Among the three protein sensors, IRE1 is the most evolutionarily conserved ER stress transducer protein (20, 28). IRE1 exists in two structurally related isoforms, IRE1α, the ubiquitously expressed isoform, and IRE1β, which has been primarily identified in the intestinal epithelium of the gut and respiratory tract (29, 30). Upon sensing the misfolded or unfolded proteins, Bip protein dissociates from IRE1α and facilitates the activation of IRE1α through homodimerization and trans-autophosphorylation or other mechanisms (31, 32). Activated IRE1α uses its endoribonuclease activity to remove a 26-bp pair segment from an unspliced mRNA encoding the transcription factor X-box binding protein 1 (XBP1u), which in turn causes a shift in the reading frame and generates a spliced and functionally active isoform of XBP1 (XBP1s) (33, 34). XBP1s is a potent CREB/ATF basic leucine zipper (bZIP) transcription factor that can induce the expression of genes involved in protein folding, secretion, maturation, the ERAD signaling and the synthesis of phospholipids (26, 35). Several lines of studies have shown that XBP1s is also implicated in various biological processes, such as lipid metabolism (36), pro-inflammatory cytokines synthesis (37), the hypoxia response signaling pathway (38), cellular differentiation (39), and the hexosamine biosynthetic pathway (40), indicating a critical role for IRE1-XBP1 UPR signaling and cellular response. Unlike XBP1s, XBP1u is a short-lived protein that lacks the transactivation domain. XBP1u inhibits the translocation of XBP1s from cytoplasm into the nucleus, therefore serving as a dominant-negative regulator to block the transactivation of XBP1 downstream targets under certain conditions (41–43). Recent studies have shown that IRE1α can target other mRNAs for degradation, therefore inhibiting the synthesis of nascent proteins through the regulated IRE1-dependent decay (19, 44, 45), indicating an additional level of regulation to cope with ER stress.

It should be noted that the functional role of IRE1α on various biological processes, such as proliferation, metabolism, inflammation, autophagy, and apoptosis, can be mediated in an XBP1-independent manner (46). First, IRE1α can bind to and activate TNFα receptor associated factor 2 in the cytoplasm, which in turn activates c-Jun N-terminal kinase or nuclear factor-κB (NF-κB), thus participating in inflammatory response or proapoptotic signaling in response to ER stress (12, 17, 18). Second, IRE1α can directly modulate p38 MAPK and ERK1/2, two critical protein kinases related to stress response, indicating a link between UPR signaling and cellular response (47). Third, IRE1α has been reported to interact with the proapoptotic BCL-2 (B-cell lymphoma 2) family proteins, BAX (BCL-2-associated X protein), or BAK (BCL-2 antagonist/killer), therefore contributing to apoptotic cell death (48). In addition to transcription regulation and protein interactions, IRE1α can cause degradation of selective microRNA that normally represses translation of caspase-2, therefore leading to activation of the mitochondrial apoptosis pathway (49–51). More studies are required to elucidate how this epigenetic regulation is implicated in and contribute to intestinal homeostasis. In contrast to IRE1α, IRE1β has a broader endoribonuclease activity, which can lead to the degradation of a large array of transcripts (52). However, the underlying mechanisms are incompletely understood.

PERK has structural similarities with IRE1 transmembrane protein (21, 53). Disengagement from Bip upon sensing misfolded or unfolded protein in the ER activates PERK by homodimerization and trans-autophosphorylation (11, 54, 55). Activated PERK phosphorylates the α subunit of eukaryotic translation initiation factor 2 (eIF2α), a component of the translation initiation complexes, therefore attenuating global protein synthesis and alleviating the overload of misfolded proteins (53, 56). Interestingly, the phosphorylated eIF2α selectively enhances translation of ATF4, a transcription factor regulating the expression of genes implicated in protein folding, oxidative stress response, and ER stress-induced apoptosis (57, 58). CHOP (CCAAT/enhancer-binding protein homologous protein, also known as GADD153) has been shown to be a critical mediator responsible for ER stress-induced cell death through different mechanisms (19, 59). First, CHOP can suppress the antiapoptotic protein BCL-2 or enhance numerous proapoptotic proteins, such as Bim, telomere repeat binding factor 3, GADD34 (growth arrest and DNA damage 34), or death receptor 5 (60–62). Second, CHOP induces the transcriptional expression of ER oxidase 1α, which in turn resulted in generation of reactive oxygen species (ROS) and release of Ca²⁺ from the ER, thereby conferring to apoptosis (12, 18, 19). Third, CHOP can interact with ATF4 and activate genes involved in protein synthesis machinery, thus causing energy depletion and apoptosis in ER-stressed cells (63). In addition to inducing cell death, activation of the PERK pathway can activate antioxidant reactions to avoid accumulation of ROS in response to ER stress (64). This effect of PERK is mediated by ATF4-induced phosphorylation of nuclear factor-erythroid-derived 2-related factor 2, which can activate enzymes with an ability to remove oxidants. These enzymes include NAD(P)H-quinone oxidoreductase, heme oxygenase 1, and glutathione S-transferase (65, 66).

Besides PERK, eIF2α can be phosphorylated by other protein kinases, including PKR (double-stranded RNA activated protein kinase) (67), general control non-depressive kinase 2 (68), and heme-regulated inhibitor kinase (69). All the protein kinases have similar kinase catalytic domains, and therefore possess a capability to phosphorylate eIF2α at its Ser 51 residue to regulate protein synthesis (70). Because of the presence of different regulatory domains in the kinases, they can be activated by different stress stimuli (70). Despite the diversity of stress stimuli and activated protein kinases, these signaling cascades converge on the phosphorylation of eIF2α, indicating a critical functional role of eIF2α in determining cell fate decision. These biochemical roles of eIF2α have been highlighted in several review papers (70, 71). Additional studies are needed to elucidate how the kinases-activate eIF2α interacts with ER stress signaling and contribute to cell survival and apoptosis under specific conditions.

Importantly, eIF2α can be dephosphorylated by protein phosphatase, such as protein phosphatase 1 regulatory subunit 15A (growth-arrest DNA damage-inducible protein 34, also known as GADD34) and subunit 15B (known as CReP) (72–74), thus forming a negative feedback regulation on PERK–eIF2α signaling. All these data indicate that activation of eIF2α acts as a molecular switch either to induce cell death or to promote cell survival by attenuating protein synthesis in a context-dependent manner (63).

Activating transcription factor 6 is a key transcription factor that helps intestinal epithelial cells cope with ER stress (22). Two homologous ATF6 proteins, such as ATF6α and ATF6β, have been identified in mammalian cells (75). Upon sensing the misfolded or unfolded proteins in the ER, ATF6 that is released from Bip migrates from the ER to the Golgi apparatus, where it subsequently undergoes cleavage by site-1 protease (S1P) and site-2 protease (S2P) in its luminal domain and transmembrane region, respectively, leading to the release of the cytosolic domain of ATF6, ATF6 fragment (ATF6f) (22, 24). ATF6f then translocates to the nucleus to bind DNA and transcriptionally upregulates target genes involved in protein folding, or ERAD to restore ER homeostasis or induce cell death in response to severe or prolonged ER stress (76–79).

In addition to the canonical ER membrane-bound proteins, IRE1, PERK, and ATF6 as abovementioned, novel types of ER stress transducers sharing a region of high sequence similarity with ATF6 have been identified (80, 81). These proteins possess a transmembrane domain, which allows them to associate with the ER, and have a transactivation domain and a basic leucine zipper (bZIP) domain. They are collectively known as old astrocyte specifically induced substance (OASIS) family members, which consist of CREB3L1/OASIS (82), CREB3L4/CREBH (RE-Bip H) (83), CREB3L2/BBF2H7 (box B-binding factor 2 human homolog on chromosome 7) (84), AIbZIP/Tisp40/CREB3L4/CREB4 (cyclic AMP responsive element Bip 4) (85), and Luman/LZIP/CREB3 (86, 87). Most of these ATF6-related bZip factors are processed at the Golgi as described for ATF6, but their functions are tissue specific due to the unique cell or tissue specific expression patterns of these transducers (88, 89).

Paneth cells and goblet cells in the gastrointestinal tract can produce large amounts of proteins, which undergo protein folding and posttranslational modifications before being secreted from the cells (90, 92). This feature of secretory cells requires a fine monitoring and management of the ER to avoid the accumulation of unfolded or misfolded proteins. Both clinical and experimental animal studies show that impaired UPR is associated with development of colitis in humans and animals as abovementioned (11, 100, 123). Several lines of studies show that dysfunction of genes involved in protein secretion, such as Anterior gradient 2 (Agr2) and MUC2, is associated with IBD through various mechanisms (98, 125).

Agr2 is an ER-resident protein expressed in secretory IECs such as goblet, Paneth, and enteroendocrine cells in the small intestine (126). This protein responsible for the formation of correctly arranged disulfide bonds in mature proteins (127, 128). By using inducible Agr2^−/− mice, Fang et al. showed that deletion of Agr2 is associated with decreased goblet cells and MUC2, dramatic expansion of the Paneth cell compartment, abnormal Paneth cell localization, elevated ER stress, and severe colitis (125). This finding, along with previous observation that both CD and UC patients have decreased AGR2 (129), indicates that AGR2 is essential for intestinal homeostasis. Deficiency of this gene impairs the secretion of proteins in intestinal epithelial cells and activates uncontrolled immune response, ultimately contributing to IBD.

MUC2 is the major component of mucin that goblet cells secrete into the intestinal lumen (130, 131). An appropriate secretion of MUC2 requires extensive O-glycosylation of central mucin repeats and intra- and inter-chain disulfide bond formation in the cysteine-rich N and C-terminal domains within the ER (132). Mutations in the Muc2 gene lead to ER stress, inflammation, and spontaneously colitis due to accumulation of MUC2 precursor in the ER and reduction in mucin secretion (98). Similar observations have been described in UC patients (131, 133). Interestingly, administration of interleukin-10 (IL-10), an anti-inflammatory cytokine, has been reported to attenuate intestinal inflammation and enhance mucin production in ER-stressed epithelial cells through currently unknown mechanisms (99).

Besides the genetic factors and secretion-related proteins, various environmental factors have been implicated in ER stress signaling and contribute to intestinal inflammation in IBD (99, 134–136). Several inflammatory mediators have been shown to be able to influence ER stress response. For example, anti-inflammatory cytokine IL-10 can modulate ATF6 nuclear recruitment to the Bip promoter, therefore blocking ER stress (99). This finding provides a plausible explanation for the development of colitis in IL-10 knockout mice (137). By contrast, pro-inflammatory cytokine tumor necrosis factor α (TNF-α) can enhance ER stress and UPR signaling by inducing ROS production and its accumulation in the ER (134). Moreover, increased expression of inflammatory cytokines, such as IL-1β, TNF-α, and IFN-γ, has been observed in aberrant mucin assembly induced colitis mice (98). These findings highlight a role of immune response in the development of experimental colitis or IBD. More studies are required to uncover how the antagonistic cytokines are finely coordinated under physiological conditions and ER-stressed conditions.

Intestinal microbiota is considered to be another environmental factor highly correlated with IBD (138, 139). Intestinal microbiota is well known to produce inflammatory molecules by regulating expression of genes involved in immune response, which in turn triggers ER stress (55). In addition, the effect of microbiota on ER stress can also be mediated by various microbial metabolites. For example, trierixin, a macrocyclic lactam derived from Streptomyces sp., has been identified as an inhibitor of XBP1 splicing (135). Interestingly, two structurally related compounds, such as mycotrienin II and trienomycin A, were isolated from the culture broth of a trierixin-producing strain and reported to inhibit the induction of XBP1 (135). These compounds might be potential pharmacological tools for the functional analysis of XBP1 signaling in response to ER stress. However, key enzymes involved in the production of these metabolites with an ability to suppress XBP1, as well as their effects on the intestinal homeostasis, are still largely unknown. Paton et al. reported that oral infection with AB₅ a cytotoxin-producing Escherichia coli resulted in pathologic UPR in mice (136). Further studies showed that the A subunit of AB₅ cytotoxin specifically cleaved Bip by the serine protease activity and activated it in eukaryotic cells (140). Notably, proteins with significant sequence homology to the A and B subunits of this cytotoxin have been reported in a wide variety of microorganisms (136), which can interfere with the ER stress-related signaling cascade and inflammatory responses in the IECs.