Roles for ER-associated Degradation (ERAD) and the Novel ER Stress Response Gene, Derlin-3, in the Ischemic Heart

Abstract

Rationale

Stresses, such as ischemia, impair folding of nascent proteins in the rough endoplasmic reticulum (ER), activating the unfolded protein response (UPR), which restores efficient ER protein folding, thus leading to protection from stress. In part, the UPR alleviates ER stress and cell death by increasing the degradation of terminally mis-folded ER proteins via ER-associated degradation (ERAD). ERAD is increased by the ER stress modulator, activating transcription factor 6 (ATF6), which can induce genes that encode components of the ERAD machinery.

Objective

Recently, it was shown that the mouse heart is protected from ischemic damage by ATF6; however, ERAD has not been studied in the cardiac context. A recent microarray study showed that the Derlin-3 (Derl3) gene, which encodes an important component of the ERAD machinery, is robustly induced by ATF6 in the mouse heart.

Methods and Results

In the present study, activated ATF6 induced Derl3 in cultured cardiomyocytes, and in the heart, in vivo. Simulated ischemia (sI), which activates ER stress, induced Derl3 in cultured myocytes, and in an in vivo mouse model of myocardial infarction, Derl3 was also induced. Derl3 overexpression enhanced ERAD and protected cardiomyocytes from sI-induced cell death, while dominant-negative Derl3 decreased ERAD and increased sI-induced cardiomyocyte death.

Conclusions

This study describes a potentially protective role for Derl3 in the heart, and is the first to investigate the functional consequences of enhancing ERAD in the cardiac context.

Introduction

Synthesis and folding of secreted, membrane-bound, and organelle-targeted proteins takes place in the ER.^1,2 The ER environment is sensitive to stresses that impair folding of ER proteins.² The aggregation of mis-folded proteins can be toxic and has been implicated in nearly sixty diseases,³ including desmin-related cardiomyopathy.⁴ In addition, a recent study has shown that transgenic overexpression of preamyloid oligomers (PAO) is toxic to cardiomyocytes.⁵

An accumulation of mis-folded ER proteins triggers the ER stress response (ERSR), also known as the unfolded protein response (UPR),⁶ in many different cell and tissue types.^7-9 Initial ERSR signaling is oriented toward resolving the stress and reducing the protein folding load on the ER, leading to survival. Such pro-survival responses include ER-associated protein degradation (ERAD), through which terminally mis-folded proteins that accumulate in the ER are degraded. Terminally mis-folded ER proteins are translocated out of the ER lumen to be degraded by ERAD.¹⁰ While the importance of ERAD in the heart has been discussed,^11,12 few studies have analyzed the functionality of ERAD in the myocardium.

Although many details of ERAD remain to be elucidated, it is believed that in part, the ATF6 branch of the ERSR induces some of the ERAD genes. ATF6 is an ER-transmembrane protein that is cleaved upon ER stress; the N-terminal fragment of ATF6 translocates to the nucleus and acts as a transcription factor to induce numerous protective ERSR genes. ATF6 can bind to cis regulatory elements in ERSR genes, including the ER-stress-response element (ERSE), ER-stress-response element II (ERSE-II), and to a lesser extent, the unfolded protein response element (UPRE)¹³.

To investigate the effects of ATF6 and ERAD in the heart we generated a line of transgenic (TG) mice featuring cardiac-restricted expression of a novel protein in which the mutant mouse estrogen receptor (MER) was fused to the C-terminus of the active fragment of ATF6. This fusion protein, ATF6-MER, is constitutively expressed in ATF6-MER TG mouse hearts, but is active only upon tamoxifen administration.¹⁴ Activation of ATF6 in TG mouse hearts enhances function and reduces necrosis and apoptosis during ex vivo I/R,¹⁴ suggesting that ATF6 contributes to reducing damage and enhancing recovery from I/R. It is possible that some of these effects might be exerted through ATF6-mediated enhancement of ERAD.

Whole-genome microarray analyses of ATF6-MER TG mouse hearts showed that 381 transcripts exhibited ATF6-dependent increases.¹⁵ One of the most induced genes was Derlin-3 (Derl3), a member of a family comprised of Derl1, 2 and 3. The Derlins were named ‘der’ for ‘degradation in the ER’.¹⁶ Derl1 facilitates the retro-translocation of mis-folded proteins from the ER lumen to the cytosolic face of the ER.¹⁷ Thus, it is possible that the other Derl family members, including Derl3, exert similar functions.

Although studies of Derl1, 2 and 3 have been done with non-cardiac cell types,¹⁸ the mechanism of Derl3 induction in any cell type is unknown, and the Derlin family has not been studied in the heart. Accordingly, we examined the mechanism of induction and function of Derl3 in the heart.

Methods

Animals

The transgenic mice used in this study have been described previously.¹⁴ Approximately 100 neonatal rats and 24 adult male mice were used in this study. All procedures involving animals were in accordance with the San Diego State University Institutional Animal Care and Use Committee.

Cultured Cardiac Myocytes

Primary neonatal rat ventricular myocyte cultures (NRVMC) were prepared and maintained in culture, as previously described.¹⁹

For the expanded Materials and Methods, see the Online Data Supplement (http://circres.ahajournals.org).

Results

Promoter Sequence Analysis of ATF6-Regulated Genes in the Heart Identifies Derl3

We previously identified ATF6-regulated genes in ATF6-MER TG mouse hearts.¹⁵ To determine which of the genes from the microarray analyses of ATF6-MER TG mouse hearts might be direct targets of ATF6, the regulatory regions of each were analyzed for the ATF6 binding sites, ERSE, ERSE-II and UPRE.^20-22 ERSE, ERSE-II and UPRE sequences were found 9.6-, 8.4- and 2.1-fold more frequently, respectively, than in the whole mouse genome (Fig. 1). Amongst the 607 ATF6-regulated genes in the heart, 16 have canonical ERSEs, ERSEIIs and/or UPREs (Online Tables 1-III), while 211 have elements with 1 mis-match (Online Tables IV-VI); mismatches of 1bp have been shown in other studies to be potential ATF6-binding elements.^13,20 Thus, 227 genes are likely to be direct targets of ATF6.

Figure 1. ATF6-MER TG Mouse Microarray is Enriched in Genes with ERSEs, ERSEIIs and/or UPREs.

To determine the enrichment of the numbers of ERSEs in the ATF6-regulated gene list, a bootstrapping analysis was performed, as described in the Methods. The number of times these elements were found in each search was plotted as a function of the number of searches, resulting in a representation of the frequency with which each was found in the whole genome (Fig. 1A, 1C and 1E, black bars) and in the ATF6-regulated gene list (Fig. 1A, 1C and 1E, white bars). The data used to generate each bar graph are shown in the tables in Figure 1B, 1D and 1F. For example, 417 of 1,000 searches of the whole genome resulted in the identification of zero ERSEs (row 1, Fig. 1B). The average number of times each element was found per search (i.e. average frequency) is shown as AveGenome and Ave_ATF6. The average frequency with which each element was found in the ATF6-regulated gene list was divided by the average in whole genome, and defined as the fold-enrichment of the array.

One of only two ATF6-regulated genes with two canonical ERSEs is Derlin-3 (see Online Table I, Derl3, NM_024440). Since it may play a role in ERAD, we investigated the mechanism of induction and the function of Derl3 in the heart and cultured cardiac myocytes. To examine transcriptional regulation cultured cardiac myocytes were transfected with several luciferase reporter constructs (Fig. 2A) + subsequent infection with either a control (AdVCon) or an adenovirus that encodes activated ATF6 (AdVATF6). Luciferase activation in cultures infected with AdVATF6 and transfected with reporter Construct 1 was 200-fold of control (Fig. 2B, Bar 4). ATF6-mediated luciferase induction was reduced by 75 to 80% in Constructs 2 and 3, which each contain one mutated ERSE (Fig. 2B, Bars 5, 6). The ERSR activator, tunicamycin (TM), conferred robust induction of Construct 1 (Fig. 2B, Bar 7); however, Constructs 2 and 3 exhibited decreased induction (Fig. 2B, Bars 8, 9). Thus, maximal ATF6- or TM-mediated induction of the Derl3 promoter was dependent upon both ERSE1 and ERSE2.

Figure 2. Effect of ATF6 Overexpression and Tunicamycin on Derl3 Promoter Activity in Cultured Cardiac Myocytes.

Panel A: The mouse Derl3 promoter (−1359 to +26, Construct 1) and versions harboring mutations in putative ERSE elements located at either −183 to −165 (Construct 2) or −285 to −267 (Construct 3), were cloned into a luciferase expression construct.

Panel B: NRVMCs were transfected with luciferase constructs 1, 2, or 3, and CMV-β-gal and then infected with AdV-Con or AdV-ATF6, as previously described ¹⁹. Twenty-four hours after infection, cultures were treated ± TM; 16h later, cultures were extracted and luciferase and β-galactosidase reporter enzyme activities were determined, as previously described ¹⁹. Shown are the mean relative luciferase activities (luciferase/β-galactosidase), expressed as the fold of construct 1 treated with AdVCon (bar 1) ± SE for each treatment (n = 3 cultures per treatment, sum of 3 separate experiments). *, + = p≤0.05 from all other values by ANOVA.

ATF6 Induces Derl3 in Mouse Hearts

The ability of ATF6 to induce the other Derlin family members was also examined. While neither Derl1 nor Derl2 were induced by tamoxifen in the ATF6 TG mouse hearts (Fig. 3A, Bars 1-8), Derl3 was induced by 400-fold by tamoxifen, but only in the TG mouse hearts (Fig. 3A, Bars 9-12); thus, only Derl3 was ATF6-inducible in the heart, consistent with the lack of ERSEs in the Derl1 and Derl2 genes (Online Table I).

Figure 3. Effect of Activated ATF6 on Derlin Family Member Induction in ATF6-MER TG Mouse Hearts.

Panel A: NTG and ATF6-MER TG mice were treated ± vehicle or tamoxifen, and RNA was extracted from hearts, as previously described ¹⁴. RNA samples were subjected to RT-qPCR to examine the levels of mRNAs for Derl1, Derl2, Derl3, and GAPDH. Shown are the mean values of each target gene/GAPDH mRNA, expressed as the fold of NTG vehicle ± SE for each target gene (n = 3 mice per treatment). TG = ATF6-MER transgenic; NTG = non-transgenic; Veh = vehicle; Tx = tamoxifen. ** = p≤0.01 from all other values by ANOVA.

Panels B and C: TG mice were treated ± Veh (Panel B) or Tx (Panel C), and hearts were sectioned for immunofluorescence confocal microscopy. Sections were co-stained with Derl3 and actin, as described in the Methods.

Derlin levels were relatively low in sections from vehicle-treated ATF6-MER TG hearts (Figure 3B, Derl3) and tamoxifen-treated NTG mouse hearts (not shown). In contrast, tamoxifen-treated ATF6-MER TG mouse hearts exhibited robust Derl3 expression that co-localized primarily with actin-postive cardiomyocytes (Fig. 3C, overlay). Thus, ATF6 induces Derl3 expression in cardiac myocytes, in vivo.

Derl3 is an ATF6-Inducible ER Stress Response Gene

The effects of ATF6 or ER stress on Derlin mRNA in cultured cardiac myocytes were examined. Derl3 was the only family member that exhibited ATF6-inducibility, in cultured cardiac myocytes (Fig. 4A, Bars 1-6). When cells were treated with TM, Derl1 and Derl2 mRNA levels were 3- and 8-fold of control, respectively (Fig. 4B, Bars 2, 6); however, neither was affected by dominant-negative ATF6 (Fig. 4B, Bars 4, 8). In contrast, upon TM treatment, Derl3 mRNA was 200-fold of control (Fig. 4B, Bar 10), and this induction was attenuated by more than half by dominant-negative ATF6 (Fig. 4B, Bar 12).

Figure 4. Derl3 mRNA is Induced by ATF6 and TM in Cultured Cardiac Myocytes.

Panel A: NRVMCs were infected with either AdV-Con or AdV-ATF6 (n = 3 cultures per treatment). Forty eight hours after infection, cultures were extracted and the RNA was subjected to RT-qPCR to examine the levels of mRNA for the target genes described in Figure 3. Shown are the mean ± SE for each target gene (n = 3 cultures per treatment). * = p≤0.05 different from all other values by ANOVA.

Panels B and C: NRVMC were infected with either AdV-Con or AdV-DNATF6 (n = 3 cultures per treatment). Twenty four hours after infection, cultures were treated with TM for 16h (10mg/ml, Panel B) or sI for 20h (Panel C). Cells were extracted and subjected to RT-qPCR to examine the levels of mRNA for the target genes described in Panel A. Shown are the mean ± SE for each target gene (n = 3 cultures per treatment).

Simulating ischemia (sI) activates ER stress in cardiac myocytes;¹⁹ accordingly, the effect of sI on Derlin expression was examined. While Derl1 and Derl2 have been shown to be inducible by TM in XBP1 knockout MEFs¹⁸, in the current study in NRVMCs, sI had no effect on Derl1, and Derl2 was only slightly increased, but this increase did not reach significance (Fig. 4C, Bars 2, 6). In contrast, sI significantly increased Derl3 to ~2.5-fold of control (Fig. 4C, Bar 10); moreover, sI-mediated Derl3 induction was completely blocked by dominant-negative ATF6 (Fig. 4C, Bar 12), which was different than the partial blockage of Derl3 induction following TM treatment. This could be due to the higher dynamic range of induction experienced with TM, or due to other potential mediators of Derl3 induction during TM but not sI. Thus, Derl1 and Derl2 were induced by the prototypical ER stressor, TM, independently of ATF6. Moreover, only Derl3 was induced by the physiological ER stressor, sI, in an ATF6-dependent manner.

ATF6 is Activated and Derl3 Is Induced in an In Vivo Model of Myocardial Infarction

We previously showed that ER stress is activated in the mouse heart by myocardial infarction (MI).¹⁹ During ER stress, full-length, 90KD ATF6 is converted to 50KD ATF6 which is an active transcription factor.^6,23 Accordingly, the effects of ischemia on ATF6 and Derl3 upon MI were examined in vivo. An MI time course indicated that he 50KD form of ATF6 increased significantly after 16h of MI, remained elevated after 1d and 3d, and decreased at 4d of MI, consistent with activation of ATF6²⁴ (Fig. 5 Panels A-B). The 50KD band migrated slightly further than a 3xFlag-tagged form of cleaved, N-terminal ATF6, consistent with its identity as cleaved ATF6 (Online Fig. I Panel A).

Figure 5. Derl3 is Up-Regulated by Myocardial Infarction in Mouse Hearts.

Panels A-B: NTG mice were subjected to sham infarct surgery or to permanent occlusion myocardial infarction for the indicated times. Animals were sacrificed and hearts were used to prepare tissue extracts for western blot analysis, as previously described,²⁹ n=3 animals per time point. * = p<0.05 different from sham, determined by two-way T-Test.

Panel C: NTG mice were subjected to sham infarct surgery or to permanent occlusion myocardial infarction for 4d. Animals were then sacrificed and hearts were used to prepare tissue extracts for RT-qPCR, as previously described,²⁹ sham n = 4; MI n= 7 to 9. * = p<0.05 different from all other values determined by ANOVA.

Panels D-E: Mice were subjected to surgeries as in Panel C, and then scarified and hearts were sectioned for confocal immunocytofluorescence as previously described,²⁹ n=3 mice per treatment, one heart/treatment shown in this figure. Heart sections were stained for Derl3 (green) (1), or tropomyosin (red) (2), and an overlay is shown (3). Samples were viewed by laser scanning confocal immunofluorescent microscopy as previously described ¹⁹. Arrows 1 point to Derl3-positive cardiac myocytes in the infarct border zone, and arrows 2 point to Derl3-positive non-cardiac myocytes in the infarct zone.

The mRNA levels of all Derl family members increased in 4d MI mouse hearts, compared to sham, with Derl1 and Derl3 reaching significance, and Derl3 exhibiting the most robust up-regulation of ~6-fold (Fig. 5C, bars 2, 4, 6). An MI time course showed that while Derl3 mRNA showed a trend of being increased after 1d MI, it was significantly increased after 4d and 7d of MI (Online Fig. 1 Panel B, Online Supplement). The increase in p50 ATF6 by 16h of MI, and continued elevation after 1 and 3d of MI, were consistent with the possibility that ATF6 could induce Derl3 mRNA as early as 1d after MI. However, since the elevation of p50 ATF6 after 1 and 3d of MI, and the elevated Derl3 mRNA after 1d of MI did not reach statistical significance, it is formally possible that ATF6 activation/inactivation may precede Derl3 induction, suggesting that ATF6 may induce Derl3 expression indirectly. For example, ATF6 is known to induce the ER stress response transcription factor, XBP1, which could induce Derl3. These limitations leave open the question of whether ATF6 directly induces Derl3 mRNA in the heart, which we intend to investigate in future studies; nonetheless, ATF6 is a potent inducer of Derl3 gene expression in the myocardium.

Derl3 was low in sham mouse hearts (Fig. 5D, Panel 1), but elevated in surviving myocytes in the infarct zone (Fig. 5E, Panels 1 and 3, Arrow 1), as well as other tropomyosin-negative cells, which were most likely non-myocytes (Fig 5E, Panels 1 and 3, Arrow 2). Derl3 staining was perinuclear in the myocytes, consistent with expression in the ER.

Derl3 protects Cardiac Myocytes from Cell Death

ERAD reduces ER stress by degrading mis-folded ER proteins; accordingly, the effect of Derl3 on ER stress was examined. Following sI or sI/R, control cells exhibited a significant increase in the prototypical ER stress response proteins, GRP94 and GRP78 (Fig. 6A Lanes 1-3 vs 7-9 and 11-13; Fig. 6B and 6C, Bars 1, 3, 5). In contrast, cells infected withAdV-Derl3 exhibited reduced GRP94 and GRP78 (Fig. 6A Lanes 4-6, 10-12 and 14-16; Fig. 6B and 6C, Bars 2, 4, 6). GRP94 and GRP78 mRNA levels were also assessed following sI and sI/R (Online Fig. II, Panels A-B). Since neither GRP94 nor GRP78 mRNA levels were elevated following 20h of sI (Online Fig. II Panels A-B bars 3-4), we hypothesized that these mRNAs were increased at shorter times of sI. As indicated in Online Fig. II Panels C and D, GRP94 and GRP78 mRNAs increased at shorter times of sI, reaching a maximum at 12h, and AdV-Derl3 reduced the induction seen at these sI time points, consistent with the reduced levels of these proteins following sI seen in Fig. 6A-C. In addition, sI/R significantly increased GRP78 and GRP94 mRNA in AdVCon infected cells by ~15 and 17 fold, respectively (Online Fig. II Panels A-B bar 5). AdVderl3 infected cells exhibited a significant reduction in sI/R mediated GRP78 and GRP94 mRNA (Online Fig. II Panels A-B bar 5 vs 6). Together, these results suggest that overexpression of Derl3 attenuated sI- and sI/R-activated ER stress.

Figure 6. Derl3 Overexpression Attenuates ER Stress Activation, Caspase Activity, and Cell Death Following sI and/or sI/R.

Panels A-C: NRVMCs were infected + AdV-Con or AdV-Derl3 and 24h later, cultures were treated + sI or sI/R. Cultures were then extracted and cell lysates were analyzed for the levels of the prototypical ERSR proteins, GRP78 and GRP94, using an anti-KDEL antibody (Panel A), as previously described. Panels B and C display the relative blot intensities of each protein/GAPDH, expressed as the fold of control (bar 1) ± SE for each treatment (n = 3 cultures per treatment).

Panels D and E: NRVMCs were treated as in Panel A, and cultures were extracted and analyzed for the levels of CHOP using an anti-CHOP antibody in Panel D and quantified in Panel E.

Panel F: NRVMCs were treated as in Panel A. Cultures were then assayed for caspase-3 activation, as described in the Methods.

Panel G: NRVMCs were treated as in Panel A. Cultures were then stained with Hoescht and propidium iodide, and images of 5 randomly chosen fields per culture were viewed at 10x magnification on a fluorescent microscope. The numbers of Hoescht-positive (total) cells and propidium iodide-positive (dead) cells were then quantified using NIH ImageJ software (n = 3 cultures per treatment, sum of 3 separate experiments). Shown is the relative amount of cell death, expressed as fold of AdVCon Con, ± S.E for each treatment. For all panels, *, +, ! = p≤0.05 different from all other values by ANOVA.

Prolonged ER stress activates apoptosis;^6,23,25,26 accordingly, we assessed whether Derl3 overexpression decreased the ER stress-inducible, pro-apoptotic protein, C/EBP homologous protein (CHOP) and caspase-3 activation. AdV-Con cells exhibited ~25- and 5-fold increase in CHOP following sI and sI/R, respectively (Fig. 6D Lanes 1-2 vs 5-6 and 9-10; Fig. 6E Bars 1, 3, 5). In contrast, AdV-Derl3 cells exhibited decreased CHOP induction following sI and sI/R (Fig. 6D Lanes 3-4, 7-8, and 11-12; Fig. 6E Bars 2, 4, 6). sI did not activate caspase-3 in AdV-Con- or AdV-Derl3-infected cells, which is consistent with the lack of ATP required to activate caspase during sI (Fig. 6F Bars 1-2 vs 3-4). However, sI/R resulted in a ~2-fold increase in caspase-3 in AdV-Con-infected cells, which was attenuated in AdV-Derl3-infected cells (Fig. 6F Bars 5-6). In addition, AdV-Con-infected cells exhibited a ~2-fold and 3.5-fold increases in cell death following sI and sI/R, respectively (Fig. 6G Bars 1 vs 3 and 5). In contrast, AdV-Derl3-infected cells exhibited significantly less cell death in response to sI/R (Figure 6G Bar 6). Thus, overexpression of Derl3 attenuated ER stress and apoptosis in cardiac myocytes subjected to these stressors.

Derl3 Enhances ERAD and Reduces ER Stress

The effect of Derl3 on ERAD was examined using a mutant form of the alpha-1 antitrypsin protein (A1ATmut), which is constitutively mis-folded in the ER,²⁷. HeLa cells were co-transfected with plasmids expressing A1ATmut and Derl3-encoding plasmid. Derl3 overexpression decreased A1ATmut (Fig. 7A and 7B), indicating that Derl3 augmented clearance of mis-folded proteins in the ER during ERAD.

Figure 7. Derl3 Overexpression Enhances Mis-folded Protein Clearance and Attenuates ER Stress Activation.

Panels A-B: HeLa cells were co-transfected with plasmids encoding mutated α-1 antitrypsin (A1ATmut) along with either an empty vector, or increasing concentrations of a plasmid expressing Derl3. Forty eight hours later, cultures were extracted and lysates were analyzed for A1ATmut by immunoblot. Shown are the relative blot intensities of A1ATmut/GAPDH, expressed as the fold of control (bar 1) ± SE for each treatment (n = 3 cultures per treatment). **, ++ = p≤0.01 different from all other values by ANOVA.

Panel C: NRVMCs were transfected with a GRP78 promoter/luciferase construct and CMV-β-gal as described previously¹⁹, along with plasmids encoding A1ATwt or A1ATmut and Derl3. Forty eight hours later, cells were extracted and luciferase and β-galactosidase reporter enzyme activities were determined, as previously described¹⁹. Shown are the mean relative luciferase (luciferase/β-galactosidase), expressed as the fold-of-control (bar 1) ± SE for each treatment (n = 3 cultures per treatment, sum of 3 separate experiments). ** = p≤0.01 different from all other values by ANOVA.

The effect of A1ATmut overexpression on ER stress was determined by measuring GRP78 promoter activation. Cells were transfected with plasmids expressing a GRP78-luciferase promoter and A1ATwt, which folds properly or A1ATmut. While A1ATwt had no effect, promoter activity was ~2-fold of control in cells transfected with A1ATmut (Fig. 7C Bars 2, 3). Moreover, co-transfecting Derl3 decreased A1ATmut-mediated GRP78 promoter activation (Fig. 7C Bars 3 vs 5). Thus, Derl3 enhanced the removal of A1ATmut and, in so doing, decreased ER stress.

Derl3-DN and miDerl3 Enhance sI/R-mediated Cardiac Myocyte Death

An expression construct encoding an inactive, dominant-negative (DN) form of Derl3 was designed based on previous studies using Derl1 and 2 dominant-negative constructs.¹⁷ HeLa cells were transfected with A1ATmut, and control, Derl3 or Derl3-DN plasmids. Derl3 conferred an approximate 70 percent reduction in the level of A1ATmut, while Derl3-DN did not significantly change the levels of A1ATmut (Fig. 8A and 8B ). Thus, unlike wild-type Derl3, Derl3-DN does not increase the clearance of A1ATmut.

Figure 8. Overexpression of Derl3-DN Attenuates Misfolded Protein Clearance and Enhances Cell Death.

Panels A and B: HeLa cells were transfected with plasmids encoding A1ATmut along with either GFP, Derl3, or Derl3-DN. Forty eight hours post-transfection, cells were extracted and analyzed for A1ATmut levels by immunoblot. Shown are the relative blot intensities of A1ATmut/GAPDH, expressed as the fold of control (bar 1) ± SE for each treatment (n = 3 cultures per treatment, sum of 3 separate experiments).

Panel C: NRVMCs were transfected with plasmids encoding GFP alone or Derl3-DN, which encodes a Derl3-GFP fusion protein. Twenty four hours post-transfection, cultures were subjected to sI or sI/R, after which cells were collected and analyzed by flow cytometry. Cells were gated for GFP expression to identify transfected cells, and then the transfected cells were assessed for cell death by propidium iodide (PI) incorporation, as described in the Methods. The gate for GFP fluorescence intensity was established by comparing both groups of transfected cells to non-transfected, naïve cells. The same GFP intensity gate was used for the cells transfected with the control plasmid, as well as those transfected with Derl3-DN. Shown are the percentage of GFP-positive cells that are positive for PI, expressed as fold of control (bar 1), (± SE, n = 3 cultures per treatment, sum of 3 separate experiments). * = p≤0.05 different from all other values by ANOVA.

Panel D: NRVMCs were infected with either an adenovirus encoding a non-specific miRNA (AdVmiCon) or an miRNA directed to Derl3 (AdVmiDerl3). Forty eight hours after infection, cultures were extracted and the RNA was subjected to RT-qPCR to examine the levels of Derl3 mRNA. Shown are the mean ± SE for each target gene (n = 3 cultures per treatment). * = p≤0.05 different from AdVmiCon by two-way T-Test.

Panels E-F: NRVMCs were infected as in Panel D and 24h later, cultures were treated + sI. Cells were lysed and analyzed for Derl3 protein levels as Described in the Methods (n = 3 cultures per treatment, sum of 2 separate experiments). Shown is the relative amount of Derl3 protein, expressed as fold of AdVCon Con, ± S.E for each treatment. *, + = p≤0.05 different from all other values by ANOVA.

Panel G: NRVMCs were infected as in Panel D and 24h later, cultures were treated + sI or sI/R. Cultures were then stained with Hoescht and propidium iodide, and analyzed for cell death as in Figure 6G (n = 3 cultures per treatment, sum of 3 separate experiments). Shown is the relative amount of cell death, expressed as fold of AdVCon Con, ± S.E for each treatment. *, + = p≤0.05 different from all other values by ANOVA.

Cultured cardiac myocytes were transfected with Derl3-DN, subjected to sI, or sI/R, then analyzed for cell death by flow cytometry. Compared to control, cells transfected with Derl3-DN exhibited slight increases in cell death under control conditions, although this did not reach significance (Fig. 8C Bars 1 vs. 2). Cells transfected with Derl3-DN exhibited significant increases in cell death upon sI and sI/R (Fig. 8C Bars 3 vs. 4 and 5 vs 6).

To examine roles for Derl3 on cell survival, recombinant adenovirus encoding Derl3-targeted miRNA (AdVmiDerl3) was generated. Compared to control, AdVmiDerl3 decreased basal Derl3 mRNA levels by about 70% (Fig. 8D). To determine whether sI increased Derl3 protein, and whether AdVmiDerl3 could attenuate this increase, NRVMCs were treated with AdVmiDerl3, subjected to sI, and lysates were examined for Derl3 protein. Derl3 was increased with sI (Fig. 8E bars 1-3 vs 4-6 and Fig. 8F bar 1 vs 2). This increase was significantly attenuated by AdVmiDerl3 (Fig. 8E lanes 4-6 vs 10-12 and Fig. 8F bar 2 vs 4). AdVmiDerl3 slightly increased cell death under basal conditions, although this increase did not reach significance (Fig. 8G, bars 1 vs. 2); however, compared to control, AdVmiDerl3 significantly increased cell death during sI (Fig. 8G, bars 3 vs. 4) and sI/R (Fig 8G, bars 5 vs. 6).

Thus, Derl3-DN or Derl3 knock-down attenuated clearance of mis-folded ER proteins, and augmented sI and sI/R-mediated cell death, suggesting that the ability to clear mis-folded proteins from the ER is especially critical during ischemic stress.

Discussion

ER stress is activated in the heart during ischemia and that ATF6 protects the heart from ischemic damage, in vivo. ATF6 induces numerous known and novel ER stress response genes, however, only recently have potential ATF6 gene targets been identified in the heart;¹⁵ however, it is unclear which genes contribute to myocardial protection. In the current study, a detailed promoter analysis identified 16 ATF6-regulated genes which contained canonical ATF6-binding elements (Online Tables I-III). Only 2 of these genes contain two canonical ER stress response elements: GRP94, the well-known ER stress responsive chaperone, and Derl3, which encodes a protein likely to be involved in ERAD. Neither Derl3 nor the functional consequences of ERAD have been studied in the cardiac context; therefore, the present study focused on this gene and roles for ERAD in cultured cells and in the mouse heart, in vivo.

The Derlins are all ER-transmembrane proteins that associate with other proteins that participate in the degradation of mis-folded proteins in the ER. Derl1 and Derl2 associate with the p97 AAA ATPase and VIMP,¹⁷ Derl2 and Derl3 associate with proteins known to be involved in ERAD, EDEM and p97¹⁸ and the Derl3 yeast homologue, Der3p/Hrd1p interacts with Hrd3p and the retro-translocon pore complex protein Sec61p.²⁸ Overexpression of Derl2 and Derl3 accelerates degradation of known mis-folded substrates in HEK293 cells, while knockdown blocks degradation.¹⁸ In addition, the IRE1/XBP1 pathway was implicated to play a role in induction of Derl1 and Derl2; however, the specific mechanism of Derl3 induction was not determined.¹⁸

In the present study, it was shown that Derl3 was strongly induced by ATF6 in vivo, a finding that was replicated in cultured cardiac myocytes. Derl3 was also induced in the infarct border zone in an in vivo mouse model of myocardial infarction, and by simulated ischemia in cultured cardiac myocytes.

Since Derl3 enhances ER-associated degradation of mis-folded proteins in other cell types, we investigated the possibility that Derl3 may serve a potentially beneficial role during physiologically relevant stresses in cardiac myocytes, such as ischemia and/or reperfusion, We found that overexpressing Derl3 attenuated long-term ER stress response signaling and cell death in response to sI and sI/R, suggesting that enhancing elements of the ERAD machinery in the heart can protect from ischemic injury. While the current study indicates that Derl3 is induced in the heart, in vivo, by physiological stress, such as myocardial infarction, additional studies must be carried out using overexpression or knockdown of Derl3 or other ERAD components, in vivo, to determine whether enhancing or inhibiting ERAD in the heart can result in attenuation or exacerbation of ischemic damage in cardiac tissue.

Abbreviations

A1ATmut

α-1 antitrypsin mutant

AdV

Adenoviral

ATF6

Activating transcription factor 6

CHOP

C/EBP homologous protein

Derl1

Derlin-1

Derl2

Derlin-2

Derl3

Derlin-3

DN

Dominant-negative

ER

Endoplasmic reticulum

ERAD

Endoplasmic reticulum-associated degradation

ERSE

ER-stress-response element

ERSE-II

ER-stress-response element II

ERSR

Endoplasmic reticulum stress response

GRP78

Glucose-regulated protein 78

GRP94

Glucose-regulated protein 94

MER

Mutant mouse estrogen receptor

miCon

microRNA control (non-specific)

miDerl3

microRNA directed to Derl3

PAO

Preamyloid Oligomers

PI

Propidium Iodide

sI

Simulated ischemia

sI/R

Simulated ischemia/reperfusion

TG

Transgenic

TM

Tunicamycin

UPR

Unfolded Protein Response

UPRE

Unfolded protein response element