1326 Diabetes Volume 63, April 2014

Richard A. Zuellig,1 Thorsten Hornemann,2,3,4 Alaa Othman,2,4 Adrian B. Hehl,5 Heiko Bode,2,3 Tanja Güntert,3,6 Omolara O. Ogunshola,6 Enrica Saponara,7 Kamile Grabliauskaite,7 Jae-Hwi Jang,7 Udo Ungethuem,7 Yu Wei,2,3,4 Arnold von Eckardstein,2,3,4 Rolf Graf,7 and Sabrina Sonda7

Deoxysphingolipids, Novel Biomarkers for Type 2 Diabetes, Are Cytotoxic for Insulin- Producing Cells

Irreversible failure of pancreatic b-cells is the main 1-deoxysphinganine contribute to its cytotoxicity. culprit in the pathophysiology of diabetes, a disease Analyses of signaling pathways identified Jun that is now a global epidemic. Recently, elevated N-terminal kinase and p38 mitogen-activated plasma levels of deoxysphingolipids, including protein kinase as antagonistic effectors of cellular 1-deoxysphinganine, have been identified as a novel senescence. The results revealed that biomarker for the disease. In this study, we analyzed 1-deoxysphinganine is a cytotoxic lipid for insulin- whether deoxysphingolipids directly compromise producing cells, suggesting that the increased levels the functionality of insulin-producing Ins-1 cells and of this sphingolipid observed in diabetic patients primary islets. Treatment with 1-deoxysphinganine may contribute to the reduced functionality of PATHOPHYSIOLOGY induced dose-dependent cytotoxicity with pancreatic b-cells. Thus, targeting senescent, necrotic, and apoptotic characteristics deoxysphingolipid synthesis may complement the and compromised glucose-stimulated insulin currently available therapies for diabetes. secretion. In addition, 1-deoxysphinganine altered Diabetes 2014;63:1326–1339 | DOI: 10.2337/db13-1042 cytoskeleton dynamics, resulting in intracellular accumulation of filamentous actin and activation of the Rho family GTPase Rac1. Moreover, In the past 3 decades, the prevalence of diabetes has 1-deoxysphinganine selectively upregulated risen worldwide at a dramatic rate, with incidence pro- ceramide synthase 5 expression and was converted jections approaching 8% of the population by 2030 (1,2). to 1-deoxy-dihydroceramides without altering This remarkable increase is largely due to the epidemic normal ceramide levels. Inhibition of intracellular spread of type 2 diabetes (T2DM), which accounts for 1-deoxysphinganine trafficking and ceramide 90% of all cases of diabetes worldwide (3,4). Given the synthesis improved the viability of the cells, level of complexity associated with the pathophysiology indicating that the intracellular metabolites of of T2DM, understanding the mechanisms underlying this

1Division of Endocrinology, Diabetes and Clinical Nutrition, University Hospital Corresponding author: Sabrina Sonda, [email protected]. Zurich, Zurich, Switzerland Received 3 July 2013 and accepted 14 December 2013. 2Institute for Clinical Chemistry, University Hospital Zurich, Zurich, Switzerland This article contains Supplementary Data online at http://diabetes 3Centre for Integrative Human Physiology, University of Zurich, Zurich, Switzerland .diabetesjournals.org/lookup/suppl/doi:10.2337/db13-1042/-/DC1. 4Competence Centre for Systems Physiology and Metabolic Diseases, Zurich, Switzerland R.A.Z. and T.H. contributed equally to this work. 5Institute of Parasitology, University of Zurich, Zurich, Switzerland © 2014 by the American Diabetes Association. See http://creativecommons 6Institute of Veterinary Physiology, University of Zurich, Zurich, Switzerland .org/licenses/by-nc-nd/3.0/ for details. 7Swiss Hepato-Pancreatico-Biliary (HPB)-Center, Division of Surgical Research, See accompanying article, p. 1191. Department of Visceral and Transplantation Surgery, University Hospital Zurich, Zurich, Switzerland diabetes.diabetesjournals.org Zuellig and Associates 1327 disease is necessary to design alternative strategies to Animal Experiments limit its progression. Recently, substantial improve- Wistar rats, leptin-deficient ob/ob mice on C57BL/6J ments occurred in the detection of early stage or un- background (B6.V-Lep/OlaHsd), and wild-type (WT) diagnosed T2DM, thus allowing appropriate treatments C57BL/6J (Harlan Laboratories) were kept under a light- in high-risk populations. One of the latest biomarkers dark regimen (16:8 h) at constant temperature and given identified in patients with diabetes and metabolic free access to food and water. All animal experiments syndromes are increased plasma levels of deoxy- were performed in accordance with Swiss federal animal sphingolipids (1-deoxySLs) (5,6), a type of sphingolipid regulations and approved by the cantonal veterinary of- characterized by an initial condensation of alanine or fice of Zurich. Islets were harvested from pancreata of glycine instead of serine with palmitic acid and the male Wistar rats (250–300 g) by collagenase (NB8 col- resultant absence of the hydroxyl group in position C1. lagenase; Serva, Heidelberg, Germany) followed by tryp- Consequently, although these deoxysphingoid bases can sin digestion to dissociate them into single cells as be acylated to deoxy-dihydroceramides, they cannot be previously described (13). further metabolized to complex sphingolipids or effi- ciently degraded by the canonical degradation pathway; Insulin Secretion thus, they tend to accumulate once produced. Impor- Dissociated islets cells were seeded in 12-well extracel- – tantly,1-deoxySLsdisplaytoxicpropertiesinvitro lular matrix (ECM) coated plates (Novamed, Jerusalem, m toward several cell lines (7–9), and in vivo, 1-deoxySLs Israel); treated for 24 h with 5 mol/L sphinganine, are believed to impair neuronal functionality in patients 1-deoxysphinganine, or BSA; and incubated in RPMI with the hereditary sensory and autonomic neuropathy medium containing 3.3 mmol/L glucose for 1 h. Follow- type I (10). In light of the increased plasma levels of ing sequential 1-h incubations with low (3.3 mmol/L), 1-deoxySLs found in diabetic patients and of the high (16.7 mmol/L), low (3.3 mmol/L) glucose concen- reported cytotoxic effects associated with the exposure trations, insulin secretion was measured by radioimmu- to increased 1-deoxySL concentrations, we investigated noassay (Insulin-CT; CIS Bio International, Schering AG, ’ whether these atypical sphingolipids directly compro- Baar, Switzerland) according to the manufacturer s mise pancreatic b-cells, the dysfunction of which plays instructions. an important role in the pathogenesis of both type 1 Quantitative RT-PCR diabetes and T2DM. Total RNA was extracted from Ins-1 cells cultured in m RESEARCH DESIGN AND METHODS 1 mol/L sphinganine or 1-deoxysphinganine for 24 h. Quality of RNA was assessed by a 2100 bioanalyzer Biochemical Reagents (Agilent Technologies, Basel, Switzerland). cDNA was Unless otherwise stated, all chemicals were purchased obtained with the RT2 First Strand Kit and profiled by from Sigma and cell culture reagents from Gibco-BRL. the Rat Cell Death Pathway Finder PCR Array (both from Inhibitor stock solutions were freshly diluted to the SABiosciences, Hombrechtikon, Switzerland) according concentrations required for the individual experiment to the manufacturer’s instructions. Ceramide synthase indicated in the figure legends. Lipid stock solutions were (CerS) primers for SYBR green quantitative PCR are prepared as previously described (10) as a bovine serum listed in the Supplementary Data. albumin (BSA) complex and added to the cells at the concentrations indicated in the figure legends. BSA was Immunohistochemistry and Flow Cytometry Analyses fi used as control. Pancreas specimens were xed in 4% formalin and par- affin embedded according to standard procedures (14). In Vitro Cell Culture Ins-1 cells were fixed in 3.6% formaldehyde and per- The Ins-1 rat insulinoma cell clone 832/13, provided by meabilized with 0.2% Triton X-100 in PBS. Primary C. Wollheim, was maintained in RPMI 1640 medium as antibodies used in this study are listed in the Supple- previously described (11,12). Cell metabolic activity was mentary Data. Apoptosis detection was performed with tested with the 0.5% tetrazolium salt solution 3-(4,5- an ApopTag peroxidase kit (MP Biomedicals, Illkirch, dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide France). Immunofluorescence analysis and image data (MTT) or WST-1 (Roche) according to the manufacturer’s collection were performed on a Zeiss Axioplan 2 imaging instructions. Cell death was quantified by trypan blue fluorescence microscope (Carl Zeiss Microimaging, exclusion or lactate dehydrogenase (LDH) release in the Göttingen, Germany) or on a Leica SP2 AOBS confocal medium (Roche). Cellular senescence was quantified with laser-scanning microscope (Leica Microsystems, Wetzlar, the b-galactosidase assay kit (Cell Biolabs). Adenovirus- Germany) using a glycerol immersion objective lens expressing p21 (Adp21) (rat) and adenovirus-containing (Leica, HCX PL APO CS 633 1.3 Corr). Image z-stacks green fluorescent protein (AdGFP) were purchased from were collected with a pinhole setting of Airy 1 and two- Vector Biolabs (Philadelphia, PA). Rac1 activity was fold oversampling. Image stacks of optical sections were measured with G-LISA Rac1 activation assay (Cytoskeleton, processed with the Huygens deconvolution software Denver, CO). package version 2.7 (Scientific Volume Imaging, 1328 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014

Hilversum, the Netherlands). Three-dimensional (3D) (5,6), we analyzed whether 1-deoxySLs can directly reconstruction, volume rendering, and quantification of affect the viability of insulin-secreting cells. To this signal overlap in the 3D volume model were done with aim, the rat insulinoma cell line Ins-1 was treated at the Imaris 7.2.1 software suite (Bitplane, Zurich, Swit- 50% confluence (Fig. 1A,whereLindicateslowdensity) zerland). The degree of signal overlap in the 3D volume for 24 h with 1-deoxysphinganine or sphinganine as models is shown graphically as scatterplots by plotting control. 1-Deoxysphinganine incubation reduced both the intensity of two fluorescent signals in each voxel of the metabolic activity, as measured by MTT (Fig. 1A) the 3D model. Voxels with similar signal intensity for and WST-1 reduction (Supplementary Fig. 1A), and the both signals appear in the area of the diagonal. Single-cell number of live cells (Fig. 1B) in a dose-dependent quantification of stained cells by flow cytometry was manner. Treatment with 5 mmol/L caused cell roundup performed with a FACSDiva flow cytometer (BD Bio- (Fig. 1C) and death, as shown by robust trypan blue sciences, Allschwil, Switzerland). inclusion (Fig. 1D)andLDHrelease(Fig.1E). However, cells treated with 1 mmol/L 1-deoxysphinganine did not Western Blotting increase in number compared with the initial seeding m Ins-1 cells cultured in 1 mol/L sphinganine, 1-deoxy- but showed modest levels of lethality (Fig. 1D and E)or sphinganine, or BSA for 24 h were lysed as previously upregulation of genes involved in cell death pathways m described (15). Aliquots corresponding to 35 gofproteins (Supplementary Fig. 1B–D), suggesting a cytostatic ef- were separated by SDS-PAGE electrophoresis, blotted, and fect of the lipid at this concentration (Fig. 1B). When probed overnight at 4°C. Primary antibodies used in this the lipid treatment was performed on 90% confluent study are listed in the Supplementary Data. Immunoreactive cells (Fig. 1A, where H indicates high density), the bands from at least three independent experiments were metabolic activity was reduced only at the highest fi b quanti ed by densitometry and normalized to -actin or concentration of 1-deoxysphinganine tested, further glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels. indicating that 5 mmol/L 1-deoxysphinganine is cyto- Analysis of Sphingoid Bases and Ceramides by Liquid toxic for both dividing and quiescent cells, whereas Chromatography-Tandem Mass Spectrometry lower lipid concentrations are cytostatic. In addition, The sphingoid base profile was analyzed as described ear- treatment for only 1 h followed by washout and sub- fi lier (5). Ceramide species were extracted by adding 1 mL sequent 23-h incubation was suf cient to reduce the methanol/chloroform (2:1) [including the addition of 200 metabolic activity of the cells comparably to a continu- F pmol C12 ceramide internal standard (Avanti Polar Lipids)] ous 24-h incubation with 1-deoxysphinganine (Fig. 1 ), to 100 mL resuspended cells followed by 0.5 mL chloro- suggesting a rapid effect of the lipid. form and 200 mL alkaline water (10). Interfering phos- pholipids were hydrolyzed by re-extracting the dried lipids 1-Deoxysphinganine Triggers p21-Mediated with methanol-KOH:chloroform (4:1) as described earlier Senescence and Multiple Cell Death Pathways in Ins-1 Cells (10). Lipids were separated on a C18 column (Uptisphere 120 Å, 5 mm, 125 3 2 mm; Interchim, Montluçon, France) We then investigated whether the reduction of replication and analyzed on a TSQ Quantum Ultra mass spectrometer following low 1-deoxysphinganine concentration is medi- (Thermo Fisher Scientific, Reinach, Switzerland) using at- ated by induction of senescence. 1-Deoxysphinganine at m b A mospheric pressure chemical ionization (10). Ceramides 1 mol/L increased -galactosidase activity (Fig. 2 )and WAF1/Cip1 B C and deoxyceramides were identified by precursor ion scan nuclear p21 expression (Fig. 2 and ), a hall- (20 mV collision energy) with fragments of (m/z 264.3) mark (16) and inducer of senescence (17), respectively. To fi and (m/z 268.3), respectively. Levels were normalized to investigate whether increased p21 expression was suf - ISTD and cell numbers. cient to trigger the senescence pathway in Ins-1 cells, we used Adp21 or AdGFP as control (Fig. 2D). Adp21 in- Statistical Analyses fection decreased Ins-1 cell replication and increased Results are expressed as mean 6 SEM. Significance was b-galactosidase activity, whereas both parameters were assessed with Student unpaired, two-tailed t tests or one- unchanged following AdGFP incubation (Fig. 2E and F). In way ANOVA. P , 0.05 was considered significant. For addition, the increased b-galactosidase activity after overall P , 0.05, the Bonferroni multiple-comparison 1 mmol/L 1-deoxysphinganine treatment or Adp21 infection test was used to determine whether there was a signifi- was accompanied by increased MTT reduction per cell cant difference between values of control (reference (Supplementary Fig. 2), suggesting increased mitochon- sample) and samples of interest. drial dehydrogenase activity, a parameter associated with cellular senescence (18). These data suggest that upregu- RESULTS lation of p21 induced by 1 mmol/L 1-deoxysphinganine 1-DeoxySL Treatment Is Cytostatic and Cytotoxic for treatment contributes to the decreased Ins-1 replication Ins-1 Cells by activating a senescence program. Because 1-deoxySLs were found to be elevated in the Next, we further explored the cytotoxic effect of high plasma of diabetic patients in the low micromole range 1-deoxysphinganine concentrations. The 5 mmol/L diabetes.diabetesjournals.org Zuellig and Associates 1329

Figure 1—DeoxySLs decrease the replication of Ins-1 cells. A: Metabolic activity tested by MTT assay of Ins-1 cells treated at 50% (L) and 90% (H) confluence with SA and dSA at the indicated concentrations and incubated for 24 h. B: Enumeration of live cells treated with the indicated dSA concentrations for 24 h. Note that the number of cells treated with 1 mmol/L dSA is comparable to the initial seeding density. C: Bright-field images showing cell rounding up upon 24-h dSA treatment. D: Quantification of trypan blue–positive cells after treatment with the indicated concentrations of dSA for 24 h. Data are percentage of total cell number. E: Quantification of LDH released in the medium after 24-h dSA treatment. Data are normalized to the number of cells. F: MTT assay of Ins-1 cells treated with dSA for the indicated time and assessed 24 h after adding the lipid. Data are mean 6 SEM (n = 3). Scale bars: 50 mm. *P < 0.05. Cntl, control; dSA, 1-deoxysphinganine; SA, sphinganine; tot, total.

1-deoxysphinganine abrogated the expression of p21 and addition, fluorescence-activated cell sorter (FACS) anal- induced cells with condensed pyknotic nuclei and high yses of cells costained with annexin V and propidium levels of activated caspase-3, hallmarks for the execution iodide (PI) revealed that 5 mmol/L 1-deoxysphinganine phase of apoptosis (Fig. 3A, Supplementary Fig. 3A). In treatment increased the number of both apoptotic and 1330 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014

Figure 2—dSA triggers senescence in Ins-1 cells. A: Quantification of b-gal activity following incubation for 24 h with 1 mmol/L SA, 0.5 and 1 mmol/L dSA, or BSA as control. Values are normalized to the number of live cells. B: Immunofluorescence imaging showing nuclear expression of p21 following incubation for 24 h with 1 mmol/L SA, 1 mmol/L dSA, or BSA as control. Nuclei are stained with DAPI (blue). The lower panels show nuclear localization of p21. C: Western blot and densitometric quantification of p21/GAPDH levels following in- cubation for 24 h with 1 mmol/L SA, 1 mmol/L dSA, or BSA as control. D: Immunostaining of p21 (upper panels) or live imaging (lower panels) of Ins-1 cells infected with Adp21 and AdGFP at MOI50. E: Enumeration of live Ins-1 cells 24 h after Adp21 or AdGFP infection. Note how Adp21 decreased the replication of Ins-1 cells in a dose–response manner without inducing cell death. F: Quantification of b-gal activity 24 h after Adp21 or AdGFP infection. Note how senescence is induced only in the presence of Adp21. Results are mean 6 SEM (n = 3). Scale bars: 50 mm. *P < 0.05. b-gal, b-galactosidase; cntl, control; dSA, 1-deoxysphinganine; MOI, multiplicity of infection; SA, sphinganine.

necrotic cells (Fig. 3B and C, Supplementary Fig. 3B) and 1-Deoxysphinganine Intracellular Metabolites induced the cells to arrest in the G0/G1 phase of the cell Contribute to Cytotoxicity in Ins-1 Cells cycle (Supplementary Fig. 4), suggesting that the lipid Because treatment with exogenous 1-deoxy-dihydroceramides triggers multiple cell death pathways in Ins-1 cells. (m18:0,24:1 and m18:0,16:0) and 1-deoxy-methylsphinganine, diabetes.diabetesjournals.org Zuellig and Associates 1331

Figure 3—dSA triggers apoptosis and necrosis in Ins-1 cells. A: Quantification of CC-3 and p21-positive cells after 24-h incubation with 5 mmol/L SA, 1 and 5 mmol/L dSA, or BSA as control (left panel). Immunofluorescence imaging showing cytosolic expression of CC-3 (right panel). Note the pyknotic nuclei in CC-3–positive cells (arrows). Nuclei are stained with DAPI (blue). Results are mean 6 SEM (n = 3). Scale bars: 50 mm. *P < 0.05. B: FACS analyses of Ins-1 cells after 24-h incubation with 1 and 5 mmol/L dSA or BSA as control and staining with PI and AnnV. C: Quantification of cells that are PI/AnnV negative (live cells), PI positive (necrotic cells), AnnV positive (ap- optotic cells), and double positive. Total population comprised 20,000 cells. Note the increased lethality of cells treated with 5 mmol/L dSA. AnnV, annexin V; CC-3, cleaved caspase 3; cntl, control; dSA, 1-deoxysphinganine; FITC, fluorescein isothiocyanate; SA, sphin- ganine; tot, total. 1332 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014 where alanine is replaced by glycine, reduced the (Fig. 4D, Supplementary Fig. 5C), whereas inhibition of cell replication similarly to 1-deoxysphinganine either the first step of sphingolipid synthesis with myr- treatment (Fig. 4A), we tested whether exogenous iocin or glucosylceramide synthesis with PPMP (1-phenyl- 1-deoxysphinganine was also metabolized by the cells 2-palmitoylamino-3-morpholino-1-propanol) had no to the deoxy forms of ceramide. Incubation with effect on cell viability (data not shown). Collectively, these 1-deoxysphinganine significantly increased the cellular data indicate that intracellular uptake followed by meta- levels of deoxy-dihydroceramide with various acyl chain bolic conversion to 1-deoxy-dihydroceramides is re- lengths (Fig. 4B) without altering normal ceramide levels sponsible, at least in part, for 1-deoxysphinganine toxicity. (Supplementary Fig. 5A) and selectively upregulated the expression of ceramide synthase 5 (CerS5) (Fig. 4C), sug- 1-Deoxysphinganine Increases the Phosphorylation of gesting that 1-deoxysphinganine is readily metabolized in Selected Kinases in Ins-1 Cells Ins-1 cells and accumulates in the acylated form. In addi- To further characterize the biochemical components of tion, pretreatment of Ins-1 cellswiththeclass2amphiphile 1-deoxysphinganine-induced cytotoxicity, we analyzed U18666A, a well-established inhibitor of NPC1 (Niemann- the phosphorylated status of key proteins involved in Pick type C) protein that prevents intracellular trafficking of major signaling pathways. Western blot quantification sphingolipids and cholesterol (19,20), partially rescued revealed increased phosphorylation of Jun N-terminal 1-deoxysphinganine-mediated cytotoxicity (Fig. 4D,Sup- kinase (JNK) and p38 mitogen-activated protein kinase plementary Fig. 5B). Moreover, pharmacological inhibition (MAPK) in 1-deoxysphinganine-treated Ins-1 cells. In of ceramide synthesis with fumonisin B1 (FB1) was unique addition, the ratio of AKT phosphorylation was un- in reducing the toxicity of 5 mmol/L 1-deoxysphinganine changed with lipid treatment, but the total protein

Figure 4—1-Deoxy-dihydroceramide contributes to dSA-induced cytotoxicity in Ins-1 cells. A: MTT assay of Ins-1 cells incubated for 72 h with SA, SO, dSA, dmethSA, and deoxy-dihydroceramides (m18:0,24:1 and m18:0,16:0). B: Mass spectrometry quantification shows the increased formation of 1-deoxy-dihydroceramide with various acyl chain lengths upon 24-h treatment with 3 mmol/L dSA. SA, SO, and BSA incubation were used as control. C: RNA expression levels of various CerS isoforms following lipid incubation as in B. Transcript levels were normalized using GAPDH RNA as a reference. D: Enumeration of live cells following incubation for 24 h with dSA in the presence or absence of 5 mmol/L U18666A or 35 mmol/L FB1. Results are mean 6 SEM (n = 3). *P < 0.05. au, arbitrary unit; cntl, control; dmethSA, deoxy-methylsphinganine; dSA, 1-deoxysphinganine; SA, sphinganine; SO, sphingosine. diabetes.diabetesjournals.org Zuellig and Associates 1333 amount increased in the presence of 1-deoxysphinganine treatment partially rescued the JNK inhibitor effect at (Fig. 5A–C). Because kinase activation by phosphory- 1 mmol/L 1-deoxysphinganine incubation. These data lation suggests a possible role of these proteins in the suggestthatp38MAPKactivationisaneffectorof 1-deoxysphinganine-mediated phenotype, we tested this 1-deoxysphinganine cytotoxicity and senescence, whereas hypothesis by treating the cells with specific kinase inhib- JNK activation plays a protective role in Ins-1 cells. itors. Pretreatment with the p38 MAPK inhibitor Birb796 partially rescued cell replication (Fig. 5D) and reduced 1-Deoxysphinganine Treatment Promotes 1-deoxysphinganine-induced senescence (Fig. 5E). Con- Reorganization of the Actin Cytoskeleton in Ins-1 Cells versely, the JNK inhibitor SP600125 potentiated cytotox- The observed changes in cell morphology observed fol- icity and senescence (Fig. 5D and E), and the combined lowing 1-deoxysphinganine treatment (Fig. 1C)

Figure 5—dSA increases the phosphorylation of selected kinases in Ins-1 cells. A: Western blot of phosphorylated and total levels of JNK, p38, and AKT following incubation for 24 h with 1 mmol/L SA, 1 mmol/L dSA, or BSA as control. B: Densitometric quantification of protein phosphorylation, total protein, and actin levels. Data are percentage of control. C: Densitometric quantification of total protein and actin levels. Data are percentage of control. D: Enumeration of live cells treated with the indicated dSA concentrations for 24 h in the presence of 10 mmol/L of inhibitors of p38 MAPK (Birb), JNK (SP), or a combination (B/SP). Cells were pretreated for 1 h with the indicated inhibitors before the addition of dSA. E: Quantification of b-gal activity following incubation for 24 h with 1 mmol/L dSA in the presence of 10 mmol/L of inhibitors of p38 MAPK (Birb), JNK (SP), or a combination (B/SP). Values are normalized to the number of live cells. Data are mean 6 SEM (n = 3). *P < 0.05. b-gal, b-galactosidase; Birb, Birb796; B/SP, Birb796 and SP600125 combination; cntl, control; dSA, 1-deoxy- sphinganine; phospho, phosphorylated; SA, sphinganine; SP, SP600125; tot, total. 1334 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014 prompted us to further analyze whether the lipid rearrangement of actin cytoskeleton with the resulting induced cytoskeletal alterations. To detect early cyto- accumulation of actin in intracellular punctated struc- skeletal rearrangements, cells were imaged after 5 h of tures (Fig. 7D). 1-Deoxysphinganine did not alter the treatment. Phalloidin-stained actin filaments were cellular content of insulin (Supplementary Fig. 7B)or mainly concentrated in the cortical area and in filopodia selectively reduce the number of insulin-producing cells following control BSA or sphinganine treatment. How- (Supplementary Fig. 7C). However, after incubation with ever, on 1-deoxysphinganine incubation, actin staining the lipid, isolated b-cells were unable to regulate insulin accumulated in punctated structures mainly located in secretion in response to glucose stimulation (Fig. 7E, the perinuclear area of the cells. Different from the Supplementary Fig. 7D). Collectively, these data indicate actin phenotype, tubulin staining showed a similar that 1-deoxysphinganine is cytotoxic to isolated primary microtubule pattern in all treatments, suggesting that islets and compromises both functionality and 1-deoxysphinganine preferentially interferes with the cytoarchitecture of b-cells. organization of actin cytoskeleton in Ins-1 cells (Fig. 6A). Quantitative analysis of members of the Rho family of Increased Glucose Levels Potentiate 1-Deoxysphinganine guanosine triphosphatases (Rho-GTPases) that regulate Toxicity intracellular actin dynamics (21) showed increased levels The in vitro results showed 1-deoxySL-mediated cyto- of Rac1 activity (Fig. 6B) and expression of Rac1 and toxicity in insulin-producing cells, suggesting that raised RhoA, albeit the latter did not reach significance upon levels of these lipids may contribute to the failure of 1-deoxysphinganine treatment (Fig. 6C), suggesting that b-cells during the development of diabetes. However, Rho-GTPases may be involved in 1-deoxysphinganine- elevated 1-deoxySL levels were also found in the plasma mediated changes in actin cytoskeleton reorganization. of patients with metabolic syndrome who did not present Of note, actin alterations were not accompanied by ap- with hyperglycemia and overt diabetes (5), raising the optosis at this time point (data not shown). In addition, question that an increased amount of 1-deoxySLs is not pretreatment with JNK, p38 MAPK, and CerS inhibitors the sole cause of b-cell toxicity and that additional fac- did not prevent actin remodeling (Supplementary Fig. tors contribute to the diabetic phenotype. To further 6A), suggesting that cytoskeletal and cell cycle effects are investigate the causal network of atypical sphingolipids modulated by different signaling pathways. and b-cell toxicity, we analyzed pancreatic islets in leptin- To further characterize whether aberrantly localized deficient ob/ob mice. Kept on a normal chow diet, these actin may intersect with the secretory apparatus of the mice develop obesity and a mild hyperglycemia that cells, we performed costaining for actin and insulin. reverts with aging because pancreatic b-cell compensa- FACS-based single-cell quantification revealed that the tion occurs and increased insulin levels improve glucose lipid treatment did not alter the cellular content of actin homeostasis (22,23). Sixty-week-old ob/ob mice had only and insulin (Supplementary Fig. 6B). However, confocal slightly higher plasma glucose levels than age-matched analyses showed that filamentous actin accumulated in- WT animals but a robust increase in plasma HDL, cho- tracellularly in proximity to and partially colocalized with lesterol, and alanine aminotransferase levels (Supple- insulin-containing vesicles (Fig. 6D). mentary Fig. 8A), the latter reflecting steatotic liver damage (Supplementary Fig. 8B). As previously shown 1-Deoxysphinganine Reduces Metabolic Activity and (23), ob/ob pancreata were characterized by pronounced Insulin Secretion and Modulates Actin Cytoskeleton in islet hyperplasia, vascularization, and robust insulin Primary Islets production (Supplementary Fig. 8B), suggesting that To confirm the relevance of the results in primary cells, b-cells could compensate for the increased insulin de- we tested whether 1-deoxysphinganine affects the func- mand without reaching exhaustion. Despite the evident tionality of isolated islets. Lipid delivery to the cells was hyperplasia, islets did not show active replication at the improved by dissociation of 1,400 islets from six Wistar analyzed age (Supplementary Fig. 8B). Of note, quantifica- rats into single cells; these were seeded in ECM-coated tion of the plasma lipid profiles in ob/ob mice revealed plates and treated for 24 h with 5 mmol/L sphinganine, amoderate,butsignificant increase of 1-deoxysphinganine 1-deoxysphinganine, or BSA as control. Of note, disso- levels (Fig. 8A), which was not associated with increased ciated islets plated on ECM are virtually quiescent senescence or apoptosis (Supplementary Fig. 8B). In ad- (R.A.Z., personal communication). Under these experi- dition, like in human samples (5), sphingosine was the mental conditions, 1-deoxysphinganine treatment in- most abundant species in the mouse plasma and signif- duced cellular vacuolization (Fig. 7A) and reduced the icantly increased in ob/ob mice (Supplementary Fig. 8C). metabolic activity (Fig. 7B) and, to a lesser extent, the Although plasma 1-deoxySLs likely did not reach a critical number of live cells (Fig. 7C). However, senescence was concentration threshold to induce b-cell failure, absence not observed, as indicated by comparable b-galactosidase of other pathological parameters, including hyperglyce- activities (Supplementary Fig. 7A) and absence of p21 mia, may account for the normal viability of b-cells. To staining (data not shown). Similarly to the effect in Ins-1 test this hypothesis, we incubated Ins-1 cells with cells, 1-deoxysphinganine treatment induced 1-deoxysphinganine in the presence and absence of diabetes.diabetesjournals.org Zuellig and Associates 1335

Figure 6—dSA incubation induced actin cytoskeleton rearrangements. A: Ins-1 cells treated with 5 mmol/L dSA, SA, or BSA as control for 5 h and stained with phalloidin (green, left panels) or antitubulin antibody (red, right panels). Nuclei are stained with DAPI (blue). Note the actin staining in punctated structures following dSA treatment (arrows in left panel). In the right panels, arrows indicate tubulin midbody in recently divided cells and the arrowhead shows the mitotic spindle. Scale bars: 50 mm. B: Rac1 activity following 5-h treatment with 5 mmol/L dSA, 5 mmol/L SA, or BSA as control. RCCA was used as a positive control. C: Western blotting quantification of Rac1 and RhoA expression upon 24-h treatment with 1 mmol/L dSA, 1 mmol/L SA, or BSA as control. Data are mean 6 SEM (n = 3). *P < 0.05. D: Confocal images of Ins-1 cells treated with 5 mmol/L dSA or BSA as control for 5 h and stained with phalloidin (green) or anti-insulin antibody (red). Nuclei are stained with DAPI (blue). Signal intensities representing the voxel space of the reconstructed image stacks are depicted as scatterplots showing partial signal overlap of actin and insulin in cntl- and dSA-treated samples. Robust overlap of actin and b-catenin signals (dots accumulated in the mid-diagonal of the plot) and limited overlap of actin and DAPI (dots preferentially distributed along the axes) were used as a control. Nuclei are stained with DAPI (blue). Scale bars: 10 mm. Cntl, control; dSA, 1-deoxysphinganine; NC, negative control; RCCA, constitutively active Rac1; SA, sphinganine. 1336 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014

Figure 7—dSA treatment is cytotoxic for rat primary islets. Dissociated islets were plated on ECM plates and treated for 24 h with 5 mmol/L SA, 5 mmol/L dSA, and BSA as control. A:Quantification of vacuolized cells following lipid treatment. Data are percentage of total cell number. Bright-field images of treated cells showing cell vacuolization (right panels, arrows and inset) upon dSA treatment. B:Metabolic activity tested by MTT assay. Data are percentage of BSA-treated control cells. C: Enumeration of live cells following lipid treatment. D: Quantification of insulin secretion stimulated with 3.3 mmol/L (low I), 16.7 mmol/L (high), and 3.3 mmol/L (low II) glucose following incubation for 24 h with the lipids. Secreted insulin was normalized to the cell number. E: Filamentous actin (phalloidin, green) and insulin (red) costaining of treated islets. Nuclei are stained with DAPI (blue). Data are mean 6 SEM (n = 3). Scale bars: 50 mm. *P < 0.05. Cntl, control; dSA, 1-deoxysphinganine; SA, sphinganine; tot, total.

glucose. Treatment with 30 mmol/L glucose for 24 h DISCUSSION B potentiated 1-deoxysphinganine-induced toxicity (Fig. 8 , Elevated levels of 1-deoxySLs have been found in the Supplementary Fig. 8D), suggesting that hyperglycemia blood of patients with metabolic syndrome and diabetes and 1-deoxySLs synergize to induce glucolipotoxicity in (5,6), raising the question of the role of atypical lipids in insulin-producing cells. the development of these pathologies. The present diabetes.diabetesjournals.org Zuellig and Associates 1337

Figure 8—A: Mass spectrometry quantification of dSA serum levels of 60-week-old WT and ob/ob mice. Data are mean 6 SEM (n = 5). *P < 0.05. B: Enumeration of live Ins-1 cells treated with the indicated dSA concentrations for 24 h in the presence or absence (control) of 30 mmol/L glucose. Data are mean 6 SEM (n = 3). *P < 0.05. Cntl, control; dSA, 1-deoxysphinganine; gluc, glucose.

results show that exposure to 1-deoxySLs in the low In the case of b-cells, the early cellular morphological micromole range compromised insulin secretion and alterations may impair cellular functionality, including triggered senescence and cell death in insulin- insulin secretion. Indeed, earlier studies demonstrated producing cells, indicating that 1-deoxySLs are indeed that insulin secretion in pancreatic b-cells is coupled with toxic for these cells. In our experimental approach, we reorganization of the filamentous actin web located be- elucidated 1-deoxysphinganine toxicity at three distinct neath the plasma membrane, thus allowing docking of levels, namely, 1) changes in cellular structure, 2)en- insulin-containing granules to the cell membrane and gagement of signaling molecules, and 3) activation of consequent secretion. Importantly, glucose stimulation effector proteins. A major finding of this study is that directly induces rearrangement of the actin web (24,25) 1-deoxysphinganine-mediated toxicity is a complex to which insulin-containing granules are in tight contact phenomenon and triggers multiple pathways, including (26). Thus, impaired remodeling of actin cytoskeleton cytoskeletal remodeling, senescence, necrosis, and resulting from 1-deoxysphinganine treatment may con- apoptosis. tribute to the defective insulin secretion observed in the b 1-Deoxysphinganine Triggers the Reorganization of present primary -cell cultures. However, we cannot Actin Cytoskeleton in Insulin-Producing Cells exclude that additional lipid-induced changes, including altered gene transcription, may contribute to the 1-Deoxysphinganine treatment selectively altered cytoskel- phenotype. eton organization both in Ins-1 cells and in primary islets, inducing accumulation of filamentous actin in intracellular 1-Deoxy-dihydroceramide Contributes to punctated structures juxtaposed to insulin-containing 1-Deoxysphinganine-Induced Cytotoxicity vesicles. Similar but transient actin fiber alterations have Ceramide is a key intracellular signaling molecule in- been observed previously in 1-deoxysphinganine-treated volved in several cellular functions, including cell death Vero cells (7). In addition, 1-deoxysphinganine impaired (27). Importantly, both cell-permeant analogs of ceram- cytoskeleton dynamics in sensory and motoneurons ide (28) and de novo ceramide synthesis (29) impair in- without inducing cell death (10). The present study sulin secretion and mitogenesis in pancreatic b-cells and shows not only that the lipid induced actin rearrange- induce apoptosis (30), supporting a critical regulatory ments before the appearance of apoptotic markers but role for ceramide in the metabolic dysfunction of these also that inhibitors shown to mitigate cytotoxicity did cells. In the search for the molecular signaling generated not prevent cytoskeletal remodeling. These data suggest by 1-deoxysphinganine treatment, we explored the hy- that cytoskeleton rearrangements are a direct effect of pothesis that 1-deoxysphinganine uptake and conversion 1-deoxysphinganine incubation rather than a conse- to 1-deoxy-dihydroceramide is necessary to exert its quence of cell lethality and that cytoskeleton and cell toxicity. Our mass spectrometry and inhibitor analyses cycle effects are regulated by distinct signaling pathways. support this hypothesis. In addition, 1-deoxysphinganine In this context, whether alterations in Rho-GTPase acti- increased the expression of CerS5, whereas b-cell lipo- vation, reported by Cuadros et al. (7) and in the present toxicity resulting from palmitate treatment stimulated study, are the key conserved molecular mechanisms also the synthesis of CerS4 (31), suggesting that different behind cytoskeletal alteration in neurons is currently lipids stimulate specific CerS isoforms. Collectively, the under investigation. results suggest that some of the cytotoxic effects of 1338 Deoxysphingolipids and Diabetes Diabetes Volume 63, April 2014

1-deoxysphinganine occur after its intracellular uptake that patients with hereditary sensory and autonomic and metabolism to 1-deoxy-dihydroceramide. However, neuropathy type I and metabolic syndrome have elevated we cannot exclude that 1-deoxysphinganine triggers death 1-deoxysphinganine levels without overt diabetes, sug- receptors on the cell surface. Given the similarity of the gests that the raised amount of 1-deoxysphinganine ob- molecular structure of 1-deoxysphinganine and sphingo- served in vivo is not sufficient to directly induce b-cell sine, it is worthy of further investigation to determine failure but would require additional pathological param- whether 1-deoxysphinganine can engage or antagonize the eters, such as an established chronic hyperglycemic state, same membrane receptors, as previously suggested (8). to promote b-cell toxicity (39). In this context, targeting 1-deoxySL synthesis as a combination therapeutic strat- 1-Deoxysphinganine Activates Multiple Intracellular egy for T2DM warrants further investigation. Pathways In addition to cytoskeletal remodeling, 1-deoxysphinganine treatment induced a complex dose-dependent pat- Acknowledgments. The authors thank Heidi Seiler for technical as- sistance and Amedeo Caflisch for providing the Birb796 inhibitor. tern of toxicity in Ins-1 cells characterized by the appearance of p21-induced senescence at low doses Funding. This research was supported by the Zurich Center for Integrative and apoptosis and necrosis at high doses. Of note, the Human Physiology and Rare Disease Initiative Zurich, Clinical Research Priority Program for Rare Diseases, University of Zurich. senescence growth arrest was limited to replicating cells, fl as demonstrated previously (32–34), and quiescent pri- Duality of Interest. No potential con icts of interest relevant to this mary islets were devoid of senescence markers upon lipid article were reported. treatment. These data imply that 1-deoxysphinganine Author Contributions. R.A.Z., T.H., A.O., A.B.H., H.B., T.G., O.O.O., triggers multiple signaling pathways. Indeed, the lipid E.S., K.G., J.-H.J., U.U., Y.W., A.v.E., R.G., and S.S. contributed to the study was shown to selectively activate JNK, MAPK, extracel- design; data acquisition, analysis, and interpretation; and drafting and critical revision of the manuscript. R.A.Z., T.H., A.O., A.B.H., H.B., T.G., O.O.O., E.S., lular signal–related kinase 1/2, and protein kinase C but K.G., J.-H.J., U.U., and Y.W. researched data and reviewed and edited the not AKT in NIH-3T3, RH-7777, PC-3, and LNCaP cell manuscript. A.v.E. and R.G. contributed to the discussion and reviewed and lines (8,9), whereas we show that JNK, p38 MAPK edited the manuscript. S.S. researched data and wrote the manuscript. S.S. is phosphorylation, and AKT levels increased in Ins-1 cells, the guarantor of this work and, as such, had full access to all the data in the indicating that the intracellular signaling effectors stim- study and takes responsibility for the integrity of the data and the accuracy of ulated by 1-deoxysphinganine depend on the cellular the data analysis. context. Because JNK and MAPK are known to be acti- Prior Presentation. Parts of this study were presented at the 45th vated by intracellular ceramide (35), it is possible that the Annual Meeting of the European Pancreatic Club, Zurich, Switzerland, 26–29 increased 1-deoxy-dihydroceramide synthesis observed June 2013. in Ins-1 cells activates these kinases. In addition, the fact that the CerS inhibitor FB1 was the only compound able References to rescue the toxicity of high 1-deoxysphinganine doses 1. Chen L, Magliano DJ, Zimmet PZ. The worldwide epidemiology of type 2 suggests that increased ceramide (or its deoxy form) diabetes mellitus—present and future perspectives. Nat Rev Endocrinol synthesis is upstream of or a prerequisite for the acti- 2012;8:228–236 vation of different signaling effectors. 2. Ashcroft FM, Rorsman P. Diabetes mellitus and the b cell: the last ten Although a more in-depth analysis of the set of years. Cell 2012;148:1160–1171 kinases activated by 1-deoxysphinganine is needed to 3. Prentki M, Nolan CJ. Islet beta cell failure in type 2 diabetes. J Clin Invest elucidate the precise downstream signaling cascade, the 2006;116:1802–1812 present inhibitor studies reveal an intriguing antagonis- 4. Donath MY, Ehses JA, Maedler K, et al. Mechanisms of beta-cell death in tic role of JNK and p38 MAPK in the context of type 2 diabetes. Diabetes 2005;54(Suppl. 2):S108–S113 1-deoxysphinganine-induced senescence. Of note, a simi- – 5. Othman A, Rütti MF, Ernst D, et al. Plasma deoxysphingolipids: a novel lar dose-dependent senescence apoptosis transition and class of biomarkers for the metabolic syndrome? Diabetologia 2012;55: opposite effect of p38 MAPK and JNK on senescence has 421–431 been reported in endothelial progenitor cells upon 6. Bertea M, Rütti MF, Othman A, et al. Deoxysphingoid bases as plasma doxorubicin treatment (36), further confirming that markers in diabetes mellitus. Lipids Health Dis 2010;9:84 different kinases activated by the same stimulus may exert opposite cellular effects. 7. Cuadros R, Montejo de Garcini E, Wandosell F, Faircloth G, Fernández- In conclusion, this work shows that 1-deoxysphinganine Sousa JM, Avila J. The marine compound spisulosine, an inhibitor of cell fi treatment and its conversion to 1-deoxy-dihydroceramide proliferation, promotes the disassembly of actin stress bers. Cancer Lett – compromises the viability of insulin-producing cells 2000;152:23 29 through multiple pathways, indicating that similar to free 8. Salcedo M, Cuevas C, Alonso JL, et al. The marine sphingolipid-derived fatty acids (30), 1-deoxySLs induce lipotoxicity but with compound ES 285 triggers an atypical cell death pathway. Apoptosis 2007; higher efficiency [low micromoles per liter for 1-deoxySLs 12:395–409 vs. low millimoles per liter for free fatty acids (31,37,38)]. 9. Sánchez AM, Malagarie-Cazenave S, Olea N, Vara D, Cuevas C, Díaz- However, the present in vivo study, together with the fact Laviada I. Spisulosine (ES-285) induces prostate tumor PC-3 and LNCaP diabetes.diabetesjournals.org Zuellig and Associates 1339

cell death by de novo synthesis of ceramide and PKCzeta activation. Eur 25. Tomas A, Yermen B, Min L, Pessin JE, Halban PA. Regulation of pancreatic J Pharmacol 2008;584:237–245 beta-cell insulin secretion by actin cytoskeleton remodelling: role of gel- 10. Penno A, Reilly MM, Houlden H, et al. Hereditary sensory neuropathy type 1 solin and cooperation with the MAPK signalling pathway. J Cell Sci 2006; – is caused by the accumulation of two neurotoxic sphingolipids. J Biol Chem 119:2156 2167 2010;285:11178–11187 26. Howell SL, Tyhurst M. Interaction between insulin-storage granules and – 11. Asfari M, Janjic D, Meda P, Li G, Halban PA, Wollheim CB. Establishment of F-actin in vitro. Biochem J 1979;178:367 371 2-mercaptoethanol-dependent differentiated insulin-secreting cell lines. 27. Zheng W, Kollmeyer J, Symolon H, et al. Ceramides and other bioactive Endocrinology 1992;130:167–178 sphingolipid backbones in health and disease: lipidomic analysis, 12. Hohmeier HE, Mulder H, Chen G, Henkel-Rieger R, Prentki M, Newgard CB. metabolism and roles in membrane structure, dynamics, signaling and Isolation of INS-1-derived cell lines with robust ATP-sensitive K+ channel- autophagy. Biochim Biophys Acta 2006;1758:1864–1884 dependent and -independent glucose-stimulated insulin secretion. Diabe- 28. Sjöholm A. Ceramide inhibits pancreatic beta-cell insulin production and tes 2000;49:424–430 mitogenesis and mimics the actions of interleukin-1 beta. FEBS Lett 1995; 13. Cavallari G, Zuellig RA, Lehmann R, Weber M, Moritz W. Rat pancreatic 367:283–286 islet size standardization by the “hanging drop” technique. Transplant Proc 29. Kelpe CL, Moore PC, Parazzoli SD, Wicksteed B, Rhodes CJ, Poitout V. 2007;39:2018–2020 Palmitate inhibition of insulin gene expression is mediated at the tran- 14. Silva A, Weber A, Bain M, et al. COX-2 is not required for the development scriptional level via ceramide synthesis. J Biol Chem 2003;278:30015– of murine chronic pancreatitis. Am J Physiol Gastrointest Liver Physiol 30021 2011;300:G968–G975 30. Lang F, Ullrich S, Gulbins E. Ceramide formation as a target in beta-cell 15. Antoniou X, Gassmann M, Ogunshola OO. Cdk5 interacts with Hif-1a in survival and function. Expert Opin Ther Targets 2011;15:1061–1071 neurons: a new hypoxic signalling mechanism? Brain Res 2011;1381: 31. Véret J, Coant N, Berdyshev EV, et al. Ceramide synthase 4 and de novo 1–10 production of ceramides with specific N-acyl chain lengths are involved in 16. Rodier F, Campisi J. Four faces of cellular senescence. J Cell Biol 2011; glucolipotoxicity-induced apoptosis of INS-1 b-cells. Biochem J 2011;438: 192:547–556 177–189 17. Romanov VS, Pospelov VA, Pospelova TV. Cyclin-dependent kinase inhib- 32. Michaloglou C, Vredeveld LC, Soengas MS, et al. BRAFE600-associated itor p21(Waf1): contemporary view on its role in senescence and onco- senescence-like cell cycle arrest of human naevi. Nature 2005;436:720–724 genesis. Biochemistry (Mosc) 2012;77:575–584 33. Braig M, Lee S, Loddenkemper C, et al. Oncogene-induced senescence as 18. Kaplon J, Zheng L, Meissl K, et al. A key role for mitochondrial gatekeeper an initial barrier in lymphoma development. Nature 2005;436:660–665 pyruvate dehydrogenase in oncogene-induced senescence. Nature 2013; 34. Collado M, Serrano M. Senescence in tumours: evidence from mice and 498:109–112 humans. Nat Rev Cancer 2010;10:51–57 19. Lloyd-Evans E, Platt FM. Lipids on trial: the search for the offending 35. Ruvolo PP. Intracellular signal transduction pathways activated by ce- metabolite in Niemann-Pick type C disease. Traffic 2010;11:419–428 ramide and its metabolites. Pharmacol Res 2003;47:383–392 20. Ikonen E, Hölttä-Vuori M. Cellular pathology of Niemann-Pick type C dis- 36. Spallarossa P, Altieri P, Barisione C, et al. p38 MAPK and JNK antago- ease. Semin Cell Dev Biol 2004;15:445–454 nistically control senescence and cytoplasmic p16INK4A expression in 21. Heasman SJ, Ridley AJ. Mammalian Rho GTPases: new insights into their doxorubicin-treated endothelial progenitor cells. PLoS ONE 2010;5:e15583 functions from in vivo studies. Nat Rev Mol Cell Biol 2008;9:690–701 37. Lupi R, Dotta F, Marselli L, et al. Prolonged exposure to free fatty acids has 22. Genuth SM, Przybylski RJ, Rosenberg DM. Insulin resistance in genetically cytostatic and pro-apoptotic effects on human pancreatic islets: evidence obese, hyperglycemic mice. Endocrinology 1971;88:1230–1238 that beta-cell death is caspase mediated, partially dependent on ceramide – 23. Höppener JW, Oosterwijk C, Nieuwenhuis MG, et al. Extensive islet amyloid pathway, and Bcl-2 regulated. Diabetes 2002;51:1437 1442 formation is induced by development of type II diabetes mellitus and 38. Baldwin AC, Green CD, Olson LK, Moxley MA, Corbett JA. A role for contributes to its progression: pathogenesis of diabetes in a mouse model. aberrant protein palmitoylation in FFA-induced ER stress and b-cell death. Diabetologia 1999;42:427–434 Am J Physiol Endocrinol Metab 2012;302:E1390–E1398 24. Nevins AK, Thurmond DC. Glucose regulates the cortical actin network 39. Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC. Beta-cell through modulation of Cdc42 cycling to stimulate insulin secretion. Am deficit and increased beta-cell apoptosis in humans with type 2 diabetes. J Physiol Cell Physiol 2003;285:C698–C710 Diabetes 2003;52:102–110