The serine/threonine-protein kinase/endoribonuclease IRE1α protects the heart against pressure overload–induced heart failure

Heart failure is associated with induction of endoplasmic reticulum (ER) stress and the unfolded protein response (UPR). The serine/threonine protein kinase/endoribonuclease IRE1α is a key protein in ER stress signal transduction. IRE1α activity can induce both protective UPR and apoptotic downstream signaling events, but the specific role for IRE1α activity in the heart is unknown. A major aim of this study was to characterize the specific contribution of IRE1α in cardiac physiology and pathogenesis. We used both cultured myocytes and a transgenic mouse line with inducible and cardiomyocyte-specific IRE1α overexpression as experimental models to achieve targeted IRE1α activation. IRE1α expression induced a potent but transient ER stress response in cardiomyocytes and did not cause significant effects in the intact heart under normal physiological conditions. Furthermore, the IRE1α-activated transgenic heart responding to pressure overload exhibited preserved function and reduced fibrotic area, associated with increased adaptive UPR signaling and with blunted inflammatory and pathological gene expression. Therefore, we conclude that IRE1α induces transient ER stress signaling and confers a protective effect against pressure overload–induced pathological remodeling in the heart. To our knowledge, this report provides first direct evidence of a specific and protective role for IRE1α in the heart and reveals an interaction between ER stress signaling and inflammatory regulation in the pathologically stressed heart.

IRE1␣ activation, was found to be protective against hypoxia and myocardial infarction also by inducing BiP (30) and coupling with hexosamine biosynthetic pathway (25). BiP expression can inhibit apoptotic signal CHOP and reduces apoptosis in cardiomyocytes (31). However, ER stress can also activate cell death signaling pathways (32)(33)(34). In particular, IRE1␣ binds to TRAF2 and activates downstream ASK1, resulting in JNK and p38 MAPK activation and apoptotic cell death (14,35,36). In contrast, other reports indicate that interactions between IRE1␣ and TRAF2 and JNK activate autophagy to promote cell survival during ER stress (37). PERK can also contribute to apoptotic signaling by promoting expression of ATF4 (17,38,39), a transcriptional activator for CHOP (40,41). Angiotensin II, along with tunicamycin and thapsigargin, can induce both adaptive protein folding chaperones and apoptotic signal CHOP (6). ER stress can also induce hypertrophy gene expression, including atrial natriuretic factor (ANF) and brain natriuretic factor (BNF), suggesting that ER stress contributes to both adaptive and pathological remodeling. Finally, ER stress is implicated in cancer therapy Imatinib-induced cardiomyopathy (26). Although evidence is clear about the importance of ER stress regulation in cardiac pathogenesis, the specific contribution of IRE1-mediated signaling in the heart has not yet been directly investigated.
In the current study, we investigated the direct impact of IRE1␣ activation in the heart. In cardiomyocytes in culture, IRE1␣ overexpression induced significant IRE1␣ autophosphorylation and Xbp1 activation, as expected. However, there was no evidence of induction in cell death, stress signaling, or hypertrophy upon prolonged IRE1␣ expression. Using an animal model with heart-specific, tamoxifen-inducible IRE1␣ overexpression, we investigated the direct impact of IRE1␣ expression in vivo. Overexpression of IRE1␣ did not lead to any pathological phenotype at baseline. Following pressure overload induced by trans-aortic constriction, IRE1␣ expression in transgenic hearts showed better preserved function, blunted pathologic marker gene induction, and better preserved ER stress signaling. More interestingly, the expression of inflammatory cytokines TNF␣ and IL-6 was significantly blunted by IRE1␣ expression, suggesting that IRE1␣-mediated ER stress signaling inhibits inflammatory cytokine induction in the heart. Thus, IRE1␣ activity has a direct protective effect in the heart against pressure overload-induced cardiac pathology, involving suppression of pro-inflammatory gene expression.

IRE1␣ induces adaptive and transient UPR in cardiac myocytes
To investigate the functional impact of IRE1␣ activity, we generated adenoviral vectors expressing WT IRE1␣. To validate the functionality of this expression vector, we tested the adenoviral vector in INS-1 cells. As shown in Fig. 1, IRE1␣ expression in INS-1 cells led to significant cell death 2 days after transfection (Fig. 1A). Xbp1 splicing was induced in a dose-dependent manner by IRE1␣ expression, similar to tunicamycin (TM) treatment (Fig. 1B). In addition, IRE1␣ autophosphorylation, BiP expression, phospho-EIF2␣ (Fig. 1C), and JNK kinase activity (Fig. 1D) were also induced in a dose-dependent fashion by IRE1␣ expression in INS-1 cells. These data are consistent with previous literature where IRE1␣ expression promotes cell death, associated with elevated RNase and stress kinase signaling via oligomerization in INS-1 cells (42,43). In contrast, neonatal rat ventricular myocytes (NRVMs) with IRE1␣ overexpression did not show significant difference in morphology or cell sizes ( Fig. 2A), although autophosphorylation and downstream Xbp1 activation were detected as in INS-1 cells (Fig. 2, B

IRE1␣ function in heart
and C). Other ER stress effectors, such as BiP expression, were also significantly induced by IRE1␣ overexpression, whereas the pro-apoptotic ER stress gene CHOP was not (p ϭ 0.09) (Fig.  2D). As expected, expression of a kinase-dead IRE1␣-KA mutant did not affect any of the ER stress molecules (Fig.  2E). These results revealed an unexpected myocyte-specific response to IRE1␣ expression, where only a subset of cytoprotective ER stress response was triggered in primary cardiomyocytes without activating the pro-apoptotic downstream pathway. Remarkably, prolonged IRE-1␣ expression (5 days) failed to sustain downstream Xbp1 activation or other UPR gene expression, although IRE1␣ autophosphorylation remained elevated (Fig. 2, F-H). These observations have revealed two specific features in IRE1␣ function in cardiomyocytes; one is a selective induction of IRE1␣-mediated cytoprotective response versus stress signaling, and another involves the transient nature of the IRE1␣-induced downstream Xbp1 activation, most likely due to a yet to be identified negative feedback inhibition.

Heart-specific and inducible expression of IRE1␣ in mice
To elucidate IRE1␣ function in intact heart, we generated an animal model where IRE1␣ expression was both inducible and heart-specific using a Cre-loxP-mediated gene switch strategy ( Fig. 3A) (44). Cre-dependent expression of IRE1␣ was demonstrated in 293 cells, validating the efficacy of the construct (Fig.  3B). As shown in vitro, the Flox-GFP-IRE1␣ transgenic mice showed widespread expression of GFP but not the IRE1␣ transgene. After cross-breeding with ␣MHC-Mer-Cre-Mer mice (45)(46)(47), the transgenic gene IRE1␣ was induced only in ventricular tissue only after tamoxifen treatment (Fig. 3, C-E). Four weeks after transgene induction, no abnormal phenotype was observed in the IRE1␣-expressing hearts compared with the littermate controls, including non-transgenic or singletransgenic mice, based on morphometric, histological, and functional analysis (Fig. 4). Furthermore, mRNA levels of sXbp1, BiP, and CHOP ( Fig. 5A) were not affected by IRE1␣ expression, consistent with our in vitro observation after longterm IRE1␣ expression. The molecular profiles of ANF,

IRE1␣ function in heart
␤-MHC, and TNF␣ were all unchanged following IRE1␣ overexpression (Fig. 5B). Together, these observations suggest that IRE1␣ expression was not sufficient to sustain ER stress response and did not exert any detrimental effect in the adult mouse heart.

IRE1␣ preserves heart function after transverse aortic constriction (TAC)
To test the effect of IRE1␣ expression on pathological stress response in heart, IRE1␣ transgenic mice were subjected to pressure overload implemented by TAC (48,49). Pressure overload led to significant loss of cardiac function over time in the control mice. In contrast, the IRE1␣ transgenic mice demonstrated preserved function even at 4 weeks post-TAC (Fig. 6A), associated with reduced chamber dilation at the systole ( Fig.  6B) based on echocardiograph measurements. Furthermore, the induction of pathological marker genes, including ANF and ␤MHC, was significantly blunted in response to TAC in the IRE1␣ transgenic heart versus the controls (Fig. 6C). However, cardiac hypertrophy as measured from tissue weight and histology was induced by TAC to the same levels in the control versus the IRE1␣ transgenic hearts (Fig. 6, D and E). Interestingly, fibrotic induction in myocardium was also significantly blunted in response to TAC in the IRE1␣ transgenic heart versus the controls (Fig. 6, E and F).
To investigate the potential mechanism, we analyzed the expression of UPR genes and stress-signaling molecules in post-TAC hearts. Although there was a trend for lower BiP expression in the post-TAC hearts, the IRE1␣ heart showed a modestly but significantly higher level of BiP expression than controls after TAC (Fig. 7A). In contrast, the mRNA level of CHOP was not affected by IRE1␣ expression, whereas there was trend for lower expression of ATF4 in the post-TAC IRE1␣ transgenic heart (38) (Fig. 7A). Therefore, IRE1␣ expression in the heart leads to better preserved heart function following the TAC with only modest impact on ER stress downstream molecules. ER stress signaling by IRE1␣ and TRAF2 has been reported to activate inflammatory signals (50). As shown in Fig.  7B, IRE1␣ transgenic hearts showed significantly reduced TAC-induced mRNA expression for both TNF␣ and IL-6 cytokines after comparing with the controls. Furthermore, whereas NF-B induction was not affected by IRE1␣ expression in the post-TAC hearts, the IkB kinase ␤ induction was significantly blunted in the IRE1␣ transgenic hearts, suggesting that attenuated NF-B signaling may in part contribute to reduced TNF␣/ IL-6 induction in the IRE1␣ TAC heart. In summary, IRE1␣ expression in the heart protects cardiac dysfunction associated with blunted induction of pro-inflammatory genes.

Discussion
This study provided first comprehensive characterization of the specific function of IRE1␣ as one of the three major ER stress-signaling branches in pressure overload-induced cardiac remodeling and dysfunction. In cultured cardiomyocytes,

IRE1␣ function in heart
we observed transient induction of canonical Xbp1 activation following IRE1␣ expression without significant effects on cell death and hypertrophy. In intact heart, IRE1␣ expression did not affect pressure overload-induced cardiac hypertrophy but significantly attenuated cardiac dysfunction and pathological marker gene expression. Interestingly, IRE1␣ expression markedly blunted the induction of pro-inflammatory cytokine gene expression following TAC. Therefore, our results provide direct in vitro and in vivo evidence that IRE1␣-mediated signaling is cardioprotective, at least in the setting of pressure overload-induced cardiac dysfunction.
ER stress-signaling pathways and the ensuing unfolded protein response are important cellular responses to various insults and have been found to be activated in human congestive heart failure (6), ischemic heart disease (21), and heart failure in response to the cancer drug imatinib (26). Experimental models have also recapitulated these observations (6,24), but it is unclear whether ER stress signaling plays protective (24,30,31) or pathological roles in heart (22,26,51,52). Expression of a constitutively active ATF6 shows a strong protection against ischemia reperfusion injury in mice (24). CHOP inhibition in NRVMs was also protective against apoptosis in the setting of ER stress by proteasome inhibition (31), and CHOP inactivation in vivo protected mouse hearts against apoptosis following ischemia/reperfusion injury (52). Among the three major UPR pathways (i.e. IRE1, PERK, and ATF6), IRE1 is unique in terms of diverse downstream outcome. IRE1 activation is capable of both protective and apoptotic signaling through Xbp1 or TRAF2 and stress-activated MAPK signaling cascades, respectively, as demonstrated previously (14,18,42,43,53,54) (Fig. 1). However, the molecular nature of IRE1␣ function and regulation appears to be highly cell type-specific. Whereas constitu-

IRE1␣ function in heart
tive expression of IRE1␣ led to cell death in nonmyocytes, such as INS-1 cells (42), as well as other cell types (54), similar expression in cardiomyocytes achieved only transient induction of Xbp1 but without cell death induction. More remarkably, prolonged expression led to sustained autophosphorylation of IRE1␣ in cardiomyocytes, but only transient induction of Xbp1, suggesting the presence of a negative inhibitory feedback mechanism for IRE1 signaling in cardiomyocytes. It is well established that the RNase activity of IRE1 is induced by oligomerization and trans-molecular autophosphorylation (55). However, its inhibition is poorly characterized but may involve Bax inhibitor 1 (56), Hsp90 -Cdc37 complex (57), or an ER-specific protein phosphatase (PP2Ce) (58). In cardiomyocytes, Xbp1-splicing activity is attenuated over time, whereas the IRE1␣ phosphorylation level remains elevated, suggesting that the underlying molecular mechanism may involve uncoupling of its autophosphorylation status and the Xbp1-splicing activity. The molecular basis of such negative feedback regulation should be further studied.
We observed that IRE1␣ transgenic mice had preserved heart function in response to pressure overload. The underlying mechanism is yet to be uncovered. It is known that inflammatory cytokines are induced in stressed myocardium, contrib-

IRE1␣ function in heart
uting to cardiac dysfunction and pathological remodeling. Indeed, both TNF␣ and IL-6 expressions were markedly reduced in the post-TAC IRE1␣ transgenic hearts. It is reported that both TNF␣ and IL-6 are induced by stress signaling, such as p38 MAPK (59,60) and NF-B pathway. Although we did not observe any effects on p38 activation, IKK␤ induction was significantly attenuated by IRE1␣ expression. The data are consistent with the notion that IRE1␣ expression in cardiomyocytes promotes cytoprotective ER stress signaling but not downstream stress signaling and offer a potential mechanistic link between ER stress signaling and anti-inflammatory gene induction in pathologically stressed myocardium. Establishing the cardioprotective role of IRE1␣ in the intact heart illustrates the potential contribution of ER stress signaling to the pathogenesis of heart failure as well as the feasibility of targeted manipulation of ER stress signaling for treating the disease.

Animal models and surgical procedures
The investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (Publication number 85-23, revised 1985). All procedures were performed in accordance with UCLA animal welfare guidelines and approved by the UCLA institutional animal care and use committee. IRE1␣ was cloned into a vector for generation of transgenic animals with Cre-regulated expression of the transgene of interest (61,62). Transgenic animals were generated in C57/Bl6 background through collaboration with the UCLA Molecular Genetics Technology Center. Founder animals were identified by PCR with transgene-specific primers.
Animals with heart-specific, inducible IRE1␣ overexpression were generated by crossing transgenic founder animals with previously established ␣MHC-Mer-Cre-Mer (MCM) transgenic mice (47) (45). IRE1␣ transgene overexpression was induced by intraperitoneal injection of tamoxifen citrate salt (Sigma), 20 mg/kg body weight/day for 5 days (45). WT and floxed single-transgenic littermate animals treated with tamoxifen or double-transgenic flox-GFP/CRE animals treated with vehicle were also used as controls. Both male and female mice age 12-16 weeks were included in this study. TAC was performed as described previously with modifications (49). Mice were anesthetized with ketamine (80 mg/kg)/ xylazine (20 mg/kg) by intraperitoneal injection. Respiration was provided by mechanical ventilation with 95% O 2 (tidal volume 0.5 ml, 130 breaths/min). Left parasternal thoracotomy was performed to access the transverse aorta, which was tied with a 5-0 nylon suture on a 27-gauge needle. The needle was removed, leaving in place a 65-70% constriction of the aortic lumen. Constriction of the aorta was confirmed by measuring differential blood flow through the right and left carotid arteries 1 week after surgery.
Animals were continuously anesthetized with 1.5% isoflurane and 95% oxygen. VisualSonics Vevo 770 and Vevo 2100 imaging systems and a 30-mHz scan head (Toronto, Canada) were used to collect short and long axis B-mode and M-mode

IRE1␣ function in heart
views. Reported values refer to short-axis measurements and calculations.

Histology
Hearts were perfused and fixed in 10% formalin before embedding in paraffin. All short-axis sections were prepared from mid-ventricle. Sections of 4-m heart were deparaffinized and rehydrated before staining by hematoxylin and eosin (H&E) or Masson trichrome and Verhoeff-Van Gieson stain. Stained tissue sections were recorded as digital images by the Aperio XT whole-slide scanning system, and snapshot images were taken using the ImageScope software (Aperio Technologies, Vista, CA).

Cell culture
293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Lipofectamine reagent (Life Technologies, Inc.) was used according to the manufacturer's protocol to achieve overexpression of the flox-GFP-IRE1␣ construct. INS-1 cells were cultured in RPMI1640 according to published methods (57). NRVMs were harvested from 1-3-day-old Sprague-Dawley rat pups as described previously (63) and cultured in serum-free Dulbecco's modified Eagle's medium supple-mented with 1% penicillin/streptomycin and 1ϫ Insulin-Transferrin-Selenium (ITS). NRVMs were infected with adenovirus (multiplicity of infection of 10) and incubated for 2 days before additional treatment with 5 g/ml TM for 4 h. Experiments with prolonged IRE1␣ expression were incubated for 5 days before RNA or protein analysis.

RNA and RT-PCR analysis
Total RNA was isolated from heart or cells with TRIzol (Life Technologies). For animal studies, cDNA was prepared using iScript Reverse Transcription Supermix and amplified with SsoFast EvaGreen Supermix on a CVX96 thermal cycler (all from Bio-Rad). For cell studies, cDNA was prepared using Superscript II (Invitrogen) and amplified with SYBR Green supermix on a MyIQ system (Bio-Rad). Melt curves were generated for each primer set during each experiment, and analysis was performed using the ⌬⌬CT method (Table 2).

IRE1␣ function in heart
AACTTGTCC-3Ј). Products were amplified with the cycling protocol of 95°C for 30 s, followed by 35 cycles of 95°C for 30 s, 60°C for 30 s, and 72°C for 25 s, followed by 72°C for 10 min. PCR products were separated on 4% agarose gel.

Statistical analysis
Data are presented as mean Ϯ S.D. Means of two groups were compared by two-tailed Student's t test. Means of more than two groups were compared by analysis of variance and Turkey post hoc test. Differences between groups were considered statistically significant when p Ͻ 0.05.