Endoplasmic Reticulum Stress Stimulates Heme Oxygenase-1 Gene Expression in Vascular Smooth Muscle

Heme oxygenase-1 (HO-1) is a cytoprotective protein that catalyzes the degradation of heme to biliverdin, iron, and carbon monoxide (CO). In the present study, we found that endoplasmic reticulum (ER) stress induced by a variety of experimental agents stimulated a time- and concentration-dependent increase in HO-1 mRNA and protein in vascular smooth muscle cells (SMC). The induction of HO-1 by ER stress was blocked by actinomycin D or cycloheximide and was independent of any changes in HO-1 mRNA stability. Luciferase reporter assays indicated that ER stress stimulated HO-1 promoter activity via the antioxidant response element. Moreover, ER stress induced the nuclear import of Nrf2 and the binding of Nrf2 to the HO-1 antioxidant response element. Interestingly, ER stress stimulated SMC apoptosis, as demonstrated by annexin V binding, caspase-3 activation, and DNA laddering. The induction of apoptosis by ER stress was potentiated by HO inhibition, whereas it was prevented by addition of HO substrate. In addition, exposure of SMC to exogenously administered CO inhibited ER stress-mediated apoptosis, and this was associated with a decrease in the expression of the proapoptotic protein, GADD153. In contrast, the other HO-1 products failed to block apoptosis or GADD153 expression during ER stress. These results demonstrated that ER stress is an inducer of HO-1 gene expression in vascular SMC and that HO-1-derived CO acts in an autocrine fashion to inhibit SMC apoptosis. The capacity of ER stress to stimulate the HO-1/CO system provides a novel mechanism by which this organelle regulates cell survival.

Heme oxygenase (HO) 1 catalyzes the rate-limiting step in the oxidative degradation of heme to biliverdin, iron, and car-bon monoxide (CO) (1). Subsequently, biliverdin is rapidly converted to bilirubin by biliverdin reductase. Three distinct isoforms of HO, HO-1, HO-2, and H0 -3, have been described (2)(3)(4). However, recent work suggests that HO-3 may be a pseudogene derived from HO-2 transcripts (4). The HO-2 isoform is constitutively expressed and is present in high concentrations in the brain and testes (5). In contrast, HO-1 is ubiquitously distributed and strongly induced by oxidative, nitrosative, osmotic, and hemodynamic stress (6 -10). Induction of HO-1 by these biochemical and biomechanical stimuli provides an important cellular defense mechanism against tissue injury (7,9,(11)(12)(13)(14). Several postulated mechanisms account for the beneficial effects of HO-1. HO-1 results in the catabolism of the pro-oxidant heme to bile pigments biliverdin and bilirubin, which are potent antioxidants capable of scavenging peroxy radicals and inhibiting lipid peroxidation (15,16). In addition, HO-1 induction is accompanied by increased ferritin activity, which exerts an additional antioxidant effect by chelating free iron (17). Finally, HO-1 liberates the diatomic gas CO that possesses significant anti-inflammatory and antiapoptotic properties (18 -20).
The endoplasmic reticulum (ER) is a dynamic membranous organelle that plays a critical role in the folding, transport, and processing of newly synthesized proteins (21,22). Numerous infectious agents, environmental toxins, and adverse metabolic conditions interfere with protein folding leading to ER stress. To protect against ER stress, the ER has evolved highly specific signaling pathways collectively termed the unfolded protein response (UPR). Three ER-resident transmembrane proteins, the kinase and endoribonuclease IRE1, the PERK kinase, and the basic leucine-zipper transcription factor ATF6, have been identified as proximal sensors of ER stress (23). Activation of these sensors results in the up-regulation of genes encoding ER chaperone proteins such as GRP78 that increase protein folding activity and prevent protein aggregation. In addition, the UPR results in translation attenuation to reduce the load of protein synthesis and to prevent further accumulation of unfolded proteins. Finally, when ER function is severely impaired, apoptosis is induced. This apoptotic event is triggered by the activation of the proapoptotic transcription factor GADD153 and/or by the activation of ER-associated caspase-12 (24 -26).
Because HO-1 is induced by several forms of cellular stress, the present study examined whether ER stress regulates HO-1 expression. We now report that ER stress is a novel inducer of HO-1 in vascular smooth muscle cells (SMC). Interestingly, the activation of the HO-1 gene by ER stress does not occur via the conventional UPR-responsive element but rather through the antioxidant-responsive element (ARE). In addition, we have shown that HO-1-derived CO functions in an autocrine manner to suppress apoptosis. The HO-1/CO system may represent an important adaptive mechanism to promote cell survival during periods of ER stress.
Cell Culture-Vascular SMC were isolated by elastase and collagenase digestion of rat thoracic aortas and characterized as described previously (27). SMC were cultured serially in minimum essential medium supplemented with 10% serum, 2 mM L-glutamine, 5 mM Tes, 5 mM HEPES, 100 units/ml penicillin, and 100 units/ml streptomycin.
mRNA and Protein Analysis-mRNA and protein levels were determined by Northern and Western blotting, respectively. Total RNA (30 g) was fractionated by electrophoresis, blot-transferred to Gene Screen Plus membranes (PerkinElmer Life Sciences), and prehybridized for 4 h at 68°C in hybridization buffer (rapid-hyb buffer; Amersham Biosciences). Membranes were hybridized overnight at 68°C in hybridization buffer containing [ 32 P]DNA probes (1 ϫ 10 8 cpm) for HO-1, GADD153, glyceraldehyde-3-phosphate dehydrogenase, or 18 S RNA. DNA probes were generated by RT-PCR (8,28,29) and labeled with [␣-32 P]dCTP using a random priming kit (Amersham Biosciences). After hybridization, membranes were washed and then exposed to x-ray film at Ϫ70°C in the presence of intensifying screens.
HO-1 Promoter Analysis-The sequences for the wild-type enhancer SX2 coupled to a minimum HO-1 promoter as well as the mutant enhancer SX2-M739 that had its three ARE core sequences mutated have been described previously (30,31). These promoter/luciferase constructs (1 g/ml), pCMV␤-galactosidase (1 g/ml), and a plasmid expressing Nrf2 (pEF-F2/Nrf2) were transfected into SMC using Lipofectamine (Invitrogen), cultured for an additional 24 h, and then treated with various agents. Cells were then collected, lysed, and luciferase activity was measured using a luciferase assay system (Promega) and a TD-Zoe luminometer (Turner Designs Inc., Mountain View, CA) and normalized with respect to ␤-galactosidase activity.
Electromobility Shift Assay-A consensus ARE oligomer containing the sequence 5ЈTTTATGCTGTGTCATGGTT-3Ј (core ARE sequence is underlined) was end-labeled with [␥-32 P]ATP and incubated with 3 g of nuclear extract for 30 min at room temperature in a reaction mixture containing 20 mM HEPES (pH 7.9), 0.1 mM EDTA, 40 mM KCl, 1 mM MgCl 2 , 0.5 mM dithiothreitol, 10% glycerol, and 1 g of poly(dI-dC). DNA-protein complexes were resolved by electrophoresis in 5% polyacrylamide gels run at 220 V for 3 h. In antibody supershift assays, 1 g of Nrf2 antibody was added to the reaction mixture and incubated for 3 h at 4°C before the addition of the probe.
Apoptosis-Apoptosis was monitored by measuring DNA fragmentation, caspase-3 activation, and the movement of phosphatidylserine from the internal to the external surface of the cell membrane as we have described previously (20).
CO Exposure-SMCs were exposed to CO via a previously described environmental chamber (28). Gas from stock tanks containing 1% CO in air and 5% CO 2 in air were combined in a stainless steel mixing cylinder and delivered into a humidified 37°C environmental chamber. Flow into the chamber was at 1 liter/min; CO levels in the chamber were continuously monitored by electrochemical detection using a CO analyzer (Interscan, Chatsworth, CA).
Statistics-Results are expressed as the means Ϯ S.E. Statistical analysis was performed with the use of a Student's two-tailed t test and an analysis of variance when more than two treatment regimens were compared. p Ͻ0.05 was considered statistically significant.

RESULTS
Treatment of vascular SMC with brefeldin A, a well established inducer of ER stress that inhibits the translocation of proteins from the ER to the Golgi apparatus (32), stimulated HO-1 mRNA and protein expression in a concentration-dependent manner (Fig. 1A). The induction of HO-1 by brefeldin A was delayed, with marked increases in HO-1 mRNA and protein appearing after 24 h of brefeldin A exposure (Fig. 1B). Homocysteine, which promotes protein misfolding in the ER by interfering with disulfide bond formation (21), also induced HO-1 protein expression ( Fig. 2A). Similarly, thapsigargin, which induces ER stress by inhibiting ER Ca 2ϩ -ATPase (33), stimulated the expression of HO-1 protein ( Fig. 2A). In addition, proteasomal inhibition, a known stimulus for UPR activation (34), induced HO-1 expression. Incubation of SMC with proteasome inhibitor MG-132 and proteasome inhibitor I increased HO-1 protein expression (Fig. 2B). The induction of ER stress by all of these agents was confirmed by the significant increase in the level of GRP78 (Fig. 2, A and B).
In the next series of experiments the molecular mechanism by which ER stress induces HO-1 expression was examined. Incubation of SMC with the protein synthesis inhibitor cycloheximide or the transcriptional inhibitor actinomycin D blocked the induction of HO-1 mRNA and protein by brefeldin A (Fig. 3A). To test for a possible role of mRNA stability in the induction of HO-1 by brefeldin A, SMC were treated with actinomycin D, and the decay of HO-1 message followed. These actinomycin D chase experiments revealed that HO-1 mRNA has a half-life of ϳ2 h (Fig. 3B). However, the stability of HO-1 transcripts was unaffected by brefeldin. To determine whether increases in HO-1 expression in response to ER stress involve transcriptional activation of the gene, SMC were transiently transfected with a HO-1 promoter construct and promoter activity was monitored. Cadmium, an established activator of the HO-1 promoter (31), induced an ϳ6-fold increase in promoter activity (Fig. 3C). Brefeldin A and the proteasome inhibitor MG-132 also stimulated more than a 2-fold increase in promoter activity (Fig. 3C). Furthermore, the expression of Nrf2 induced a significant increase in HO-1 promoter activity (Fig.  3C). Interestingly, mutation of the AREs attenuated basal activity and abolished the response to cadmium, brefeldin A, proteasome inhibition, and Nrf2 (Fig. 3C). These results indicated that ER stress activates HO-1 transcription via the ARE.
Electromobility shift assays using an ARE probe corresponding to the mouse HO-1 promoter detected a complex that was significantly increased in intensity by the ER stress inducers, brefeldin A, thapsigargin, and homocysteine (Fig. 4A). Because the transcription factor Nrf2 appears crucial for ARE-mediated gene expression (Fig. 3C) (35), we investigated whether ER stress stimulated the binding of Nrf2 to the ARE. Indeed, supershift electromobility shift assay reactions using an antibody directed against Nrf2 retarded the migration of the ARE complex (Fig. 4A), indicating the presence of Nrf2 in the AREnuclear protein complex (Fig. 4B). Moreover, the activation of Nrf2 by homocysteine (Fig. 4B) and other ER stress agents was corroborated by the nuclear accumulation of Nrf2.
In subsequent experiments the functional role of HO-1 during ER stress was examined. Treatment of SMC with homocysteine resulted in a significant increase in the rate of apoptosis as reflected by annexin V binding (Fig. 5). However, the addition of the HO inhibitors, zinc protoporphyrin-IX or tin protoporphyrin-IX, increased homocysteine-mediated apoptosis by ϳ2-fold. These findings suggested that the induction of HO-1 by ER stress may function in an autocrine manner to inhibit SMC apoptosis. Subsequently, we determined which of the HO-1 products mediates this cytoprotective effect. Incubation of vascular SMC with brefeldin A or thapsigargin for 24 h resulted in a significant increase in caspase-3 activity that was reversed by the exogenous administration of CO (200 ppm) (Fig. 6A). Similarly, the stimulation of DNA fragmentation by brefeldin A, thapsigargin, or homocysteine was prevented by CO (Fig. 6, B and C). In contrast, the addition of the bile pigments, biliverdin or bilirubin, or free iron had no effect on DNA laddering (Fig. 6D). Furthermore, increasing endogenous

FIG. 3. Mechanism of HO-1 induction by brefeldin A in vascular SMC.
A, effect of cycloheximide (Cx) and actinomycin D (ActD) on brefeldin A (BF)-induced HO-1 expression. SMC were treated with BF (1.0 M) in the presence or absence of Cx (5 g/ml) or ActD (2 g/ml) for 24 h, and HO-1 protein and mRNA were determined by Northern and Western blotting, respectively. Similar findings were observed in four separate experiments. B, effect of BF on HO-1 mRNA stability. SMC were treated with BF (0.5 M) for 24 h and then exposed to actinomycin D (5 g/ml). Total RNA was collected at the indicated time points, subjected to Northern blotting, and autoradiograms were subjected to densitometric analysis and corrected for the internal control (GAPDH). Open circles represent SMC treated with ActD alone, and solid symbols represent cells treated with ActD and BF. Results are representative of three separate experiments. C, effect of BF on HO-1 transcription. SMC were co-transfected with a mouse HO-1 promoter construct (SX2) or a mutated mouse HO-1 promoter construct (M739) and a ␤-galactosidase construct (pCMV␤gal) to control for transcription efficiency. After 24 h SMC were exposed to cadmium (Cd; 10 M), BF (0.5 M), or MG132 (10 M) for 6 h and then analyzed for luciferase and ␤-galactosidase activity. Results are expressed as the means Ϯ S.E. of three to five separate experiments. *, statistically significant increase in promoter activity. CO production via the exogenous addition of hemin blocked brefeldin A-mediated DNA fragmentation (Fig. 6E).
Finally, we investigated the effect of CO on specific pathways that promote apoptosis during ER stress. In particular, we examined whether CO regulates the activity of caspase-12 or the expression of GADD153 because both these proteins have been implicated in the induction of apoptosis by ER stress (24 -26). Incubation of vascular SMC with brefeldin A stimulated the processing of procaspase-12. Western blot analysis demonstrated a significant decline in procaspase-12 levels following 24 h of brefeldin exposure (Fig. 7A) that paralleled the onset of apoptosis. However, the application of CO (200 ppm) had no effect on the activation of caspase-12 by brefeldin A (Fig.  7A). In contrast, CO markedly attenuated the increase in GADD153 protein expression induced by brefeldin A or homocysteine (Fig. 7B). Moreover, CO blocked the time-dependent induction of GADD153 mRNA by homocysteine (Fig. 7C). The other HO-1 products failed to inhibit GADD153 expression (data not shown). DISCUSSION In the present study, we identified ER stress as a novel inducer of HO-1 gene expression in vascular SMC. We found that HO-1 is induced by a diverse set of conditions that cause ER stress, including inhibition of protein transport, calcium depletion, homocysteine exposure, and protein accumulation. Furthermore, we discovered that the induction of HO-1 inhibits SMC apoptosis during ER stress via the release of CO. These findings implicate the HO-1/CO system as a critical modulator of cell survival during periods of ER stress.
Our finding that ER stress stimulates HO-1 gene expression in SMC extended an earlier study showing that thapsigargin induces a marked rise in HO-1 mRNA in primary neuronal cells (36). In addition, the induction of HO-1 by homocysteine in cultured cells is also consistent with the increase in HO-1 expression observed in a murine model of hyperhomocysteine- mia (37). In contrast, ER stress fails to stimulate HO-1 expression in cultured hepatoma cells (38). The dissimilar results between these studies may reflect differences in cell type, culture conditions, or duration and extent of ER stress. Interestingly, a recent report (39) found that ER stress promotes hyaluronan deposition and leukocyte adhesion to SMC, demonstrating that SMC represent a physiologically relevant cellular target of ER stress.
ER stress induces HO-1 expression in a concentration-and time-dependent manner. The induction of HO-1 by ER stress is dependent on de novo RNA synthesis and does not involve alterations in HO-1 mRNA stability. Transient transfection experiments indicate that ER stress stimulates the transcriptional activation of the HO-1 gene. However, the activation of HO-1 transcription by ER stress does not occur via the classical mammalian UPR element (5Ј-CCAAT-N 9 -CCACG-3Ј). A computer search of the entire 15 kb of the 5Ј-flanking region of the mouse HO-1 gene failed to detect any candidate UPR elements. Similarly, computer analysis of a BAC clone sequence (Gen-Bank TM accession number Z82244) containing the entire human HO-1 gene and greater than 20 kb of the 5Ј-flanking region did not reveal a UPR element (38). Instead, ER stress induces HO-1 transcription via the ARE. Electromobility shift assays found increased ARE binding in nuclear extracts from SMC exposed to ER stress. Moreover, when the ARE element is mutated, the stimulation of HO-1 promoter activity by ER stress is completely abrogated.
Many nuclear transcription factors have been reported to bind to the ARE. However, recent work indicates that the transcription factor Nrf2 plays a critical role in ARE-dependent HO-1 gene expression (35). Consistent with this, we found that stimulation of HO-1 promoter activity in vascular SMC by Nrf2 is dependent on the presence of AREs in the promoter. Furthermore, we detected Nrf2 binding to the HO-1 ARE in SMC following exposure to ER stress. Under normal conditions, Nrf2 is retained in the cytoplasm via its binding to the cytosolic protein Keap1 (40). However, changes in cellular redox potential and/or phosphorylation of Nrf2 lead to the release of Nrf2 from Keap1 and its translocation to the nucleus, resulting in ARE-mediated gene transcription (40,41). Interestingly, we observed that ER stress induces the accumulation of Nrf2 in the nucleus of vascular SMC. This finding is supported by a recent study showing that ER stress triggers the nuclear import of Nrf2 in embryonic fibroblasts (42). The activation of Nrf2 appears to involve the ER signaling protein, PERK, which has been shown to phosphorylate and activate Nrf2 in response to ER stress conditions (43). Collectively, these data suggest a model where PERK activation in response to ER stress stimulates the phosphorylation and nuclear translocation of Nrf2 where it complexes with the ARE in the HO-1 promoter to trigger gene expression.
HO-1 exerts important cytoprotective actions in response to numerous cellular insults, (7,9,(11)(12)(13)(14). In the present study, we have shown that HO-1 also counteracts the damaging effect of ER stress. The induction of HO-1 by ER stress contributes to cell survival because inhibition of HO activity augments apoptosis, whereas addition of a HO substrate prevents apoptosis during ER stress. The anti-apoptotic effect of HO-1 is likely mediated via the release of CO because the exogenous administration of CO mimics the cytoprotection afforded by HO-1. In contrast, the other HO-1 products do not protect cells from undergoing apoptosis. CO has no effect on the activation of the ER-specific caspase-12 by ER stress, but it markedly blunts the expression of the proapoptotic transcription factor, GADD153. Interestingly, cells with a targeted deletion of Nrf2 exhibit elevated levels of GADD153 and reduced survival following ER stress relative to wild-type cells (41,42). It is tempting to speculate that HO-1-derived CO synthesis contributes to the repression of GADD153 expression and increased survival in the wild-type cells.
Recent studies indicate that ER stress may also activate the intrinsic (mitochondrial) and extrinsic (death receptor) signaling pathways to trigger apoptosis (44 -46). Given that CO has been shown to block both these signaling pathways (19,20,47), it is likely that CO inhibits ER stress-mediated apoptosis via multiple mechanisms. Thus, the induction of HO-1 and the release of CO during ER stress may represent an important survival response in providing cells additional time to execute the UPR and restore ER function. In addition, HO-1-derived CO leads to cell cycle arrest (28,48), which might allow the cell to concentrate resources on alleviating stress rather than on growth and division.
The ability of HO-1-derived CO to block ER stress-mediated apoptosis may be of clinical significance. ER stress has been implicated in the pathogenesis of a variety of diseases. Cell death following ER stress is believed to contribute to the progression of neurodegenerative disorders such as Alzheimer and Parkinson disease and to the loss of pancreatic ␤ cells in diabetes (49 -51). Furthermore, ER stress may underlie cell death following ischemia-reperfusion (52). In addition, damage to vascular cells by ER stress in hyperhomocysteinemic states promotes the development of atherosclerosis (53)(54)(55)(56). Thus, approaches that target HO-1 or CO to sites of ER stress offer a promising therapeutic modality in treating a variety of disorders related to ER stress.
In conclusion, the present study has demonstrated that ER stress induces HO-1 gene expression in vascular SMC and that HO-1-catalyzed CO formation inhibits apoptosis during ER stress. In addition, it has shown that the anti-apoptotic effect of CO is associated with a decrease in GADD153 expression. The HO-1/CO system represents a potentially new therapeutic target in ameliorating diseases associated with ER stress.