Ischemia activates the ATF6 branch of the endoplasmic reticulum stress response

Abstract

Stresses that perturb the folding of nascent endoplasmic reticulum (ER) proteins activate the ER stress response. Upon ER stress, ER-associated ATF6 is cleaved; the resulting active cytosolic fragment of ATF6 translocates to the nucleus, binds to ER stress response elements (ERSEs), and induces genes, including the ER-targeted chaperone, GRP78. Recent studies showed that nutrient and oxygen starvation during tissue ischemia induce certain ER stress response genes, including GRP78; however, the role of ATF6 in mediating this induction has not been examined. In the current study, simulating ischemia (sI) in a primary cardiac myocyte model system caused a reduction in the level of ER-associated ATF6 with a coordinate increase of ATF6 in nuclear fractions. An ERSE in the GRP78 gene not previously shown to be required for induction by other ER stresses was found to bind ATF6 and to be critical for maximal ischemia-mediated GRP78 promoter induction. Activation of ATF6 and the GRP78 promoter, as well as grp78 mRNA accumulation during sI, were reversed upon simulated reperfusion (sI/R). Moreover, dominant-negative ATF6, or ATF6-targeted miRNA blocked sI-mediated grp78 induction, and the latter increased cardiac myocyte death upon simulated reperfusion, demonstrating critical roles for endogenous ATF6 in ischemia-mediated ER stress activation and cell survival. This is the first study to show that ATF6 is activated by ischemia but inactivated upon reperfusion, suggesting that it may play a role in the induction of ER stress response genes during ischemia that could have a preconditioning effect on cell survival during reperfusion.

FIGURE 1.

Effect of sI/R on grp78 mRNA and protein, pgk mRNA, and HIF-1α protein. A, cultured cardiac myocytes were subjected to sI for the times shown. The levels of rat grp78, pgk, and gapdh (glyceraldehyde-3-phosphate dehydrogenase) mRNA were assessed by reverse transcription quantitative PCR. Shown are the mean values of grp78 or pgk/gapdh mRNA expressed as -fold control level (control = 0 h sI/0 h sR) ± S.E. *, #, and §, p ≤ 0.05 different from control and all other values. B, cultured cardiac myocytes were treated with or without sI for 8 or 20 h, and then extracts were examined for rat HIF-1α and GAPDH by immunoblotting. The migration positions of 112-, 60-, and 37-kDa molecular mass markers are shown, as is the approximate expected migration position for full-length HIF-1α (n = 3 cultures/treatment). C, cultured cardiac myocytes were subjected to 20 h of sI/R for the times shown. The levels of rat grp78 and gapdh mRNA were assessed by reverse transcription quantitative PCR. Shown are the mean values of grp78/gapdh mRNA expressed as -fold control (control = 0 h sI/0 h sR) ± S.E. *, #, and §, p ≤ 0.05 different from control and all other values. D, cultured cardiac myocytes were subjected to sI or sI/R for the times shown. The levels of rat GRP78 and GAPDH were determined by immunoblotting (n = 3 cultures/treatment). The GRP78 levels were normalized to GAPDH and are shown as -fold control ± S.E.

FIGURE 2.

Effect of oxygen and/or glucose deprivation on GRP78- and GRP94-luciferase. A, cultured cardiac myocytes were transfected with human GRP78(+7)-Luc (GRP78(+7)) or human GRP78(+221)-Luc (GRP78(+221)) and a β-galactosidase reporter and then subjected to sI or sI/R for the times shown, followed by extraction and reporter enzyme assays. *, p ≤ 0.05 different from control and all other values. B, cultured cardiac myocytes were transfected with GRP78(+221)-Luc or GRP94-(−800 to +105)-Luc (GRP94(+105)) and subjected to oxygen and/or glucose deprivation for 20 h, followed by extraction and reporter enzyme assays. Shown are the mean relative luciferase values (luciferase/β-galactosidase), expressed as the percentage of maximum for GRP78(+221)-Luc ± S.E. *, &, and #, p < 0.01 different from all other values for GRP78(+221)-Luc; §, p ≤ 0.05 different from all other values for GRP94-(−800 to +105)-Luc. *, #, §, and &, p ≤ 0.05 different from control and all other values.

FIGURE 3.

Effect of sI/R on GRP78-luciferase. A, cultured cardiac myocytes were transfected with GRP78(+221)-Luc and a β-galactosidase reporter and then subjected to various sI, followed by extraction and reporter enzyme assays. B, cardiac myocytes transfected as described in A were subjected to 20 h of sI, followed by various times of simulated reperfusion (sI/R), followed by extraction and reporter enzyme assays. Shown are the mean relative luciferase values (luciferase/β-galactosidase), expressed as -fold control (no sI; no sR) ± S.E. *, p ≤ 0.05 different from control and all other values.

FIGURE 4.

Effect of dominant negative ATF6 and ERSE mutations on sI-activated GRP78-Luc. A, cultured cardiac myocytes were transfected with GRP78-(+221)-Luc and a β-galactosidase reporter and later infected with AdV-Con or AdV-DN-ATF6. After 24 h, they were subjected to 20 h of sI, followed by extraction and reporter enzyme assays. Shown are the mean relative luciferase values (luciferase/β-galactosidase), expressed as -fold control (no sI) ± S.E. *, p ≤ 0.05 different from control and all other values. B, ERSEs 1, 2, and 3 are shown; the CCAAT-like motifs required for NF-Y and ATF6 binding are designated in boldface type. The mutations in the CCAAT-like motif and NFY-binding sites are shown in lowercase type below each ERSE. These mutations replicated those previously studied in the human GRP78 promoter (32). C, shown are the eight constructs used in this study. Construct 1 is the plasmid encoding the GRP78 promoter from −284 to +221 containing three ERSEs. Constructs 2–8 contain mutations that disrupt the ERSEs, as shown. D, cultured cardiac myocytes were co-transfected with either an empty vector control or with one of the eight constructs and β-galactosidase and, 48 h later, subjected to 20 h of sI, extracted, and then analyzed for reporter activities. Shown are the mean relative luciferase values (luciferase/β-galactosidase), expressed as -fold control (no sI) ± S.E. *, #, §, 196, and &, p ≤ 0.05 different from control and all other values.

FIGURE 5.

Electromobility shift assays and ATF6 immunoblots. A, to examine ATF6 binding to the GRP78 ERSE 2, EMSA analysis was carried out as described under “Materials and Methods.” Recombinant ATF6-(116–373), used in the reactions analyzed in lanes 2, 4, and 6, was prepared by in vitro transcription/translation and then added to neonatal rat ventricular myocyte nuclear extracts, as previously described (28). The ³²P-labeled GRP78 ERSE 2 probe was added to initiate the binding reactions. Complex 1 is due to direct binding of nuclear extract-derived proteins (e.g. NF-Y and YY1) to the ERSE and has been shown to be required before ATF6 will bind under these conditions. For the supershift, EMSA was carried out as described, except for the addition of either preimmune or ATF6 antiserum to lanes 3, 4, 5, and 6, as shown. B, competition binding EMSA was carried out as described above, except for the addition of 10× or 50× unlabeled GRP78 ERSE 2 oligonucleotide to lanes 3 and 4 or 10× and 50× mutated GRP78 ERSE 2 oligonucleotide to lanes 6 and 7. C and D, cultured cardiac myocytes were treated with sI or sI/R, extracted, and subjected to subcellular fractionation, as described under “Materials and Methods” (n = 2 cultures/treatment; ∼20 × 10⁶ cells/culture). The ER fraction (C) and the nuclear fraction (D) were then analyzed by SDS-PAGE and immunoblotting for ATF6β. The region of the gel that includes the full-length endogenous ATF6 is shown in C, whereas the region of the gel that includes the cleaved active form of ATF6β is shown in D. Also shown are VDAC and NP-62, which are used as loading controls for ER and nuclear fractions, respectively. The reason for the ATF6β doublet is not known; however, it is possible that this could represent variable cleavage by S1P and/or S2P. E, immunoblots shown in C and D were quantified by densitometry. Shown are ATF6β/VDAC (ER) or ATF6β/NP-62 (Nuclear) as relative band intensity. ER and nuclear sI and sI/R values were normalized to ER and nuclear control values, respectively. n = 2 cultures/treatment, as described in C and D.

FIGURE 6.

Effects of ATF6-targeted siRNA or miRNA on GRP78 mRNA and cell survival. A, cultured cardiac myocytes were transfected with a control or ATF6 siRNA. After 24 h, cultures were subjected to sI for 20 h plus sR for a subsequent 24 h. Cultures were then extracted and subjected to reverse transcription quantitative PCR to determine the levels of rat atf6 mRNA (bars 1 and 2) or rat grp78 mRNA, relative to gapdh. Shown are the mean values of grp78/gapdh mRNA, expressed as -fold control (bar 1) ± S.E. * and #, p ≤ 0.05 different from all other values. B, cultured cardiac myocytes were infected with AdV encoding a control, nontargeted miRNA (Con, bar 1) or with AdV encoding ATF6-targeted miRNA. Cultures were then extracted and subjected to reverse transcription quantitative PCR to determine the levels of ATF6 mRNA (bars 1 and 2) relative to GAPDH. Shown are the mean values of rat ATF6 or rat grp78/gapdh mRNA, expressed as -fold control (bar 1) ± S.E. for each treatment (n = 3 cultures/treatment; three experiments compiled). Alternatively, cultures were infected with control or ATF6-targeted miRNA, subjected to sI or sI/R, as shown, and then examined for cell death, as described under “Materials and Methods.” Shown are the mean values of cell death ± S.E., expressed as the percentage of dead cells. *, #, and §, p ≤ 0.05 different from control and different from each other.