|Home | About | Journals | Submit | Contact Us | Français|
The endoplasmic reticulum (ER) is a cellular compartment responsible for multiple important cellular functions including the biosynthesis and folding of newly synthesized proteins destined for secretion, such as insulin. A myriad of pathological and physiological factors perturb ER function and cause dysregulation of ER homeostasis, leading to ER stress. ER stress elicits a signaling cascade to mitigate stress, the Unfolded Protein Response (UPR). As long as the UPR can relieve stress, cells can produce the proper amount of proteins and maintain ER homeostasis. If the UPR, however, fails to maintain ER homeostasis, cells will undergo apoptosis. Activation of the UPR is critical to the survival of insulin-producing pancreatic β-cells with high secretory protein production. Any disruption of ER homeostasis in β-cells can lead to cell death and contribute to the pathogenesis of diabetes. There are several models of ER stress-mediated diabetes. In this review, we outline the underlying molecular mechanisms of ER stress-mediated β-cell dysfunction and death during the progression of diabetes.
Proteins form the basic building blocks of cells, tissues, enzymes, and organs. Continuous production of proteins and selective degradation of defective or excessive proteins are essential processes for the maintenance of cellular homeostasis and for the regulation of cell function. Newly produced proteins are not immediately functional. To become functional, they must fold into their proper three-dimensional structures. The endoplasmic reticulum (ER) has an essential function in the protein folding process for secretory proteins, such as insulin, as well as cell surface receptors and integral membrane protein.
Protein folding in the ER is essential to cell survival, and is preserved during evolution in both unicellular organisms and eukaryotes. But as secretion is the basis of multicellularity, ER protein folding homeostasis powerfully impacts the physiology of higher eukaryotes. The sensitive folding environment in the ER can be perturbed by pathological insults such as viral infections, environmental toxins, inflammatory cytokines, and mutant protein expression, as well as by physiological processes such as aging and the large biosynthetic load placed on the ER due to insulin production in response to food uptake (Figure 1). When the demand that a load of proteins places on the ER exceeds its folding capacity, the result is ER stress.
The cytoprotective, adaptive response to ER stress is the unfolded protein response (UPR), also called ER stress signaling (Figure 2). As long as the UPR can mitigate ER stress, cells maintain protein homeostasis and generate protein relative to the need for it. Under these conditions, ER stress is beneficial. There are four distinct responses of the UPR: 1) up-regulation of molecular chaperones to increase the folding activity and reduce protein aggregation, 2) translational attenuation to reduce ER workload and prevent further accumulation of unfolded proteins, 3) ER-associated protein degradation (ERAD) to promote clearance of unfolded proteins, and 4) apoptosis when function is extensively impaired [1,2]. The UPR must maintain a balance between its downstream targets, which are anti-apoptotic and those which are pro-apoptotic (Figure 1). When the cell encounters ER stress and the UPR is properly balanced, the cell is then primed for a future insult. However, when activation is tipped in the favor of pro-apoptotic components due to UPR dysfunction or chronic and high ER stress, the cell undergoes irreversible damage which leads to cell death.
There are three master regulators of this signaling pathway: inositol requiring 1 (IRE1), PKR-like kinase (PERK), and activating transcription factor 6 (ATF6) (Figure 2). The signaling from downstream effectors of these pathways merges in the nucleus to activate UPR target gene transcription.
IRE1, a central regulator of the UPR, is a type I ER transmembrane kinase. Its N-terminal luminal domain acts as a sensor for ER stress signaling . Upon sensing the presence of unfolded or misfolded proteins, IRE1 dimerizes and autophosphorylates to become active. Activated IRE1 splices X-box binding protein 1 (XBP1) mRNA [4–6]. Spliced XBP1 mRNA encodes a basic leucine zipper transcription factor that upregulates UPR target genes, including genes that function in ERAD such as ER-degradation-enhancing-α-mannidose-like protein (EDEM) , as well as genes that code for folding proteins such as protein disulfide isomerase (PDI) . High levels of chronic stress lead to the recruitment of TNF-receptor-associated factor 2 (TRAF2) by IRE1 and the activation of apoptosis-signaling-kinase 1 (ASK1). Activated ASK1 activates c-Jun N-terminal protein kinase (JNK) and leads to apoptosis [9–11].
PERK, a second regulator of the UPR, is responsible for regulating protein synthesis during ER stress [12,13]. It, too, is a type I ER transmembrane kinase. Like IRE1, its N-terminal luminal domain is sensitive to ER stress. When activated by ER stress, PERK oligomerizes, autophosphorylates and then directly phosphorylates Ser51 on the α-subunit of eukaryotic initiation factor 2 (eIF2α) . This in turn inhibits protein translation by reducing the formation of ribosomal initiation complexes and recognition of AUG initiation codons. This leads to a reduction in ER workload and protects cells from ER stress-mediated apoptosis . Concomitant with the inhibition of general translation is the increased selective translation of UPR target genes, as these polycistronic mRNAs have inhibitory upstream open reading frames (uORFs) and are thus preferentially translated by the ribosome. One of these mRNAs is that of activating transcription factor 4 (ATF4), a b-ZIP transcription factor that regulates UPR targets such as C/EBP-homologous protein (CHOP) and growth arrest and DNA damage inducible gene 34 (GADD34), as well as genes involved in redox balance and amino acid synthesis .
A third regulator of ER stress signaling and mediator of transcriptional induction, ATF6, is a type II ER transmembrane transcription factor . Upon sensing stress in its N-terminal luminal domain, ATF6 transits to the Golgi where it is cleaved by S1 and S2 proteases, generating an activated b-ZIP factor . This processed form of ATF6 translocates to the nucleus to activate UPR target genes responsible for protein folding and ERAD [17,18]. There are two isoforms of ATF6, ATF6α and ATF6β, with fairly ubiquitous tissue distribution. The α-isoform has been shown to be solely responsible for transcriptional induction of ER chaperones [19*]. It has been reported that unprocessed ATF6 is unstable and quickly degraded by the ubiquitin-proteasome pathway to prevent hyper-activation of the UPR .
Diabetes mellitus is characterized by chronic high blood glucose levels. The two main forms of diabetes mellitus are type 1 and type 2 diabetes; they are characterized by an absolute or relative deficiency, respectively, in insulin production. This disease is associeted with severe peripheral effect and lead to heart disease, kidney failure, and blindness, thus it can be very devastating. The pathogenesis of this disease involves pancreatic β-cell dysfunction (i.e. insufficient insulin production). Recent evidence suggests that defective ER stress signaling contributes to β-cell dysfunction and eventual β-cell death. One main reason behind this is the highly specialized secretory function of β-cells to produce insulin in response to fluctuations in blood glucose levels. This constant demand from the body for insulin biosynthesis and secretion, has made β-cells dependent on a greatly efficient UPR. Indeed, baseline ER stress levels are high in these cells [21**].
Due to these observations, ER stress signaling is becoming more and more a focus in diabetes research. These efforts are supported by the recent identification of genetic links between ER stress and patients with type 1 and type 2 diabetes, a focus for this review.
The relationship between ER stress and diabetes was first revealed in a rare autosomal recessive form of juvenile diabetes, Wolcott-Rallison syndrome. In 1972, Wolcott and Rallison described two brothers and a sister with infancy-onset diabetes mellitus and multiple epiphyseal dysplasia . In this syndrome, mutations have been reported in the EIF2AK3 gene encoding PERK . Because these mutations are within the catalytic domain of PERK, it is likely that they cause a loss-of-function of PERK kinase activity. This loss of PERK kinase activity leads to the reduction in the phosphorylation of eIF2α, a substrate of PERK. When a high workload is placed on the ER of the beta cell, for example when insulin demand increases following meal intake, phosphorylation of eIF2α is essential in mitigating ER stress and thereby promotes cell survival . Therefore, a loss-of-function of PERK and a consequent disruption in translational attenuation during ER stress via decreased eIF2α phosphorylation, could directly attribute to beta cell apoptosis. This has been illustrated in several animal models. Indeed, PERK −/−mice develop diabetes due to excessive ER stress in their β-cells causing β-cell apoptosis . Additionally, mutant mice carrying a heterozygous mutation in the phosphorylation site of eIF2α (Eif2s1+/tm1Rjk) become obese and, due to beta cell dysfunction, diabetic when fed a high-fat diet . Collectively, these observations suggest that beta cell apoptosis in Wolcott-Rallison patients is caused by excessive, unresolved ER stress and a defect in the UPR (i.e. PERK signaling).
The negative regulator of PERK signaling, protein kinase inhibitor of 58 kDa (P58IPK), also functions in maintaining ER homeostasis in β-cells. P58IPK is an important component of a negative feedback loop used by these cells to inhibit eIF2α signaling and attenuate the UPR . P58IPK knockout mice show a gradual onset of glucosuria and hyperglycemia associated with increased apoptosis of islet cells , thus this UPR component may be involved in the pathogenesis of diabetes in humans.
Permanent neonatal diabetes may also be attributed to excessive ER stress in the β-cell. Neonatal diabetes is defined as insulin-requiring hyperglycemia within the first month of life. This is typically associated with slowed intrauterine growth and is a rare disorder. Permanent neonatal diabetes, considered a genetic disorder, can be caused by several types of mutations, including mutations in insulin promoter factor 1 (IPF-1), and results in lifelong dependence on insulin injections. It has recently been shown that mutations in the human insulin gene, specifically mutations occurring in critical regions of preproinsulin can also cause this kind of pathology [28*]. This is an autosomal dominant disorder, with mutations primarily occurring in critical regions of preproinsulin. This presumably leads to improper folding of insulin, triggering the UPR and ultimately leads to β-cell apoptosis. A mouse model of this disease, the Munich and Akita mouse, has a dominant missense mutation in the Ins2 gene [29,30]. In the Munich mouse, there is a cysteine95-to-serine substitution, leading to a loss in an inter-chain disulphide bond of proinsulin . In the Akita mouse, there is a cysteine96-to-tyrosine substitution . This mutation also leads to disruption of disulphide formation between the A and B chain of proinsulin, causing insulin to misfold and accumulate in the ER of the β-cell . This accumulation of misfolded insulin leads to severe ER stress, β-cell apoptosis, and consequently diabetes .
Increasing evidence supports the role of ER stress-mediated β-cell death in the pathogenesis of type 1 diabetes (i.e. autoimmune diabetes). The baseline of ER stress in β-cells is higher than that of other cell types due to their exposure to frequent energy fluctuations, as well as their high client load, insulin. It is therefore possible that any additional ER stress applied to these cells by genetic or environmental factors can lead to cell death. β-cells that undergo apoptosis as a consequence of this additional, unresolved ER stress contain misfolded proteins that can act as “neo-autoantigens” – dendritic cells in the islets engulf ER stress-induced apoptotic β-cells and stimulate the maturation of β-cell-reactive T cells that mediate autoimmune destruction of remaining β-cells . Viral infections, environmental stresses, as well as nitric oxide (NO), are several insults to the β-cell that can lead to excessive, unresolved ER stress, triggering an apoptotic cascade and leading to the production of “neo-autoantigens”.
NO plays an important role in β-cell apoptosis in type 1 diabetes . Inflammatory cytokines such as γ-interferon (IFN-γ) and interleukin-1β (IL-1β) in beta cells induce the production of NO, which leads to β-cell failure and consequently cell death. There is evidence that this process is mediated by ER stress . Production of NO leads to the attenuation of the sarco-endoplasmic reticulum pump Ca2+ ATPase2b (SERCA2b) and consequently the reduction of calcium in the ER. This calcium depletion leads to severe ER stress and the induction of the pro-apoptotic transcription factor CHOP [35,36**]. It has been shown that CHOP is induced by a NO donor, S- nitroso-N-acetyl-D,L-penicillamine (SNAP), and pancreatic islets from CHOP−/− mice are resistant to NO-induced apoptosis  (Figure 3).
Activating transcription factor 3 (ATF3), a pro-apoptotic transcription factor of the ATF/CREB family, may also contribute to ER stress-mediated apoptosis in type 1 diabetes. ATF3 is induced by pro-inflammatory cytokines and NO (Figure 3). ATF3 knockout mouse islets are partially protected from NO- and cytokine-induced beta cell apoptosis, while over-expression of this transcription factor in mouse islets leads to β-cell dysfunction .
One of the main contributing factors to the pathogenesis type 2 diabetes is the reduction of β-cell mass . Resistance to insulin action in peripheral tissues (i.e. adipose, muscle, and liver) is one of the primary presenting features of this disorder. This insulin resistance leads to the hyper-production of insulin (i.e. hyperinsulinemia) in the β-cell. Hyperglycemia and type 2 diabetes develops only in patients that are unable to sustain this compensatory response of the β-cell . This increase in insulin biosynthesis overwhelms the folding capacity of the ER, leading to chronic activation of the UPR. This chronic, hyperactivation of ER stress signaling can lead to β-cell dysfunction and death. There are several components of ER stress signaling that could contribute to this β-cell loss: IRE1-JNK signaling, CHOP, and glycogen synthase kinase 3β (GSK3β) (Figure 4). The IRE1 pathway is important in insulin biosynthesis, where transient increases in insulin production lead to IRE1 activation . Chronic activation of IRE1 during prolonged increases in insulin biosynthesis, however, may lead to β-cell death through IRE1-mediated activation of JNK . CHOP is also an important player of ER stress-mediated β-cell death and may promote the progression of type 2 diabetes . A third signaling component, GSK3β, also plays a role in β-cell death caused by ER stress. GSK3β is a substrate of the survival kinase, Akt , and it has been demonstrated that attenuation of Akt phosphorylation during ER stress mediates dephosphorylation of GSK3β, leading to ER stress-mediated apoptosis .
Insulin resistance, a feature of type 2 diabetes, leads to β-cell exhaustion, glucotoxicity, and hyperinsulinemia which place a massive strain on the ER of the β-cell . This, however, is not the only source of stress for the ER. It has recently been shown that free fatty acids (FFAs), specifically long-chain FFAs, also induce β-cell apoptosis [43**,44] (Figure 4). Treatment of β-cell lines with the long-chain FFA, palmitate, increases levels of ER stress markers such as ATF4 and spliced XBP-1. In addition, it has been shown that circulating FFAs lead to β-cell lipotoxicity and consequently excessive ER stress .
Recent studies suggest an involvement of ER stress in insulin resistance of liver, muscle, and adipose tissues. IRE1-JNK signaling plays an important role in the insulin-resistant liver tissue of type 2 diabetes patients. Obesity leads to hyper-activation of JNK signaling due to severe ER stress, leading to serine phosphorylation of insulin receptor substrate-1 (IRS-1), which inhibits insulin action . Like β-cells, hepatocytes have a high baseline ER stress level, and therefore may be sensitive to additional ER stress. It has been shown that high ER stress in liver cells can be resolved via overexpression of the ER-resident chaperone oxygen-regulated protein 150 (ORP150), while suppression of this chaperone in mice inhibits insulin sensitivity .
Wolfram syndrome, a rare autosomal recessive disorder characterized by diabetes mellitus and optical atrophy, was first described by Wolfram and Wagener in 1938 . This syndrome is also described as DIDMOAD (Diabetes Insipidus, Diabetes Mellitus, Optical Atrophy, and Deafness), as patients also present with secondary symptoms in addition to diabetes and optic atrophy. Postmortem studies reveal a non-autoimmune-linked selective loss of pancreatic β-cells . The nuclear gene responsible for this syndrome was identified by two separate groups in 1998 and named WFS1 [50,51].
WFS1 has been shown to be mutated in 90% of patients with Wolfram syndrome . More than 100 mutations of the WFS1 gene have been identified, most of which are inactivating mutations and located in exon 8 which encodes the protein’s transmembrane and C-terminal domains [53,54]. The WFS1 protein has been shown to be localized to the ER, and while ubiquitously expressed, it is highly expressed in the pancreas [55,56]. In the pancreas, it is localized to the β-cell [57**].
WFS1 has also been shown to be a downstream target of IRE1 and PERK signaling, induced transcriptionally and translationally in response to ER stress, and when suppressed, causes high levels of ER stress in the β-cell [57**], suggesting that the pathogenesis of Wolfram syndrome can be attributed to chronic, unresolved ER stress due to the lack of functional WFS1 protein in the β-cell.
Although other endocrine and exocrine cells of the pancreas are active in protein secretion, WFS1 expression levels in these cells are not detectable, as compared with β-cells which are specialized in insulin biosynthesis and secretion. WFS1 expression has been shown to be induced during insulin secretion [21,57**], suggesting that WFS1 is an important component of proinsulin folding and processing in the ER of the β-cell.
The high levels of ER stress and pancreatic β-cell death in patients with Wolfram syndrome may be related to the β-cell dysfunction in patients with type 2 diabetes. The pathogenesis of type 2 diabetes is a result of the peripheral resistance to the action of insulin, which leads to the prolonged increase in insulin biosynthesis. Because the folding capacity of the ER is then overwhelmed, this peripheral resistance to insulin activates ER stress signaling pathways . For this reason, chronic ER stress in β-cells may lead to β-cell apoptosis in patients with type 2 diabetes who are genetically susceptible to ER stress. Indeed, recent genome studies show a link between WFS1 single nucleotide polymorphisms (SNPs) and an increased risk for type 2 diabetes [58**,59].
Our understanding of ER stress signaling in β-cells is far from complete. This is because interactions of the three UPR pathways regulated by the master regulators, IRE1, PERK, and ATF6, have not been studied extensively. System biology approaches using genomics, transcriptomics, and proteomics are necessary to the complete understanding of ER stress signaling. However, based on the information that we have today, predictions can be made about the future of research on ER stress signaling in β-cell biology and the consequences for the pathogenesis of diabetes.
Increasing evidence suggests that crosstalk exists between ER stress signaling and other signaling pathways, such as mTOR signaling, JNK signaling, and insulin receptor signaling pathways. The complete understanding of this crosstalk will lead to the discovery of unexpected links between ER stress and biological outcomes, such as cell proliferation and regulation of glucose metabolism.
It has been shown that mild activation of ER stress signaling or specific activation of pro-apoptotic components of ER stress signaling has a beneficial effect on β-cell function and survival . Glucagon-like peptide 1 (GLP-1) is a gut-derived peptide secreted from intestinal L-cells after a meal and has numerous physiological actions, including enhancement of β-cell growth and survival . Interestingly, GLP-1 is a physiological activator of ER stress signaling. Activation of GLP-1 signaling improves β-cell function and survival through the activation of the PERK-ATF4 pathway . We predict that other endogenous factors, like GLP-1 or chemical compounds that can activate ER stress signaling will be discovered and used to increase the viability of β-cells.
Genome-wide association studies have identified numerous single nucleotide polymorphisms (SNPs) associated with increased risk for type-2 diabetes. WFS1 is one of such genes. In the future, human genetics will identify more ER stress-related genes as markers of susceptibility to type 1 and type 2 diabetes. These markers will be potentially useful in targeting patients who would benefit from screening for early detection of diabetes.
To understand the transition from the latent to the overt diabetes state, we need to define the dynamics of β-cell function at the system level. Recent evidence strongly suggests that unresolvable severe ER stress leads to β-cell dysfunction and death during the progression of type1 and type 2 diabetes, as well as genetic forms of diabetes such as Wolfram syndrome. A complete understanding of the states of protein homeostasis regulated by ER stress signaling pathways will have a direct impact on future therapies for diabetes.
This work was supported by grants from NIH-NIDDK (R01DK067493), the Diabetes and Endocrinology Research Center at the University of Massachusetts Medical School (5 P30 DK32520), the Juvenile Diabetes Research Foundation International, and the Worcester Foundation for Biomedical Research.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Papers of particular interest, published within the period or review, have been highlighted as:
* of special interest
** or outstanding interest