Apoptosis Repressor With Caspase Recruitment Domain Ameliorates Amyloid-Induced β-Cell Apoptosis and JNK Pathway Activation Free

In type 2 diabetes, a number of factors, including hyperglycemia, elevated free fatty acids, and islet amyloid, contribute to the dysfunction and death of islet β-cells, and this in turn contributes to insufficient insulin secretion (1–3). Islet amyloid is found in the vast majority of individuals with type 2 diabetes and contains as its unique peptide component human islet amyloid polypeptide (hIAPP) (4,5), which is cosecreted with insulin by the pancreatic β-cell (6). Islet amyloid formation is linked to β-cell apoptosis and loss in humans (7), and this process has also been observed in transgenic animal models of islet amyloidosis (8–10).

To gain insight into the mechanisms of hIAPP toxicity, in vitro studies have used two general approaches: 1) application of synthetic hIAPP to β-cell lines or islets, and 2) transgenic expression of hIAPP in the β-cell. Both of these approaches have been used to show hIAPP aggregation activates c-Jun N-terminal kinase (JNK) pathway signaling in the β-cell (11–13). JNK pathway activation is commonly characterized by increased phosphorylation of JNK and its downstream target c-Jun as well as increased c-Jun target gene expression (14–16). Notably, inhibition of JNK signaling almost completely abrogates β-cell apoptosis both when hIAPP is applied exogenously (11) or when hIAPP is expressed endogenously (13), indicating the central importance of this pathway in amyloid-induced β-cell toxicity.

Apoptosis repressor with caspase recruitment domain (ARC) is an inhibitor of apoptosis that has been shown to reduce cell death in cardiac and skeletal muscle (17,18), neurons (19,20), hepatocytes (21,22), and more recently, in islet β-cells (23). In these tissues, ARC antagonizes several proapoptotic signaling pathways implicated in amyloid-induced cell death (24,25), including the JNK pathway (21,22). Physical interaction between ARC’s caspase recruitment domain (CARD) and proapoptotic molecules has been shown to govern its antiapoptotic effects (24,26).

Because studies have recently demonstrated that ARC is expressed in islet β-cells and that many of the molecules it can modulate are involved in islet amyloid-induced toxicity (19,24,25), we hypothesized that ARC may decrease amyloid-induced β-cell apoptosis and that it may do so by suppressing the JNK signaling pathway. In this study, we used islets from transgenic mice that express hIAPP in their β-cells to examine these hypotheses and demonstrate for the first time the potential for ARC to ameliorate amyloid-induced β-cell apoptosis and loss.

This study was approved by the Veterans Affairs Puget Sound Health Care System Institutional Animal Care and Use Committee.

This study used hemizygous hIAPP transgenic mice (27) on an F1 C57BL/6 × DBA/2J background, with nontransgenic littermates as controls. PCR was used to determine transgene presence, as previously described (28).

Islets were isolated from the pancreata of 10- to 12-week-old male and female mice by collagenase digestion and cultured overnight in RPMI-1640 medium containing 11.1 mmol/L glucose, 10% FBS, 100 units/mL penicillin, and 100 μg/mL streptomycin at 37°C in a humidified 5% CO₂/95% air incubator to recover from the isolation procedure (29). After overnight recovery, hIAPP transgenic and nontransgenic control islets were cultured for 144 h in media containing 16.7 mmol/L glucose (amyloid-inducing conditions in hIAPP transgenic islets). Media were changed every 48 h.

For studies involving adenoviral transduction of islets, islets were incubated immediately after isolation with adenovirus particles in RPMI-1640 medium containing 11.1 mmol/L glucose, 0.2% BSA, 100 units/mL penicillin, and 100 μg/mL streptomycin for 20 h. Thereafter, islets were cultured for 144 h in media as described above.

To overexpress ARC, islets were transduced with 4.0 × 10⁶ plaque-forming units/mL (multiplicity of infection = 100) adenovirus encoding human ARC (AdV-hARC) (30) or green fluorescent protein (Adv-GFP; a gift from Dr. Christopher Rhodes, Chicago, IL). Adv-hARC and Adv-GFP contained the same vector backbone (Ad5CMV). To decrease ARC expression, islets were transduced with 4.0 × 10⁷ plaque-forming units/mL (multiplicity of infection = 1,000) AdV encoding ARC short hairpin (sh)RNA (AdV-shARC) or scrambled shRNA control (AdV-shCTL) (31).

Islets were fixed in 10% (wt/vol) neutral-buffered formalin, embedded in agar and then paraffin, and sectioned at 10-μm intervals. Sections were stained with thioflavin S (0.5% vol/vol in water) or Congo Red (1% wt/vol in water) to visualize amyloid deposits, incubated with mouse monoclonal anti-insulin antibody (1:5,000; #I2018; Sigma-Aldrich, St. Louis, MO), followed by goat anti-mouse Cy3 (1:250; #115-165-146; Jackson ImmunoResearch, West Grove, PA) to visualize β-cell area, and Hoechst 33258 to visualize nuclei. Islet area, β-cell area, and thioflavin S– or Congo Red–positive areas were determined using Image-Pro Plus software (Media Cybernetics, Bethesda, MD) (32).

To determine the rate of β-cell apoptosis, propidium iodide (PI) and insulin costaining was performed to visualize condensed, fragmented apoptotic nuclei and β-cells, respectively. Sections were first incubated with insulin antibody (1:5,000), then goat anti-mouse Alexa 350 (1:100; #A-11045; Invitrogen, Carlsbad, CA) as the secondary antibody, and subsequently incubated with PI (9 μg/mL) combined with ribonuclease A (900 μg/mL) in PBS for 20 min at 37°C. The percentage of β-cells that were apoptotic was determined by manual counting of condensed nuclei in insulin-positive cells.

At least 30 islets per condition (an average of 33.8 islets per replicate per condition) were analyzed for each experiment for quantification of amyloid area, insulin area, and PI-positive nuclei, with the observer blinded to the genotype and culture conditions of the islets (33).

Islets (200 per condition) were washed twice with PBS, then resuspended in lysis buffer containing 0.1% NP-40, 0.05% deoxycholate, 0.1% SDS, 0.2% sarkosyl, 10% glycerol, 1 mmol/L dithiothreitol, 1 mmol/L EDTA, 10 mmol/L NaF, 50 mmol/L Tris, and protease and phosphatase inhibitors (#04693116001 and #04906837001, respectively; Roche, Indianapolis, IN). After centrifugation at 10,000 rpm for 10 min at 4°C, the supernatant was collected and protein concentration determined using the bicinchoninic acid assay. Islet protein (20 μg) was separated on a Tris-glycine gel by electrophoresis and transferred to a polyvinylidene difluoride membrane. Blots were blocked with 5% skim milk, 5% BSA, or Superblock (Invitrogen) and probed with the following primary antibodies: rabbit polyclonal ARC (1:500; #160737; Cayman Chemicals, Ann Arbor, MI), mouse monoclonal ARC (1:200; #sc-374177; Santa Cruz Biotechnology, Dallas, TX), rabbit polyclonal phospho-JNK1/2 (1:500; #ab4821; Abcam, Cambridge, U.K.), mouse monoclonal JNK (1:200; #sc-7345, Santa Cruz Biotechnology), rabbit polyclonal phospho-c-Jun (1:500; #9261; Cell Signaling Technology, Danvers, MA), rabbit polyclonal c-Jun (1:500; #9165, Cell Signaling), rabbit polyclonal HSP90 (1:500; #4874, Cell Signaling), rabbit polyclonal tumor necrosis factor (TNF)-α (1:500; #ab9739, Abcam), or mouse monoclonal β-actin (1:500; #A2228, Sigma-Aldrich). Primary antibodies were detected with goat anti-rabbit horseradish peroxidase or goat anti-mouse horseradish peroxidase secondary antibodies. Representative immunoblots are shown. Immunoblot band density was quantified using ImageJ software (National Institutes of Health, Bethesda, MD), then reported as the ratio of total protein to the housekeeping protein or as the ratio of phosphorylated to total protein.

After culture, islet protein lysates (500 μg) were collected and immunoprecipitated with ARC or IgG rabbit polyclonal antibodies (2 µg) for 1 h. Immunoprecipitated protein was bound to microMACS protein G microbeads (Miltenyi Biotec, San Diego, CA), separated using MACS magnetic columns (Miltenyi), and washed with lysis buffer, then 20 mmol/L Tris-HCl using gravity flow. After washing, bound protein was eluted in 2% SDS buffer, and immunoblot analysis was performed as described above.

Total RNA from islets was recovered using the High Pure RNA Isolation Kit (Roche), reverse transcribed, and quantitative (q)RT-PCR was performed as previously described (29). Data were normalized to 18S rRNA expression levels. All qRT-PCR data represent means of triplicate determinations from at least three independent experiments. TaqMan probes were used to quantify relative mRNA expression of Jun (Mm00495062_S1), Tnf (Mm00443260_g1), Ddit3/Chop (Mm00492097_m1), Ins2 (Mm00731595_gH), and 18S (HS99999901_s1).

Data are presented as mean ± SE. Statistical differences were determined using the Student t test or Mann-Whitney U test (for nonparametric data), with a value of P < 0.05 considered significant.

To understand the role ARC plays in regulating β-cell apoptosis in the setting of islet amyloid formation, we examined the effect of AdV-mediated knockdown of ARC. Amyloid-forming hIAPP transgenic and nontransgenic control islets were transduced with AdV to knockdown islet ARC expression (AdV-shARC) or AdV-shCTL. AdV-mediated ARC knockdown resulted in an ∼50% reduction in ARC protein expression in islets (1.15 ± 0.14 vs. 0.58 ± 0.05, P < 0.01, n = 4) (Fig. 1A). Islet amyloid formation led to a significant increase in β-cell apoptosis (0.11 ± 0.03% vs. 0.25 ± 0.02% apoptotic β-cells, P < 0.01, n = 5) (Fig. 1B) and a decrease in β-cell area (69.5 ± 4.7% vs. 57.9 ± 2.5% insulin area/islet area, P = 0.05, n = 5) (Fig. 1C) in AdV-shCTL islets. Knockdown of ARC resulted in a near-doubling of the rate of amyloid-induced β-cell apoptosis (0.46 ± 0.01% vs. 0.25 ± 0.02% apoptotic β-cells, P < 0.0001, n = 5) (Fig. 1B) and greater β-cell loss (47.4 ± 1.8% vs. 57.9 ± 2.5% insulin area/islet area, P < 0.01, n = 5) (Fig. 1C). Interestingly, ARC knockdown did not alter islet amyloid deposition in hIAPP transgenic islets (1.5 ± 0.1% vs. 1.8 ± 0.2% amyloid area/islet area, P = 0.36, n = 4) (Fig. 1D), suggesting that ARC acts on mechanisms downstream of amyloid deposition.

To determine whether increased ARC expression can diminish amyloid-induced β-cell apoptosis and thus loss, we next examined the effect of AdV-mediated overexpression of human ARC in islets. Amyloid-forming hIAPP transgenic and nontransgenic control islets were transduced with AdV-hARC or AdV-GFP. As illustrated in Fig. 2A, AdV-hARC delivery resulted in approximately fivefold overexpression of ARC protein (4.79 ± 0.49 vs. 1.00 ± 0.23, P < 0.001, n = 5). Again, islet amyloid formation increased β–cell apoptosis (0.42 ± 0.09% vs. 0.10 ± 0.01% apoptotic β-cells, P < 0.01, n = 5) (Fig. 2B) and β-cell loss (63.9 ± 3.4% vs. 76.7 ± 2.5% insulin area/islet area, P < 0.05, n = 5 (Fig. 2C) in GFP-treated control islets. Overexpression of ARC reduced amyloid-induced β-cell apoptosis by 57% (0.18 ± 0.05% vs. 0.42 ± 0.09% apoptotic β-cells, P < 0.05, n = 5) (Fig. 2B) and preserved β-cell area (73.0 ± 1.8% vs. 63.9 ± 3.4% insulin area/islet area, P < 0.05, n = 5) (Fig. 2C) compared with islets expressing GFP. Again, this reduction in β-cell apoptosis and loss was not associated with any change in islet amyloid deposition in hIAPP transgenic islets (2.2 ± 0.4% vs. 1.8 ± 0.5% amyloid area/islet area, P = 0.56, n = 4) (Fig. 2D).

Because we observed that ARC represses amyloid-induced β-cell apoptosis, we next sought to identify mechanisms through which ARC may act. We first confirmed that the JNK pathway was activated in our 144-h model of in vitro islet amyloid formation, in line with our previous findings after 48 h of culture (13). Nontransduced hIAPP transgenic islets developed significant amyloid deposition (1.9 ± 0.3% of islet area, P < 0.01, n = 5; 56 ± 15% of islets contained amyloid deposits, P < 0.01, n = 5) (Fig. 3A and B, respectively) and displayed increased rates of apoptosis (0.37 ± 0.08% vs. 0.09 ± 0.02% apoptotic β-cells, P < 0.01, n = 5) (Fig. 3C) compared with nontransgenic islets. As expected, JNK phosphorylation was significantly increased in hIAPP transgenic versus control islets (1.8 ± 0.3 vs. 1.1 ± 0.03, P = 0.02, n = 3) (Fig. 3D and E). Further, c-Jun phosphorylation (4.0 ± 1.2 vs. 1.3 ± 0.1, P = 0.03, n = 4) (Fig. 3D and F) and TNF-α protein expression (1.6 ± 0.2 vs. 1.0 ± 0.1, P = 0.04, n = 4) (Fig. 3D and G) were also elevated in amyloid-forming islets. Conversely, ARC protein expression was not altered by islet amyloid deposition (Fig. 3D and H). Consistent with JNK pathway activation in hIAPP transgenic islets, we also observed increased mRNA expression of c-Jun (Jun: 1.49 ± 0.12 vs. 1.03 ± 0.02, P < 0.01) (Fig. 3I) and TNFα (Tnf: 20.08 ± 0.37 vs. 1.04 ± 0.03, P < 0.05) (Fig. 3J), two genes previously reported to be c-Jun targets (11,34,35). Alternatively, mRNA expression of the endoplasmic reticulum stress marker C/EBP homologous protein (Chop) and insulin 2 (Ins2) (Fig. 3K and L) were not altered in amyloid-forming islets.

Having established that the JNK pathway is activated by islet amyloid formation, we next determined whether endogenous ARC limits JNK pathway signaling in islets. AdV-mediated ARC knockdown increased JNK phosphorylation in amyloid-forming islets by ∼60% (1.6 ± 0.2 vs. 1.05 ± 0.02, P = 0.02, n = 4) (Fig. 4A and B). Furthermore, c-Jun phosphorylation (2.0 ± 0.4 vs. 1.1 ± 0.1, P < 0.05, n = 4) (Fig. 4A and C) and TNF-α protein expression (1.5 ± 0.2 vs. 1.0 ± 0.1, P = 0.04, n = 4) (Fig. 4A and D) were also increased by ARC knockdown. To further assess JNK pathway activation, we next asked whether ARC knockdown increases mRNA expression of the c-Jun target genes Jun and Tnf. Expression of both genes was increased by ARC knockdown (Jun: 1.80 ± 0.26 vs. 0.99 ± 0.04, P < 0.05, n = 6; Tnf: 1.85 ± 0.30 vs. 0.96 ± 0.05, P < 0.05, n = 6) (Figs. 4E and F). Expression of Chop mRNA was also increased by ARC knockdown (2.20 ± 0.38 vs. 0.93 ± 0.04, P = 0.02, n = 4) (Fig. 4G), consistent with previous work showing ARC can modulate ER stress signaling (23). Ins2 mRNA expression was unaffected by ARC knockdown (Fig. 4H). These data indicate that endogenous ARC acts to reduce apoptotic signaling in the islet, in part through reduction of amyloid-induced JNK signaling.

Next, we determined whether expression of exogenous ARC can diminish JNK and c-Jun phosphorylation in amyloid-prone islets. Human ARC overexpression halved JNK phosphorylation (0.5 ± 0.1 vs. 1.0 ± 0.1, P = 0.03, n = 3) (Fig. 5A and B) as well as c-Jun phosphorylation (0.5 ± 0.1 vs. 1.0 ± 0.2, P < 0.05, n = 4) (Fig. 5A and C) in amyloid-laden hIAPP transgenic islets. ARC overexpression also decreased islet TNF-α protein expression in hIAPP transgenic islets (0.5 ± 0.2 vs. 1.0 ± 0.1, P = 0.04, n = 3) (Fig. 5A and D). To determine whether the observed changes in c-Jun phosphorylation resulted in changes to c-Jun target gene expression, we again examined Jun and Tnf mRNA abundance. Overexpression of human ARC in hIAPP transgenic islets reduced levels of both c-Jun target genes (Jun: 0.68 ± 0.08 vs. 0.94 ± 0.05, P = 0.03, n = 4; Tnf: 0.71 ± 0.07 vs. 1.07 ± 0.05, P < 0.01, n = 4) (Fig. 5E and F). However, ARC overexpression did not alter Chop (Fig. 5G) or Ins2 (Fig. 5H) gene expression. These data indicate that exogenous expression of human ARC reduces amyloid-induced JNK pathway activation in amyloid-prone islets.

Because ARC is known to bind the proapoptotic molecules that it antagonizes (24,26), we asked whether ARC binds JNK in islets. Using lysates from hIAPP transgenic islets with ARC overexpression, we immunoprecipitated ARC protein and found that it binds both the p54 and p46 isoforms of JNK (Fig. 6). This finding suggests that human ARC may reduce amyloid-induced apoptosis by binding JNK and limiting its activity in the β-cell.

To determine whether alteration of ARC expression affects islet function, we next examined glucose-stimulated insulin secretion in response to 20 mmol/L glucose and insulin content in hIAPP transgenic islets with AdV-mediated ARC knockdown or ARC overexpression. Neither ARC knockdown (P = 0.29, n = 5) nor ARC overexpression (P = 0.44, n = 5) changed glucose-stimulated insulin secretion compared with control transduced islets (Fig. 7A). Similarly, insulin content was not altered by ARC knockdown (P = 0.53, n = 4) or ARC overexpression (P = 0.27, n = 4) (Fig. 7B). These data indicate that ARC does not significantly affect insulin secretion or abundance in amyloid-forming islets.

In this study, we used primary islets with β-cell hIAPP expression to demonstrate for the first time that endogenous ARC is a physiological repressor of islet amyloid-induced β-cell apoptosis and that overexpression of ARC is capable of diminishing amyloid-induced β-cell loss. We found that the effect of ARC on β-cell apoptosis is mediated through JNK and c-Jun phosphorylation as well as c-Jun target gene expression. Furthermore, we showed that altering ARC protein expression did not change islet insulin secretion or content. These findings suggest that ARC restrains apoptotic signaling in the β-cell partly by suppressing JNK pathway signaling and may represent a therapeutic target for prevention of β-cell loss in type 2 diabetes.

The ability of ARC to reduce cell death was first established in cardiac muscle, skeletal muscle, and neurons (17,20,22). The work in these cell types showed its antiapoptotic effects were determined to be mediated by binding proapoptotic molecules at its N-terminal CARD and inactivating their proapoptotic functions (24,26). More recently, ARC was shown to be present in β-cells (23), and exogenous ARC was shown to bind JNK and block its activation in hepatocytes, a cell type that does not normally produce ARC (21,22). Given that exogenous ARC has been shown to reduce JNK signaling (21,22) and that JNK inhibition reduces amyloid-induced β-cell apoptosis (13), we determined whether ARC reduced the toxic effects of islet amyloid by inhibiting JNK pathway activation. Here, we provide the first evidence that endogenous ARC can limit JNK activation and that overexpression of ARC is capable of reducing amyloid-induced JNK pathway activation in islets.

We found that endogenous ARC reduces apoptosis and JNK pathway signaling in hIAPP transgenic islets. In support of endogenous ARC having an effect to diminish JNK signaling and β-cell death, we found that ARC knockdown resulted in 1) increased phosphorylation of JNK and its most proximal downstream target, c-Jun; 2) increased c-Jun target gene expression (Jun and Tnf); and 3) increased β-cell apoptosis and loss under conditions of amyloid formation. Interestingly, we also found that ARC knockdown resulted in increased expression of the ER stress marker Chop, consistent with previous findings showing that ARC represses CHOP protein expression in β-cells (23). Thus, in the setting of ARC knockdown, pathways other than JNK may contribute to the increased cell death we observed. However, neither islet amyloid deposition nor ARC overexpression altered Chop mRNA expression, the former a finding we have previously reported (36). Therefore, although ARC appears capable of affecting ER stress signaling pathways, these pathways are not activated by islet amyloid in our model and do not explain the reduction of β-cell apoptosis we observed with ARC overexpression. Furthermore, because ARC expression is not decreased by islet amyloid deposition, the β-cell loss observed with amyloid is not the result of a loss of ARC expression but is more likely caused by an inability of endogenous ARC to fully inhibit signaling activated by islet amyloid.

We also showed that ARC overexpression reduced JNK and c-Jun phosphorylation in hIAPP transgenic islets. This resulted in decreased expression of the c-Jun target genes Jun and Tnf, which we and others have previously shown are enhanced by amyloid formation (11,34). ARC overexpression also ameliorated amyloid-induced β-cell apoptosis and loss. However, ARC overexpression did not alter Chop expression. Together, our data indicate that reduction of JNK pathway signaling is a major mechanism through which ARC prevents amyloid-induced β-cell death.

Our finding that ARC binds JNK suggests that the physical interaction between ARC and JNK may mediate the protective effects of ARC to reduce JNK pathway activation in amyloid-induced β-cell apoptosis. Such a mechanism is in line with the known role of ARC’s CARD domain to interact with and suppress the function of proapoptotic molecules (24,26). Moreover, previous reports have provided evidence that ARC binds JNK and reduces JNK signaling in hepatocytes (21,22). These reports, and our own findings, are consistent with JNK activity being regulated by physical interaction with other proteins, a mechanism that has been described previously (37,38).

Modulation of ARC expression in hIAPP transgenic islets did not alter islet glucose-stimulated insulin secretion or insulin content (Figs. 7A and B). These findings (along with the Ins2 mRNA data in Figs. 3L, 4H, and 5H) suggest that ARC expression itself does not significantly affect insulin synthesis or secretion. Given our findings that ARC ameliorates amyloid-induced β-cell apoptosis and preserves β-cell area, it is interesting that we have not observed an effect of ARC on β-cell function. One explanation is that we have not altered β-cell abundance sufficiently to cause a β-cell secretory defect and that enough β-cells remain to respond appropriately to the glucose stimulus. Our data do support the concept that increasing ARC expression in β-cells may represent a viable therapeutic strategy to prevent β-cell loss in type 2 diabetes, a condition in which the long duration of the disease process results in greater loss of β-cells than we observed in this in vitro study.

We found that altering ARC expression did not change the degree of islet amyloid deposition. Thus, all of the protective effects of ARC in the islet would appear to occur downstream of amyloid formation per se. Based on previous studies and our own observations, we propose that ARC overexpression ameliorates amyloid-induced apoptosis in the β-cell by reducing JNK phosphorylation and kinase activity, thereby leading to reduced c-Jun phosphorylation and expression of inflammatory mRNA transcripts. This effect is likely at least partly mediated by direct interaction between ARC and JNK. Because β-cell JNK signaling is activated in response to a number of other diabetogenic stimuli, including interleukin-1β and endoplasmic reticulum stress (39,40), ARC may be capable of reducing apoptotic signaling in response to multiple β-cell insults.

Although we believe inhibition of JNK signaling is a major mechanism by which ARC prevents β-cell apoptosis, ARC may also inhibit amyloid-induced apoptosis by acting on additional proapoptotic molecules. Some molecules reportedly affected by ARC include procaspase 8, Bcl-2–associated death promoter (Bad), Bcl-2–associated X protein (Bax), and Puma, many of which are involved with amyloid-induced cell death (19,25,41,42). Furthermore, previous studies have shown that ARC alters tumor necrosis factor receptor 1 (TNFR1) signaling (43) and Fas signaling (24), two pathways upstream of JNK signaling (44). Thus, future studies will explore the relationship of these molecules to JNK signaling in the islet.

Our study has some limitations. Because the AdV constructs used in this study are not β-cell specific, islet viral transduction may have affected ARC expression in non–β-cells and that this may have contributed to the observed effects of ARC. Although endogenous expression of ARC has been observed in nearly all β-cells, its expression is also present to a lesser degree in δ-, α-, and PP cells (23). Therefore, future studies with β-cell–specific knockdown or overexpression of ARC will be required to determine whether non–β-cells contribute to the phenotype we observed. However, apoptosis was quantified only in insulin-positive islet cells to determine the effect of altered islet ARC expression specifically on β-cell apoptosis. Thus, we believe that altered ARC expression affects β-cell death primarily through its effect on β-cell JNK signaling.

Blocking amyloid-induced β-cell loss is one potential therapeutic approach to preserve β-cells in type 2 diabetes. Thus far, approaches to do so have focused largely on reduction of amyloid formation per se (45–49). Here, we show that it is also possible to reduce amyloid-induced β-cell apoptosis by targeting intracellular events downstream of amyloid formation. Whether such an approach will be effective at preventing β-cell loss and maintaining β-cell function in vivo will require further studies involving animal models of islet amyloid formation. These studies are also imperative because high levels of ARC expression have been observed in cancer cells (50). Enhancing ARC activity in β-cells may thus pose a risk for the development of insulinomas. In this respect, the inherently low proliferative rate of the β-cell may make it a good candidate for therapeutics that enhance ARC function, but the safety of such an approach remains to be determined.

In summary, we have demonstrated that ARC is a physiological regulator of amyloid-induced β-cell death and that increasing its expression inhibits islet amyloid-induced apoptosis independent of any change in amyloid deposition. This effect of ARC is likely mediated through reduction of JNK pathway activation. These observations with ARC provide a new approach to reduce β-cell apoptosis and imply that strategies aimed at increasing ARC expression or activity may help prevent the β-cell loss that is a characteristic of type 2 diabetes.

Acknowledgments. The authors thank Josh Willard, Michael Peters, Breanne Barrow, Daryl Hackney, Phil Bergquist, and Jessica Wilkins-Gutierrez (Seattle Institute for Biomedical and Clinical Research, Seattle, WA) for excellent technical assistance provided during the performance of these studies and Christopher Rhodes (Department of Medicine, University of Chicago, Chicago, IL) for providing the GFP-expressing adenoviral construct.

Funding. This work was supported by funding from the National Institute of Diabetes and Digestive and Kidney Diseases (P30-DK-017047 to the Cell Function Analysis Core, Cellular and Molecular Imaging Core, and Viral Vector and Transgenic Mouse Core of the University of Washington Diabetes Research Center, F32-DK-107022 to A.T.T., F32-DK-109584 to M.F.H., and DK-098506 to S.Z.), the University of Washington (Dick and Julia McAbee Fellowship to D.T.M.), the Department of Veterans Affairs (BX001060 to S.E.K.), and the American Diabetes Association Mentor-Based Fellowship (7-11-MN-28 to S.E.K.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. A.T.T. participated in study design, performed research, analyzed data, and wrote the manuscript. T.S. participated in study design, performed research, analyzed data, and reviewed and edited the manuscript. D.T.M., M.F.H., M.M., R.N.K., S.Z., R.L.H., and S.E.K. participated in study design, helped interpret data, and reviewed and edited the manuscript. M.T.C. designed and contributed adenoviral vectors. A.T.T. 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.

Prior Presentation. Parts of this study were presented at the 75th Scientific Sessions of the American Diabetes Association, Boston, MA, 5–9 June 2015.