Global profiling of genes modified by endoplasmic reticulum stress in pancreatic beta cells reveals the early degradation of insulin mRNAs
Pancreatic beta cells respond to endoplasmic reticulum (ER) stress by activating the unfolded protein response. If the stress is prolonged, or the adaptive response fails, apoptosis is triggered. We used a ‘homemade’ microarray specifically designed for the study of beta cell apoptosis (the APOCHIP) to uncover mechanisms regulating beta cell responses to ER stress.
Materials and methods
A time course viability and microarray analysis was performed in insulin-producing INS-1E cells exposed to the reversible ER stress inducer cyclopiazonic acid (CPA). Modification of selected genes was confirmed by real-time RT-PCR, and the observed inhibition of expression of the insulin-1 (Ins1) and insulin-2 (Ins2) genes was further characterised in primary beta cells exposed to a diverse range of agents that induce ER stress.
CPA-induced ER stress modified the expression of 183 genes at one or more of the time points studied. The expression of most of these genes returned to control levels after a 3 h recovery period following CPA removal, with all cells surviving. Two groups of genes were particularly affected by CPA, namely, those related to cellular responses to ER stress, which were mostly upregulated, and those related to differentiated beta cell functions, which were downregulated. Levels of Ins1 and Ins2 mRNAs were severely decreased in response to CPA treatment as a result of degradation, and there was a concomitant increase in the level of IRE1 activation.
In this study we provide the first global analysis of beta cell molecular responses to a severe ER stress, and identify the early degradation of mRNA transcripts of the insulin genes as an important component of this response.
KeywordsApoptosis CPA Cyclopiazonic acid Diabetes mellitus Endoplasmic reticulum stress Insulin Interleukin-1 Pancreatic beta cells
aminoallyl antisense cRNA
activating transcription factor
immunoglobulin heavy chain binding protein
C/EBP homologous protein
eukaryotic translation initiation factor 2α
ER-associated degradation pathway
inducible nitric oxide synthase
- INS 1/2
inositol-requiring ER-to-nucleus signal kinase 1
oxygen-regulated protein 150
- PCSK 1/2
proprotein convertase subtilisin/ kexin type 1/2
protein kinase RNA-dependent-like ER kinase
sarcoendoplasmic reticulum Ca2+-ATPase
unfolded protein response
X-box binding protein 1
Pancreatic beta cells function as nutrient sensors, responding to increases in circulating nutrients by synthesising and releasing insulin. During functional stimulation by glucose, insulin synthesis represents nearly 50% of the total protein biosynthesis by beta cells . This places an enormous burden on the endoplasmic reticulum (ER), the organelle responsible for the synthesis and proper folding of secretory proteins , rendering beta cells vulnerable to agents that perturb ER function . Disturbances in normal ER function lead to the accumulation of misfolded proteins and activate the unfolded protein response (UPR) . The UPR is mediated through three ER membrane proteins, namely, protein kinase RNA-dependent-like ER kinase (PERK; also known as eukaryotic translation initiation factor 2-alpha kinase 3, EIF2AK3), activating transcription factor 6 (ATF6) and inositol-requiring ER-to-nucleus kinase 1 (IRE1; also known as ER to nucleus signalling 1, ERN1), the latter responsible for Xbp1 mRNA splicing and activation of the protein. The UPR decreases ER protein levels and restores ER function by (1) attenuating translation via activation of PERK and subsequent phosphorylation of the eukaryotic translation initiation factor 2α (eIF2α); (2) activating the transcription factors ATF6 and XBP1, which upregulate the expression of genes encoding ER chaperones and thus increase the folding capacity of the ER; and (3) degrading misfolded proteins via the ER-associated degradation pathway . A fourth pathway has been recently described in Drosophila, namely, the degradation of mRNAs encoding ER-targeted proteins by IRE1, thus halting the production of new proteins in the organelle . Failure of the UPR to solve ER stress leads to activation of the apoptosis pathway . The mechanisms linking ER stress and apoptosis remain to be clarified, but may involve activation of the transcription factors C/EBP homologous protein (CHOP; also known as DNA-damage-inducible transcript 3, Ddit3) and ATF3 and the kinase JNK [4, 6, 7, 8, 9].
Homozygous mutations in the gene encoding PERK cause beta cell death and diabetes in Wolcott–Rallison syndrome in humans  and in a transgenic mouse model . Even a partial reduction in eIF2α phosphorylation, when coupled with a high-fat diet, triggers beta cell failure and diabetes . Prolonged inhibition of eIF2α dephosphorylation also triggers beta cell death , suggesting that an effective modulation of ER function is necessary for beta cell viability. In line with this, we and others have recently shown that NEFA and cytokines, agents that may contribute to beta cell death in type 2 and type 1 diabetes, respectively, cause beta cell apoptosis, at least in part, by triggering ER stress [13, 14, 15, 16, 17, 18]. Moreover, beta cells are particularly sensitive to agents that deplete ER Ca2+, such as the sarcoendoplasmic reticulum Ca2+-ATPase (SERCA) blockers cyclopiazonic acid (CPA) and thapsigargin [15, 18, 19]. We have recently shown that exposure of INS-1E cells to CPA for 6–12 h triggers severe ER stress, characterised by increased Chop, Bip (also known as heat-shock 70 kDa protein 5, Hspa5) and Xbp1 spliced mRNA expression and activation of a UPR-luciferase reporter construct .
Against this background, we have combined high-throughput and conventional experimental approaches to reveal the mechanisms regulating beta cell responses to ER stress. For this purpose, we used an ‘in-house’ spotted rat oligonucleotide microarray, the APOCHIP, containing 60-mer probes for nearly 600 genes selected for the study of beta cell apoptosis . The APOCHIP provides reproducible data, and compares well with the Affymetrix GenChip . Time-dependent profiles of the genes available in the APOCHIP were measured in INS-1E cells exposed to the reversible SERCA blocker CPA to create an ER stress compendium of several thousand measures. Elucidation of the global pattern of beta cell gene expression in response to ER stress involved statistical analysis of this large data set and confirmation by real-time RT-PCR of some genes. The results of this experiment indicated that ER stress causes an early and marked decrease in Ins1/2 mRNA expression, and this was independently tested in insulin-producing cells and primary beta cells exposed to CPA, thapsigargin or cytokines (IFN-γ+IL-1β).
Materials and methods
Cell culture and treatments
INS-1E cells were cultured in RPMI 1640 medium (Invitrogen, Paisley, Scotland) , while primary rat beta cells were FACS purified and cultured in Ham’s F-10 medium (Invitrogen) . Cells were plated and cultured for 48 h prior to the addition of test agents. The SERCA blockers—CPA and thapsigargin—were solubilised in DMSO and were used at concentrations of 25 μmol/l and 1 μmol/l, respectively. DMSO (0.125%) was used as the control condition in the experiments involving SERCA blockers. This concentration of DMSO does not affect gene expression or survival in INS-1E or primary beta cells ([15, 18]; data not shown). Recombinant rat IFN-γ (R&D Systems, Abingdon, UK) was used at a concentration of 0.036 μg/ml (±100 U/ml), and human recombinant IL-1β (a gift from C.W. Reinolds, National Cancer Institute, Bethesda, MD, USA) was used at a concentration of 10 U/ml for the INS-1E cell experiments and 50 U/ml for the primary beta cell experiments. The inducible NO synthase (iNOS) blocker NG-methyl-l-arginine (LMA; Sigma, Steinheim, Germany) was used at 1.0 mmol/l. CPA, thapsigargin, cytokines and LMA concentrations were selected based on our previous time course and dose–response studies [15, 18, 22].
Assessment of INS-1E cells viability
INS-1E cells were exposed to CPA or DMSO (control) for the time periods indicated. CPA was then removed by changing the medium, and the cells were subsequently incubated for a further 24 h before assessment of viability. For comparison, cells in some experiments were also treated with CPA for the last 9 and 12 h of the experiment. The percentage of viable, apoptotic and necrotic cells was determined as previously described [15, 18].
Preparation of the target for array analysis
INS-1E cells were exposed to CPA or DMSO (control) for 2, 6 or 12 h, or for 6 h followed by a 3 h recovery period in the absence of CPA. Cells were then collected, and total RNA was isolated using the RNeasy Mini Kit (Qiagen, Venlo, The Netherlands). Double-stranded cDNA was prepared and was used as template for aminoallyl antisense cRNA (aa-cRNA) synthesis . The aa-cRNA was purified and coupled to a fluorophore, either indocarbocyanine (Cy3) or indodicarbocyanine (Cy5) (Amersham, Diegem, Belgium). The labelled aa-cRNA was fragmented, precipitated and resuspended in RNAse-free water for spectrometric quantification .
Hybridisation and data analysis
An aliquot (1 μg) of control target was mixed with 1 μg of CPA target labelled with the other fluorophore; all hybridisations were replicated with dyes swapped . Microarrays were scanned with a GenePix 4000B scanner (Axon Instruments, Union City, CA, USA). Spot and surrounding local background intensities were quantified with GenePix Pro 5.0 (Axon Instruments) using the irregular spot feature. The gene expression log2-ratios were calculated from the spot intensities (with background intensity subtracted) and used for normalisation. Housekeeping gene-based Lowess normalisation of the microarray data was carried out using the tRMA package  written in R code (http://cran.r-project.org, last accessed in January 2007). The normalised log2-ratios of each dye swap were averaged. The means for the three independent experiments were then used for a Student’s t test. To address the problem of multiple hypothesis testing, a method controlling for a false discovery rate (FDR) was applied [24, 25] using the Q-value package (http://faculty.washington.edu/~jstorey/qvalue/, last accessed in January 2007). Genes were considered as significantly modified by CPA if they had a q value equal or less than 0.06 (FDR of 6% among the significant genes) for at least one of the four time points studied.
Real-time RT-PCR, RNA stability and Ins1 promoter activity
For assessment of Ins1 promoter activity, INS-1E cells were transfected using lipofectamine (Invitrogen) with 250 ng of the rat Ins1 promoter-luciferase construct (a kind gift from G. Holz, New York University School of Medicine, New York, NY, USA) , and the pRL-CMV plasmid (50 ng; with Renilla luciferase used as an internal control for transfection efficiency). Twenty-four hours after transfection the cells were exposed to CPA or DMSO (control) for 6 or 12 h. Luciferase activities in the cell lysates were determined and expressed as firefly/Renilla (relative luciferase activity).
Data are shown as means±SEM, and comparisons between groups were made by the paired Student’s t test or by ANOVA followed by the Student’s t test with the Bonferroni correction, as indicated. A p value of ≤0.05 was considered statistically significant.
Viability of INS-1E cells exposed to a CPA-induced ER stress
In the presence of CPA, the percentage of apoptotic INS-1E cells had increased by the 4 h time point, and increased with time to study end, with 17% of cells apoptotic by the 12 h time point (Fig. 1). There was no significant increase in the proportion of necrotic cells (data not shown), confirming our previous observations that SERCA blockers kill beta cells by apoptosis [15, 18]. When beta cells were exposed to CPA for 9 h, and then washed and cultured for an additional 24 h without the SERCA blocker, the percentage of apoptotic cells was similar to that observed in beta cells exposed to CPA alone for 9 h (12.5 ± 1.3% vs 12.7 ± 1.2%, respectively; means±SEM, n = 4). Similar results were observed after exposure to CPA for 12 h, with or without a further 24 h incubation in medium alone (data not shown). This indicates that the effects of CPA are reversible, and suggests that beta cells trigger defence mechanisms that allow them to survive and recover once ER stress is removed. Based on these and previous experiments , we selected periods of 2, 6 and 12 h exposure to CPA for the subsequent array analysis. An additional group of cells exposed to CPA for 6 h, followed by a 3 h recovery period, was included to allow the identification of genes expressed in the early phase of ER stress recovery.
Identification of ER stress-modified genes by microarray analysis
Confirmation by real-time RT-PCR of CPA-induced modifications in mRNA expression of a selected set of changed genes from the microarray experiment
2.5 ± 0.2**
5.7 ± 0.9**
2.6 ± 0.2*
7.6 ± 0.6***
2.6 ± 0.3*
3.3 ± 0.5**
3.1 ± 0.3**
6.2 ± 1**
0.9 ± 0.02*
1 ± 0.1
0.8 ± 0.03*
0.9 ± 0.1
1.3 ± 0.08*
2.6 ± 0.4*
1.5 ± 0.1*
4.4 ± 0.4**
1.5 ± 0.3
2.8 ± 0.6*
2.1 ± 0.3*
4.9 ± 1.1*
1.3 ± 0.05*
3.6 ± 0.5**
1.6 ± 0.1**
4.1 ± 0.4**
Another group of genes modified by ER stress were those involved in beta cell differentiated functions (ESM Table 1). Thus, we observed a decrease in the expression of Ins2, Pcsk1, Pcsk2, Rab3a, Pdx-1, Isl-1, HNF-1α and HNF-3α. Inhibition was most marked for Ins2, the expression of which was decreased by 60–80% after 6–12 h (ESM Table 1). We also observed a 50% decrease in levels of Ins1 transcript after 6–12 h (data not shown); however, this gene failed to pass our statistical test.
ER stress induces a marked decrease in insulin mRNA expression in INS-1E cells and primary beta cells
In the present study we used the APOCHIP, a rat oligonucleotide microarray , to determine the molecular mechanisms regulating beta cell responses to ER stress. Two groups of genes were particularly sensitive to CPA: those involved in the response to ER stress, which were mostly upregulated, and those related to differentiated beta cell functions, which were mostly downregulated.
The genes encoding the ER chaperones Bip, calreticulin and Grp94, the ER stress-related transcription factors ATF4, ATF3 and Chop, the pro-apoptotic caspase 12 and the ER-protein translocator Sec61 were all upregulated by CPA, while the gene for cyclin D1 was inhibited during ER stress as a mechanism to prevent cell division . Bip, Grp94 and calreticulin—three of the most abundant chaperones in the ER—are involved in the correct folding of nascent protein. They also carry out the following important functions in this organelle: (1) Bip and calreticulin are involved in the ER ‘quality control’ system, which leads to the destruction of malfolded proteins via the ER-associated degradation pathway (ERAD) via the proteasome [31, 32]; (2) Grp94 and calreticulin have a high calcium binding capacity, and are therefore involved in ER Ca2+ homeostasis [32, 33]; (3) Bip also plays a key role in the regulation of the three UPR transducer proteins ATF6, PERK and IRE1 via a binding-release mechanism . Sec61 is involved in the formation of channels across the ER membrane, permitting the evacuation of malfolded protein from the ER lumen to proteasomes in the cytosol . The transcription factor ATF4 regulates a set of protective genes related to amino acid import/metabolism and resistance to oxidative stress , and is also involved in the regulation of the pro-apoptotic protein Chop [36, 37]. ATF3, Chop and caspase 12 are three potential mediators of ER stress-triggered beta cell apoptosis, as suggested by (1) partial protection against cell death in Chop knockout islets exposed to NO  and in ATF3-deficient islets exposed to IFN-γ plus IL-1β ; (2) Delayed diabetes onset in Akita mice with an homozygous deletion of Chop ; (3) decreased expression of Bcl2 in Chop-overexpressing cells , an effect suggested in the present study by the inverse relationship between Chop and Bcl2 expression (ESM Table 1) (4) resistance of renal tubular epithelium and neurons from caspase 12 null mutant mice to ER stress-induced apoptosis ; (5) diabetic phenotype secondary to beta cell loss in mice overexpressing ATF3 .
In the present study, expression of ATF3 and Chop returned to basal levels after the 3 h recovery period, while expression of the Bip and Grp94 remained elevated. This pattern of gene expression may explain how beta cells endure 9–12 h of severe ER stress without reaching the ‘point of no return’ for cell death; indeed, as soon as CPA was removed, the induction of apoptosis was arrested in the surviving cells.
CPA induced a marked decrease in the expression of several genes involved in the maintenance of a beta cell differentiated phenotype. There was a decrease in the expression of Ins2, Pcsk1/2 (whose protein products, the prohormone convertases 1 and 2, are responsible for the conversion of proinsulin to insulin ), Rab3a (which encodes Ras-related small GTP binding protein 3A, involved in insulin exocytosis ) and Pdx-1, Isl-1, HNF-1α and HNF-3α (which encode transcription factors involved in the maintenance of a differentiated beta cell phenotype ). Beta cell exposure to the cytokines IFN-γ and IL-1β, which induce ER stress , results in the inhibition of expression of Pcsk1/2, impairing the conversion of pro-insulin into insulin [43, 44], and the decreased expression of Pdx-1 and Isl-1 . These cytokines also decrease the expression of SNAP-25, VAMP-2 and Rab3a, whose protein products are involved in granule fusion with the beta cell membrane [22, 45], preferentially decreasing the first phase of exocytosis . The present data suggest that these effects of cytokines are, at least in part, secondary to ER stress.
The most marked effect of CPA was the decrease in the expression of Ins1 and Ins2 mRNAs. Similar findings were observed in INS-1E cells exposed to thapsigargin or cytokines (IFN-γ plus IL-1β) and in primary beta cells exposed to the same agents. Cytokines inhibit Ins1/2 mRNA expression and pro-insulin biosynthesis in rat and mouse islets [46, 47, 48]. The decrease in Ins1/2 levels is secondary to NO production , a finding confirmed in the present study. Cytokine-induced ER stress is prevented by iNOS blockers , suggesting that the inhibitory effects of cytokines on Ins1/2 mRNA expression are, at least partly, secondary to ER stress.
It has been recently shown that ER stress causes the rapid degradation of mRNAs targeted for translation at the ER in Drosophila cells. This degradation is mediated by IRE1 activation, and complements other UPR mechanisms by selectively halting the production of non-vital proteins at the ER . The two insulin gene transcripts are the most abundant mRNAs directed to the ER of rat beta cells, and a rapid decrease in Ins1/2 mRNA expression could represent a novel mechanism of beta cell adaptation to ER stress. Against this background, we evaluated whether CPA-induced ER stress activates the IRE1 pathway in INS-1E cells, as measured by Xbp1 splicing, and whether the observed decrease in Ins1/2 mRNA is due to mRNA degradation. Our results indicated that this was indeed the case. Thus, CPA induced early and intense Xbp1 splicing, which was paralleled by the early degradation of both Ins1 and Ins2 without affecting the activity of the Ins1 promoter.
Acute ‘physiological’ and prolonged ‘pathological’ exposure to high glucose triggers several markers of the UPR; however, while short-term (1–3 h) exposure to high glucose favours insulin biosynthesis, long-term (3–7 days) exposure decreases Ins1/2 mRNA expression . Phosphorylation of IRE1α is observed under both conditions, but Xbp1 splicing is only detectable after a long term exposure to high glucose . Xbp1 splicing therefore seems to be a marker that differentiates between ‘physiological’ and ‘pathological’ activation of the UPR. In this study, CPA exposure clearly induced Xbp1 splicing (Fig. 6). In addition, we have recently completed another study on the regulation of Chop mRNA expression by diverse ER stresses in insulin-producing cells , where the effects of CPA, at the concentration used in the present study, induced ER stress similar to that induced by cytokines and by the NEFA palmitate. Moreover, it was recently described that ER stress markers are present in islets obtained from diabetic db/db mice and patients with type 2 diabetes . These observations suggest that the present data reflects a ‘pathological’ beta cell situation of potential relevance to the diabetic state.
The present study reports the first global analysis of beta cell molecular responses to severe ER stress, and identifies the early degradation of Ins1 and Ins2 mRNAs as an important component of this response. We suggest that massive degradation of Ins1/2 mRNA, the most prevalent ER-targeted mRNA in beta cells, alleviates functional demand on the ER. This, together with a vigorous upregulation of ER chaperones, contributes to beta cell survival and recovery once the source of ER stress is removed.
This work was supported by grants from the European Union (STREP SAVEBETA, contract no. 036903 in the Framework Programme 6 of the European Community), the Fonds National de la Recherche Scientifique (FNRS; National Fund for Scientific Research; Belgium) and the ‘Communauté française de Belgique-Actions de Recherche Concertées’ (ARC; French Community of Belgium-Concerted Research Actions) to D. L. Eizirik, and a JDRF Transition Award Grant to A. K. Cardozo. We gratefully acknowledge G. Høj for advice on the microarray procedure and G. Pirot for developing the handmade tools to dry the arrays. We also thank M. A. Neef, G. Vandenbroeck and R. Leeman for excellent technical support. This paper is dedicated to the memory of Professor Claes Hellerström, who made many crucial contributions to our understanding of pancreatic beta cell function and dysfunction.
Duality of interest
The authors declare that they have no duality of interest in connection with this study.
- 37.Pirot P, Ortis F, Cnop M et al (2007) Transcriptional regulation of the endoplasmic reticulum (ER) stress gene Chop in pancreatic insulin producing cells. Diabetes DOI: 10.2337/db06–1253
- 50.Laybutt DR, Preston AM, Akerfeldt MC et al (2007) Endoplasmic reticulum stress contributes to beta cell apoptosis in type 2 diabetes. Diabetologia. DOI 10.1007/s00125-006-0590-z