Transcriptional Regulation of Activating Transcription Factor 4 under Oxidative Stress in Retinal Pigment Epithelial ARPE-19/HPV-16 Cells

Abstract

Purpose.: Oxidative stress plays an important role in the pathogenesis of various ocular diseases such as retinopathy, glaucoma, and age-related macular degeneration. Activating transcription factor 4 (ATF4) is induced by various stressors, including endoplasmic reticulum (ER) and oxidative stress, and ATF4 expression is regulated translationally through the PERK pathway of eIF2α phosphorylation. Transcriptional regulation of the ATF4 gene under oxidative stress was investigated in human papillomavirus 16 (HPV-16)–transformed retinal pigment epithelial ARPE-19/HPV-16 cells.

Methods.: Retinal pigment epithelial cells, trabecular meshwork cells, and corneal endothelial cells were treated with anoxia and thapsigargin (TG). Gene expression of ATF4 and nuclear factor (erythroid-derived 2)-like 2 (Nrf2) and transcription factors was investigated by Western blot analysis, reporter assays, chromatin immunoprecipitation (ChIP) assays, and small interfering (si)RNA strategies. Cellular sensitivity to oxidative stress was determined.

Results.: The expression of two transcriptional factors, ATF4 and Nrf2, was significantly induced by anoxia and TG. The Nrf2 regulator Keap1 was downregulated by anoxia. Downregulation of Nrf2 abolished ATF4 expression. On the other hand, downregulation of Keap1 enhanced the expression of both Nrf2 and ATF4. The promoter activity of ATF4 was transactivated by the co-transfection of Nrf2 expression plasmids and reduced by the transfection of Nrf2-specific siRNA. The ChIP assays demonstrated that Nrf2 bound to the promoter of the ATF4 gene. Nrf2 downregulation nearly abolished the ATF4 induction by anoxia and TG. Consistent with these findings, the promoter activity of ATF4 was augmented by treatment with TG, HCA, H₂O₂, and anoxia. However, stress induction of ATF4 promoter activity was observed, even when a mutation was introduced into the antioxidant-responsive elements site. Furthermore, stress induction of the ATF4 promoter was completely abolished when the 5′ untranslated region of the ATF4 gene was deleted. Downregulation of ATF4 rendered ARPE-19/HPV-16 cells sensitive to oxidative stress.

Conclusions.: These results suggest that the stress induction of ATF4 is significantly regulated transcriptionally through a Nrf2-dependent mechanism and may be a double-edged sword in the pathogenesis of various retinopathies.

The eye is protected against oxidative stress by several mechanisms. We have shown that transcription factor Foxo3a functions to protect trabecular meshwork cells from oxidative stress. ¹ We have also reported that activating transcription factor 4 (ATF4) is upregulated in drug-resistant cells and protects against oxidative stress through glutathione biosynthesis. ^2,3 Thus, in addition to the transcription factor Foxo3a, the ATF4 gene is thought to be one of the critical transcriptional factors for a cellular response against oxidative stress. 

Cellular stress, including oxidative stress, activates the PERK pathway, and phosphorylation of the translational initiation factor eIF2α subsequently causes ATF4 expression to be translationally activated. ^{4 –6} Phosphorylation of eIF2α leads to a general inhibition of translation ^{7 –10} ; however, translation of ATF4 mRNA is selectively processed, the mechanism of which has been elucidated in the general amino acid control response in yeast. ^4,11 The translational upregulation of the transcription of the ATF4 gene remains to be elucidated. 

Nuclear factor (erythroid-derived 2)-like 2 (Nrf2) is essential for inducible and constitutive expression of several phase II detoxification enzymes and antioxidant enzymes. During an oxidative stress condition, Nrf2 dissociates with Keap1, which is involved in cytoplasmic sequestration of Nrf2, and translocates into the nucleus, activating target genes by heterodimerizing with the small Maf family and binding to antioxidant-responsive elements (AREs). ^{12 –16} Both Nrf2 and ATF4 are coordinately activated by oxidative stress. ^{17 –19}  

In the present study, we showed that Nrf2 is involved in transcriptional ATF4 gene expression under oxidative stress and that the ATF4 gene is one of the Nrf2 target genes. 

Methods

Cell Culture

The human retinal pigment epithelium cell line ARPE-19/HPV-16 (CRL-2502) ²⁰ and the bovine corneal endothelial cell line BCE C/D-1b (CRL-2048) ²¹ were obtained from the American Type Culture Collection (Manassas, VA). ARPE-19/HPV-16 cells were cultured in a 1:1 mixture of Dulbecco's modified Eagle's medium and Ham's F12-medium (DMEM:F12). BCE C/D-1b cells and an immortalized TM cell line (i.e., NTM5 cells derived from normal trabecular meshwork), ^1,22 were cultured in Dulbecco's modified Eagle's medium. All cell lines were maintained in a 5% CO₂ atmosphere at 37°C. The primary human retinal pigment epithelial cells, HRPEpiC, was purchased from Sciencell Research Laboratories (San Diego, CA). Early passages (3–4) of HRPEpiCs were used in the study. 

Antibodies and Drugs

Antibodies against Nrf2 (sc-30915), ATF4 (sc-200), Keap1 (sc-15,246), and PCNA (sc-56) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-Nrf2 antibody (RB10471P), anti-β-actin antibody (AC-15), and anti-HA-peroxidase (3F10) were purchased from Thermo Scientific, Inc. (Bremen, Germany), Sigma-Aldrich (St. Louis, MO), and Roche Molecular Biochemicals (Mannheim, Germany), respectively. Thapsigargin (TG) and homocysteine (HCA) were purchased from Sigma-Aldrich. Drug concentrations in this study corresponded with those used in clinical practice. 

Plasmid Construction

To obtain full-length complementary (c)DNA for human Nrf2, we performed PCR on a cDNA library (SuperScript; Invitrogen Life Technologies, Carlsbad, CA), using the following primer pairs (italic indicates the start and stop codons): 5′-ATGATGGACTTGGAGCTGCCGCC-3′ and 5′-CTAGTTTTTCTTAACATCTGGCTTCTTAC-3′. This PCR product was cloned into a vector (pGEM-T Easy; Promega, Madison, WI). To construct a plasmid expressing HA tagged Nrf2, N-terminal HA-tagged Nrf2 cDNA was ligated into a pcDNA3 vector (Invitrogen Life Technologies). The luciferase reporter constructs containing the core promoter and the partial first exon with either the short (−314 to +81) or long (−314 to +877) 5′ untranslated region (UTR) of ATF4 were generated by using the following primers: primer 1, 5′-AGATCTGCGCTGACACCGGAAGCGAGGCG-3′; primer 2, 5′-AGATCTGCGCGGACACCATAAGCGAGGCG-3′; primer 3, 5′-AAGCTTGGCCGTGGACCCTGAGGGC-3′; and primer 4, 5′-AAGCTTTGCTGGAATCGAGGAATGTGC-3′. Italic and bold indicate restriction enzyme cleavage sites and mutations of ARE, respectively. To obtain ATF4-Luc WS, ATF4-Luc WL, ATF4-Luc MS, and ATF4-Luc ML, PCR using placental DNA was performed with primers 1 and 3, primers 1 and 4, primers 2 and 3, and primers 2 and 4, respectively. These PCR products were cloned and ligated into the BglII-HindIII sites of the pGL3-basic vector (Promega). 

Knockdown Analysis with siRNAs

The following double-stranded RNA 25-bp oligonucleotides were commercially generated (Invitrogen): ATF4 small interfering RNA (siRNA), 5′-GGGCAGUGAAGUGGAUAUCACUGAA-3′ (sense) and 5′-UUCAGUGAUAUCCACUUCACUGCCC-3′ (antisense), ATF4-1 siRNA; 5′-GGGCUGAAGAAAGCCUAGGUCUCUU-3′ (sense) and 5′-AAGAGACCUAGGCUUUCUUCAGCCC-3′ (antisense), ATF4-2 siRNA; 5′-UCACUUUGCAAAGCUUUCAACCAAA-3′ (sense) and 5′-UUUGGUUGAAAGCUUUGCAAAGUGA-3′ (antisense), Nrf2-1 siRNA; 5′-GACACACUACUUGGCCUCAGUGAUU-3′ (sense) and 5′-AAUCACUGAGGCCAAGUAGUGUGUC-3′ (antisense), Nrf2-2 siRNA; 5′-CCAACCAGUUGACAGUGAACUCAUU-3′ (sense) and 5′-AAUGAGUUCACUGUCAACUGGUUGG-3′ (antisense), Nrf2-3 siRNA; 5′-GGAAACAGAGACGUGGACUUUCGUA-3′ (sense) and 5′-UACGAAAGUCCACGUCUCUGUUUCC-3′ (antisense), Keap1-1 siRNA; 5′-UCCACGUCAUGAACGGUGCUGUCAU-3′ (sense) and 5′-AUGACAGCACCGUUCAUGACGUGGA-3′ (antisense), Keap1-2 siRNA; and 5′-GCCCGGGAGUACAUCUACAUGCAUU-3′ (sense) and 5′-AAUGCAUGUAGAUGUACUCCCGGGC-3′ (antisense), Keap1-3 siRNA. siRNA transfections were performed as described previously. ^1,3,23 Briefly, 250 picomoles of the indicated siRNAs or control siRNA (Invitrogen) were transfected to 1 × 10⁶ NTM5 cells, 2 × 10⁵ cells were used for the luciferase assay, and 2.5 × 10³ cells were used for the WST-8 assay described in the following sections. The remaining cells were subjected to Western blot analysis after 72 hours culture as described below. 

Western Blot Analysis

The preparation of whole nuclear extracts is described elsewhere. ^3,23 The indicated amounts of whole nuclear extracts were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) microporous membranes (Millipore, Billerica, MA), by using a semidry blotter. The blotted membranes were treated with 5% (wt/vol) skimmed milk in 10 mM Tris, 150 mM NaCl and 0.2% (vol/vol) Tween 20, and were incubated for 1 hour at room temperature with primary antibody. The following antibodies and dilutions were used: a 1:500 dilution of anti-Nrf2, a 1:1,000 dilution of anti-Keap1, a 1:5,000 dilution of anti-ATF4, a 1:5,000 dilution of anti-PCNA, and a 1:20,000 dilution of anti-β-actin. Membranes were then incubated for 40 minutes at room temperature with a peroxidase-conjugated secondary antibody and were visualized with chemiluminescence (ECL kit; GE Healthcare Bio-Science, Piscataway, NJ). Membranes were then exposed to autoradiograph film (X-OMAT film; Eastman Kodak, Rochester, NY). The reproducibility was confirmed by several Western blot analyses, and the resulting blots are shown in the figures. For the correlation assay, the intensity of each signal was quantified by the NIH Image program (NIH, Bethesda, MD; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html). 

Luciferase Assay

Transient transfection and a luciferase assay were performed as described elsewhere. ^1,3,23 In preparing expression plasmids, 2 × 10⁵ ARPE-19/HPV-16 cells per well were seeded into 12-well plates. The following day, indicated amounts of reporter plasmids (see Figs. 3 and 5) were transfected with or without expression plasmids (Superfect reagent; Qiagen, Valencia, CA). After 48 hours, the luciferase activity was detected with a kit (PicaGene; Toyo Ink, Tokyo, Japan). At 24 hours after transfection with indicated siRNAs described earlier, indicated amounts of reporter plasmids were transfected with the reagent (Superfect; Qiagen). After 48 hours, the luciferase activity was detected using the assay (PicaGene; Toyo Ink). For treatment with TG or HCA, 1 μM TG or 1 mM HCA were directly added into the culture medium at 36 hours after transfection. After 6 hours, the luciferase activity was detected with the kit (PicaGene; Toyo Ink). For H₂O₂ or anoxic treatment conditions, at 42 or 40 hours after transfection, cells were cultured with 500 μM H₂O₂ for 2 hours or under anoxic conditions (Anaerocult A mini; Merck, Darmstadt, Germany) creating oxygen-free circumstances for 4 hours, respectively. Cells treated with H₂O₂ or anoxic conditions and mock treated cells were further cultured with a new normal culture medium and normal conditions for 4 hours by recovery/reoxygenation. The light intensity was measured with a luminometer (Luminescencer JNII RAB-2300; ATTO). The results shown are normalized to the protein concentrations measured using the Bradford method and are representative of at least three independent experiments. 

Chromatin Immunoprecipitation Assay

The chromatin immunoprecipitation assay (ChIP) assay was performed as described previously. Briefly, soluble chromatin from ARPE-19/HPV16 cells was incubated with 2 μg anti-Nrf2 antibody (sc-722), anti-Nrf2 antibody (RB10471P), and normal rabbit immunoglobulin G (IgG). Immunoprecipitated DNAs with protein A/G-agarose were purified and dissolved in 20 μL of distilled H₂O. Each 2 μL of DNA was used for PCR analysis with the following primer pairs; ATF4 (−308 to +81), 5′-GCGCTGACACCGGAAGCGAGGCG-3′ (forward) and 5′-GGCCGTGGACCCTGAGGGC-3′ (reverse); YB-1 (+409 to +936), 5′-GCCCGGCACTACGGGCTGCG-3′ (forward) and 5′-GTGTGCGCAGGCCGCGGACG-3′ (reverse). The PCR products were separated by electrophoresis on 2% agarose gels and stained with ethidium bromide. 

Cytotoxicity Analysis with WST-8

The water-soluble tetrazolium salt (WST-8) assay was performed according to published methods. ^1,3,23 Briefly, 2.5 × 10³ ARPE-19/HPV-16 cells per well that had been transfected with the indicated amount of siRNA were seeded into 96-well plates. The following day, the cells were incubated with the indicated concentrations of H₂O₂ or HCA (see Fig. 6). For H₂O₂, cells were treated with H₂O₂ in the serum-free medium for 40 minutes, and subsequently the medium was changed to a normal culture medium. After 72 hours, surviving cells were stained (TetraColor ONE; Seikagaku Corp., Tokyo, Japan) for 90 minutes at 37°C. The absorbance was then measured at 450 nm. 

Cell Viability Analysis

The ARPE-19/HPV-16 cells, that were transfected with the indicated amount of siRNA were seeded into six-well plates at 2.0 × 10⁴ per well. The following day, the cells were incubated with the indicated concentrations of H₂O₂ (see Fig. 6) in serum-free medium for 4 hours, and subsequently the medium was changed to normal culture medium. After 72 hours, the cells were harvested with trypsin. Living cells were counted in a cell viability assay (Adam-MC; NanoEnTek Inc., Seoul, Korea), according to the manufacturer's instructions. Briefly, the cells in each condition were suspended with 100 μL PBS, and each 40 μL was mixed with 40 μL solution T containing propidium iodide (PI) and detergent or 40 μL solution N containing PI without detergent. The number of living cells was calculated as (T number − N number)/T number. The number of living cells without cisplatin was set at 100%. 

Statistical Analysis

The Pearson correlation method was used for statistical analysis, and significance was set at the 5% level. 

Results

Stress Induction of ATF4

We first investigated endoplasmic reticulum (ER) stress induction of three cell types derived from the eye with TG. ATF4 expression was upregulated by TG treatment in these cell lines. Among these cell lines, ARPE-19/HPV-16 cells showed the highest level of ATF4 expression after treatment with TG (Fig. 1A). In this study, we investigated cellular sensitivity against oxidative stress with the expression of transcription factors, but not physiological properties of retinal pigment epithelium. ARPE-19/HPV-16 cells can be easily maintained and thought to be suitable for our purpose. ARPE-19/HPV-16 cells were then used in the following experiments. We subsequently confirmed the stress induction of ATF4 expression. As shown in Figure 1B, ATF4 expression was upregulated more than threefold after anoxia. 

Figure 1.

 

(A) The effect of TG on expression of ATF4. ARPE-19/HPV-16 (retina), NTM5 (trabecular meshwork), and BCE C/D-1b (cornea) cells were incubated with or without 1 μM TG for 4 hours. The nuclear and cytosolic extracts containing 50 μg of proteins were prepared and subjected to Western blot analysis with ATF4 antibodies. Relative intensity is shown under each blot. CBB, Coomassie brilliant blue. (B) ATF4 is upregulated in response to anoxic conditions. ARPE-19/HPV-16 cells were incubated in anoxic conditions for 6 hours, and the cells were lysed either immediately or after reoxygenation. The nuclear extracts containing 50 μg of proteins were prepared and subjected to Western blot analysis with the indicated antibodies. Immunoblot analysis of PCNA is shown as a loading control. Relative intensity is shown under each blot.

Stress Induction of Nrf2

We previously showed that an E-box was located in the core promoter region of the ATF4 gene and E-box binding circadian transcriptional factor (clock) positively regulated the ATF4 gene. ³ We also found one ARE located −300 bp from the transcription start site of the ATF4 gene promoter. Nrf2 binds to ARE to transactivate the target genes. We investigated whether TG treatment can enhance Nrf2 expression. As shown in Figure 2A, nuclear expression of Nrf2 was increased by TG treatment. Consistent with the result shown in Figure 1A, ARPE-19/HPV-16 cells showed the highest level of Nrf2 expression after treatment with TG. On the other hand, the expression of Keap1, involved in the sequestration of Nrf2, was varied in the cells. Furthermore, Nrf2 expression was induced by anoxia (Fig. 2B). In accordance with the anoxia-dependent induction of Nrf2 expression, Keap1was reduced by anoxia. To investigate whether ATF4 expression requires Nrf2 expression, we performed a knockdown experiment using siRNA for both Nrf2 and Keap1. As shown in Figure 2C, knockdown of Nrf2 expression downregulated ATF4 expression in both transformed (ARPE-19/HPV-16) and nontransformed (HRPEpiC) RPE cells. On the other hand, knockdown of Keap1 upregulated both Nrf2 and ATF4 expression, suggesting that ATF4 is a target gene of Nrf2 (Fig. 2D). 

Figure 2.

 

(A) The effect of TG on expression of Nrf2 expression. ARPE-19/HPV-16, NTM5, corneal endothelial cells were incubated with or without 1 μM TG for 4 hours. Nuclear extracts (100 μg for Nrf2) and cytosolic extracts (50 μg for Keap1) were subjected to SDS-PAGE, and Western blot analysis was performed using the indicated antibodies. Relative intensity is shown under each blot. (B) Nrf2 is upregulated in response to anoxic conditions. ARPE-19/HPV-16 cells were incubated in anoxic conditions for 6 hours, and cells were lysed either immediately or after reoxygenation. Nuclear extracts (100 μg for Nrf2) and cytosolic extracts (50 μg for Keap1) were prepared and subjected to Western blot analysis with the indicated antibodies. Immunoblot analysis of β-actin is shown as a loading control. Relative intensity is shown under each blot. (C) Nrf2 expression is involved in ATF4 gene expression in transformed RPE (ARPE-19/HPV-16) cells (left) and nontransformed RPE (HRPEpiC) cells (right). Control siRNA (100 picomoles) or Nrf2 siRNA-1, -2, and -3 (100 picomoles) were transfected into indicated cells. The nuclear extracts containing 100 μg of proteins were subjected to SDS-PAGE. The transfer membrane was blotted with indicated antibodies. Immunoblot analysis of β-actin is shown as a loading control. Relative intensity is shown under each blot. (D) Keap1 expression is involved in Nrf2 and ATF4 gene expression. Control siRNA (100 picomoles) or Keap1 siRNA-1, -2, and -3 (100 picomoles) were transfected into ARPE-19/HPV-16 cells. Nuclear extracts (100 μg for Nrf2 and ATF4) and cytosolic extracts (50 μg for Keap1 and β-actin) were subjected to SDS-PAGE. Transferred membrane was blotted with the indicated antibodies. Immunoblot analysis of β-actin is shown as a loading control. Relative intensity is shown under each blot. CBB, Coomassie brilliant blue.

Transcriptional Regulation of the ATF4 Gene

To confirm the transcriptional expression of the ATF4 gene, we constructed reporter genes (Fig. 3A). We introduced the mutation into ARE. We also constructed the reporter with either a short or long 5′ UTR to test the role of 5′UTR. The promoter activity of the ATF4 gene was enhanced by co-transfection with the Nrf2 expression plasmid (Fig. 3B). Interestingly, the promoter activity of the ATF4 gene with short 5′UTR was strongly transactivated by the co-transfection with the Nrf2 expression plasmid. However, Nrf2-dependent transactivation of the ATF4 promoter was completely abolished when the mutation was introduced into ARE. Downregulation of Nrf2 expression also reduced the basal promoter activity of the ATF4 gene (Fig. 3C). To determine whether transcriptional activation of ATF4 gene was a result of direct recruitment of Nrf2 to the promoter in vivo, we performed ChIP assays using two different anti-Nrf2 antibodies. ChIP assays demonstrated that Nrf2 bound to the promoter of the ATF4 gene but not to that of the unrelated YBX1 gene (Fig. 3D). ²⁴  

Figure 3.

 

(A) Representations of ATF4-Luc WS, WL, MS, and ML are shown. The reporter plasmids containing wild-type ARE with long or short UTR indicate WL and WS, respectively. The reporter plasmids containing the mutant type of ARE with long or short UTR indicate ML and MS, respectively. (B) Left: Nrf2 transactivates the promoter activity of the ATF4 gene. Indicated amounts of Nrf2 expression plasmid was transiently co-transfected with the indicated reporter plasmids into ARPE-19/HPV-16 cells. The results were normalized to the protein concentration measured using the Bradford method. All values are representative of at least three independent experiments. The luciferase activity of each ATF4-Luc with transfection of HA vector corresponds to 1. Bars, SD. Right: HA-Nrf2 expression of ARPE-19/HPV-16 cells was detected by Western blot analysis with anti-HA-antibody. (C) Knockdown of Nrf2 downregulates the promoter activity of the ATF4 gene. ARPE-19/HPV-16 cells were transiently transfected with the indicated amounts of control siRNA or Nrf2 siRNAs followed by transfection with 0.75 μg of the indicated reporter plasmids at intervals of 12 hours. The results shown are normalized to protein concentrations measured using the Bradford method and are representative of at least three independent experiments. The luciferase activity of each ATF4-Luc with transfection of control siRNA corresponds to 1. Bars, SD. (D) Top: Schematic representations of the promoter region and 5′ end of ATF4 gene. Arrow, black box, and gray boxes: primer, ARE, and exon, respectively. Bottom: The ChIP assay was performed with ARPE-19/HPV16 cells. Soluble chromatin (lanes 1 and 4) and immunoprecipitated DNAs (lanes 2 and 5; normal rabbit IgG, lane 3; anti-Nrf2 antibody (sc-722), and lane 6; anti-Nrf2 antibody RB10471P) were amplified by polymerase chain reaction with specific primer pairs for the ATF4 promoter. Lane M: DNA size marker; arrowhead: PCR products (389 bp for Nrf2 and 528 bp for YB-1-5).

Role of Nrf2 in Stress Induction of ATF4

Next, we performed reporter assays under oxidative stress. The expression of both Nrf2 and ATF4 was enhanced reproducibly by anoxia. However, ATF4 induction by anoxia was almost cancelled when Nrf2 was downregulated (Fig. 4A). Similar results were also observed when the cells were treated with TG (Fig. 4B). Next, we performed reporter assays under oxidative stress. The assays showed that the promoter activity with wild-type ARE was significantly increased when the cells were treated with H₂O₂ and anoxia (Fig. 5A). Interestingly, the luciferase activity was also enhanced by treatment with H₂O₂ and anoxia, even when reporter constructs containing a long 5′UTR were used. Furthermore, luciferase activity with a long 5′UTR was much higher than that with short a 5′UTR. The stress induction of promoter activity was completely abolished when reporter constructs containing ARE mutations and a short 5′UTR were used. Similar results were observed when the cells were treated with both an ER stress inducer, TG, and an oxidative inducer, HCA (Fig. 5B). 

Figure 4.

 

Nrf2 regulated the expression of ATF4 under anoxic conditions and ER stress. (A) ARPE-19/HPV-16 cells were transiently transfected with 100 picomoles of control siRNA or Nrf2 siRNA-1. After 48 hours, the cells were incubated in anoxic conditions for 6 hours and were cultured for the indicated recovery times in normal conditions. Nuclear extracts (100 μg for Nrf2, ATF4, and PCNA) and cytosolic extracts (50 μg for Keap1 and β-actin) were subjected to SDS-PAGE. Transferred membrane was blotted with indicated antibodies. PCNA is a loading control for the nuclear fraction, and β-actin is a loading control for the cytosolic fraction. Relative intensity is shown under each blot. (B) ARPE-19/HPV-16 cells were transiently transfected with 100 picomoles of control siRNA or Nrf2 siRNA-1. After 48 hours, the cells were incubated with 1 μM TG for 4 hours, and the cells were lysed. Nuclear extracts (100 μg for Nrf2, ATF4, and PCNA) and cytosolic extracts (50 μg for Keap1 and β-actin) were subjected to SDS-PAGE. Transferred membrane was blotted with indicated antibodies. PCNA was a loading control for nuclear fraction, and β-actin was a loading control for cytosolic fraction. Relative intensity is shown under each blot. CBB, Coomassie brilliant blue.

Figure 5.

 

The transcriptional activity of ATF4 is upregulated in response to ER stress, oxidative stress, and anoxic conditions. (A) ARPE-19/HPV-16 cells were transiently transfected with 0.5 μg ATF4-Luc reporter plasmids. At 42 hours after transfection, the cells were incubated at 500 μM H₂O₂ for 2 hours and were further cultured with a new culture medium for 6 hours. At 40 hours after transfection, the cells were cultured in anoxic conditions for 4 hours and were further cultured with a new culture medium and in normal conditions for 4 hours. The results shown are normalized to protein concentrations measured using the Bradford method and are representative of at least three independent experiments. The luciferase activity of each ATF4-Luc in normal conditions set to 1. Bars, SD. (B) ARPE-19/HPV-16 cells were transiently transfected with 0.5 μg ATF4-Luc reporter plasmid. At 36 hours after transfection, the cells were incubated in normal conditions, treated with either 1 μM TG or 1 mM HCA for 6 hours, and subjected to the luciferase assay. The results shown are normalized to protein concentrations measured using the Bradford method and are representative of at least three independent experiments. The luciferase activity of each ATF4-Luc under normal conditions set to 1. Bars, SD.

Role of ATF4 Expression in Cellular Sensitivity against Oxidative Stress

ARPE-19/HPV-16 cells were modestly sensitive to H₂O₂ when ATF4 expression was downregulated (Fig. 6A). HCA is thought to induce oxidative stress. Our results showed that the expression of both Nrf2 and ATF4 was induced by HCA treatment (Fig. 6B). The cells were sensitive to HCA even without knockdown of ATF4 expression. There was simply a slight increase in HCA sensitivity after the knockdown of ATF4. To confirm these results, we used an alternative cell viability assay (Adam-MC; NanoEnTek Inc.). The WST-8 measures only the activity of dehydrogenase enzymes derived from living cells (Figs. 6A, 6B). On the other hand, the assay can detect both PI-permeable apoptotic or dead cells and PI nonpermeable living cells. Knockdown of ATF4 expression again rendered the cells sensitive to H₂O₂ (Fig. 6C). 

Figure 6.

 

(A) Downregulation of ATF4 sensitizes ARPE-19/HPV-16 cells to oxidative stress. Left: ARPE-19/HPV-16 cells were transiently transfected with 100 picomoles of control, ATF4 siRNAs. After 72 hours, nuclear extracts (100 μg) were subjected to SDS-PAGE, and Western blot analysis was performed with the indicated antibodies. Immunoblot analysis of β-actin is shown as a loading control. Relative intensity is shown at the bottom of the panel. Right: 2.5 × 10³ ARPE-19/HPV-16 cells were transfected with 40 nM of control siRNA, ATF4–1 siRNA, and ATF4–2 siRNA. The following day, for induction of oxidative stress, the cells were incubated with the indicated concentrations of H₂O₂ in the serum-free medium for 40 minutes. Then, the medium was replaced with fresh culture medium. After 72 hours, cell survival was analyzed by WST-8 assay. All values are the means of at least three independent experiments. Bars, SD. (B) HCA upregulated the expression of Nrf2 and ATF4 in a time-dependent manner. Downregulation of ATF4 sensitized ARPE-19/HPV-16 cells to HCA. Left: ARPE-19/HPV-16 cells were incubated with 100 μM HCA for the indicated time. Nuclear extracts (100 μg) were subjected to SDS-PAGE, and Western blot analysis was performed with the indicated antibodies. Immunoblot analysis of PCNA is shown as a loading control. Relative intensity is shown under each blot. Right: 2.5 × 10³ ARPE-19/HPV-16 cells were transfected with 40 nM of control siRNA, ATF4–1 siRNA, ATF4–2 siRNA. The following day, to induce oxidative stress, cells were incubated with the indicated concentrations of HCA. After 72 hours, cell survival was analyzed by a WST-8 assay. All values are the means of at least three independent experiments. Bars, SD. (C) ATF4 is involved in oxidative stress. (A) ARPE-19/HPV-16 cells (2.0 × 10⁴ per well) that were transfected with the indicated amount of siRNA were seeded into 6-well plates. The following day, the cells were incubated with the indicated concentrations of H₂O₂. For H₂O₂, the cells were treated with H₂O₂ in the serum-free medium for 4 hours, and subsequently the medium was changed to a normal culture medium. After 72 hours, surviving cells were measured with a cytotoxicity assay. The results with statistical analyses are presented.

Discussion

Oxidative stressors have been implicated in the pathogenesis of various ocular diseases such as diabetic retinopathy (DR), ^25,26 retinopathy of prematurity (ROP), ^27,28 age-related macular degeneration (AMD), ^{29 –31} and glaucoma. ^{32 –35} The ability to respond to cellular stressors, including ER and oxidative stress, is fundamentally critical to cellular survival. Cellular stressors can initiate a program of both transcriptional and translational regulation. 

We have shown that ATF4 is overexpressed in drug-resistant cells and is involved in glutathione biosynthesis, ³ suggesting that ATF4 expression can protect cells from oxidative stress. We have also shown that the ATF4 gene is one of the clock-regulated genes indicating that the protective system against oxidative stress may be regulated periodically. In the present study, we showed that activation of both Nrf2 and ATF4 transcription factors coordinates and regulates the convergence of ER and oxidative stress signaling. Furthermore, in addition to translational control, these stressors initiate a transcriptional program for ATF4 expression via Nrf2 activation. 

We initially found that ARPE-19/HPV-16 cells derived from retina show a high ATF4 expression in comparison with cells derived from trabecular meshwork and corneal endothelium (Fig. 1A). The retina has the highest metabolic rate and oxygen consumption in the body. The highest metabolic rate and oxygen consumption is usually accompanied by generation of reactive oxygen species. Therefore, the RPE is a primary target of oxidative stress. ³⁶ This finding indicates that an antioxidative system may be highly developed in the RPE to protect the retina. This protective mechanism, via ATF4 activation, is compatible because oxidative stress may occur in infants with retinopathy of prematurity. 

Cells respond to various stressors through the upregulation of genes that function specifically to alleviate stress. The activation of these pathways causes increased transcription. On the other hand, translational upregulation of specific mRNA was evident in the ER stress pathway, whereas, ER stress leads to a general inhibition of translation. It has been well studied that ATF4 expression is translationally regulated during conditions of stress. ^{4 –6} However, transcriptional regulation of the ATF4 gene under oxidative stress remains to be elucidated. Our study clearly showed that constitutive expression of the ATF4 gene is regulated by Nrf2 (Fig. 2C). Inducible expression of ATF4 under conditions of stress is also upregulated by Nrf2 (Figs. 4A, 4B). In addition to transcriptional regulation, reporter assays suggest that translational regulation of ATF4 expression is present in a 5′UTR-dependent manner (Fig. 5). We also showed that ATF4 expression can protect cells from oxidative stress (Fig. 6). Thus, ATF4 expression is critical to protect cells from oxidative stress. Nrf2 has been shown to regulate target genes through the interaction with ATF4. ¹⁷ These data suggest that Nrf2-dependent ATF4 expression may function to augment the expression of Nrf2 target genes via mutual protein–protein interaction. The cellular level of Nrf2 is also regulated by Keap1. ¹² Consistent with a previous report, ³⁷ the decrease of Keap1 by anoxia-reoxygenation was also observed, probably due to the increased degradation of Keap1, as shown in Figure 4A. Interestingly, the cellular level of Keap1 during anoxia-reoxygenation was constant when Nrf2 was downregulated, suggesting that the cellular level of Keap1 is regulated by an Nrf2-dependent protein-degradation pathway. On the other hand, oxidative stress, such as treatment with HCA, increases the expression of vascular endothelial growth factor by an ATF4-dependent mechanism in ARPE-19/HPV-16 cells, suggesting that ATF4 is also involved in angiogenic retinopathy. ^{38 –40} Both arsenite and HCA treatment produce a transient phosphorylation of eIF2α followed by an increase in ATF4 protein levels. This study produced the novel findings that HCA induced the expression of both Nrf2 and ATF4 transcription factors (Fig. 6C). It also showed that downregulation of Keap1 upregulates ATF4 expression (Fig. 2D), indicating that stress induction of ATF4 expression may be regulated mainly by the activation of Nrf2. Consistent with our results, it has been recently shown that Nrf2 is a positive regulator of ATF4. ⁴¹  

In summary, the results of our study indicate that ATF4 is transcriptionally regulated under oxidative stress and may be a double-edged sword in the pathogenesis of various retinopathies. Further studies are needed to investigate ATF4 target gene expression in eye tissues under various types of ER and oxidative stress. 

Footnotes

 Supported in part by Grants-in-Aid for Scientific Research from the Ministry for Education, Culture, Sports, Science and Technology of Japan (17016075), UOEH Grant for Advanced Research, and The Vehicle Racing Commemorative Foundation.

Footnotes

 Disclosure: N. Miyamoto, None; H. Izumi, None; R. Miyamoto, None; H. Bin, None; H. Kondo, None; A. Tawara, None; Y. Sasaguri, None; K. Kohno, None