRX system is primarily exerted by peroxiredoxins, which are reduced by TRX to enable scavenging of ROS.12 The TRX system is reproduced in distinct cellular compartments in the cytosol and nucleus is found the traditionally described thioredoxin-1 (TRX1) isoform, whereas in the mitochondria is found the more recently identified thioredoxin-2 (TRX2) isoform. Homozygous knockout of either isoform in mouse is embryonically lethal13, 14 and TRX molecules are retained throughout evolution, indicating that the TRX system is essential for life.15–18 Both isoforms are ubiquitously expressed and contain the highly conserved (Trp-Cys-Gly-Pro-Cys) catalytic motif that reduces oxidized proteins and ROS. In concert with these oxidoreductases are the selenium-dependent enzymes, thioredoxin reductase-1 (TRXR1) and thioredoxin reductase-2 (TRXR2), which utilize the electron donor NADPH to regenerate the TRX isoforms (Figure 1A). The ability of both TRX1 and TRX2 to undergo reversible oxidation may be modulated by TXNIP, which interacts with the catalytic motif to form a stable mixed disulfide that inhibits TRX activity. This property has led to the notion that TXNIP is the endogenous inhibitor of the TRX system, a finding that has been confirmed in multiple cell types in vitro19–21 albeit not consistently recapitulated in vivo.22, 23

Figure 1. The Thioredoxin System.

Figure 1

(A) Thioredoxin (TRX) in the oxidized (TRX-S2) or reduced (TRX-(SH)2) state. In the reduced state, TRX directly reduces disulfides in oxidized substrate proteins (PR--S2). The resultant oxidation of TRX in this process is reversible and maintained by thioredoxin reductase (TRXR) and the electron donor NADPH. (B) Thioredoxin interacting protein (TXNIP) can form a mixed disulfide with reduced thioredoxin-1 (TRX1), inhibiting the ability of TRX1 to reduce disulfides of other protein substrates and/or undergo reversible oxidation.

The ability of TRX1 and TRX2 to interact with cysteine residues across a host of proteins entrenches the TRX system in a diverse range of cellular processes supra to its canonical role as a cytoprotective measure against oxidative stress. These include protein structure/folding, reductive and metabolic enzymes, energy utilization, transcription factors and immune modulation.24 The report that TRX1 was associated with increased tumor angiogenesis in a murine xenograft model11 has since led a number of investigators to explore the role of the TRX system in angiogenesis. Recent studies involving transgenic animal models and whole genome microarray analyses have enhanced our understanding of the role for the TRX system in angiogenesis, although significant gaps in our knowledge also remain.

Thioredoxin-1 and Angiogenesis

TRX1is a 12 kDa ubiquitous oxidoreductase originally isolated in Escherichia coli as a hydrogen donor for ribonucleotide reductase16 that is the rate-limiting enzyme in DNA synthesis (Figure 2). Homozygous knockout of Trx1 results in absorption of conceptuses after implantation, with cells derived from the inner cell mass unable to proliferate ex vivo.13 Conversely, mice overexpressing Trx1 develop normally, although they exhibit greater resistance to ischemic neuronal injury25 and adriamycin-induced cardiotoxicity.26 These data suggest that TRX1 expression conveys a cytoprotective effect to a variety of stressors, enhancing cell survival and function. TRX1may also regulate key EC activities relevant to physiological and tumor angiogenesis, including EC migration, proliferation and vascular network formation, and apoptosis and cell survival.

Figure 2. The Thioredoxin System and Angiogenesis.

Figure 2

Thioredoxin-1 (TRX1) and its endogenous inhibitor Thioredoxin Interacting Protein (TXNIP) are involved in multiple signaling pathways and cellular processes that confluence in mediation of angiogenesis. The TRX1/TXNIP system can modulate cell growth and proliferation by transcriptional mechanisms such as redox factor-1 (REF1) and NF-kB, JAB1/p27kip1 translocation, DNA synthesis via ribonucleotide reductase and energy metabolism/glycolysis through reductive inhibition of PTEN. Akt signaling facilitates cell survival. Apoptosis is modulated through the interaction of either cytosolic TRX1 or mitochondrial thioredoxin-2 (TRX2) with apoptosis signaling kinase-1 (ASK1) and competitive inhibition by TXNIP. In endothelial cells TRX1 prevents von Hippel-Lindau (pVHL) mediated degradation of the transcription factor hypoxia inducible factor-1α (HIF1α) leading to induction of vascular endothelial growth factor (VEGF) expression. The VEGF signaling cascade results in the activation of endothelial nitric oxide synthase (eNOS) and nitric oxide (NO) release that facilitates key angiogenic events.

TRX1affects a number of transcriptional pathways involved in EC angiogenic function (Figure 2). Activation of the transcription factor NF-κB modulates a variety of angiogenesis-related proteins and genes, including Akt activation of VEGF that is critical for EC migration, proliferation and survival.27, 28 In the cytoplasm TRX1 blocks NF-κB activation by preventing its nuclear translocation, whereas in the nucleus TRX1 increases NF-κB transactivation through reduction of the Cys62 residue of the p50 DNA binding motif.29, 30 Thus TRX1 can modulate NF-κB activity to up-regulate or downregulate angiogenic gene transcription. Redox factor-1 (REF1) is a transcriptional co-activator in the HIF1α complex that binds to the VEGF promoter31, 32 and TRX1 has been demonstrated to enhance the binding activity of REF1 to the transcription factors fos, jun and AP1.33–35 The action of HIF1α is dependent on TRX1 redox activity with overexpression of TRX1 leading to increased HIF1α, VEGF, and nitric oxide synthase-2 expression in cancer cell lines.11 Nitric oxide (NO) production from VEGF signaling and endothelial nitric oxide synthase activation directly facilitates pro-angiogenic events including migration and proliferation.36, 37 The production of NO further enhances TRX activity by S-nitrosylation at Cys69, thereby increasing post-transcriptional pro-angiogenic signaling.38 The ability of TRX1 to interact with these transcription factors and undergo favorable post-translational modification is consistent with a central role for the TRX system in angiogenesis signaling.

Apoptosis and cell survival are important components of angiogenesis and vascular remodeling that can be modulated by the TRX system and ROS. TRX1interacts with apoptosis signaling kinase-1 (ASK1), preventing its homodimerization and suppressing its pro-apoptotic function by flagging ASK1 for ubiquitin-mediated degradation (Figure 2).39, 40 The interaction between TRX1 and ASK1 is dependent on TRX1 existing in the reduced state, with oxidation of TRX1 causing dissociation from ASK1.39, 40 Activation of ASK1 is reinforced by sustained elevated concentrations of ROS that induce cathepsin-D degradation of TRX1, while short-term exposure to low concentrations of ROS induces TRX1 expression and is anti-apoptotic.41 In response to ROS and oxidative damage, ASK1 activates JNK and p38 mitogen activated protein kinases leading to EC apoptosis and death. Indeed, when ASK1was overexpressed in human ECs, VEGF-induced cell migration and vascular network formation were inhibited.42 However, disparate phenotypes are reported when ASK1 is deleted in mouse. In one study, Ask1 null mice exhibited increased limb perfusion and VEGF receptor-2 signaling in a hindlimb ischemia model.43 By contrast another study demonstrated Ask1 deletion to reduce ischemia-induced neovascularization and VEGF expression in the hindlimb ischemia model.44 To date, the reasons underlying these conflicting observations are unclear.

TRX1and VEGF are overexpressed in a variety of cancers including mesothelioma,45 lung,46 colorectal,47, 48 liver,49 prostate,50 breast11 and leukemia.51 Elevated levels of TRX1 correlate with poor prognosis.46, 48, 52 When MCF7 breast carcinoma cells overexpressing TRX1 cells were xenotransplanted in mouse there was increased formation of solid tumors, while cells overexpressing a redox-impaired C32S/C35S TRX1 mutant had reduced tumorigenicity.53 Murine lymphoma xenografts overexpressing TRX1 also formed solid tumors with enhanced capillary networks and increased VEGF levels.11 The influence of TRX1 on hypoxia-mediated EC angiogenic signaling in the cancer milieu is illustrated by pharmacologic modulation of TRX activity. The TRX1 inhibitor AW464 was found to impair TRX1 and HIF1α activity, and specifically reduce EC proliferation and vascular network formation in vitro.54 Treatment of breast and colon cancer cells with the TRX1 signaling inhibitors, pleurotin and PX-12, decreased TRX1 levels and activity, inhibited increases in HIF1α protein and reduced levels of VEGF - ultimately resulting in decreased tumor capillary density.55 Blockade of TRX1 using the putative anticancer drug PX-12 has shown promising results in phase I clinical trials with decreased plasma concentrations of the angiogenic growth factor VEGF noted in PX-12 treated patients.56

Targeted pharmacological induction of TRX1 activity may prove useful in therapeutic modulation of angiogenesis in cardiovascular disease. For example, in a rat myocardial infarct model the red-wine compound resveratrol was demonstrated to enhance ischemia-mediated angiogenesis via TRX1 induction.57 Subsequent increases in expression of VEGF and the stress response enzyme heme oxygenase-1 were associated with enhanced angiogenesis and improved cardiac function.57 Resveratrol or exogenous TRX1 administration alone also increased EC vascular network formation in a Matrigel assay in vitro. In a similar myocardial infarction model in diabetic rats, enforced expression of TRX1 in the myocardium by adenovirus led to a reduction in oxidative stress and apoptosis, with improvements in cardiac function.58 Interestingly, these findings were associated with increased angiogenesis, arteriogenesis, and heme oxygenase-1 and VEGF expression. These initial studies indicate that modulation of the TRX system in ischemic myocardium represents an avenue of future investigation.

Mitochondrial Thioredoxin-2 and Ischemia-Induced Neovascularization

Mitochondrial dysfunction and excessive ROS generation are apparent in cardiovascular disease, leading to perturbations in endothelial function and apoptotic signaling, in part by inhibiting NO bioavailability. As deletion of TRX2 is embryonically lethal coinciding with mitochondrial maturation stage14 TRX2 may play a regulatory role in the spectrum of endothelial function, including angiogenesis. Indeed, in EC-specific Trx2 transgenic mice, both NO bioavailability and mitochondrial ROS scavenging were increased and there was an overall reduction in oxidative stress.59 When crossed with Apolipoprotein E null mice atherosclerotic lesions were reduced. Subsequent investigations demonstrated that EC-specific Trx2 transgenesis enhanced both angiogenesis and arteriogenesis in a murine model of hindlimb ischemia.8 Induction of ASK1-dependent apoptosis by ROS was reduced in ECs isolated from Trx2 transgenics. When Trx2 transgenics were crossed with eNOS knockout mice (that have a severely impaired ischemia-induced angiogenic response) Trx2 overexpression dramatically abrogated the deleterious eNOS null phenotype. These data demonstrate a role for TRX2 in facilitating ischemia-induced angiogenesis. The recent observation that TRX2 reduces hypertension and ROS-generating NADPH oxidases emphasizes the importance of the TRX system in the broader context of endothelial biology.60

Thioredoxin Reductases

Thioredoxin reductase-1 is a selenium dependent enzyme localized to the cytosol. In mouse, the TrxR1 null phenotype is embryonically lethal61 and siRNA-mediated knockdown of TRXR1 renders ECs susceptible to oxidative stress and apoptosis.62 The selenoprotein is also implicated in angiogenesis. Low levels of selenium have been shown to increase VEGF expression and tumor angiogenesis in vivo,63 while inhibition of TRXR1 activity in selenium replete ECs increases VEGF production leading to enhanced migration, proliferation and tubulogenesis in vitro. (this seems inconsistent with the angiogenic effects of TRXR1) 64 Congruent with these observations, the TRXR1 inhibitor PX-916 displays marked anti-cancer activity with reductions in TRX1 activity, HIF1α and VEGF levels in tumors.65 The TRXR1 inhibitor auranofin has been implicated in reduced (?)EC damage and anti-angiogenic activity in rheumatoid arthritis,66, 67 while other gold compounds inhibited angiogenesis in zebrafish (mention mechanism).68 No studies have explicitly examined the role of mitochondrial TRXR2 in angiogenesis, although it is speculated to have roles in the endothelium.69

Thioredoxin Interacting Protein as a Modulator of Angiogenesis Signaling

Thioredoxin interacting protein (TXNIP) was isolated as a 50 kDa Vitamin-D3 inducible protein (Vitamin-D3 upregulated protein-1) by yeast two-hybrid screen for TRX1 interacting proteins in cervical cancer cells.70 This interaction requires that TRX1 is in the reduced state71, 72 and is significant in that TXNIP is the only currently known endogenous cofactor that inhibits TRX1 and TRX2 redox activity.72–74 Patwari et al, elucidated that an intra-molecular disulfide bond between Cys63 and Cys247 in TXNIP is susceptible to disulfide exchange with reduced TRX1 at Cys247, leading to a stable mixed disulfide (Figure 1B).71, 75 While yet to be confirmed by crystallography this mixed disulfide at the catalytic motif would significantly impair the ability of TRX1 to interact with a number of pro-angiogenic molecules, leading to modulation of angiogenic signaling.

In cancer biology, TXNIP has been described as a metastasis suppressor76 that is strongly downregulated in a variety of tumor tissues and cell lines.77–80 Mice deficient in TXNIP also have increased incidence of hepatocellular carcinoma.81 Conversely, when TXNIP was over-expressed in the HEp-2 epithelial cell line, xenotransplants displayed reduced tumorigenesis in mouse.82 TXNIP is known to interact with a variety of molecules to modulate cell cycle progression and growth (Figure 2). Overexpression of TXNIP leads to an enhanced interaction with α-importin-1 (RCH1) and nuclear translocation of the complex suppressing growth activity in MCF7 breast cancer cells.83 Thioredoxin interacting protein can also competitively inhibit Jab1-mediated nuclear export of p27kip1 to the cytoplasm leading to cell cycle arrest84 (Figure 2), which is critical to halting tumor progression.85, 86 In the cytoplasm TXNIP interacts with TRX1 and inhibits its nuclear translocation, blocking TRX1-dependent gene transcription87 and further influencing the expression of cell death and survival genes.88 The cell cycle inhibitor and tumor suppressor p21(WAF1) also suppresses TXNIP gene transcription leading to increases in TRX activity, EC migration, vascular network formation and invasion.89

Recent work from our laboratory revealed a central role for TXNIP in angiogenic growth factor-mediated EC migration.10 To gain insights into the transcriptional events necessary for angiogenesis, we used whole genome cDNA microarray analysis to identify and characterize concordant gene expression programs triggered by administration of three angiogens: VEGF, basic fibroblast growth factor and nicotine. In human microvascular ECs, TXNIP was consistently repressed by all three angiogenic factors. Co-repression of TXNIP was associated with stimulation of TRX activity. Interestingly, hexamethomium, a nicotinic acetylcholine receptor antagonist, abrogated growth factor-related effects on TXNIP and TRX1, suggesting a role for the nicotinic acetylcholine receptor in growth factor signaling via modulation of TXNIP/TRX1. Gene knockdown of TRX1 by siRNA abrogated the stimulatory effects of all three angiogenic factors on EC migration. Gene silencing of TXNIP alone, without angiogenic factors, induced TRX activity and profoundly stimulated EC migration a pro-angiogenic effect. More recently, we explored the effect of TRX1/TXNIP modulation on EC migration and vascular network formation using a Matrigel assay. Downregulation of TRX1 by siRNA in human coronary artery ECs inhibited migration and vascular network formation in a Matrigel assay, while siRNA inhibition of TXNIP increased migration and vascular network formation.90

Fluid shear stress is a principal trigger for arteriogenesis. Normal laminar flow has been demonstrated to inhibit TXNIP expression and increase TRX activity in rabbit aorta and ECs.91 Shear stress also enhances S-nitrosylation (and thereby the activity) of TRX1 and increases the production of NO in ECs92 leading to suppression of TXNIP expression in smooth muscle cells.20 These observations implicate the TRX system, particularly TXNIP, as a regulator of biomechanical signaling in the vasculature. Interestingly, ischemia was found to induce TXNIP expression in a rat myocardial infarction model, while DNAzyme to TXNIP led to reduced apoptosis and improved cardiac function.93 The mechanisms underlying the relationship between TXNIP and endothelial function require further investigation.

Despite the demonstration that TXNIP inhibits TRX activity in multiple cell types in vitro including ECs,10 vascular smooth muscle cells,20 breast cancer cells,21 lens epithelial cells94 and mesangial cells,95 and in aortic tissue ex vivo91 this observation has not been consistently demonstrated in vivo. This is particularly true of the total and tissue-specific TXNIP knockout mice,22, 23, 96, 97 and TXNIP-deficient mouse models available,98 which demonstrate no significant changes in TRX activity. The inability of current techniques to specifically and sensitively measure TRX redox moieties in vivo could explain this remaining controversy. Indeed, the glutathione system represents the primary redox buffer, and competes with the TRX system for reducing equivalents (NADPH).99 While some investigations in the TXNIP deficient and knockout mice have noted minor changes in the reduced to oxidized glutathione ratio and NADP/NADPH ratio,96, 100 others have not22, 23 and these discrepancies highlight a major outstanding issue in the field that needs to be addressed.

In diabetes mellitus, hyperglycemia induces ROS that contribute to the pathogenesis of diabetic vascular complications. Reactive oxygen species have profound effects on the vasculature, leading to endothelial dysfunction, accelerated atherosclerosis, microvascular and peripheral arterial disease.101 A striking characteristic of the diabetic state is the heterogeneity of angiogenic dysregulation. For example, VEGF is upregulated in the diabetic eye whereas VEGF signaling is impaired in the peripheral arterial and microvascular environments.102 To date, the mechanisms underlying dysregulated angiogenesis and impaired VEGF signaling in diabetes mellitus remain unclear.

Microarray studies identified TXNIP as the gene most strongly induced by glucose in pancreatic beta cells.103 A carbohydrate response element within the TXNIP promoter underlies this striking feature.104 Thioredoxin interacting protein is induced by sugars in a variety of other cell types including vascular smooth muscle cells,105 ECs,90, 106 hepatoma,107 and breast cancer.21 It is associated with increased apoptosis in pancreatic beta islets108, 109 and cardiomyocytes by competitive inhibition of the TRX1/ASK-1 complex.93, 110, 111 Hyperglycemic induction of TXNIP in vascular smooth muscle cells and ECs in vitro represses TRX activity, induces ROS accumulation,105 and inhibits vascular network formation.90 Knockdown of TXNIP by siRNA or adenoviral overexpression of TRX1 abrogates glucose-induced ROS accumulation in smooth muscle cells.105 Furthermore, in streptozotocin-induced diabetic rats, TXNIP expression was increased in the vasculature and TRX activity reduced.105 As impaired angiogenesis and endothelial dysfunction are the hallmark of diseases such as diabetes mellitus, hyperglycemia-mediated induction of TXNIP could have important consequences for ROS-induced endothelial dysfunction, impaired angiogenesis and dysregulated biomechanical signaling.

We recently identified a critical role for hyperglycemia-mediated induction of TXNIP in diabetes-related impairment of ischemia-mediated angiogenesis using a murine model of hindlimb ischemia.(Dunn et al, TBA). Moreover, TXNIP knockdown to non-diabetic levels rescued diabetes-related impairment of ischemia-mediated angiogenesis and limb functional recovery. Diabetes-related inhibition of VEGF was also attenuated by TXNIP knockdown. The demonstration that adenoviral overexpression of TRX1 in the infarcted myocardium of diabetic rats enhances angiogenesis and improves cardiac function58 is consistent with our proposal that dysregulation of the TRX system is a critical mechanism underlying disordered angiogenesis in diabetes mellitus. Interestingly, there appears to be a link between diabetes mellitus and TRX2. In rats with streptozotocin-induced diabetes, aortic expression of TRX2 is reduced and in vitro siRNA knockdown of TRX2 in ECs results in increased glucose toxicity with up-regulation of the endothelial glucose transporter, GLUT1.112 As previously mentioned, the interaction between TXNIP and TRX2 within the mitochondria is particularly intriguing78 given the metabolic aberrations of diabetes mellitus.

The Thioredoxin System: Redox Modulation of Angiogenesis?

Despite structural and functional similarities re-iterated between the TRX systems such as redox metabolism and apoptosis, ablation of either the respective TRX1/2 or TRXR1/2 isoforms results in cessation of embryogenesis,13, 14, 61, 113 surprisingly demonstrating little functional redundancy between the two systems. Thioredoxin-1 has been shown to stabilize HIF1α protein by causing the dissociation of von Hippel-Lindau protein thereby preventing HIF1α ubiquitination and degradation (Figure 2).9 Yet opposing roles for TRX1 and TRX2 have been described in HEK-293 cells submitted to hypoxic treatment.114, 115 Overexpression of TRX1 resulted in increased HIF1α accumulation and activity under hypoxia, normoxia and NO treatments. Conversely, when TRX2 was overexpressed, hypoxia-induced HIF1α accumulation and activity were reduced with increases in mitochondrial ROS.114 Overexpression of TRXR2 also reduced NO-induced HIF1α accumulation and activity.115 As mitochondria facilitate ATP production the investigators went on to demonstrate that TRX1 expression enhances ATP levels thereby increasing HIF1α protein translation, while TRX2 or TRXR2 overexpression attenuated ATP levels and protein translation.115 These findings suggest an association between TRX2 and the mitochondrial respiratory chain, with ROS entering the cytosol and counterbalancing TRX1 enhanced translation of HIF1α protein.115 Recent studies in pancreatic beta cells identify TXNIP as a mechanism underlying the TRX1/TRX2 balance of power.74 Under normal conditions Saxena et al, found TXNIP to be strongly localized to the nucleus. However, under conditions of oxidative stress TXNIP is shuttled to the mitochondria where it out-competes ASK1 binding to TRX2 leading to induction of apoptosis via the mitochondrial pathway.

The divergent effects of the reductases TRXR1 and TRXR2 strengthen the concept that cellular homeostasis requires a delicate balance between TRX1 and TRX2. In cardiac development a similar balancing act is evident for TRXR1 and TRXR2. Embryos null for TrxR1 are not viable in vivo and ECs isolated from embryos do not proliferate in vitro.61 However, isolated cardiomyocytes are not affected and cardiac specific deletion does not affect embryo development.61 By contrast, cardiac-specific TrxR2 deletion leads to cardiomyopathy and death in the embryo.116 The independent features of the two TRX systems are intriguing and suggests that redox control of hypoxia- and metabolically-mediated cellular function extends to development as well as cellular physiology and disease.

Conclusion

The paradigm of TRX1 as a simple redox regulator has shifted in recent times with TRX activity proven to extend to many aspects of vascular function, including angiogenesis. In the EC, TRX1 exerts pro-angiogenic effects throughout a broad range of cellular activities relevant to angiogenic function, including transcription, post-translational modification, migration, proliferation, vascular network formation, apoptosis and intracellular signaling. Mitochondrial TRX2 also plays a critical role in ischemia-induced angiogenesis and arteriogenesis. These potent effects may be modulated by TXNIP in vitro although the effect of TXNIP on TRX activity in vivo requires further clarification. Intriguingly, normalization of hyperglycemia-induced TXNIP expression to non-diabetic levels rescues diabetes-related impairment of ischemia-mediated angiogenesis. The ability of TXNIP to further regulate cell cycle progression and metastasis clearly identifies TXNIP as an important target for therapeutic angiogenesis. The onus of future investigations rests on precisely delineating the inter-relationship of these TRX system members so that effective modulation of this system as an angiogenic therapy can be realized.

Acknowledgments

The authors would like to acknowledge funding support from the National Health and Medical Research Council of Australia (GRANT No. 512299). Dr Louise Dunn is the recipient of a National Health and Medical Research Council of Australia Postdoctoral Training Fellowship (GRANT No. 537537). Dr. Cooke is supported by grants from the National Institutes of Health (RC2HL103400 and 1U01HL100397), and the Tobacco Related Disease Research Program of the University of California (18XT-0098).

References