IRE1/XBP1 and endoplasmic reticulum signaling — from basic to translational research for cardiovascular disease

Introduction

Protein synthesis and folding are indispensable for physiological cellular function. During protein synthesis, ribosomes direct nascent polypeptides to the endoplasmic reticulum (ER). This transfer takes place upon translation of a short peptide (16–30 amino acids long) signal peptide or targeting peptide at the N-terminus (or occasionally C-terminus) of the newly synthesized protein. The signal peptide is present in proteins that reside in the lumen of the organelles that make up the secretory pathway (endomembrane system), as well as in proteins secreted into the extracellular space, and inserted into most cellular membranes [1]. The ER represents a highly specialized cellular compartment, which, due to its oxidizing environment, is especially well-suited for expression and oxidative folding of the ER-resident proteins, proteins of the endomembrane system, and secreted proteins [1]. The process of controlling polypeptide influx into the ER, posttranslational modification, protein folding, protein export, and degradation of misfolded proteins is controlled tightly by the highly coordinated complex regulation of protein homeostasis or proteostasis.

The endoplasmic reticulum unfolded protein response (UPR) plays a central role in proteostasis, and is activated during physiological (cell growth, mitosis) and pathological stress stimuli (heat, hypoxia, and ischemia), to combat increased misfolded protein burden or declining folding capacity. The mammalian UPR, as known to date, consists of three main signaling protein branches, mediated by three ER transmembrane receptors: inositol-requiring kinase 1 (IRE1), protein kinase R-like endoplasmic reticulum kinase (PERK), and activating transcription factor 6 (ATF6).

When subjected to ER stressors, cells react within as early as 15–30 min to halt their translational machinery via PERK activation and phosphorylation of its substrate, the eukaryotic translation initiation factor-2α [2,3]. IRE1 activation (see the following chapter) occurs in the same timeframe [4]. Around 4 h after exposure, downstream transcription factors, such as the activating transcription factor 4 (ATF4), which acts downstream of PERK, are actively translated and translocated into the nucleus. ATF6 is an ER-membrane resident protein that is transported to the Golgi apparatus and cleaved into its active form starting at ~1 h after the onset of ER stress, although earlier activation has been reported upon disruption of disulfide bonds of ER proteins caused by the reducing agent dithiothreitol [4]. The transcriptional targets of ATF6, mainly chaperones that aid in increasing protein-folding capacity in the ER, emerge at ~6 h after ER stress stimulus [5]. This initial burst of signaling from all three UPR branches is regarded as the protective, adaptive phase of UPR. If ER stress persists, the ATF6 and IRE1 pathways are turned off, despite continuous stimulation of UPR signaling, while the PERK pathway remains active [6], thereby switching to the maladaptive phase, where cell death programs are activated, preventing further damage to the organism (Figure 1a).

Disruption of proteostasis is evident in many human diseases [7–9]. Understanding the molecular mechanisms of the UPR may benefit the diagnosis and therapeutic treatment of disease. Since its discovery, studies utilizing various in vitro and in vivo models of disease stress have led to an increasing understanding of the complex signaling pathways of the UPR [10]. As a result, chemical compounds enhancing proteostasis have been developed and tested, and exciting preclinical results have led to promising clinical trials [11] and even to the introduction of some of the compounds to the clinical setting [12]. Challenges remain to resolve the problem of targeting such a ubiquitous and integrative cellular network precisely in specific tissues and organs [13]. Moreover, modifying the UPR is potentially a double-edged sword. While it is an adaptive biochemical process that regulates cell homeostasis and maintenance of normal physiological function, prolonged ER stress can push the UPR past beneficial functions, such as reduced protein production and increased folding and clearance, to apoptotic signaling [14]. Furthermore, recent studies demonstrate cell-type-specific aspects of the UPR, which are mainly driven by tissue-specific differences in the differential demand on ER capacity and function [15]. Therefore, detailed understanding of the UPR in specific cell types and disease models is critical for adopting treatment options in a clinical setting. The IRE1 pathway is considered the most evolutionarily conserved branch, orchestrating the UPR in yeast [16,17] and most other eukaryotes such as plants, fungi, and most metazoans [18]. This is reflected in the plethora of its downstream signaling pathways, which are switched on and off, depending on the stress phase, reflecting opposing effects from cell survival to mediating apoptosis (Figure 1a and c). This review will focus on the details of the complexities of the IRE1 branch of the UPR, as well as on our current knowledge of the roles of the IRE1 branch in cardiovascular disease.

Inositol-requiring kinase 1 stress sensing and activation

While yeast IRE1 only has one isoform, the mammalian IRE1 consists of two isoforms: IRE1α and IRE1β. IRE1α, encoded by the ERN1 gene, is expressed ubiquitously, while IRE1β, encoded by the ERN2 gene, is expressed in lung and intestinal epithelial cells [19] (for simplicity, we will use the term ‘IRE1’ in the rest of this review to refer to the yeast IRE1 and the mammalian IRE1α). Three models of stress sensing have been proposed to lead to initiation of the IRE1 pathway. In the original (competitive) model, it was proposed that in resting state, IRE1 is associated with the ER-resident chaperone, binding-immunoglobulin protein also named 78 kDa glucose-regulated protein (BiP, GRP78). When high amounts of misfolded proteins are present in the ER, BiP preferentially binds to these proteins, thereby dissociating from IRE1 [20]. This release allows IRE1 to multimerize and autophosphorylate its cytosolic domain, triggering the activation of the IRE1 RNase activity, initiating XBP1 mRNA splicing and UPR responses.

In addition to this competitive model, BiP has been suggested to bind to IRE1 via its ATPase domain and undergo allosteric changes, thereby dissociating from IRE1 when unfolded proteins are bound to its substrate-binding domain during UPR [21]. Mutagenesis studies of the BiP-binding site on yeast IRE1 have shown that IRE1 retains its ER stress-sensing ability, despite deletion of the BiP-binding site [22]. Evidence of direct binding to unfolded proteins suggests that this is sufficient to activate yeast IRE1 in response to unfolded protein accumulation [23]. Similar results have been shown in mammalian cells [24–26], where IRE1 activation seems to be more complicated. Recent studies not only indicate that preformed oligomers of ER stress sensors, such as IRE1, exist in a resting state [24], but also hetero-oligomers consisting of IRE1 and other regulatory proteins modify IRE1 activity during resting conditions and stress exposure [27]. While extensive biochemical characterization of the IRE1 interactome is required to determine the dynamics of its protein composition under conditions of mild and chronic ER stress, studies to date suggest that IRE1 is involved in cellular signaling, protein synthesis, and targeting [28] (Table 1).

Signal transduction by inositol-requiring kinase 1

Upon activation, IRE1 dimerizes, trans-autophosphosphorylates, and oligomerizes (Figure 1b) [29], forming the kinase-extension nuclease domain [30], which possesses site-specific RNase activity to cleave HAC1 (homologous to ATF/CREB 1; yeast)/XBP1 (mammalian cells) [31]. IRE1 transduces signals in part through the excision of a 26-nucleotide intron of XBP1, resulting in spliced XBP1 (XBP1s), encoding a highly active transcription factor to enforce UPR [32]. Translated unspliced XBP1 (XBP1u) targets its own mRNA–ribosome-nascent chain complex to the vicinity of IRE1 at the membrane of ER, facilitating its splicing [33]. IRE1 RNAse also targets a subset of ER-membrane-localized mRNAs as well as some noncoding RNAs, in a mechanism termed regulated IRE1-dependent mRNA decay (RIDD) [34]. RIDD is sequence-specific and degrades hundreds of RNA targets upon ER stress, alleviating the load of proteins entering the ER and taking part in cell-fate decisions upon unresolvable ER stress [35,36]. Target selection and mechanisms of recruitment of specific RNA targets to IRE1 remain nebulous. Recently, by integrating nascent and global RNA transcriptomes in an IRE1-knockout model, IRE1α-dependent decay of a set of mRNAs lacking the RIDD sequence was identified: this was termed “RIDD lacking endo-motif” (RIDDLE). While XBP1 and RIDD targets are targeted by IRE1 dimers, it appears that only higher-order polyphospho-oligomers can perform RIDDLE [37]. Increasing knowledge of IRE1 rheostatic RNAse target selection may uncover a unique master switch governing UPR and cell fate [38].

Besides its endonuclease effector activity, IRE1 associates with TNF receptor-associated factor 2 (TRAF2) and apoptosis signal-regulating kinase 1 (ASK1) in a hetero-oligomer manner, inducing ASK1 to activate the c-Jun N-terminal kinase pathway (JNK–c-Jun) with its pleiotropic effector properties [39,40]. Extracellular signal-regulated kinase activation has also been found to be IRE1-dependent during azetidine-2-carboxylic acid (nonprotein amino acid proline homolog used to induce ER stress) and tunicamycin- (inhibitor of protein glycosylation) induced ER stress (Figure 1c) [41].

The proximal part of the IRE1–XBP1 pathway is induced early and attenuates its canonical signaling during a relatively early phase of ER stress compared with ATF6 and PERK (Figure 1a) [6]. XBP1u mediates deactivation of the IRE1–XBP1 signaling by binding to spliced XBP1 mRNA. The XBP1u protein–XBP1s mRNA complex is then degraded by the proteasome through a degradation domain on XBP1u [42]. XBP1u also functions as a negative regulator of XBP1s and activated ATF6, by targeting these proteins for destruction by the proteasome via direct association [43]. Moreover, XBP1u directly binds and promotes the degradation of forkhead box protein O1 (FOXO1), thereby suppressing autophagy [44]. Aside from targeting binding partners for degradation, XBP1u has been shown to reduce shear stress-induced oxidative stress in endothelial cells by complexing with histone deacetylase 3 and AKT1, promoting AKT1 phosphorylation and activating the phosphoinositide-3-kinase/AKT pathway [45].

Nevertheless, non-XBP1s-mediated signaling has been the focus of recent studies, revealing previously unknown functions of the IRE1–XBP1 pathway. By employing photoactivatable ribonucleoside-specific UV cross-linking, close physical association between IRE1, with a novel subset of structural noncoding RNAs and proteins associated with these RNAs, was identified, revealing a link between IRE1, the ribosome machinery, and the signal-recognition particle ribonucleoprotein complex [46]. This new interaction may shed light on a novel surveillance function of IRE1 on global translation. Another surprising association has been found in mitochondria-associated ER membranes (MAMs), a specialized ER subdomain within close physical proximity to the mitochondria (Figure 1b), where IRE1 enhances mitochondrial calcium uptake, increases mitochondrial ATP levels, and energetics, especially in response to ER stress, independent of the RNAse activity of IRE1. Observations in IRE1-knockout mouse hepatocytes showed altered mitochondrial morphology and significant reduction of InsP3R at MAMs. Interestingly, these mice developed decreased global glucose tolerance, indicating a cell nonautonomous effect, in which phenotypes were observed outside of the IRE1-knockout cells, posing possible therapeutic consequences [47]. Possible involvement of nonautonomous effects of IRE1 and XBP1 has also been demonstrated in C. elegans, where enforced neuronal XBP1s expression was sufficient to induce intestinal UPR through neurohumoral means and prolonged lifespan by increasing resistance to ER stress in adulthood, where UPR is constitutively downregulated [48]. These novel discoveries take IRE1 beyond a simple binary signaling switch to an interesting target in senescence and its associated diseases.

The unfolded protein response in the heart

Adult human cardiomyocytes are highly differentiated specialized cells without the ability of meaningful cell division or regeneration [49], highlighting the importance of maintaining proteostasis and cell integrity in the myocardium. It is no surprise that the UPR has been shown to be involved in the initiation and progression of cardiovascular disease [50], which is the most prominent disease burden in the elderly worldwide [51]. During disease progression in the heart, activation of all three branches of the UPR is evident. However, distinct differences in signaling output, depending on the pathophysiological stimulus, separate the activation of the different pathways. ATF6 is regarded as generally protective through its ability to induce the adaptive phase of the UPR by boosting the folding capacity of the ER when misfolded protein accumulation occurs in the diseased heart. This is evident in cardiac ischemia/reperfusion (I/R) models where ATF6 is activated during ischemia and ATF6 was shown to protect from I/R damage by upregulating ER chaperones and antioxidant and alleviating toxic cytosolic calcium levels [52,53]. In physiological and pathological growth of the heart, activation of ATF6 is required to drive adaptive myocyte hypertrophy [54]. Strikingly, although different stresses, such as I/R, pressure overload, or exercise, can activate ATF6, the consequent transcriptional reprogramming seems to be a stress-specific response [54]. PERK activation is an early event observed in the setting of I/R and promotes survival, but prolonged PERK activation is detrimental to myocyte survival by activating apoptosis through the ATF4/C/EBP homologous protein (CHOP) axis [55]. Similarly, PERK is cardioprotective during the early phases of pressure overload. In contrast, chronic heart failure results in downstream CHOP activation and apoptosis [55]. In contrast to ATF6 and PERK, the known impact of IRE1/XBP1 pathway on cardiovascular diseases is highly tissue-specific. In addition, multiple novel noncanonical functions of XBP1u/XBP1s have been uncovered to be important in cardiovascular diseases, further enlarging the network of possible signaling by IRE1.

The inositol-requiring kinase 1–X-box- binding protein-1 axis in cardiac remodeling and heart failure

Cardiovascular disease increases the risk of developing heart failure with reduced ejection fraction (HFrEF). Heart failure defines a clinical state where cardiac output is insufficient to meet oxygen demand [56]. The progression to heart failure is associated with myocardial remodeling, an adaption of the myocardium, extracellular matrix, and myocardial vasculature to disease stress [57]. Myocardial fibrosis, an important part of cardiac remodeling, is associated with adverse outcome of heart-failure patients and as such is key for diagnosis and therapeutic potential in heart failure [58]. Fibrosis is mediated by dynamic cardiac fibroblast cell states, which underlie the heart’s fibrotic response to maintain structural integrity in the short term, and result in cardiac fibrosis and maladaptive remodeling in the long term [57]. Activation of the IRE1–XBP1 pathway guides ER biogenesis and enhances ER capacity by expanding ER size through transcriptional activity of XBP1 [59]. Moreover, XBP1 is required for the development of specialized secretory cell types such as antibody-secreting plasma cells and pancreatic acinar cells [59]. Given that protein secretion is a feature of cardiac fibroblasts, it is plausible that the IRE1 branch of the UPR serves important roles in the regulation of the dynamic cell states of cardiac fibroblasts and, therefore, fibrosis. In a recent study, inducible, cardiomyocyte-specific overexpression of IRE1 resulted in potent but transient IRE1 activation [60]. Nevertheless, in response to pressure overload, the transgenic hearts exhibited preserved function and reduced fibrotic area, increased adaptive UPR signaling, and blunted inflammatory and pathological gene expression. However, XBP1s has been shown to act downstream of calreticulin expression in mice to induce cardiac fibrosis by TGFβ1/Smad2/3 signaling pathway activation, which was reversed by tauroursodeoxycholic acid [61]. Moreover, activation of Sigma-1 receptor by fluvoxamine inhibited cardiac fibrosis, apparently through downregulation of the IRE1–XBP1 pathway [62].

Upregulation of the UPR is observed in HFrEF patients [63,64]. Unloading the failing heart by implantation of a left ventricular assist device improves cardiac function and ameliorates the UPR [65]. In a rat model of pressure overload-induced hypertrophy via abdominal aortic constriction, XBP1s expression was dynamically regulated, with increasing expression in the compensatory hypertrophy phase up to 4 weeks after induction of pressure overload, and a subsequent decrease during the heart failure phase. This regulation was mediated by microRNA-214 and the microRNA-30* family, which reduce the expression of XBP1 and its targets during the late heart-failure stage [66,67]. Analogous to in vitro models where the dynamics of IRE1 signaling is extensively studied, this may reflect a protective function of IRE1 signaling during a relatively early phase of UPR in heart disease, while in a later maladaptive phase, IRE1 signaling is switched off. Spliced XBP1 is associated with the expression of brain natriuretic peptide (BNP) in neonatal rat cardiomyocytes by binding to an AP1/CRE-like element in the BNP promoter region and increasing its activity [68]. Conversely, recent evidence in heart failure with preserved ejection fraction (HFpEF) patients and rodent models of HFpEF unveiled inhibition of the IRE1–XBP1 pathway [69,70]. In the most widely accepted rodent model of HFpEF, feeding mice a high-fat diet and N^ω-nitro-L-arginine methyl ester (L-NAME) resulted in a model of metabolic syndrome and hypertension and led to increased left ventricular filling pressure and decreased longitudinal myocardial strain. In that study, increased filling pressure led to pulmonary congestion and hypertrophy of cardiomyocytes, which resulted in exercise intolerance. Moreover, the HFpEF mice displayed reduction in cardiac IRE1 phosphorylation and decreased levels of XBP1s, BiP, and CHOP, in contrast to a HFrEF model undergoing transverse aortic constriction. Most interestingly, while IRE1–XBP1 signaling was downregulated, the PERK and ATF6 pathways were not, suggesting a possible UPR-independent regulation of the IRE1 pathway. Overexpression of XBP1s was sufficient to partially rescue the phenotype. Mechanistically, the authors found that application of l-NAME induced systemic inflammation, upregulating the inducible nitric oxide synthase in the heart, and inducing S-nitrosylation of IRE1, preventing XBP1 splicing. XBP1 splicing was restored in iNOS-knockout mice that were subjected to the HFpEF feeding protocol, while the HFpEF phenotype was partially rescued [69]. These findings suggest multi-organ regulation of the IRE1–XBP1 pathway as a pathological driver of HFpEF.

The inositol-requiring kinase 1–X-box- binding protein-1 axis in atherosclerosis and myocardial ischemia/reperfusion

Atherosclerosis develops through endothelial injury and subendothelial infiltration of monocytes and lipoproteins through a leaky, dysfunctional endothelium [71]. These lipoproteins are oxidized by nitric oxide (NO), produced by macrophages, and subsequently internalized to give rise to foam cells, which undergo apoptosis, forming the fatty streaks. Following advanced immune response, smooth muscle cells migrate to the site of inflammation, proliferate, and synthesize the extracellular matrix, which stabilizes the plaque while narrowing the vessel lumen, forming the fibrous plaque [71,72]. Acute coronary syndrome and stroke are caused by disruption of a vulnerable atherosclerotic fibrous plaque. A vulnerable plaque is the result of chronic inflammation, which leads to decreased collagen production and increased expression of collagenases, decreasing the width of the plaque cap and rendering it unstable, therefore prone to rupture and thromboembolic events [71,73].

Evidence of upregulated UPR in human atherosclerosis has been gathered in autopsies involving patients who succumbed to acute coronary syndrome and noncardiac causes, as well as biopsies from patients with stable or unstable angina pectoris. Immunostaining revealed increased expression of ER chaperones GRP78 and GRP94 as well as CHOP in smooth muscle cells and macrophages in patients with more severe disease (unstable angina, unstable thin plaques, and ruptured plaques in the deceased patients), forming a possible connection to reactive UPR signaling upon disease progression [74]. In a ApoE-knockout atherosclerosis mouse model, XBP1 splicing was found highly upregulated in the aortic endothelium in areas with disturbed blood flow, but not in regions with laminar flow. Forced expression of XBP1s induced atherosclerosis in isografted straight vessel with laminar flow, linking XBP1 splicing to the pathogenesis of atherosclerosis. Mechanistically, XBP1 splicing was found to be required for sheer stress-induced endothelial proliferation, and induced endothelial apoptosis through suppression of vascular endothelial (VE)-cadherin expression, which maintains endothelial integrity [75].

XBP1 splicing is also crucial for vascular smooth muscle cell (vSMC) proliferation, and is partially needed for vSMC migration, resulting in reduced neointima formation in a XBP1-knockout mouse model. XBP1 acts through suppression of interestingly in part by a transforming growth factor (TGF)-β-mediated paracrine mechanism, and in part through microRNA-18a and microRNA-1274B, which target calponin h1 transcript for degradation. Moreover, XBP1s forms a complex with PI3K and Akt and activates the Akt pathway and thereby, vSMC migration [76]. While XBP1s is involved in vSMC calcification downstream of bone morphogenetic protein 2-induced Runx2 expression in human coronary artery smooth muscle cells [77], a novel function of the unspliced XBP1 (XBP1u) has been uncovered in vSMC calcification. XBP1u levels were found to be reduced through proteasomal degradation in calcified vSMCs, calcified aorta of adenine diet-fed mice, and human radial arteries with chronic kidney failure. Inhibition of XBP1u promoted chondrogenic markers and overexpression of XBP1u reduced calcification in rat vSMC and XBP1 SMC-specific knockout mice fed with adenine diet. Through interactome analysis, β-catenin, a member of Wnt signaling governing calcification, was found to specifically bind to XBP1u, following which, XBP1u sequesters β-catenin for proteasome-mediated degradation [78]. In macrophages, several pro-atherogenic chemokines involved in lipid-driven inflammation have been identified to be IRE1-regulated, including interleukin-1 beta (IL-1β) and C-C motif chemokine 2 (CCL2). IRE1 or XBP1 inhibition abolished lipid-driven IL-1β through suppression of lipid-induced release of mitochondrial reactive oxygen species and NACHT, LRR and PYD domains-containing protein 3 (NLRP3) inflammasome. Treating ApoE-knockout mice with the specific IRE1 inhibitors STF-083010 or 4μ8c decreased atherosclerosis in vivo and modified plaque composition with markedly reduced macrophages, posing a possible mechanism in reducing plaque progression [79].

In the I/R model in cultured cardiomyocytes and mouse hearts, XBP1 splicing was increased during the ischemic phase [80]. While data point to a pro-atherogenic role for XBP1s, inhibition of XBP1s in the I/R setting is detrimental to cardiomyocytes [80,81]. Mechanistically, XBP1s was found to mediate transcription of genes encoding enzymes of the hexosamine biosynthetic pathway, including the rate-limiting glutamine fructose-6-phosphate aminotransferase 1 (GFAT1), all of which facilitate O-linked β-N-acetylglucosamine protein modification, which is necessary for correct protein folding. Expression of XBP1s in transgenic mice for 2 weeks followed by I/R decreased infarct size and preserved cardiac function. Suppression of GFAT1 in vitro and in vivo partially restored the protective effect [81]. XBP1s, inducible by vascular endothelial growth factor, promotes angiogenesis through PI3K/AKT/GSK3β pathway [82]. Altogether, these findings suggest cytoprotective and pro-angiogenic functions for the IRE1–XBP1 axis in the setting of ischemia and I/R in most of the cell types studied to date. Depending on cellular identity and function, the activation of this axis may preserve or exacerbate cardiovascular function.

Endoplasmic reticulum stress in diabetic cardiomyopathy

The diabetic heart is characterized by heart failure independent of coronary artery disease, hypertension, or valvular disease in diabetic patients. Early stages of diabetic cardiomyopathy consist of subcellular changes of signaling, and structural abnormalities of cellular compartments, which remain asymptomatic to the patients. Early clinical presentations are concentric hypertrophy, diastolic dysfunction, and increased myocardial fibrosis. Ultimately, systolic impairment will develop [83]. Guideline-driven diagnostic procedures or targeted treatment options are currently not available, although the emergence and success of sodium-glucose cotransporter-2 inhibitors (STLG2i) in treating HFrEF and HFpEF sheds light on the interplay of diabetes and heart failure [84]. On the molecular level, factors contributing to the still-vaguely understood pathogenesis include chronic inflammation, generation of excessive reactive oxygen species, accumulation of advanced glycation end products, and activation of signaling such as TGF-β, leading to increased myofibroblast differentiation, disturbance of extracellular matrix homeostasis, and increased cross-linking of extracellular matrix proteins [85].

ER stress is a key feature of metabolic disorders. UPR and IRE1 have been extensively studied in glucose metabolism and metabolic syndrome [86]. XBP1 deficiency is linked to β-cell dysfunction, while the IRE1–JNK pathway mediates intracellular insulin signaling in a cell-specific manner. In obesity, nuclear translocation of XBP1s is markedly decreased, and expression of XBP1 in adipocytes improves glucose tolerance and insulin sensitivity [87]. More recently, IRE1/XBP1s activation was found to promote systemic adaptive remodeling in obesity [88]. Specifically, XBP1s-selective pharmacological IRE1 activator, IXA4, transiently activated protective IRE1/XBP1s signaling in liver without inducing RIDD or TRAF2/JNK signaling. IXA4 treatment improved systemic glucose metabolism and liver insulin action through IRE1-dependent remodeling of the hepatic transcriptome that reduced glucose production and steatosis. IXA4-stimulated IRE1 activation also enhanced pancreatic function. These findings indicate that transient activation of IRE1/XBP1s signaling promotes integrative multitissue benefits that mitigate obesity-driven metabolic dysfunction.

Less data exist on possible IRE1/XBP1-related mechanisms in the diabetic heart, although IRE1 shares common pathways with key signaling cascades contributing to diabetic cardiac remodeling. The NLRP3 inflammasome is also triggered in a type-2 diabetes rat model fed with high-fat diet and injected with streptozotocin, which is specifically cytotoxic to pancreatic β-cells, leading to reduced systolic and diastolic heart function and increased fibrosis. Gene silencing of NLRP3 reversed cardiac function and fibrotic remodeling [89]. Bone morphogenetic protein-7, STLG2i, rosuvastatin, which has an anti-inflammatory effect on some cardiovascular diseases, and the blood thinner ticagrelor have been demonstrated to alleviate cardiac dysfunction in different diabetic rodent models [90–93]. IRE1 has been shown to regulate NLRP3 inflammasome activity via upregulation of thioredoxin-interacting protein during ER stress, and suppression of IRE1 via small-molecule RNase inhibitor counteracts disease progression in a diabetic mouse model [94]. However, direct evidence of possible disease-modulating effect of cardiac-specific IRE1 and possible pro-inflammatory effects in diabetic cardiomyopathy are still missing.

Summary and outlook

A plethora of studies have established IRE1 as a key regulator of cellular stress over the 30 years since its discovery by Kazutoshi Mori and Peter Walter, who independently, and according to Peter Walter, serendipitously, identified IRE1 in yeast [95,96].

Over the years, basic research has contributed to understanding stress sensing and activation of IRE1 through three models, competitive or allosteric dissociation of BiP, and direct substrate binding to IRE1 itself. It is feasible that these mechanisms are not exclusive, but mutually exist alongside each other. Understanding the contributions of these activating mechanisms in animal models and in relation to disease stress remains to be investigated. Furthermore, by employing emerging global profiling approaches in a forward-genetics setting, the extensive networks of effector signaling of the RNAse function of IRE1, through XBP1 splicing, RIDD, and RIDDLE, have revealed the mechanisms by which IRE1 regulates the UPR dependent on the stress stimulus. Moreover, IRE1 forms hetero-oligomers with other proteins, activating the JNK and ERK pathways, and directly interacts with ribosome complexes with potential implications in translational regulation and RNA decay, as well as with MAMs to dictate cellular energetics.

Clinically, regulation of the IRE1–XBP1 axis has been demonstrated in a model- and disease-dependent manner (Figure 2). While end-stage heart failure resulting from ischemic heart disease and dilated cardiomyopathy is associated with a sustained upregulation of UPR and the IRE1 pathway in patients, animal models reveal dynamic regulation of IRE1 signaling during progression to heart failure. HFpEF patients and a unique HFpEF mouse model displayed a suppression of UPR, and more specifically XBP1 splicing. Sustained XBP1 splicing is crucial to chronic progression of atherosclerosis by inducing endothelial dysfunction, vascular smooth muscle cell proliferation and migration, and macrophage-mediated inflammation. Conversely, sufficient and brief upregulation of XBP1 splicing in cardiomyocytes is essential for induction of UPR in the setting of ischemia/reperfusion and angiogenesis. In diabetic cardiomyopathy, pro-inflammatory pathways, such as the NLRP3 inflammasome, play a central role in disease progression. IRE1 has been shown to modulate the NLRP3 inflammasome in diabetic mice, although evidence of possible disease modulation by cardiac IRE1 is missing.

Genetic approaches to modify the IRE1/XBP1 pathway in humans remain challenging due to the complexity of the IRE1/XBP1 signaling cascade and the transient nature of stress signaling, requiring timely control of activation and deactivation. Recent advances in deploying small molecules specifically targeting independent functions of IRE1 have shown positive effects in animal disease models. STF-083010 and 4μ8c are RNase inhibitors often used in tumor research, where inhibition of an overactive IRE1 branch blocks tumor progression. While IRE1 kinase inhibitors (type I) such as APY29, which stabilize an active kinase-domain conformation, activate the IRE1 RNase domain, other IRE1 kinase inhibitors (type II) such as KIRA3/6/7, which stabilize an inactive ATP-binding site conformation in IRE1, keep the RNase domain in an inactive state [97]. Since kinase-driven oligomerization of IRE1 hyperactivates its RNase to trigger apoptosis, then kinase inhibitors that block oligomerization (type II) should prevent apoptosis under ER stress. More intriguingly, recent work by the Wiseman lab identified a group of IRE1-specific activators, among which IXA4 and IXA6 showed the most potent induction in XBP1 splicing while maintaining little-to-no cytotoxicity [98]. IXA4 was successfully deployed in a diet-induced obese mouse model, where a brief boosting of IRE1/XBP1 signaling led to multitissue benefits without triggering the maladaptive UPR phase, providing a potential clinically relevant tool for the cardiovascular field [99].

The IRE1 pathway does not exist independently from the ATF6 and PERK pathways, as extensive crosstalk occurs between all three (Figure 1a). In the future, complex, tissue-specific high-throughput analysis approaches, up to single-cell resolution, will enable thorough characterization of the cardiac UPRosome, including the roles of the UPR in cellular crosstalk (Figure 2).