, Volume 18, Issue 5, pp 537–546

New directions in ER stress-induced cell death


  • Susan E. Logue
    • Apoptosis Research Centre, NUI Galway
  • Patricia Cleary
    • Apoptosis Research Centre, NUI Galway
  • Svetlana Saveljeva
    • Apoptosis Research Centre, NUI Galway
    • Apoptosis Research Centre, NUI Galway
Original Paper

DOI: 10.1007/s10495-013-0818-6

Cite this article as:
Logue, S.E., Cleary, P., Saveljeva, S. et al. Apoptosis (2013) 18: 537. doi:10.1007/s10495-013-0818-6


Endoplasmic reticulum (ER) stress has been implicated in the pathophysiology of many diseases including heart disease, cancer and neurodegenerative diseases such as Alzheimer’s and Huntington’s. Prolonged or excessive ER stress results in the initiation of signaling pathways resulting in cell death. Over the past decade much research investigating the onset and progression of ER stress-induced cell death has been carried out. Owing to this we now have a better understanding of the signaling pathways leading to ER stress-mediated cell death and have begun to appreciate the importance of ER localized stress sensors, IRE1α, ATF6 and PERK in this process. In this article we provide an overview of the current thinking and concepts concerning the various stages of ER stress-induced cell death, focusing on the role of ER localized proteins in sensing and triggering ER stress-induced death signals with particular emphasis on the contribution of calcium signaling and Bcl-2 family members to the execution phase of this process. We also highlight new and emerging directions in ER stress-induced cell death research particularly the role of microRNAs, ER-mitochondria cross talk and the prospect of mitochondria-independent death signals in ER stress-induced cell death.


Endoplasmic reticulumStressUnfolded protein responseCell death


ER stress is triggered due to a loss of homeostasis in the ER causing accumulation of misfolded proteins within the ER lumen. Examples of such physiological stresses include hypoxia, glucose deprivation and oxidative stress, conditions which can also often be found associated with tumor microenvironments. Three ER transmembrane receptors IRE1α (inositol requiring enzyme/endonuclease 1), PERK (double stranded RNA-activated protein Kinase (PKR)-like ER kinase) and ATF6 (activating transcription factor 6) constantly monitor the “health” of the ER. Under normal conditions each receptor is maintained in an inactive state through binding, via their luminal domain, with the ER chaperone protein Grp78 (Bip, HspA5). Accumulation of unfolded proteins triggers dissociation of Grp78 (owing to a higher affinity for unfolded proteins) from IRE1α, PERK and ATF6 facilitating their activation. Upon Grp78 release, IRE1α dimerizes and autophosphorylates activating its kinase and endonuclease functions [1]. Likewise, PERK dimerizes and autophosphorylates, activating its kinase domain [1], while ATF6 translocates to the Golgi where site 1 protease (S1P) and site 2 protease (S2P) process it to generate an active transcription factor which subsequently translocates to the nucleus [2]. The collective signaling pathways initiated by these ER stress receptors are commonly referred to as the unfolded protein response (UPR). The UPR is a highly conserved stress pathway which functions as a short term adaptive mechanism aimed at reducing levels of unfolded proteins and restoring balance to the ER. However, if the UPR is insufficient to deal with chronic exposure to ER stress-inducing stimuli then a switch to ER stress-induced death signaling commences.

ER stress-mediated death initiation


IRE1α is an ER transmembrane protein containing a kinase and endoribonuclease (RNase) domain on its cytosolic portion [3]. Oligomerization of IRE1α by Grp78 dissociation juxtaposes the kinase domains causing transautophosphorylation which in addition to activating its kinase activity also triggers the endoribonuclease activity of IRE1α. By virtue of its RNase activity, IRE1α splices a 26 nucleotide intron from XBP1 mRNA causing a frame shift enabling translation and generation of a basic leucine zipper family transcription factor, spliced XBP1 (XBP1s) [4]. XBP1s has a diverse range of target genes which share the common aim of short term adaption and ultimately restoration of ER function. The majority of XBP1 target genes are involved in either increasing the folding capacity of the ER or associated with the degradation of accumulated proteins with the aim of reducing ER protein load. Temporal analysis of IRE1α activation in response to ER stress found it to be an early event which diminished upon prolonged stress [5]. Moreover, expression of a mutant form of IRE1α, in which RNase activity can be selectively activated, lead to an enhancement in cell survival upon treatment with ER stress inducers indicating pro-survival functions [5]. Recent reports also suggest XBP1s signaling may be able to modulate apoptotic signaling. Upon IL-3 deprivation, BaF3 cells stably expressing XBP1s exhibited increased survival which was in part attributed to modulation of Bcl-2 family members including Bim [6]. Furthermore, overexpression of XBP1s in MCF-7 cells increased Bcl-2 levels following stimulation with Tamoxifen or ethanol [7]. Currently, it remains unknown whether XBP1s can modulate Bcl-2 family member expression during ER stress. It is possible XBP1s targets stretch beyond proteins directly involved in ER function and it may be actively involved in the suppression of apoptosis through the modulation of Bcl-2 family members however future studies will be needed to verify this.

Owing to the pro-survival targets of XBP1s, IRE1α signaling is generally regarded as an adaptive response. Indeed work by Lin and colleagues, investigating the temporal activation of UPR stress sensors, found IRE1α signaling to be attenuated in cells undergoing prolonged ER stress supporting the hypothesis that this pathway does not actively participate in pro-apoptotic signals [5]. However, overexpression of IRE1α in HEK293T cells has been reported to induce death indicating there must be pro-apoptotic signaling components [3]. Indeed the recruitment of TNF receptor associated factor 2 (TRAF2) to IRE1α has been linked to several pro-apoptotic pathways the most well defined being the IRE1α-TRAF2-JNK axis [8]. The association of IRE1α with TRAF2 triggers phosphorylation cascades involving ASK1 and culminating in JNK activation. JNK-mediated phosphorylation has been demonstrated to modulate Bcl-2 family member function. For example, phosphorylation of Bcl-2/Bcl-xL by JNK can reduce their anti-apoptotic ability while phosphorylation of Bid and Bim by JNK has been demonstrated to increase their pro-apoptotic ability [912]. Therefore, IRE1α-mediated JNK activation may represent a mechanism through which IRE1α can manipulate relative levels of pro- and anti-apoptotic Bcl-2 family members thus tipping the balance in favor of apoptosis (Fig. 1). IRE1α signaling has also been suggested to modulate cellular release of TNFα which can feedback in an autocrine manner and activate death receptor signaling. Again this pathway is mediated by the adapter protein TRAF2 which recruits IKK to the IRE1α complex where it is phosphorylated and activated. Activated IKK phosphorylates IκB tagging it for degradation thereby permitting NF-κB translocation to the nucleus and upregulation of target genes such as TNFα [13] (Fig. 1). TNF receptor 1 (TNFR1) has also been implicated in IRE1α-mediated JNK signaling with JNK activation found to be deficient in TNFR1−/− MEFs exposed to ER stress [14]. It is proposed that upon ER stress TNFR1 co-localizes with RIP1 and IRE1α at the membrane of the ER and that this complex is necessary for the optimal JNK signaling upon ER stress and execution of apoptosis [14].
Fig. 1

Unfolded protein response: IRE1, PERK and ATF6 activation. Cells cope with stressful conditions by activating the unfolded protein response. This response is mediated via the dissociation of Grp78 from three ER transmembrane proteins IRE1α, PERK and ATF6. a Following dissociation of Grp78, IRE1α oligomerizes and autophosphorylates facilitating its activation. Active IRE1α induces splicing of XBP1 mRNA to XBP1s and also activates JNK via TRAF2 and ASK1. Furthermore, active IRE1α has been linked to downstream NF-κB activation and also RIDD, which can lead to the degradation of pro-survival mRNA. b Like IRE1α, PERK dimerizes and autophosphorylates following Grp78 dissociation. Active PERK mediates its response via phosphorylation of eIF2α leading to a translational block and cap independent translation of ATF4. ATF4 induces CHOP which has multiple downstream targets. c Following Grp78 dissociation, ATF6 is transported to the Golgi where it is cleaved from its membrane anchor. Little is known about ATF6 regulated pathways but it is involved in the upregulation of UPR associated genes, XBP1, CHOP, Grp78, PDI and EDEM1


The RNase activity of IRE1α has recently been linked to a process referred to as regulated IRE1-dependent decay of mRNAs (RIDD). RIDD was first described in D. melanogaster where IRE1α activity was shown to mediate the rapid decay of ER localized mRNAs [15]. Subsequent studies have also verified the existence of RIDD in mammalian cells [16]. While this process is reliant upon IRE1α RNase activity it is distinct from XBP1 splicing and is reported to selectively target and degrade mRNAs encoding secretory proteins involved in protein folding within the ER. Initial activation of RIDD would be expected to aid cell survival by reducing the protein load on the ER. However, prolonged RIDD signaling has been reported to correlate with increased apoptosis [16]. The switch between anti-apoptotic XBP1s signaling and pro-apoptotic RIDD may be dependent upon the conformational state of IRE1α. Administration of a peptide domain derived from the kinase domain of IRE1α triggered IRE1α oligomerization and XBP1 cleavage but diminished RIDD and JNK activation [16]. IRE1α mediated RIDD activation is a new phenomenon in the field of ER stress and further studies are required to identify RIDD targets and appreciate the mechanisms controlling its activation.

Regulation of IRE1α signaling

Since IRE1α can elicit pro-survival and pro-apoptotic signals, mechanisms controlling the switch between the two must exist. Recent studies have revealed that IRE1α signaling is indeed finely controlled by a complex array of protein interactions with IRE1α as the central core component [17]. Pro-apoptotic Bcl-2 family members Bax and Bak positively modulate the amplitude of IRE1α signaling by interacting at the ER with the cytoplasmic domains of IRE1α resulting in increased XBP1s and JNK phosphorylation [18]. Binding of Bax and Bak to IRE1 is negatively regulated by Bax Inhibitor 1 (BI-1), a transmembrane protein localized to the ER and nuclear envelope. Normally, BI-1 is ubiquitinated by bi-functional apoptosis regulator (BAR) leading to proteosomal degradation. Under prolonged ER stress BAR expression is downregulated, BI-1 expression maintained and IRE1α signaling attenuated [19]. Recent work by Hetz and colleagues has proposed another layer of regulation mediated by BH3-only Bcl-2 family members. Under mild ER stress BH3-only proteins Bim and PUMA bind IRE1α, via their BH3 domain, and stimulate its RNase activity. Indeed upon induction of mild ER stress an in vivo reduction in XBP1s levels was determined in Bim−/− mice. However, upon sustained or chronic ER stress BH3-only proteins resume their pro-apoptotic function and target mitochondrial mediated pathways committing the cell to death [20]. Hsp72 has also recently been demonstrated to bind to and regulate IRE1α signaling. Gupta and colleagues reported binding of Hsp72 to the cytosolic domain of IRE1α, an interaction which increased the RNase activity of IRE1α resulting in an increase in XBP1 splicing [21]. Based on the current data it appears that numerous mechanisms may regulate the amplitude of IRE1α signaling, and through this mechanism control the switch from pro-survival to pro-apoptotic signaling. Future studies are required to fully determine the complexity of IRE1α regulation.


Following dissociation from Grp78, PERK dimerizes, autophosphorylates and signals for a general translational inhibition by phosphorylating elongation initiation factor 2α (eIF2α) [22] (Fig. 1). This general block in translation promotes cell survival by providing the cell with a window of opportunity to reduce the backlog of unfolded proteins thereby alleviating ER stress. The importance of this translational block is clearly evident in the hypersensitivity to ER stress-induced death of PERK−/− MEFs and knock-in non-phosphorylatable eIF2α cells [23, 24]. However, the translational block is not absolute as genes with particular regulatory sequences in their 5′ untranslated region, such as an internal ribosome entry site (IRES) can bypass the translational block with activating transcription factor 4 (ATF4) being one such example [25]. ATF4 is a member of the CCAAT/enhancer binding protein family (C/EBP) family of transcription factors. The majority of transcriptional targets of ATF4 are associated with cell survival and include genes involved in amino acid metabolism, redox reactions, protein secretion and stress responses [26]. As such transcription of this subset of genes in conjunction with translation inhibition should reduce levels of ER stress. However, when stress cannot be alleviated ATF4 helps push the cell towards death by upregulating transcription factor C/EBP homologous protein (CHOP) [23].


CHOP upregulation is a common point of convergence for all 3 arms of the UPR with binding sites for ATF6, ATF4 and XBP1s present within its promoter clearly illustrating the importance of this transcription factor. CHOP signaling is thought to mediate cell death signaling by firstly altering the transcription of genes involved in apoptosis and oxidative stress and secondly by relieving PERK mediated translational inhibition [27]. Pro-apoptotic targets of CHOP include BH3-only members of the Bcl-2 family. Puthalakath and colleagues demonstrated Bim upregulation in MCF-7 cells specifically in response to ER stress-inducing agents. Furthermore, knockdown of Bim in MCF-7 cells significantly attenuated ER stress-induced cell death clearly highlighting a role of Bim in the execution of ER stress-induced apoptosis. Dissection of the specific pathways regulating Bim revealed that a combination of transcriptional upregulation via CHOP and post translational modification namely protein phosphatase 2a (PP2a)-mediated dephosphorylation enabled sustained Bim expression [28]. CHOP has also been reported to regulate expression of BH3 only proteins by interacting with FOXO3A (in neuronal cells treated with tunicamycin) [29] and AP-1 complex protein c-Jun leading to its phosphorylation (in saturated fatty acid treated hepatocytes) [30].

CHOP-mediated downregulation of Bcl-2 has also been reported in response to ER stress suggesting that this transcription factor may shift the balance of Bcl-2 family members in favor of pro-apoptotic thus ensuring propagation and execution of the apoptotic signal [31]. Other transcriptional targets of CHOP include endoplasmic reticulum oxidoreductin 1 (ERO1α) and tibbles related protein 3 (TBR3). Increased ERO1α expression results in a hyperoxidizing environment within the ER which may promote cell death [32]. Additionally ERO1α has been reported to activate the inositol trisphoshate receptor (IP3R) stimulating calcium release from the ER [33], concurrent uptake by the mitochondria may lead to calcium overload and apoptosis. TBR3 is an intracellular pseudokinase that modulates the activity of several signal transduction kinases. Overexpression of TRB3 has been linked to cell death onset while knockdown of TRB3 in 293 and HeLa cells was reported to attenuate tunicamycin induced death [34]. The mechanism through which TRB3 mediates death signals is not understood but it has been suggested that TRB3 promotes apoptosis through binding AKT, preventing its phosphorylation and reducing its kinase activity [35, 36]. In addition to transcriptional control of pro- and anti-apoptotic genes CHOP activation also lifts translational inhibition mediated by PERK phosphorylation of eIF2α. CHOP mediated enhancement of GADD34 permits protein phosphatase 1 (PP1) dephosphorylation of eIF2α thus lifting translational inhibition [37]. Release of this translational block permits production of pro-apoptotic proteins further committing the cell to death. Inhibition of eIF2α dephosphorylation, by treatment with salubrinal, inhibited ER stress-induced apoptosis underscoring the contribution of releasing translational inhibition to progression of cell death [38]. Indeed, the importance of GADD34 signaling for ER stress-induced apoptosis is clearly evident in knockout mice which displayed resistance to ER stress-induced kidney damage [32].

The importance of CHOP-mediated signaling to ER stress-induced apoptosis is clearly illustrated by the presence of binding sites for ATF6, ATF4 and XBP1s in its promoter region. However important CHOP signaling is to the ER stress-induced apoptosis, the requirement for it is not absolute as CHOP−/− MEF cells still undergo apoptosis in response to prolonged ER stress albeit with much slower kinetics [32].


Owing to the presence of an ER targeted hydrophobic sequence ATF6 is an ER tethered protein. Following dissociation of Grp78, ATF6 translocates to the Golgi where SP1 and SP2 proteases cleave it releasing active ATF6 into the cytosol [2]. This bZip transcription factor family member upregulates expression of genes mainly involved in adapting to ER stress such as Grp78, Protein Disulphide Isomerase (PDI) and ER degradation-enhancing a-mannosidase-like protein 1 (EDEM1) [39]. ATF6 also increases transcription of XBP1 mRNA, an important IRE1α target [40]. ATF6 signaling is largely pro-survival and adaptive, however it can also be pro-apoptotic. ATF6 has been demonstrated to upregulate levels of CHOP during sustained ER stress [40]. Although not in an ER stress context selective activation of ATF6 apoptotic myoblasts during the differentiation process has been reported and linked to the downregulation of the anti-apoptotic protein Mcl-1 highlighting a potential pro-apoptotic role for ATF6 [41]. Whether ATF6 can mediate downregulation of Mcl-1 or other anti-apoptotic Bcl-2 family members during ER stress is currently unknown.

Mitochondria-mediated death signaling

As discussed above sustained activation of UPR signals can result in the upregulation of pro-apoptotic Bcl-2 family members such as CHOP-mediated activation of Bim. Other BH3 only proteins transcriptionally regulated by ER stress include PUMA and Noxa. Puma and Noxa are pro-apoptotic BH3 family members often referred to as ‘sensitizers’ of apoptosis, with Noxa reported to interact with Mcl-1 and A1 while Puma is thought to interact with various members of the pro-survival Bcl-2 family leading to subsequent MOMP induction [42, 43].Transcriptional activation of both Puma and Noxa in response to ER stress has been reported in a p53-dependent manner [44]. Partial suppression of ER stress-induced apoptosis has been reported in p53−/− cells and attributed to defective induction of Puma and Noxa [44]. The mechanism facilitating p53 activation during ER stress has not been fully elucidated. Recent studies suggest p53 upregulation during ER stress occurs in a NF-κB dependent manner [45]. Interestingly, IRE1α, PERK and ATF6 have all been linked to the activation NF-κB signals under various circumstances. PERK-mediated translational inhibition has been reported to lower levels of the short half-life protein IκB, permitting NF-κB translocation to the nucleus [46]. IRE1α signals have also been implicated in NF-κB activation via TRAF2 recruitment of IKK permitting translocation of NF-κB [47], while ATF6 signaling has been implicated in NF-κB activation during shiga toxin treatment of rat Nrk52e renal proximal tubular cells [48]. Knockdown of p53 exerted protective effects against ER stress induced by tunicamycin or brefeldin A in MCF-7 cells indicating it may have an important role in mediating death signals [45]. Given the diverse targets of NF-κB it is likely that its activation increases expression of pro-apoptotic proteins such as BH3-only proteins thus committing the cell to apoptosis. CHOP and ATF4 have also been implicated in PUMA and Noxa induction respectively [30, 44]. The importance of BH3-only protein induction is illustrated by PUMA and Noxa null MEFs which like Bim null MEFs exhibit partial resistance to ER stress-induced apoptosis [44]. Work by Futami and colleagues in which they carried out a siRNA screen for genes regulating ER stress-induced apoptosis confirmed a functional role for Noxa and Puma [49]. In neuronal cells, Puma transcriptional induction alone is crucial for the execution of apoptosis in response to ER stress [29]. The combination of increased BH3-only protein expression, via predominately transcriptional but also post-translation modifications in the case of Bim, and repression of anti-apoptotic proteins such as Bcl-2 shifts the balance in favor of apoptosis permitting Bax-Bak homo-oligomerization and mitochondrial outer membrane permeabilization causing cytochrome c release and subsequent apoptosome formation. Overexpression of Bcl-2 reduces loss of mitochondrial membrane potential and protects cells against ER stress inducers such as thapsigargin underscoring the importance of mitochondrial mediated signals in the propagation of ER stress-induced apoptosis [50].

ER/Mito Calcium cross talk and death

In addition to mediating death signals by triggering Bax-Bak oligomerization and mitochondrial outer membrane permeabilization, Bcl-2 family members have also been implicated in the regulation of ER mitochondria calcium signaling. The ER sequesters high concentrations of calcium (1-3 mM) through a dynamic process of active uptake via sarco/endoplasmic reticulum calcium transport ATPase (SERCA) pumps and release through calcium channels inositol trisphosphate receptor (IP3R) and ryanodine receptors [51]. The maintenance of sufficient ER calcium concentrations is imperative for ER function as many chaperone proteins, such as Grp78, require calcium binding to function at their optimum capacity [52]. Therefore, low ER calcium concentrations reduce chaperone function and disrupt the protein folding capacity of the ER resulting in a backlog of unfolded proteins and ER stress.

In addition to their specialized cell death functions recent work has demonstrated Bcl-2 family Bax, Bak, Bcl-xL and Bcl-2 can associate with the ER both under basal and stress conditions [5355]. Reports indicate that ER specific overexpression of Bcl-2 and Bcl-xL lower free ER calcium concentration and increase protection against apoptosis [54, 56]. The mechanism facilitating reduced free calcium levels is thought to involve Bcl-2 Bcl-xL interactions with IP3R possibly controlling channel opening. The protective role of Bcl-2 in regulating calcium release can be inhibited by the activation of kinases such as JNK. Phosphorylation of Bcl-2 within an unstructured loop region diminishes its anti-apoptotic protection by firstly inhibiting its ability to bind and neutralize BH3-only proteins and secondly by causing increased calcium release from the ER (presumably by an inability to bind and regulate IP3R) which associated with an increase in mitochondria calcium uptake and pro-apoptotic signals [57]. Studies have also implicated the Bcl-2/Bcl-xL binding partner BI-1 in regulation of ER calcium concentration. Overexpression of BI-1 in HT1080 cells reduced Bax translocation, mitochondrial depolarization and ER calcium release in response to thapsigargin treatment. A similar dysregulation in calcium release was present in cells derived from BI-1−/− mice which displayed enhanced calcium release and increased sensitivity to tunicamycin compared to wild type BI-1+/+ counterparts suggesting BI-1 is important in the transmission of the death signal from the ER to the mitochondria [58, 59].

Pro-apoptotic Bcl-2 family members Bax and Bak also localize to the ER where they function to antagonize Bcl-2 and Bcl-xL increasing ER calcium concentration and enhancing apoptotic sensitivity. The function of Bak and Bax is nicely illustrated by Bax/Bak double knockout MEFs which exhibit lower ER calcium concentrations and increased resistance to calcium dependent apoptotic signals [60]. Bax and Bak, analogous to their role in release of mitochondrial intramembrane space proteins, can oligomerize at the ER during ER stress-induced apoptosis [55]. Recently it has been demonstrated that Bax-Bak oligomerization and insertion into the ER induces pore formation facilitating release of luminal proteins Grp78 and PDI [61]. Whether calcium can be released by this mechanism has not been determined.

Localization of BH3-only members of the Bcl-2 family to the ER has also been described. For example, Bik a primarily ER localized BH3-only protein can mediate Bax-Bak dependent calcium release that has been shown to participate in intrinsic apoptotic signals. Surprisingly Bik upregulation in response to ER stress signals has not been reported and it appears to be an event solely associated with genotoxic stress [62]. Other members of the BH3-only subfamily regulated by ER stress signals include Puma and Noxa [44, 63]. Increased Puma expression has been linked to depletion of ER calcium levels via Bax activation [64].

Bcl-2 family members help regulate both ER calcium levels and release in response to pro-apoptotic signals. Surprisingly, mitochondrial calcium transporters have a low affinity for calcium and therefore require high levels to stimulate mitochondrial uptake. Within the cell this is achieved by contact sites between the ER and mitochondria with high calcium concentrations enabling mitochondrial calcium uptake. Such regions are referred to as mitochondria associated ER membranes (MAMs). MAMs ensure the efficient shuttling of calcium between the ER and mitochondria and as a consequence of this function are enriched in IP3 receptors which are linked to voltage dependent anion channel 1 (VDAC1) by the mitochondrial chaperone protein Grp75 [65]. The importance of Grp75 in this interaction has been demonstrated by knockdown of Grp75 resulting in reduced mitochondrial calcium uptake following agonist stimulation [66]. Regulation of MAM signaling in response to ER stress has been reported. The Sigma receptor 1 (Sig1-R) is an ER localized, calcium sensitive, transmembrane chaperone which complexes with Grp78 at MAMs. Calcium depletion from the ER causes Grp78 dissociation from Sig1-R increasing their respective chaperone activities and Sig1-R binding to and stabilization of IP3 receptors. Upon conditions of chronic ER stress Sig-1R redistribute from MAMs to the entire ER (via an unknown mechanism) where presumably they attempt, via their chaperone activity, to alleviate ER stress [67]. Indeed overexpression of Sig-1Rs reduces ER stress responses whereas knockdown of Sig1-R enhances apoptosis [67].

Another ER stress-induced MAM localized protein recently implicated ER stress-induced apoptosis is sarcoplasmic reticulum calcium ATPase 1 (S1T). Upon induction of ER stress S1T expression is enhanced via PERK-eIF2α-ATF4-CHOP signaling. Increased S1T expression increases ER calcium depletion through a combination of increased ER calcium leak, increased ER mitochondria contact sites and inhibition of mitochondria movement [68]. Knockdown of S1T expression reduced ER stress, mitochondrial calcium overload and apoptosis highlighting an important role for S1T in ER stress-induced apoptosis [68]. Aside from facilitating increased ER mitochondria contact sites by controlling S1T expression recent work has proposed that PERK itself is an essential MAM component. Verfaille and colleagues recently demonstrated PERK−/− cells have weaker ER mitochondria contact sites resulting in dysregulated ER mitochondria calcium signaling [69]. Given that PERK signaling is required for the regulation of many genes upon induction of ER stress including S1T it would be expected that the kinase domain of PERK is required for maintenance of ER mitochondria interaction sites. Surprisingly expression of a kinase dead mutant of PERK was able to restore ER mitochondria interaction in PERK−/− MEFs suggesting that PERK may, in addition to regulating downstream effectors, also function as a scaffold protein [69]. Indeed significant enrichment of PERK at MAMs was identified; as yet the exact function of PERK at MAMs sites is unknown.

Alternate modes of ER stress-induced cell death

Overexpression of Bcl-2 is able to inhibit ER stress-induced apoptosis indicating an important role for mitochondrial death signals in this process. Bcl-2 overexpression antagonizes ER stress-induced regulation of BH3-only proteins preventing mitochondrial cytochrome c release and caspase activation [70]. Additionally, Bcl-2 family members are inherently important in ER mitochondria calcium signaling with overexpression of Bcl-2 lowering ER calcium levels thereby preventing mitochondrial calcium overload and apoptosis. Likewise Bax−/− Bak−/− deficient cells exhibit resistance to ER stress-induced apoptosis presumably through a combination of mitochondrial and calcium mediated processes. Several reports, using Bax−/− Bak−/− cells, have demonstrated cell death upon prolonged exposure to ER stress-inducing conditions/agents [71, 72]. In vitro studies examining important regulators of ER stress-induced apoptosis such as Bax−/− Bak−/− or caspase-9−/− MEFs rarely extend ER stress treatment times beyond 48 h as wild type cells have succumbed to death at this point and inhibition in the knockout cells is evident. However, prolonged ER stress conditions can initiate cell death in mitochondrial-mediated apoptosis compromised cells such as Bax−/− Bak−/− MEFs. This in itself is not an unexpected result as exposure to prolonged stress will at some point trigger death via an alternate mechanism. Indeed Bax−/− Bak−/− MEFs exhibit features of autophagy and cell death in response to prolonged ER stress [73]. Studies within our laboratory have demonstrated that cells deficient in the mitochondrial pathway undergo an alternate form of cell death involving aspects of autophagy (LC3 I to II conversion) and apoptosis (caspase activation) when exposed to prolonged stresses including ER stress (unpublished results). Moreover, our data indicates that caspase activation is dependent upon ATG5 indicating cross-talk between autophagy and cell death pathways in response to prolonged ER stress (unpublished results). These findings are of considerable interest when we take into account that many cancer cells are resistant to death signals propagated via the mitochondrial pathway. Furthermore in vivo such cells are exposed to stresses such as sustained glucose deprivation or hypoxia that are known to induce a robust ER stress response. Therefore, in the future it will be important, in the context of diseases such as cancer to understand how cells devoid of conventional apoptotic signaling pathway retain susceptibility to ER stress-induced death and in particular the role that autophagy may play.

microRNAs and ER stress

The regulation of ER stress-induced death pathways by microRNAs is a recent area of research with studies indicating miRNAs can either directly modulate the ER stress response or themselves be regulated by ER stress. For example, Yang and colleagues demonstrated miR-122 overexpression downregulated ER stress responses in HepG2 cells [74]. This observation is particularly interesting in the context of hepatocellular cancer where repression of miR-122 is frequently observed. The downregualtion of miR-122 would presumably lift repression on UPR responses increasing the adaptive ability of the cancer cell. Indeed in cisplatin treated Huh7 cells inhibition of miR-122 decreased cell death highlighting the benefit of miR-122 repression to cancer cells. [74]. ER stress-mediated downregulation of miR-221/222 has been reported in hepatocellular carcinoma cells where it associated with a resistance to cell death [75]. Addition of miR-221/222 mimetics restored sensitivity to ER stress-induced apoptosis via a mechanism involving upregulation of p27kip1 and G1 phase arrest suggesting mimetics directed against miR-122 or miR-221/222 maybe of therapeutic benefit particularly in hepatocellular cancer [75].

Direct regulation of miRNA expression by ER stress sensors particularly PERK has been reported and may regulate the delicate balance that exists between pro-and anti-apoptotic signaling during ER stress. PERK mediated induction of miR-30c-2* has been reported during ER stress and linked to a downregulation in XBP1 mRNA reducing pro-survival signaling and aiding commitment to cell death [76]. Additionally PERK mediated repression of the mir-106b-25 cluster and its host gene MCM-7 has been reported to result in increased Bim expression and apoptosis [77]. Conversely, recent work from Chitnis and colleagues implicates PERK facilitated miRNA regulation in pro-survival signaling. miR-211 was identified as a PERK target and demonstrated to repress CHOP expression allowing a temporal window for the pro-survival response. However, upon sustained ER stress miR-211 expression was silenced, permitting CHOP accumulation and induction of the pro-apoptotic response [78]. Based on the current literature it seems that miRNA regulation help shift the balance between survival and cell death during ER stress. Further research into ER stress-mediated regulation of miRNAs is required to fully elucidate their role and determine if they represent a viable therapeutic target.


ER stress-induced cell death is a complex and highly regulated process carefully controlled by ER localized stress receptors. Initial signaling from each stress receptor aims to reduce levels of unfolded proteins and restore cellular homeostasis. However, following sustained or excessive ER stress a switch in signaling from survival to death occurs sealing the fate of the cell. Based on the current data IRE1α and PERK signals are important in cell death commitment. Signals from each of these receptors have important roles in regulating Bcl-2 family member expression particularly the expression of BH3-only proteins. By tipping the balance in favour of pro-apoptotic Bcl-2 family members pro-apoptotic mitochondria-mediated signals are activated committing the cell to death. In addition to triggering Bax/Bak oligomerization and cytochrome c release Bcl-2 family members have recently been shown to function in ER mitochondria cross talk thereby controlling calcium movement between these two organelles. Recent studies have highlighted the complexity of ER mitochondria calcium signaling particularly the importance of MAMs in this process. In the last few years the role of ER localised proteins Sigma 1 receptor and the calcium ATPase S1T in ER mitochondria cross talk has emerged. The role of MAMs and the proteins which regulate cross talk during ER stress is one obvious area of research for the future. The role of microRNAs in regulation of ER stress-induced cell death also merits future research. It is only in the past few years that we have started to appreciate the function of microRNAs in ER dependent death signaling. Further work is required to unmask the array of microRNA targets and determine their function in ER stress-induced death.

Over the past 10 years the field of ER stress-induced death has yielded much information concerning the basic signaling mechanisms triggered. It is only now that we are beginning to both understand the delicate balance of interplay between pro-survival and pro-death signals.


Our research is supported by grants from Science Foundation Ireland (09/RFP/BIC2371), Breast Cancer Campaign (2010NovPR13). P Cleary is funded by an Irish Cancer Society Scholarship (CRS11CLE).

Copyright information

© Springer Science+Business Media New York 2013