The Apoptosis Inhibitor ARC Alleviates the ER Stress Response to Promote β-Cell Survival

Mouse insulinoma MIN6 cells (provided by Dr. Peter Arvan, University of Michigan, Ann Arbor, MI) were cultured as described (16), except for the addition of 140 μmol/L β-mercaptoethanol to the media. βTC-tet cells (provided by Dr. Norman Fleischer, Albert Einstein College of Medicine, Bronx, NY) were used as described (17). MIN6 cells with stable expression of ARC were generated by transduction with a retrovirus encoding human ARC containing a 3′ HA-tag (18). Empty vector was used as the corresponding control. Stable knockdown of ARC in MIN6 cells was carried out using lentiviruses encoding short hairpin (sh)RNAs corresponding to the coding region or 3′ untranslated region of mouse ARC from Sigma-Aldrich (St. Louis, MO). Numbers of the shRNAs in Fig. 4 correspond to the last two digits of The RNAi Consortium (TRC) number. Scrambled shRNA was used as the control. In all experiments, populations of stable transductants were studied. Cells were plated at a density of 5 × 10⁴ cells/cm² and treated with the specified concentrations of thapsigargin, tunicamycin, staurosporine, or palmitate for the specified times. Palmitate was dissolved in culture media that had been supplemented with physiologic levels of fatty acid–free BSA (600 μmol/L) at 60°C overnight, and was filtered before use in treatments.

Isolated mouse islets were collected as previously described (19). Briefly, islets were handpicked after mouse pancreata were digested with collagenase and were separated via gradient centrifugation using Histopaque from Sigma-Aldrich (St. Louis, MO). Islets were placed on matrigel-coated plates and allowed to recover in RPMI media (Invitrogen, Grand Island, NY) supplemented with 5.5 mmol/L glucose and 10% (v/v) FBS for 18 h. After 2-h serum starvation, islets were transduced with adenovirus encoding scrambled shRNA or ARC shRNA (20) at 4 × 10⁶ pfu/islet. Palmitate treatment was initiated 24 h after addition of the adenovirus.

These studies used male 8–12-week-old wild-type C57Bl/6 mice and C/EBP homologous protein knockout (CHOP^−/−) mice (B6.129S-Ddit3^tm1Dron/J), back-bred at least five generations onto a C57Bl/6 background from Jackson Laboratories (Bar Harbor, ME). Mouse maintenance and all experimental procedures were approved by the Albert Einstein College of Medicine Institute for Animal Studies.

Cultured cells and isolated islets were lysed in 50 mmol/L Tris (pH 8.0), 150 mmol/L NaCl, 1% (v/v) Triton X-100, 100 µg/mL phenyl methanesulfonyl fluoride, and 1 µg/mL aprotinin, after which cells were sonicated and isolated islets rotated for 30 min at 4°C. Mouse heart and liver tissue were flash-frozen in liquid nitrogen, lysed in RIPA buffer, and homogenized. Immunoblotting was performed with rabbit polyclonal antisera against ARC (1:4,000, Cayman Chemical, Ann Arbor, MI), active caspase-3 (1:500, Cell Signaling Technology, Danvers, MA), poly(ADP-ribose) polymerase (PARP; 1:1,000, Cell Signaling Technology), caspase-8 (1:1,000, Cell Signaling Technology), eukaryotic initiation factor 2 α-subunit (eIF2α; 1:1,000, Cell Signaling Technology), p-eIF2α (1:1,000, Cell Signaling Technology), and protein kinase RNA-like ER kinase (PERK; 1:1,000, Cell Signaling Technology), GADD 153 (CHOP; 1:100, Santa Cruz Biotechnology, Santa Cruz, CA), and a mouse monoclonal antibody against α-tubulin (1:20,000, Sigma-Aldrich). Primary antibodies were incubated overnight at 4°C in 5% milk and 0.1% (v/v) Tween-20 in PBS. IRDye 800 and 680 secondary antibodies (1:4,000, LI-COR, Lincoln, NE) of the corresponding species were used for detection of the primary antibodies using LI-COR Odyssey and quantified using ImageJ software (National Institutes of Health, Bethesda, MD).

Mouse pancreatic tissue was fixed in 10% neutral-buffered formalin and embedded in paraffin. Tissue was sectioned at 5 μm, deparaffinized, and antigen retrieval performed using Antigen Unmasking Solution (Vector Laboratories, Burlingame, CA). Islets were fixed in 4% paraformaldehyde, embedded in Tissue Tek OCT compound (Electron Microscopy Sciences, Hatfield, PA), flash-frozen on dry ice, and sectioned at 5 μm. Cells were fixed in 4% paraformaldehyde and permeabilized in 0.1% (v/v) Triton X-100. After blocking for 1 h in 10% goat or donkey serum, samples were incubated with primary antibodies diluted in blocking solution. Primary antibodies used were active caspase-3 (1:200, Cell Signaling Technology), ARC (1:400, Cayman Chemical), insulin (1:100, Abcam, Cambridge, MA), glucagon (1:2,000, Abcam), somatostatin (1:50, Santa Cruz Biotechnology), and pancreatic polypeptide (1:200, Millipore, Billerica, MA). For immunofluorescence, primary antibodies were detected using Alexa Fluor 488, 568, and 647 secondary antibodies (1:1,000, Invitrogen, Grand Island, NY) of the corresponding species. Slides were coverslipped using VECTASHIELD HardSet Mounting Medium with DAPI (Vector Laboratories) to counterstain nuclei. For immunohistochemistry, nuclei were counterstained with methyl green as described (21). All images were collected using an Axio Observer .Z1 microscope (Zeiss, Thornwood, NY).

For the live/dead assay, propidium iodide (PI) and Hoechst 33342 (Invitrogen) were directly added to media. Cells were incubated for 5 min in the dark at 37°C and immediately imaged. Three random fields (1,000–2,000 cells) were scored. Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) was performed using the Fluorescein or TMR red In Situ Cell Death Detection Kit (Roche Applied Science, Indianapolis, IN) according to manufacturer’s protocol. Six random fields (1,000–4,000 cells) were scored. When TMR red was used, it was displayed as pseudocolor green. Annexin V staining was performed using the Annexin V-FITC Apoptosis Detection Kit (Abcam) according to manufacturer’s protocol. Three random fields (1,000–3,000 cells) were scored.

RNA was isolated from cells using TRIzol reagent (Ambion, Grand Island, NY). Oligo(dT)20 primers were used to produce cDNA with SuperScript III First-Strand Synthesis System (Invitrogen). RT-PCR was performed using the following primer pairs: mouse XBP1 sense 5′-GAA CCA GGA GTT AAG AAC ACG-3′ and mouse XBP1 antisense 5′-AGG CAA CAG TGT CAG AGT CC-3′. Samples were then separated on a 3% (w/v) agarose gel.

Data are presented as mean ± SEM. The two-tailed Student t test was used to compare two groups. ANOVA, followed by a Tukey post hoc test, was used for multiple comparisons. GraphPad Prism 5 software (GraphPad, La Jolla, CA) was used to calculate statistics. P < 0.05 was considered significant.

The expression of ARC was originally reported to be restricted to cardiac and skeletal myocytes and some neurons (12,13), and more recent studies have uncovered that ARC is also highly induced in multiple cancers (18,22–25). Because tissues contain a heterogeneous mixture of cells, some of which perform unique functions, we hypothesized that ARC expression may have been overlooked in studies that examined homogenates of whole organs. Accordingly, we screened for ARC expression in subpopulations of cells thought not to express ARC. Unexpectedly, we found that cells of the mouse (Fig. 1A) and human (Fig. 2D) endocrine pancreas contain high levels of ARC. Moreover, ARC abundance in isolated islets is roughly equivalent to that in heart tissue (Fig. 1B). In contrast, ARC was not detectable in the exocrine pancreas (Fig. 1A), explaining previous studies reporting its absence in the pancreas (13). ARC levels in isolated islets from wild-type mice were similar to those from ob/ob mice (Supplementary Fig. 1D). These data demonstrate that ARC protein is abundant in cells of the islet.

To determine which cells within islets express ARC, we immunostained mouse and human pancreatic tissue. ARC is expressed primarily in β-cells and is present in >99.5% of these cells in the mouse (Fig. 2A and B). It is predominantly localized to the cytoplasm because ARC immunofluorescence overlaps with insulin but not with the nuclear DAPI staining (Fig. 2A and C). Abundant ARC was also detected in various mouse β-cell lines (Fig. 1C). In addition, ARC is found in 30% of mouse α-cells, 16% of pancreatic polypeptide (PP) cells, and in 67% of δ-cells, the latter sharing a common progenitor cell with β-cells (26) (Fig. 2B). Moreover, in the human, ARC is present in 73% of β-cells and in 64% of α-cells (Fig. 2E). These data indicate that ARC is abundant in the endocrine pancreas, predominantly in β-cells.

To determine whether ARC plays a functional role in β-cells, we used gain- and loss-of-function approaches. Stable retroviral transduction of ARC in MIN6 cells (Fig. 3A) inhibited cell death triggered by the general apoptosis inducer staurosporine. This was indicated by changes in the abundance of active caspases-3 and -8, cleavage of the caspase substrate PARP, and loss of plasma membrane integrity (PI staining), the latter a readout of overall death in cultured cells (Fig. 3B–C, F). Because of metabolic disturbances, Ca²⁺ abnormalities, and proteotoxic stress resulting from increased insulin production, ER stress is a key component of type 2 diabetes (27–31) and an initiator of β-cell death (32,33). Accordingly, we tested whether ARC inhibits ER stress-induced apoptosis in these cells. Cell death triggered by the ER stressors thapsigargin (an inhibitor of sarcoplasmic/endoplasmic reticulum Ca²⁺ATPase [SERCA], which mediates Ca²⁺ reuptake into the ER) and tunicamycin (an inhibitor of N-linked glycosylation, which is required for proper folding of proteins) was suppressed by ARC overexpression (Fig. 3D, E, G, and H). Of note, levels of endogenous ARC were not affected by any of these death inducers (Supplementary Fig. 1A–C).

Conversely, knockdown of ARC was carried out using RNAi. Five independent shRNAs were tested for their abilities to knockdown ARC after stable transduction into MIN6 cells. Compared with scrambled shRNA, ARC shRNAs 14 and 15 were most effective, with 55% and 70% knockdown, respectively (Fig. 4A). To define the importance of endogenous levels of ARC in regulating ER stress-mediated cell death, we used shRNA 15. Depletion of ARC resulted in two- to fourfold increases in cell death induced by thapsigargin (Fig. 4). This was demonstrated by activation of caspases-3 and -8 and PARP cleavage (Fig. 4C and D). In addition, there were increases in the percentage of cells exhibiting TUNEL, active caspase-3 immunofluorescence, annexin V, and PI entry (Fig. 4B and E–G). These data demonstrate that endogenous levels of ARC protect β-cells against ER stress-induced cell death.

ER stress is a critical component in the pathogenesis of type 2 diabetes (27,28). The free fatty acid palmitate, which models this stress, has been shown to induce β-cell death (34,35). Accordingly, we assessed whether ARC inhibits cell death in response to this pathophysiological stimulus. Palmitate at 1.2 and 2.4 mmol/L concentrations induced β-cell apoptosis, which was dramatically reduced by ARC overexpression (Fig. 5A and B, Supplementary Fig. 2). In addition, ARC overexpression continued to prevent cell death even at late time points (Supplementary Fig. 3). Conversely, knockdown of ARC potentiated palmitate-induced apoptosis in a concentration- and time-dependent manner (Fig. 5C–E). These data indicate that ARC regulates β-cell death elicited by a pathophysiologically relevant stimulus of ER stress.

ER stress has been linked to β-cell death (32,33), but the precise functional and mechanistic relationships between these processes remain unclear. The major role of the ER stress response is to restore ER homeostasis, including refolding of misfolded proteins (reviewed in 36–38). Upstream mediators include the kinases PERK and inositol-requiring protein 1α (IRE1α), the combined actions of which include attenuating translation and activating a complex transcriptional response that promotes protein refolding. However, when the ER stress response is prolonged, apoptosis can result (reviewed in 39,40). Although inhibition of cell death by ARC has traditionally been thought to be mediated solely through its antagonism of the central death machinery (14,41), we considered the possibility that ARC also regulates the ER stress response itself.

Using thapsigargin to stimulate the ER stress response, we observed that overexpression of ARC does not affect PERK activation (as indicated by phosphorylation of its downstream target eIF2α) or IRE1α activation, as indicated by XBP1 splicing (Fig. 6A and B). In contrast, ARC overexpression inhibited upregulation of the transcription factor CHOP, an important initiator of apoptosis in response to ER stress that has been implicated in type 2 diabetes (42,43) (Fig. 6C). This effect was even more marked when CHOP induction was elicited by the physiological ER stressor palmitate (Fig. 6D). These data indicate that ARC acts distal to activation of PERK and IRE1α to inhibit the ER stress response.

To determine whether endogenous levels of ARC are important in suppressing palmitate-induced apoptosis in the context of islet tissue, we used isolated islets. Transduction of islets with adenovirus encoding ARC shRNA knocked down ARC levels in ∼74% of cells (Fig. 7A and B). Although treatment with palmitate (1.2 mmol/L) did not result in significant cell death in islets transduced with scrambled shRNA, knockdown of ARC augmented apoptosis more than fivefold (Fig. 7A and C). Moreover, deletion of CHOP rescued the incremental cell death resulting from ARC knockdown (Fig. 7D and F, Supplementary Fig. 4) without influencing ARC levels (Fig. 7D and E, Supplementary Fig. 4). These data indicate that endogenous levels of ARC are critical in suppressing palmitate-induced cell death in islets and that this inhibition is mediated through CHOP.

These experiments reveal the unexpected presence of ARC in pancreatic islets and a novel role for ARC in modulating the ER stress response and inhibiting β-cell death. Under normal conditions, the expression of ARC was previously thought to be restricted to cardiac and skeletal myocytes and neurons. Accordingly, ARC has never been studied in the context of diabetes. Surprisingly, we found that ARC is as abundant in the endocrine pancreas as it is in the heart. Although expressed in variable percentages of cell types in islets, almost all mouse β-cells and a high proportion of human β-cells contain ARC. On the basis of studies in other cell types (21,41,44,45), the abundance of ARC in the β-cell may be mediated by both transcriptional and posttranslational mechanisms.

Although ARC has been shown to be a potent inhibitor of apoptosis in other systems, cell death is often regulated in a cell type- and stimulus-specific manner; therefore, we investigated the role of ARC in β-cell apoptosis. Overexpression of ARC inhibited cell death elicited by generic cell death stimuli and ER stressors, including palmitate, a free fatty acid relevant to the pathogenesis of type 2 diabetes. Conversely, knockdown experiments demonstrated the importance of endogenous levels of ARC in protecting against these death signals in the β-cell. These effects were demonstrated using multiple parameters indicative of apoptosis assessed at a variety of time points.

In addition to its known effects on the two central apoptosis pathways, a new finding in this study is that ARC modulates cell death by alleviating the ER stress response. This pathway plays an important adaptive role in helping the cell to resolve various stresses. When insults are of overwhelming magnitude or prolonged duration, however, this pathway is no longer sufficient to compensate, sometimes triggering cell death. This duality of the ER stress response is attributable, in part, to effectors such as CHOP, which are capable of signaling cellular adaptation and also demise. Although the mechanisms that control the transition from adaptation to death are currently not understood, our data indicate that ARC acts downstream of PERK and IRE1α and requires CHOP to ameliorate the ER stress response and inhibit β-cell death.

We observed that ER stress activates caspase-8 in MIN6 cells, an effect suppressed by ARC overexpression and exacerbated by ARC knockdown. These observations are consistent with a previous study in melanoma cells (46). It remains unclear, however, whether caspase-8 is modulated through known actions of ARC at the DISC or whether these effects of ARC involve a pool of caspase-8 known to reside at the ER (47). Interestingly, β-cell–specific deletion of procaspase-8 in mice ameliorates high fat diet–induced diabetes (9), a significant component of which involves ER stress.

This study raises a number of interesting avenues for future investigation. One relates to the mechanisms by which ARC suppresses CHOP. This potentially includes effects on CHOP transcription, mRNA or protein stability, posttranslational modifications, and nuclear localization. Moreover, indirect mechanisms may also be involved. Second, given known effects of ARC in regulating mitochondrial events in cell death, another important question is whether bidirectional ER–mitochondrial crosstalk plays a role in effects of ARC on the ER stress response. Third, and perhaps most interesting, is the possibility that ARC functions as a survival factor for β-cells exposed to the noxious milieu of type 2 diabetes in vivo. Experiments using ARC gain- and loss-of-function mice will be needed to test this possibility. In addition, because ARC is also expressed in skeletal muscle and brain, there remains the possibility that ARC modulates other aspects in the pathogenesis of type 2 diabetes.

This work was supported by grants from the National Institutes of Health (5R01-HL-060665-14 to R.N.K., 5R01-DK-063306-10 to S.C.C., 5R37-DK-033823-31 to J.E.P., 5P60-DK-020541-34 to J.E.P., R.N.K., S.C.C., and L.C., 5P30-CA-013330-39 to R.N.K., and 5T32-GM-007491-35 to W.M.M^c.K.) and from the American Heart Association (AHA 09GRANT2290168 to M.T.C.). R.N.K. is supported by the Dr. Gerald and Myra Dorris Chair in Cardiovascular Disease.

No potential conflicts of interest relative to this article were reported.

W.M.M^c.K. conceived experiments, researched data, analyzed experiments, contributed to discussion, and wrote the manuscript. J.W. researched data and contributed to discussion. L.C. and M.Z. researched data. M.C.T. contributed reagents and reviewed and edited the manuscript. J.E.P. and S.C.C. conceived experiments, contributed to discussion, and reviewed and edited the manuscript. R.N.K. conceived and analyzed experiments, contributed to discussion, and wrote the manuscript. R.N.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Parts of this study were presented in abstract form at the Islet Biology Keystone Meeting, Whistler, BC, Canada, 12–17 April 2010, and at the Gordon Research Conference: Stress Proteins in Growth, Development & Disease, Lucca (Barga), Italy, 17–22 July 2011.

The authors are most grateful to the Wilf Family for their ongoing generosity and support. The authors would also like to acknowledge Dr. Christina Medina-Ramirez and Dr. Melissa James (Albert Einstein College of Medicine, Bronx, NY) for their help in generating the ARC and ARC shRNA viruses.