Loss of Redox Factor 1 Decreases NF-κB Activity and Increases Susceptibility of Endothelial Cells to Apoptosis
Objective— The aim of this project was to test the hypothesis that redox factor 1 (Ref-1) was a critical upstream determinant of NF-κB–dependent survival signaling pathways in the vessel wall.
Methods and Results— Aortas from hemizygous transgenic mice harboring a single allele of Ref-1 exhibited a significant loss in NF-κB DNA binding activity. The NF-κB–dependent survival gene A20 was significantly downregulated in aortas of hemizygous Ref-1 mice, whereas IAP-2 was unchanged. Overexpression of A20 rescued cells from tumor necrosis factor (TNF)-induced apoptosis, suggesting that the loss of A20 in Ref-1 hemizygotes may be a rate-determining step in endothelial cell fate. Deletion of the previously defined redox-sensitive or the AP endonuclease domains of Ref-1 significantly decreased NF-κB transcriptional activation and endothelial cell survival. Furthermore, TNF-induced apoptosis was significantly potentiated in endothelial cells after delivery of Morpholino antisense oligodeoxynucleotides targeted to Ref-1. Loss of the redox-sensitive domain blocked the ability of Ref-1 to reduce p50; however, loss of the endonuclease domain did not effect p50 reduction, suggesting alternative mechanisms of action of Ref-1 on NF-κB activity.
Conclusions— These findings establish a role for Ref-1 as an upstream determinant of NF-κB and A20-dependent signaling and endothelial survival in the vessel wall.
Recent insights into the underlying mechanisms of vascular disease have highlighted the central role of redox signaling, with a particular emphasis on NF-κB.1–3 Work from our group and others has demonstrated that Redox factor 1 (Ref-1) regulates DNA binding of multiple transcription factors, including NF-κB.4–6 Specifically, upregulation of Ref-1 results in a significant increase in NF-κB activity in vascular endothelial cells.4 Moreover, recent work suggests that the reduction of the Cys-62 residue of p50 is essential for DNA binding activity of NF-κB, and that Ref-1 is involved in the reduction of this critical residue.7 Ref-1 (also known as APE, APEX, and HAP-1) was originally identified as a dual-acting transcription factor containing a redox-sensitive domain, as well as a DNA repair domain.5,6,8–12 Ref-1–null mice die during embryonic development and exhibit an abundance of fragmented and pyknotic nuclei, suggesting that Ref-1 plays a critical role in cell survival.12
Based on these findings, we hypothesized that Ref-1 was a rate-determining factor governing NF-κB activity and cell fate in the intact vasculature. To test this hypothesis, we used hemizygous mice harboring a single allele of Ref-1. We present evidence that hemizygosity for Ref-1 results in a significant reduction in NF-κB activity in the aorta in parallel with a significant loss of the NF-κB survival gene A20. A complementary Morpholino antisense approach to decrease Ref-1 expression in vitro resulted in a congruent loss of endothelial cell survival. Further analysis suggested that both the redox sensing domain and the AP endonuclease/DNA repair domain of Ref-1 are important in the regulation of endothelial cell survival. Using a biochemical approach, we demonstrate that the redox-sensing domain of Ref-1 is sufficient to reduce p50 and inhibit NF-κB activity, whereas the AP endonuclease domain partially inhibits NF-κB activity despite the fact that this region is not necessary to reduce p50. Taken together, these findings suggest that Ref-1 is an important mediator of NF-κB signaling and survival in vascular endothelial cells.
Please see http://atvb.ahajournals.org for the complete Methods section.
Bovine pulmonary endothelial cells (CPAE) (American Type Tissue Culture) were grown in DMEM, 10% fetal bovine serum, and 1% penicillin/streptomycin (Gibco, Life Technologies) and used between passages 18 to 26.
Ref-1+/− (hemizygous transgenic mice harboring a single allele of Ref-1) mice and wild-type controls in a C57Bl6 background were a kind gift from Dr Tom Curran (St Jude Children’s Research Hospital, Memphis, Tenn).12 Briefly, the genotyping protocol to identify Ref-1+/− mice included the use of 4 primers: Ref-D4, CAGCCTATGTTCCAATGCAGG; Ref-D5, TGTAAGCGTAAGCAGTGTTGGGG; Ref D18, CCGAGTGACGAGCTGTAACCG; and PGK.D1, GAGGAGTAGAAGTGGCGCGAAGG.
Real-Time Quantitative Polymerase Chain Reaction
Real-time quantitative polymerase chain reaction was performed as we have previously described.13 Primers and conditions are listed in online supplement.
All transfections with expression vectors were performed in CPAE cells using a lipid-based transfection strategy (Effectene) as previously described.4.
Morpholino antisense oligodeoxynucleotides were transfected into cells according to the manufacturer’s directions (GeneTools LLC.). Briefly, oligodeoxynucleotides (1.4 μmol/L final concentration) were bound electrostatically to weakly basic ethoxylated polyethylenimine and endocytosed into the cell. Oligodeoxynucleotides included a Ref-1 antisense sequence 5′TTTTTCCCACGCTTCGGCATTCCCG or an inverted control sequence 5′GCCCTTACGGCTTCGCACCCTTTTT.
NF-κB and AP-1 Transcriptional Activity
NF-κB and AP-1 transcriptional activity were determined by transiently transfecting CPAE cells with the pNF-κB–luciferase vector or the AP-1–luciferase vector along with pEGFP-C1, and pCB6-Ref-1, the Ref-1 redox-sensitive deletion mutant, or a control (pcDNA3.1) vector, as we have previously described.4
NF-κB DNA binding activity was determined by EMSA (Gel Shift Assay Kit; Promega) using the manufacturer’s instructions except when noted (please see online supplement).
SDS-PAGE and Western blotting were performed as we have previously described.4,13,14
Both CPAE and HUVEC cells were grown on 6-well plates to near-confluence for all apoptosis experiments. Apoptosis was assessed by staining with the nuclear chromatin dye H33342 and quantifying the percentage of apoptotic nuclei in each sample (400 cells counted/sample) as we have previously described.4,13,14
Plasma Free F2-Isoprostanes
Plasma free F2-isoprostanes (a collection of isomers) were measured by a gas chromatography–mass spectrometry-based method as described.15
Expression and Isolation of Recombinant p50 and Ref-1 Proteins and F5 mol/L (Fluorescein-5-Maleimide) Experiments
The F5 mol/L blotting was performed as described by Nishi et al.7 Briefly, recombinant p50 was treated with dithiothreitol (DTT), diamide, and recombinant wild-type Ref-1, the C-terminal deletion or N-terminal deletion Ref-1. In vitro assays were then performed on a gel and the reduced state of p50 determined by the increased binding of F5 mol/L to p50 as described by Nishi et al.7
Comparisons between 2 groups were analyzed via a Student t test (P<0.05), whereas comparisons between 3 groups were analyzed by an analysis of variance (ANOVA) with a Student-Newman-Keuls post-hoc test (P<0.05); n refers to the total number of replicates in multiple experiments. Data are presented as means±SE.
To define the role of Ref-1 in the regulation of NF-κB signaling in the intact vessel, we used a mouse harboring a single allele of Ref-1. Mice lacking both alleles of Ref-1 die in utero.12 Ref-1 mRNA and protein expression were significantly decreased in the Ref-1+/− mice compared with wild-type control animals (Figure 1) (please see online supplement for mRNA expression). F2 isoprostanes were significantly elevated in the serum of Ref-1+/− mice compared with wild-type controls, indicative of increased oxidative stress (wild-type 82.7 pg/mL±9.2 versus Ref-1+/− 133.9 pg/mL±17.3; n=7; P<0.02).
Based on our previous in vitro work, we hypothesized that loss of Ref-1 would lead to a significant reduction in NF-κB activity. Use of lipopolysaccharide (LPS) was based on preliminary work in which LPS significantly increased Ref-1 transcriptional activation as measured with a Ref-1 reporter construct (data not shown), combined with the well-described ability of LPS to stimulate NF-κB. In accord with our hypothesis, NF-κB binding in the aorta of Ref-1+/− animals was significantly reduced (Figure 2) (please see online supplement for histogram). Given the significant loss of NF-κB binding in the Ref-1+/− mice, we postulated that expression of NF-κB–dependent genes would also be suppressed. Expression levels of the cell survival gene A20 were significantly decreased in the aortas of Ref-1+/− animals (wild-type 128±22 A20/GAPDH fluorescence units versus Ref-1 38±6; n=5; P<0.01) (please see online supplement). In contrast, IL-6 levels were near zero in both groups (Ref-1, 0.01±0.00 IL-6/b-actin versus wild-type, 0.01±0.01; n=6, NS), and IAP-2 was not significantly different in the vessels of Ref-1+/− and wild-type mice (Ref-1, 23.4±10.0 IAP-2/b-actin versus wild-type, 5.2±3.2; n=6, NS).
To specifically define the role of Ref-1 in endothelial fate, we inhibited Ref-1 expression using a 25-mer Ref-1 Morpholino antisense sequence and asked if loss of Ref-1 expression was sufficient to induce endothelial cell apoptosis.4 Initial characterization revealed a 50% loss of Ref-1 protein expression using the Ref-1–specific Morpholino antisense oligodeoxynucleotide sequence (Figure 3; densitometry data; control 0.46±0.02 versus Ref-1+/− 0.25±0.03 arbitrary units of Ref-1/HSP70). In line with our hypothesis, CPAEs treated with the Ref-1 antisense sequence were significantly more susceptible to apoptosis under low serum conditions and tumor necrosis factor (TNF) compared with cells exposed to an inverted control sequence (low serum: inverted control sequence 9%±1% apoptotic nuclei; antisense to Ref-1, 18%±1% apoptotic nuclei; n=6; P<0.001) (TNF: inverted control sequence, 16%±1% apoptotic nuclei versus specific antisense sequence to Ref-1, 34%±2% apoptotic nuclei; n=6; P<0.001) (please see online supplement).
To determine whether the significant loss of the NF-κB responsive survival gene A20 in the Ref-1+/− vessel may be an important determinant governing the increase in susceptibility to endothelial cell apoptosis, we upregulated A20 and asked whether this was sufficient to block TNF-induced apoptosis. Upregulation of A20 significantly inhibited TNF-induced endothelial cell apoptosis (please see online supplement).
To delineate the region(s) of the Ref-1 gene critical for NF-κB regulation and cell survival, we used two Ref-1 deletion mutants in which the previously characterized N-terminal redox-sensitive domain (aa 1 to 116) or the C-terminal DNA repair domain (aa 191 to 318) had been deleted.11 Upregulation of wild-type Ref-1 led to significant activation of NF-κB at baseline and in response to TNF as we have previously described (Figure 4). Loss of the redox-sensitive domain abolished Ref-1–induced NF-κB activation (Figure 4). Interestingly, loss of the DNA repair domain also inhibited the ability of Ref-1 to activate NF-κB in response to TNF (Figure 4). Next, we performed a biochemical assay using fluorescein-5-maleimide (F5 mol/L) to define the ability of the deletion mutants to reduce cysteine residues in recombinant p50. These experiments were based on the elegant findings by Nishi et al demonstrating that Ref-1 reduction of p50 is a necessary step in NF-κB activation.7 As predicted, wild-type Ref-1 was able to reduce the oxidized form of p50 (Figure 5). Lack of the C-terminal AP endonuclease domain did not effect the ability of Ref-1 to reduce oxidized p50 (Figure 5). However, loss of the N-terminal redox-sensitive domain abolished the ability of Ref-1 to reduce the oxidized form of p50 (Figure 5). These findings suggest that the N-terminal (aa 1 to 116) redox-sensitive domain is important for NF-κB transcriptional activation and reduction of p50. The role of the C-terminal domain (aa 191 to 318) in the regulation of NF-κB transcriptional activity is less clear given that NF-κB transcriptional activity is inhibited in this deletion mutant, whereas the ability to reduce p50 is not impaired. This may suggest alternative mechanisms by which Ref-1 regulates NF-κB activity.
Next, we mutated residue C64 in Ref-1, a previously identified amino acid suspected to be necessary for NF-κB regulation.16 Mutation of the C64 residue to an alanine in Ref-1 did not inhibit NF-κB transcriptional activation at baseline or in response to TNF (Ref-1 WT, 2.7±0.1 luciferase units/eGFP versus Ref-1 C64A mutant, 3.6±0.1 at baseline, and Ref-1 WT, 7.2±0.1 versus Ref-1 C64A mutant, 11.4±0.2; n=6; P<0.05). In fact, mutation of C64 unexpectedly potentiated NF-κB activity.
We hypothesized that the dampened ability of the Ref-1 deletion mutants to regulate NF-κB would translate into a significant inhibition in their ability to promote survival. In accord with this hypothesis, loss of either the N-terminal redox-sensitive domain or the C-terminal repair domain of Ref-1 abolished the ability of Ref-1 to promote survival in response to TNF as we have previously shown (Figure 6). In addition, loss of the N- or C-terminal domains also inhibited endothelial cell survival at baseline (Figure 6). The C64A Ref-1 point mutation did not inhibit the ability of Ref-1 to promote survival, consistent with its inability to inhibit NF-κB activity (Figure 6).
Ref-1 has also been shown to activate AP-1. Thus, we performed a series of studies to assess the ability of Ref-1 to stimulate AP-1. Ref-1 promoted significant activation of AP-1 at baseline as measured by an AP-1 promoter–reporter construct (control, 1.00±0.03 versus Ref-1, 1.55±0.1; n=12; P<001). Deletion of either the C- or N-terminal domain significantly inhibited the ability of Ref-1 to stimulate AP-1 transcriptional activation (C-terminal deletion mutant 0.84±0 0.07 and N-terminal deletion mutant, 0.65±0.07; n=12; P<0.001). In the presence of TNF, Ref-1 also potentiated AP-1 activity (control, 1.1±0.1 arbitrary luciferase units/eGFP fluorescence versus Ref-1, 1.6±0.1; n=12; P<0.001). Lack of either the N- or C-terminal domain abolished the ability of Ref-1 to stimulate AP-1 in the presence of TNF (N-domain deletion, 0.7±0.01; C-domain deletion, 0.9±0.1; n=12; P<0.001).
We have demonstrated that loss of Ref-1 significantly attenuates NF-κB signaling in the intact vasculature. In parallel with the loss of NF-κB signaling in Ref-1–deficient mice, expression of the NF-κB–dependent survival gene A20 was significantly reduced, whereas IAP-2 and IL-6 were unchanged. Studies in isolated endothelial cells further corroborated the importance of Ref-1 in NF-κB activation in response to TNF-α and demonstrated the importance of both the redox-sensitive and AP endonuclease domains in regulating NF-κB activity. The redox-sensitive domain, but not the AP endonuclease domain, appears to stimulate NF-κB through the reduction of p50. Loss of Ref-1 also led to a significant increase in F2 isoprostanes, a marker of oxidative stress, in the serum. This finding agrees with previous work.17,18
Loss of Ref-1 led to a significant increase in the susceptibility of endothelial cells to undergo apoptosis. The mechanism appears to be mediated in part through decreased stimulation of the pro-survival NF-κB–A20 signaling cascade. Taken together, the evidence suggests that Ref-1 is an important upstream control point regulating NF-κB signaling in response to TNF-α in the intact vessel, oxidative stress, and endothelial cell survival.
Ref-1 is a multifactorial protein involved in both DNA repair and redox-mediated transcriptional events, including the activation of NF-κB.4–6,8–12,16,19,20 The N-terminal 116 amino acid domain, partially absent in its bacterial homolog exonuclease III, has been shown to be responsible for its redox regulation, whereas the C-terminal domain is largely responsible for its DNA repair activity.11 We demonstrated that both the N- and C-terminal domains for Ref-1 may be important for NF-κB activity, as measured with our NF-κB reporter assay. A limitation of this assay is the possible variation in deletion mutant expression. Thus, we chose to use a different biochemical technique using F5 mol/L to detect the ability of the Ref-1 deletion mutant recombinant proteins to reduce p50 levels. The goal was not to directly compare these two assays but to define the ability of Ref-1 itself and its mutants to reduce p50 as a potential mechanism defining how Ref-1 and the two mutants alter NF-κB activity. These studies revealed that only the redox-sensitive domain of Ref-1 was critical for reduction of p50. These studies extend the work by Nishi et al demonstrating that Ref-1 is a proximal determinant of NF-κB activity.7 Our findings demonstrating that loss of the Ref-1 C-terminal domain abolished the ability of Ref-1 to activate NF-κB despite the fact that the C-terminal domain was not required to reduce p50 suggest that Ref-1 may be regulating NF-κB through redox-independent mechanism(s).
Walker et al originally identified a cysteine residue (cysteine 65) in Ref-1 that was a critical redox-active site essential for the reactivation of DNA binding of oxidized Jun proteins. To specifically test the role of cysteine 65 (cysteine 64 in mouse) as a critical residue governing NF-κB transcriptional activation, we transiently overexpressed a Ref-1 transgene in which cysteine 64 was mutated to an alanine. In direct contrast to the inhibitor effect of this residue on Jun binding, NF-κB transcriptional activation was potentiated. These data agree with recent work by Ordway et al demonstrating that cysteine-64 of Ref-1 is not essential for redox regulation of another relevant transcription factor, AP-1.21 The unexpected potentiation of TNF-induced NF-κB by the C64A point mutant may explain the anti-apoptotic effect of this mutant in endothelial cells. The potential differences explaining the survival effects of the C64A point mutation versus the amplification of cell death in response to TNF with the repair and redox-sensitive deletion mutants are many. Ref-1 binds to multiple factors, including AP-1 and HIF-1a. Loss of the larger domains most likely impairs the ability of Ref-1 to bind to several factors while at the same time altering redox state of the cell, whereas loss of the point mutation may have a more limiting effect (inhibition of c-jun). Future studies will be needed to address this issue.
Ref-1 has been shown to regulate several transcription factors in addition to NF-κB, including AP-1, c-myb, p53, and members of the ATF/CREB family.5,6,19,20 Enhanced activation of AP-1 by Ref-1 has been described in several cell types.5,6,22 We have reconfirmed this in endothelial cells. However, the role of AP-1 in regulating endothelial cell fate is not clear.23–25 We also saw a decrease in AP-1 DNA binding in aortas of Ref-1 heterozygote mice compared with wild-type controls. Ref-1 has also been demonstrated to activate p53 and induce apoptosis in human tumor cell lines.19 Given that we clearly document an anti-apoptotic effect of Ref-1 in response to TNF, it is unlikely that a pro-apoptotic effect via p53 is a major component of its effect in endothelial cells. Recent work by Angkeow et al demonstrated that Ref-1 overexpression in the setting of hypoxia promotes endothelial cell survival.26 These findings are consistent with our previous work.4 The ability of Ref-1 to promote survival in the setting of hypoxia does not appear to be mediated through NF-κB.4,26 In fact, Angkeow et al reported that Ref-1 overexpression inhibited NF-κB in the setting of hypoxia.26 This fits with earlier work from our laboratory demonstrating that NF-κB is not involved in mediating Ref-1–induced cell survival in the setting of hypoxia.4 Work from our laboratory suggests that under hypoxic conditions, Ref-1 is likely inducing HIF-1a (unpublished findings from our laboratory). Taken together, these findings suggest that the ability of Ref-1 to bind to and regulate transcription factors is dependent on the environmental stimuli.
We found that loss of Ref-1 in the vessel wall correlated with a significant loss of LPS-stimulated NF-κB activity, as well as a loss of the NF-κB responsive survival gene, A20. Our interest in A20 stems from work by others demonstrating that A20 is a potent anti-apoptotic factor downstream of NF-κB and serves as a feedback inhibitor to block NF-κB activity.27–31 Ferran et al have discussed the potential application of A20 as a therapeutic candidate.29 Overexpression of A20 would protect endothelial cells from TNF-induced apoptosis while at the same time inhibiting NF-κB through a negative feedback loop and thereby inhibiting inflammation.29 To our knowledge, this is the first study to define a relationship between Ref-1 and A20. Future studies will be needed to further define the interplay between Ref-1, NF-κB, A20, inflammation, and apoptosis in the vascular endothelium.
In conclusion, we have demonstrated that loss of Ref-1 leads to a significant increase in the susceptibility of vascular endothelial cells to undergo apoptosis. This effect appears to be mediated through the decreased stimulation of a pro-survival cascade involving NF-κB and A20. Furthermore, previous work by our laboratory and others has demonstrated in gain-of-function experiments that increasing Ref-1 promotes survival to cytokine stress through increased NF-κB activity. Taken together, these findings support previous work demonstrating a role for redox-related mechanisms in the regulation of NF-κB and strengthen previous work defining a critical role of Ref-1 in mediating pro-survival signaling cascades in the vascular endothelium.
The authors gratefully acknowledge the help of Dr T. Curran and Dr J. Ordway for providing the Ref-1 transgenic mice, as well as Dr Myron Gross (Director) and Linda Lewis of The Molecular Epidemiology and Biomarker Research Laboratory at the University of Minnesota for the F2-isoprostane measurements. This work was supported by the Juvenile Diabetes Research Foundation (J. Hall) and the American Heart Association (Z. Guan).
Z.G. and D.B. contributed equally to this work.
- Received April 16, 2004.
- Accepted October 29, 2004.
Gibbons GH, Dzau VJ. Molecular therapies for vascular diseases. Science. 1996; 272: 689–693.
Griendling KK, Alexander RW. Oxidative stress and cardiovascular disease. Circulation. 1997; 96: 3264–3265.
Kunsch C, Medford RM. Oxidative stress as a regulator of gene expression in the vasculature. Circ Res. 1999; 85: 753–766.
Hall JL, Wang X, Van A, Zhao Y, Gibbons GH. Overexpression of Ref-1 inhibits hypoxia and tumor necrosis factor-induced endothelial cell apoptosis through nuclear factor-κb–independent and -dependent pathways. Circ Res. 2001; 88: 1247–1253.
Nishi T, Shimizu N, Hiramoto M, Sato I, Yamaguchi Y, Hasegawa M, Aizawa S, Tanaka H, Kataoka K, Watanabe H, Handa H. Spatial redox regulation of a critical cysteine residue of NF-κB in vivo. J Biol Chem. 2002; 277: 44548–44556.
Demple B, Herman T, Chen DS. Cloning and expression of APE, the cDNA encoding the major human apurinic endonuclease: definition of a family of DNA repair enzymes. Proc Natl Acad Sci U S A. 1991; 88: 11450–11454.
Robson CN, Hickson ID. Isolation of cDNA clones encoding a human apurinic/apyrimidinic endonuclease that corrects DNA repair and mutagenesis defects in E. coli xth (exonuclease III) mutants. Nucleic Acids Res. 1991; 19: 5519–5523.
Xanthoudakis S, Miao GG, Curran T. The redox and DNA-repair activities of Ref-1 are encoded by nonoverlapping domains. Proc Natl Acad Sci U S A. 1994; 91: 23–27.
Xanthoudakis S, Smeyne RJ, Wallace JD, Curran T. The redox/DNA repair protein, Ref-1, is essential for early embryonic development in mice. Proc Natl Acad Sci U S A. 1996; 93: 8919–8923.
Wang X, Xiao Y, Mou Y, Zhao Y, Blankesteijn WM, Hall JL. A role for the beta-catenin/T-cell factor signaling cascade in vascular remodeling. Circ Res. 2002; 90: 340–347.
Hall JL, Matter CM, Wang X, Gibbons GH. Hyperglycemia inhibits vascular smooth muscle cell apoptosis through a protein kinase C-dependent pathway. Circ Res. 2000; 87: 574–580.
Walker LJ, Robson CN, Black E, Gillespie D, Hickson ID. Identification of residues in the human DNA repair enzyme HAP1 (Ref-1) that are essential for redox regulation of Jun DNA binding. Mol Cell Biol. 1993; 13: 5370–5376.
Meira LB, Devaraj S, Kisby GE, Burns DK, Daniel RL, Hammer RE, Grundy S, Jialal I, Friedberg EC. Heterozygosity for the mouse Apex gene results in phenotypes associated with oxidative stress. Cancer Res. 2001; 61: 5552–5557.
Ozaki M, Suzuki S, Irani K. Redox factor-1/APE suppresses oxidative stress by inhibiting the rac1 GTPase. FASEB J. 2002; 16: 889–890.
Gaiddon C, Moorthy NC, Prives C. Ref-1 regulates the transactivation and pro-apoptotic functions of p53 in vivo. EMBO J. 1999; 18: 5609–5621.
Jayaraman L, Murthy KG, Zhu C, Curran T, Xanthoudakis S, Prives C. Identification of redox/repair protein Ref-1 as a potent activator of p53. Genes Dev. 1997; 11: 558–570.
Ordway JM, Eberhart D, Curran T. Cysteine 64 of Ref-1 Is Not Essential for Redox Regulation of AP-1 DNA Binding. Mol Cell Biol. 2003; 23: 4257–4266.
Yao KS, Xanthoudakis S, Curran T, O’Dwyer PJ. Activation of AP-1 and of a nuclear redox factor, Ref-1, in the response of HT29 colon cancer cells to hypoxia. Mol Cell Biol. 1994; 14: 5997–6003.
Moreno-Manzano V, Ishikawa Y, Lucio-Cazana J, Kitamura M. Suppression of apoptosis by all-trans-retinoic acid. Dual intervention in the c-Jun n-terminal kinase-AP-1 pathway. J Biol Chem. 1999; 274: 20251–20258.
Wang N, Verna L, Hardy S, Zhu Y, Ma KS, Birrer MJ, Stemerman MB. c-Jun triggers apoptosis in human vascular endothelial cells. Circ Res. 1999; 85: 387–393.
Ferran C, Stroka DM, Badrichani AZ, Cooper JT, Wrighton CJ, Soares M, Grey ST, Bach FH. A20 inhibits NF-κB activation in endothelial cells without sensitizing to tumor necrosis factor-mediated apoptosis. Blood. 1998; 91: 2249–2258.
Wertz IE, O’Rourke KM, Zhou H, Eby M, Aravind L, Seshagiri S, Wu P, Wiesmann C, Baker R, Boone DL, Ma A, Koonin EV, Dixit VM. De-ubiquitination and ubiquitin ligase domains of A20 downregulate NF-κB signalling. Nature. 2004.