ARTICLE

Lysosomal acid lipase and lipophagy are constitutivenegative regulators of glucose-stimulated insulin secretionfrom pancreatic beta cells

Gemma L. Pearson & Natalie Mellett & Kwan Yi Chu & James Cantley & Aimee Davenport &Pauline Bourbon & Casey C. Cosner & Paul Helquist & Peter J. Meikle & Trevor J. Biden

Received: 17 July 2013 /Accepted: 26 September 2013 /Published online: 23 October 2013# Springer-Verlag Berlin Heidelberg 2013

AbstractAims/hypothesis Lipolytic breakdown of endogenous lipidpools in pancreatic beta cells contributes to glucose-stimulated insulin secretion (GSIS) and is thought to be me-diated by acute activation of neutral lipases in the amplifica-tion pathway. Recently it has been shown in other cell typesthat endogenous lipid can be metabolised by autophagy, andthis lipophagy is catalysed by lysosomal acid lipase (LAL).This study aimed to elucidate a role for LAL and lipophagy inpancreatic beta cells.Methods We employed pharmacological and/or genetic inhi-bition of autophagy and LAL in MIN6 cells and primaryislets. Insulin secretion following inhibition was measuredusing RIA. Lipid accumulation was assessed by MS andconfocal microscopy (to visualise lipid droplets) and autoph-agic flux was analysed by western blot.Results Insulin secretion was increased following chronic(≥8 h) inhibition of LAL. This was more pronounced withglucose than with non-nutrient stimuli and was accompaniedby augmentation of neutral lipid species. Similarly, following

inhibition of autophagy in MIN6 cells, the number of lipiddroplets was increased and GSIS was potentiated. Inhibitionof LAL or autophagy in primary islets also increased insulinsecretion. This augmentation of GSIS following LAL or au-tophagy inhibition was dependent on the acute activation ofneutral lipases.Conclusions/interpretation Our data suggest that lysosomallipid degradation, using LAL and potentially lipophagy, con-tributes to neutral lipid turnover in beta cells. It also serves as aconstitutive negative regulator of GSIS by depletion of sub-strate for the non-lysosomal neutral lipases that are activatedacutely by glucose.

Keywords Autophagy . Beta cell . Insulin secretion . Islet .

Lipidmetabolism . Lipophagy . Lysosomal acid lipase

AbbreviationsAPG AutophagosomeATG Autophagy-related genesATGL Adipose triglyceride lipaseBA Bafilomycin ABCA Bicinchoninic acidCE Cholesteryl esterCQ ChloroquineGSIS Glucose-stimulated insulin secretionHBSS Hanks’ buffered saline solutionHSL Hormone-sensitive lipaseKRHB Krebs–Ringer HEPES bufferLAL Lysosomal acid lipaseLC3 Microtubule-associated protein light chain 33MA 3-MethyladeninePFA ParaformaldehydesiRNA Small-interfering RNATG Triacylglycerol

G. L. Pearson :K. Y. Chu : J. Cantley :A. Davenport :T. J. Biden (*)Diabetes and Obesity Department, Garvan Institute of MedicalResearch, 384 Victoria Street, Darlinghurst,Sydney, NSW 2010, Australiae-mail: t.biden@garvan.org.au

N. Mellett : P. J. MeikleThe Baker IDI Heart and Diabetes Institute, Melbourne, Australia

P. Bourbon : C. C. Cosner : P. HelquistDepartment of Chemistry and Biochemistry,University of Notre Dame, Notre Dame, IN, USA

Diabetologia (2014) 57:129–139DOI 10.1007/s00125-013-3083-x

Introduction

Glucose-stimulated insulin secretion (GSIS) comprises twoseparate but interconnected metabolic pathways in the betacell: the initiation pathway, which is K+

ATP channel dependentand results in increased calcium influx [1, 2], and the ampli-fication pathway, which potentiates insulin secretion indepen-dently of further changes in cytosolic calcium [3, 4]. Themechanisms underlying amplification are much less well un-derstood than those of initiation [5–8] but the turnover ofendogenous lipid stores, and consequent generation of a lipidsignalling molecule, is one possibility [5, 6, 9, 10]. Evidencefor this includes the inhibition of GSIS by deletion of neutrallipases, such as hormone-sensitive lipase (HSL) [11, 12] andadipose triglyceride lipase (ATGL) [13], or their pharmaco-logical blockade using the pan lipase inhibitor orlistat [14]. Inaddition, at least one substrate of the neutral lipases, triacyl-glycerol (TG), is hydrolysed during GSIS, as witnessed by anacute glucose-stimulated, and orlistat-inhibited, release ofglycerol from beta cells [9, 15, 16].

In contrast to the established roles of neutral lipases ininsulin secretion, the potential involvement of lysosomal acidlipase (LAL) has largely been unaddressed. Historically, thechief function of this enzyme, as established in other celltypes, was considered to be the breakdown of cholesteryl ester(CE) delivered to the cell via the LDL pathway [17, 18].However, LAL has also been implicated in the turnover ofendogenous lipid stores (lipid droplets) via the process oflipophagy, first established in hepatocytes in 2009 [19].Lipophagy is a specific form of autophagy, the process bywhich macromolecules are surrounded by the formation of anautophagosome (APG) and then degraded by fusion of theAPG with the lysosome. This process depends on a set ofautophagy-related genes (ATGs), which are responsible forthe production of the double-membrane APG structure [20,21]. Of the associated proteins, only ATG8/microtubule-associated protein light chain 3 (LC3)-II, produced fromlipidation of a precursor (LC3-I), is known to be associatedwith mature APGs and therefore serves as a marker for theirformation [21]. Lipophagy is now recognised as an alternativemethod of lipid hydrolysis, whereby lipid droplets aresurrounded by an APG and then delivered to the lysosomefor degradation [19, 22]. It has been shown in macrophagesthat LAL is the enzyme responsible for the breakdown of TGand CE delivered to the lysosome via this route [23]. In liver,LAL/lipophagy makes a quantitatively important contributionto overall TG degradation, especially following high-fatfeeding [24, 25].

This link between autophagy and lipid degradation led usto investigate whether a lysosomal pathway of lipid break-down, in concert with neutral lipid hydrolysis, might contrib-ute to the regulation of insulin secretion. In addressing theseissues we now demonstrate for the first time that LAL and

lipophagy are key negative regulators of a pool of neutral lipidin beta cells, which is mobilised as an amplification signalduring GSIS.

Methods

Reagents All tissue culture media, supplements and trypsin forMIN6 cells and isolated islets were from Gibco (Gaithersburg,MD, USA). Protease inhibitor tablets were from Roche Diag-nostics (Penzburg, Germany). Insulin RIA kits were fromLinco/Millipore (Billerica, NJ, USA). Phenylmethylsulfonylfluoride (PMSF), sodium orthovanadate, DMSO, fatty acidfree fraction V BSA and L-glucose, were from Sigma-Aldrich (St Louis, MO, USA). TaqMan reverse transcriptionkit and TaqMan PCR probes were from Applied Biosystems(Mulgrave, VIC, Australia). ATG7 (NM_028835) and LAL(NM_021460) ON-TARGETplus SMARTpool siRNA,control Non-Targeting siRNA and Dharmafect TransfectionReagent 3 were from Dharmacon (Pittsburgh, PA, USA).RNeasy kit was from Qiagen (Valencia, CA, USA). TheBCA (bicinchoninic acid assay) protein assay kit was fromPierce (Rockford, IL, USA). The pre-cast NUPAGE gels,sample buffer, reducing agent, antioxidant and electrophoresistank were from Invitrogen (Carlsbad, CA, USA). The transfersystem for immunoblotting and the protein standard markerswere from BioRad (Hercules, CA USA). The 96-well conical-bottom plates for islet insulin secretion assays were fromGrenier Bio-one (Frickenhausen, Germany). Antibodies forimmunoblotting/immunofluorescence were as follows:Anti-14-3-3 from Santa-Cruz (Santa Cruz, CA, USA),anti-ATG7 from Cell Signalling Technologies (Danvers, MA,USA), anti-LAL from Novus Biologicals (Littleton, CO,USA) and anti-LC3 from Sigma-Aldrich.

Cell culture and treatments The pancreatic beta cell lineMIN6 [26] was used at passages 26–35, as previously de-scribed [27]. Cells were grown at 37°C and 5%CO2 in DMEM(25 mmol/l glucose) supplemented with 10% FCS, 10 mmol/lHEPES, 50 U/ml of penicillin and 50 μg/ml streptomycin.

For Lalistat treatment, cells were seeded at 3×105 cells perwell in 24-well plates or 4×105 cells per well in 12-wellplates, and treated at varying doses and times as indicated,or with 1:2,000 DMSO as control.

For small-interfering RNA (siRNA) transfection MIN6cells were seeded in 12-well plates, as above. LAL ON-TARGETplus SMARTpool siRNA, ATG7 ON-TARGETplusSMARTpool siRNA or control non-targeting siRNAwere transfected using Dharmafect3 transfection reagent.siRNA constructs were present for 24 h and the media wasthen changed to complete DMEM (25 mmol/l glucose) andcells were incubated for a further 48 h. This achieved 70–80%transfection efficiency as per the manufacturer’s instructions.

130 Diabetologia (2014) 57:129–139

Cell death was determined, as previously described [28],using a Cell Death Detection ELISA (Roche Diagnostics,NSW, Australia).

Islet isolation and insulin secretion assays Islets were isolatedfrom male C57Bl6 mice as previously described [16]. Afterpancreatic digestion, islets were purified and incubated over-night in RPMI 1640 medium (11 mmol/l glucose) supple-mented with 10% FCS, 0.2 mmol/l glutamine, 10 mmol/lHEPES, 50 U/ml of penicillin and 50 μg/ml streptomycin.For inhibitor studies, islets were cultured for a further 48 hwith 5 μmol/l Lalistat, 5 mmol/l 3-methyladenine (3MA) orvehicle control, and the media was changed every 24 h.Batches of five islets were picked, with at least six replicatesper group, and insulin secretion assays were carried out asfollows: islets or MIN6 cells were preincubated for 1 h inKrebs–Ringer HEPES buffer (KRHB) containing 0.1%(wt/vol.) BSA and 2 mmol/l glucose. They were then incu-bated for 1 h at 37°C with KRHB containing either 2 or20 mmol/l glucose. An aliquot of the buffer was taken andinsulin release was measured by RIA.

Confocal microscopy MIN6 cells were seeded onto glasscoverslips at 2×105 cells per well, and grown in DMEM(25 mmol/l glucose) for 24 h. For Nile Red staining, cover-slips were washed three times with 1 ml PBS pre-warmed to37°C. Cells were fixed in 4% paraformaldehyde (PFA) for20 min at room temperature. The PFA was removed,100 ng/ml Nile Red in 37°C HBSS (Hanks’ buffered salinesolution) was added and the cells were incubated in the darkfor 15 min at room temperature. Staining solution was re-moved and the coverslips were washed three times with 1 mlPBS (37°C). Coverslips were mounted using Prolong Goldwith DAPI (Molecular Probes, Carlsbad, CA, USA) andsealed with nail polish. Slides were left to dry in thedark for at least 3 h and then visualised using confocalmicroscopy. For LAL and BODIPY 493/503 (MolecularProbes) co-staining, following washing and fixation asabove, coverslips were then blocked for 45 min with3% (wt/vol.) BSA and incubated with anti-LAL antibody (1:1,000) overnight at 4°C. Cells were then washed three timeswith PBS and incubated with anti-rabbit Alexafluor 647(Molecular Probes) and 1 μg/ml BODIPY 493/503 for 1 h atroom temperature, protected from light. Coverslips were thenmounted as described above.

Lysosomal function assays MIN6 cells were treated for 24 hwith Lalistat, bafilomycin A (BA) or chloroquine (CQ) andthe following assays were performed.

Lysosomal pH was measured by staining with 1 μmol/lLysoSensor Yellow/Blue DND-160 (Invitrogen) for 5 minfollowing treatments. The cells were then washed with warmmedium and examined under a Leica DMI 6000 SP8 confocal

microscope using 405 nm excitation wavelength with 450 or530 nm emission filters. The 530:450 nm emission ratio wascalculated using the software Photoshop CS4 (Adobe SystemsIncorporated, San Jose, CA, USA).

Lysosomal proteolysis was measured using DQ-BSA(Molecular Probes), as per reference [29]. Briefly, cells wereseeded onto coverslips, incubated overnight with DQ-BSA(10 μg/ml) and then treated with inhibitors (24 h). Cells werethen washed in PBS, fixed with 2% PFA and mounted usingProlong Gold with DAPI (Molecular Probes). Fluorescentparticles were visualised by confocal microscopy and quanti-fied using Image J software (National Institute of Health,Bethesda, MD, USA).

RNA analysis Total RNA was extracted from MIN6 cellsusing an RNeasy minikit and cDNA synthesis was performedon 1.5 μg of DNA, using the TaqMan reverse transcrip-tion kit. Real-time PCR was carried out on an HT-7900PCR machine using TaqMan probes. The following probeswere used: Lipe/Hsl (Mm00495359_m1), Pnpla2/Atgl(Mm00503040_m1), Pnpla3/Adpn (Mm00504420_m1), Lpl(Mm00434770_m1) and Lipa/Lal (Mm00498820). Analysiswas carried out using the standard curve method and relativegene expression was normalised to Tbp (Mm00446971) as anendogenous control.

Western blotting Protein extracts from MIN6 cells were re-solved on a pre-cast 12% SDS-PAGE gel and transferred topolyvinylidene fluoride membrane. After blocking for 1 h in5% skimmed milk, membranes were probed with antibodiesovernight at 4°C (anti-ATG7, anti-LAL and anti-LC3) andnormalised to a loading control (anti-pan 14-3-3). Band quan-tification was undertaken using ImageJ.

Lipidomic profiling using MS Following 24 h control orLalistat treatment MIN6 cells were pelleted in ice-cold PBS(two wells of a six-well plate per condition) and subjected tolipid MS as previously described [28, 30].

Statistical analysis All data are expressed as mean ± SEM.All statistics were performed using GraphPad Prism5 software(GraphPad Software, La Jolla, CA, USA), and subjected totwo-way, one-way or repeated measures ANOVA (with theBonferroni post hoc test), paired Student’s (two-tailed) t testor unpaired Student’s (two-tailed) t test.

Results

LAL is expressed in pancreatic beta cells Several lipasesare expressed in beta cells [13, 16], although Hsl/Lipe andAtgl/Pnpla2 have generally been considered the most abun-dant. There is, however, residual lipase activity in mouse

Diabetologia (2014) 57:129–139 131

models of HSL and ATGL depletion, so additional enzymesmay contribute to lipid hydrolysis in pancreatic beta cells [11,13]. To investigate lipase expression in the mouse pancreaticbeta cell line, MIN6, we carried out RT-PCR (Fig. 1a). LalmRNAwas present in MIN6 cells and LAL protein was alsoabundantly expressed in MIN6 cells, mouse islets and thehuman liver cell line HEPG2 (Fig. 1b). To determine whetherthis enzyme plays a functional role in beta cells we employedsiRNA to knockdown LAL protein (Fig. 1c) by 28.8±0.04%compared with control siRNA transfected cells (p <0.05, one-sample t test). No increase in insulin secretion was observedwhen correcting for protein content (Fig. 1d), although weobserved a significant decrease in total insulin content follow-ing Lal siRNA treatment (Fig. 1e). When this decrease wastaken into consideration there was a significant increase inGSIS but not in basal insulin secretion (Fig. 1f).

Inhibition of LAL increases insulin secretion To characterisethe role of LAL in beta cell function more comprehensively,and to avoid confounding effects of incomplete gene knock-down, we employed Lalistat, an inhibitor that is highlyspecific for LAL over neutral lipases [23, 31, 32]. Lalistatpretreatment for 24 h dose dependently increased insulinsecretion from MIN6 cells (Fig. 2a) during a subsequentglucose challenge. However, Lalistat was without effect whenonly present for 2 h during the glucose stimulation (Fig. 2b).This contrasts with the function of neutral lipases, whichregulate insulin secretion acutely [14]. From more extensivetime course analyses we established that 5 μmol/l Lalistat was

4

3

2

1

0

b

In

sulin

sec

retio

n (

pg [n

g co

nten

t]-1 h

-1)

d

2 mmol/l glucose

20 mmol/l glucose

*

MIN6Mouseislets HepG2

LALLoadingcontrol

In

sulin

sec

retio

n (

ng [µ

g pr

otei

n]-1

h-1

)

f

2 mmol/lglucose

20 mmol/lglucose

200

150

100

50

0

4

3

2

1

0

25

20

15

10

5

0

a

Lipa

/La

l

Lipe

/H

slP

npla

3/A

dpn

Lpl

mR

NA

exp

ress

ion

(rel

ativ

e un

its)

LAL

Loadingcontrol

ControlsiRNA

LalsiRNA

c

Pnp

la2/

Atg

l

Insu

lin (

ng/µ

g pr

otei

n)

e

**

ControlsiRNA

LalsiRNA

Fig. 1 Characterisation of LAL expression and function in pancreaticbeta cells. (a) RT-PCR analysis of lipases in MIN6 cells. (b) Western blotanalysis of MIN6 cells and mouse islets showing expression of LALprotein; HepG2 cells were included as a positive control for the antibody.(c) Representative western blot image for LAL protein and loadingcontrol, following transfection of a control pool of siRNA or Lal siRNA.(d) Insulin secretion fromMIN6 cells, treated with control (white bars) orLal (black bars) siRNA, after 2 or 20 mmol/l glucose stimulation (n =4independent experiments). (e) Insulin content of MIN6 cells followingcontrol (white bar) or Lal (black bar) siRNA treatment (n =4 independentexperiments; **p <0.01 Student’s unpaired t test). (f) Insulin secretionfromMIN6 cells following control (white bars) or Lal (black bars) siRNAtreatment, corrected for total insulin content (n =4 independent experi-ments; *p <0.05 Student’s unpaired t test)

a b

c

e

d

Lalistat (µmol/l)0

0

5

10

15

0

2

6

4

8

1 5 10 0

10

8

6

4

2

05 10

**

- +

8 h 24 h 48 h

+ - - ++ - - ++-

Glucose KCl

FSKDZX

*

**

ns

2 2 2 2 2 2 2 22020 2020- - - - + + + + + + + +- - - - - - - -+ + + +

++- - - - - - - - - -

0 1 5 10

** * *

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1)

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1)

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1)

Insu

lin (

ng/µ

g pr

otei

n)

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1)

Lalistat (µmol/l)

Lalistat (5µmol/l) Lalistat (µmol/l)

0

15

30

45

15

10

5

0

Fig. 2 Inhibition of LAL using Lalistat increases insulin secretion. (a)MIN6 cells were treated for 24 h with 1, 5 or 10 μmol/l Lalistat. Insulinsecretion over 1 h in response to 2 mmol/l (white bars) or 20 mmol/l(black bars) glucose is shown (n =3 experiments; *p <0.05 one-wayANOVA with Bonferroni post hoc test). (b ) Insulin secretion fromMIN6 cells treated for 2 h with 5 μmol/l Lalistat and 2 mmol/l glucose(white bars) or 20 mmol/l glucose (black bars) (n=3 experiments). (c)Time course of Lalistat treatment; insulin secretion in response to 2mmol/lglucose (white bars) and 20 mmol/l glucose (black bars) is shown (n=4experiments; *p<0.05, **p<0.01, Student’s paired t test). (d) Insulincontent ofMIN6 cells following various Lalistat doses (n=3 experiments).(e ) Insulin secretion in response to glucose, KCl, KCl + DZX andKCl + FSK after 24 h 5 μmol/l Lalistat (grey bars) or control treatment(white bars) (n =4 experiments; *p <0.05, **p <0.0,1 repeatedmeasures ANOVAwith Bonferroni post hoc test). DZX, diazoxide;FSK, forskolin

132 Diabetologia (2014) 57:129–139

effective by 8 h, and that this effect was sustained duringtreatment for at least 48 h (Fig. 2c). There was a slight, butnon-significant, increase in basal insulin release followingLalistat treatment (Fig. 2a, c). Insulin content was unaffectedat low Lalistat doses of 1 and 5 μmol/l although there was anon-significant increase observed at 10 μmol/l (Fig. 2d).Therefore the 5 μmol/l concentration was used thereafter toavoid content issues.

To investigate the amplification pathway [3, 4], MIN6 cellswere stimulated in the combined presence of diazoxide (tocircumvent metabolic closure of K+

ATP channels) and KCl (toprovide an independent depolarisation stimulus). Under theseconditions, raising glucose from 2 to 20mmol/l doubled insulinsecretion, indicative of the amplification pathway (Fig. 2e).Lalistat pretreatment slightly augmented the response to KClplus diazoxide at low glucose concentration but this was sig-nificantly further increased at high glucose concentration. Thuschronic LAL inhibition augments the amplification phase ofGSIS. We next employed forskolin, which can also potentiateCa2+-dependent secretion. Following Lalistat pretreatment,however, the increment in secretion, comparing KCl plusforskolin with KCl alone, was not statistically significant andless marked than the corresponding effect of 20mmol/l glucose(Fig. 2e). This is consistent with the mechanism of action offorskolin (raising cAMP), which is not thought to contribute tothe amplification pathway employed by glucose [3, 4]. Collec-tively these results suggest that LAL inhibition has a greaterimpact on the secretory response to glucose than non-nutrientsecretagogues, and acts in the amplification pathway.

Lalistat does not adversely affect general lysosomal functionin MIN6 cells To check whether chronic LAL inhibitionaffects general lysosomal function, we employed BA andCQ, which disrupt the lysosomal pH gradient. UsingLysoSensor pH probes, BA reduced the yellow/blue ratio,corresponding to the expected decrease in lysosomal acidity,but this was not observed after Lalistat treatment (Fig. 3a).Interestingly, Lalistat caused a significant increase in the totalfluorescence emission following LysoSensor administration,which was not observed with BA treatment (Fig. 3b), indica-tive of an increase in total lysosome number. To assess whetherLalistat altered lysosomal protein degradation, cells wereincubated with a fluorescently labelled BSA molecule(DQ-BSA), which when cleaved by proteolytic enzymes res-ident in the lysosome generates individual fluorescent particlesthat can be quantified to assess proteolysis. Lalistat treatmentdid not inhibit proteolytic activity but rather appeared toslightly increase the number of particles (Fig. 3c). This in-crease is most likely due to the increased lysosomal number(seen in Fig. 3b). Together, these results suggest that Lalistatdoes not cause general lysosomal disruption but rather specif-ically impacts on the lipid degradative capacity of the lyso-some, and seemingly increases lysosomal number.

We also measured insulin secretion from MIN6 cells inresponse to these general lysosomal inhibitors. Treatment withCQ completely abolished GSIS fromMIN6 cells (Fig. 3d), butincreasing doses of BA were without effect. There was alsodecreased insulin content following high-dose BA or CQtreatment (Fig. 3e) in contrast to Lalistat treatment, whichhad no effect or increased effect at high doses (Fig. 2d).Additionally, chronic Lalistat treatment had no effect on apo-ptosis in MIN6 cells, as measured by DNA fragmentationELISA (Fig. 3f), whereas CQ treatment significantly increasedapoptosis compared with no-treatment controls. Taken togeth-er, these data indicate that a general perturbation of lysosomalfunction is neither necessary nor sufficient for augmentingGSIS, suggesting that the effect of Lalistat is mediated morespecifically by inhibition of LAL.

Lalistat increases the neutral lipid pool of MIN6 cells Wenext determined the effect of LAL inhibition on lipid poolswithin MIN6 cells. Lipidomic analysis was carried out oncells treated with Lalistat for 24 h. Inhibiting LAL activityhad no effect on total ceramide, monohexosyl ceramide orsphingomyelin levels (Fig. 4a) or on major phospholipids(Fig. 4b). Total levels of the neutral lipids, TG, diacylglyeroland CE were, however, all significantly increased followingpretreatment (Fig. 4c).

Autophagy inhibit ion leads to increased insulinsecretion Because in other cell types, LAL has been implicat-ed in the process of lipophagy [23], we next assessed whetherinhibiting autophagy might recapitulate the augmentation ofinsulin secretion seen with Lalistat. We therefore knockeddown ATG7 protein, which is an important initiator of APGformation, in MIN6 cells (Fig. 5a). This was accompanied byreduced accumulation of the downstream effector proteinLC3-II (Fig. 5b), which reflects decreased APG formationcompared with control siRNA-treated cells. Under these treat-ment conditions, inhibition of autophagy significantly in-creased GSIS from MIN6 cells (Fig. 5c), similar to chronicexposure to Lalistat. Knockdown of ATG7 caused a signifi-cant increase in total insulin content in MIN6 cells (Fig. 5d),which was not necessarily unexpected when inhibiting proteinturnover by decreasing autophagy. However, when insulinsecretion was corrected for total insulin content there was stilla significant increase in GSIS from Atg7 siRNA-treated cellscompared with control siRNA-treated cells (Fig. 5e).

We next investigated any potential feedback of LAL inhi-bition on autophagy. Under steady-state conditions, Lalistattreatment had no effect on total ATG7, LAL or LC3-II : LC-Iratio (Fig. 6a, b). Autophagic flux was also evaluated usingMIN6 cells pre-exposed to Lalistat for 24 h and then acutelytreated with CQ to inhibit downstream degradation of LC3-II(Fig. 6c, d). Under control conditions CQ significantly in-creased the LC3-II to LC3-I ratio, as expected, but this was

Diabetologia (2014) 57:129–139 133

attenuated by Lalistat. These results suggest some crosstalkbetween LAL inhibition and autophagy, albeit not complete.

Inhibition of LAL and autophagy promotes lipid-dropletaccumulation Our lipidomic results (Fig. 4c) demonstrated aselective increase in neutral lipids following Lalistat treat-ment, and these lipids would be expected to be stored inpancreatic beta cells as lipid droplets [33–35]. Thus untreatedMIN6 cells contained relatively few lipid droplets, but these

were increased following 24 h Lalistat treatment (Fig. 7a–c).Moreover, inhibition of autophagy by knocking down ATG7increased neutral lipid storage (Fig. 7d–f). Compared withcontrol siRNA-treated cells,Atg7 siRNA increased lipid drop-lets approximately twofold (Fig. 7f). This indicates thatinhibiting lysosomal lipid degradation, by blocking eitherLAL or autophagy, increases neutral lipid storage.

We also investigated the site of this lipid accumulation bycomparing staining of LAL (as a marker of lysosomes) and

a2.5 40 10

8

6

4

2

0

30

20

10

0

d

BA (nmol/l) CQ (µmol/l)

0 0.5 5 0 0

0 0 0 5 50

c

Yello

w/b

lue

ratio

(A

U)

Notreatment

Lalistat BA

**

Notreatment

Lalistat CQ

*

Flu

ores

cent

par

ticle

s

Notreatment

Lalistat BA

*** ***

Tota

l flu

ores

ence

(A

U)

b

BA (nmol/l) CQ (µmol/l)

- 0.5 5 - -

- - - 5 50

e

Insu

lin(n

g/µg

pro

tein

) *

Notreatment

Lalistat CQ BAFol

d ch

ange

vs

no tr

eatm

ent

f

*

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1) 2.0 30 5

4

3

2

1

0

20

10

0

0.5

1.5

0.0

1.5

1.00.4

0.2

Fig. 3 Lalistat treatment does not adversely affect lysosomal function.MIN6 cells were treated for 24 h with or without 5 μmol/l Lalistat, orovernight with BA (5 nmol/l) or CQ (5 μmol/l). (a) Lysosomal aciditywas measured using LysoSensor (n =3 independent experiments,visualising 26 fields of view; **p <0.01, one-way ANOVA, Bonferronipost hoc test). AU, arbitrary units. (b) Total blue and yellow fluorescencefrom the LysoSensor assay was analysed (n =3 independent experiments,visualising 26 fields of view per condition; ***p <0.001 one-wayANOVA, Bonferroni post hoc test). (c) General proteosomal degradationwas measured using fluorescently labelled DQ-BSA (n=3 independent

experiments, visualising 93–152 cells per condition: *p<0.05 one-wayANOVA, Bonferroni post hoc test). (d) Insulin secretion results followingBA (0.5 or 5 nmol/l) and CQ (5 or 50 μmol/l) treatment with 2 mmol/l(white bars) or 20 mmol/l (black bars) glucose stimulation (n=3 exper-iments). (e) Insulin content of BA- or CQ-treated MIN6 cells (n =3experiments): *p <0.05, Student’s unpaired t test. (f) Apoptosis followingLalistat, CQ or BA treatment. Results are expressed as fold change vs notreatment, corrected for DNA content (n =3 experiments; *p <0.05one-sample t test)

COHCEPC PE PI PS DG TG

*

***

*

a

Lipi

d (m

mol

/mg

prot

ein)

CerSM MHC

b c

Lipi

d (m

mol

/mg

prot

ein)

Lipi

d (m

mol

/mg

prot

ein)30

20

10

0

25015050201612

210

1501005020

10

1.5

0.5

Fig. 4 Inhibition of LAL causes changes to the lipid profile of MIN6cells. MIN6 cells were treated for 24 h with 5 μmol/l Lalistat (black bars)or control DMSO (white bars). The lipids were extracted and thensubjected to MS, as described in the Methods. Results are expressed asmmol lipid per mg protein (n =4 independent experiments; *p <0.05,**p <0.01, ***p <0.001, Student’s unpaired t test). (a ) Total

sphingolipids: amount of total sphingomyelin (SM), ceramide (Cer) andmonohexosyl ceramide (MHC). (b) Total phospholipids: amount of totalphosphatidylcholine (PC), phosphatidylethanolamine (PE), phos-phatidylinositol (PI) and phosphatidylserine (PS). (c) Total neutral lipids:amount of total diacylglycerol (DG), TG, CE and non-esterified choles-terol (COH)

134 Diabetologia (2014) 57:129–139

BODIPY 493/503 (to stain neutral lipid). Under control con-ditions the few observed lipid bodies (green staining, Fig. 7g),did not localise with lysosomes (in red). Upon Lalistat treat-ment, the number of lipid bodies increased, but they remainedoutside of the lysosome (Fig. 7h), consistent with cytosoliclipid droplets.

Inhibition of LAL and autophagy in primary islets alsoincreases insulin secretion We next sought to confirm theroles of LAL and autophagy in regulating GSIS in primarybeta cells. Pretreatment of isolated mouse islets with Lalistatfor 48 h also caused a significant increase in insulin secretion(Fig. 8a), with no effect on total insulin content (Fig. 8b).Moreover, this increased GSIS due to chronic exposure to

Lalistat was completely abolished when orlistat (a pan lipaseinhibitor) was present acutely during the glucose stimulation.We next demonstrated that inhibiting autophagy in isolatedmouse islets, using the chemical inhibitor 3MA, also increasedinsulin secretion (Fig. 8c), despite reducing insulin content(Fig. 8d). Interestingly, the increment in GSIS due to 3MAwas also partially inhibited by orlistat (Fig. 8c). These resultstogether suggest that inhibiting LAL or autophagy augmentsinsulin secretion via a similar mechanism, most likely byaugmenting a neutral lipid pool, which is acutely mobilisedduring GSIS via an orlistat-sensitive lipase.

Discussion

Lipophagy, a specialised form of autophagy in which lipiddroplets are degraded via lysosomal fusion, has been recentlydescribed in several tissues [19, 23, 36]. It appears to functionas a means of mobilising lipid stores in times of nutrientdemand [37], or as a means of helping to regulate cellularcholesterol homeostasis [23]. LAL has recently been impli-cated as the important metabolising enzyme in this process[23, 38]. We now provide the first evidence of a functional

*

ControlsiRNA

Atg7siRNA

20

10

0

30

40

* *

ATG7 LC3

1.5

1.0

0.5

0.0

Insu

lin

(ng

/µg

prot

ein)

Pro

tein

(fo

ld c

hang

e vs

cont

rol s

iRN

A)

*

2 mmol/lglucose

20 mmol/lglucose

0

1

2

3

4

5

2 mmol/lglucose

20 mmol/lglucose

*200

150

100

50

0

Insu

lin s

ecre

tion

(ng

[µg

prot

ein]

-1 h

-1)

Insu

lin s

ecre

tion

(pg

[ng

cont

ent]-

1 h-

1 )

ATG7

LC3

Loadingcontrol

ControlsiRNA

Atg7siRNA

a b

c

e

d

Fig. 5 Atg7 siRNA decreases autophagy and increases insulin secretionfrom MIN6 cells. MIN6 cells were transfected with control siRNA orAtg7 siRNA and left to culture for a total of 72 h. (a) Representativewestern blot image showing ATG7 and LC3 protein levels followingcontrol siRNA or Atg7 siRNA treatment. (b) Quantification of westernblots for ATG7 and LC3-II protein for control siRNA (white bars) orAtg7siRNA (black bars) treated MIN6 cells (n=3 or 4 experiments; *p<0.05,one-sample t test). (c) Insulin secretion following control siRNA (whitebars) or Atg7 siRNA (black bars) treatment with 2 or 20 mmol/l glucose(n =5 independent experiments; *p <0.05, Student’s paired t test). (d)Insulin content after control (white bar) or Atg7 siRNA (black bar)treatment (n =5 independent experiments; *p <0.05, Student’s pairedt test). (e) Insulin secretion following control (white bars) or Atg7 siRNA(black bars) treatment, corrected for insulin content (n =4 independentexperiments; *p <0.05, Student’s paired t test)

LC3-ILC3-II

Loadingcontrol

Notreatment

Lalistat

CQ- + - +

Notreatment

Lalistat

*

**

*

1

2

3

4

0

LC3-

II:LC

3-I r

atio

LC3-IIratio

ATG7 LAL

1.5

0.5

0.0

1.0

Pro

tein

(fo

ld c

hang

e vs

no tr

eatm

ent)

LAL

LC3-ILC3-II

ATG7

Loadingcontrol

Lalistat- + - +

a b

c d

Fig. 6 There is crosstalk between LAL inhibition and autophagy. MIN6cells were treated for 24 h with 5 μmol/l Lalistat or control and subjectedto western blot analysis. (a) Representative western blot for autophagymarkers (ATG7 and LC3-II) and LAL following Lalistat treatment. (b)Quantification of western blot results following control (white bars) orLalistat (grey bars) treatment, for LC3-II:LC3-I ratio, ATG7 and LAL(n =3 independent experiments). (c) Representative western blot for LC3following 24 h Lalistat treatment and with or without acute (2 h) CQaddition. (d) Quantification of western blot results showing LC3-II : LCIratio following 24 h Lalistat treatment with (black bars) or without (whitebars) acute (2 h) CQ addition (n =4 independent experiments; *p<0.05,**p <0.01 Student’s paired t test)

Diabetologia (2014) 57:129–139 135

role for lipophagy/LAL in beta cells, by employing pharma-cological and/or genetic inhibition of autophagy and LAL, in

clonal MIN6 beta cells and isolated islets, and monitoringeffects on insulin secretion, neutral lipid accumulation andautophagic markers. The generally excellent concordance be-tween all these approaches suggests that lipophagy serves as aconstitutive negative regulator of GSIS. It will be important infuture studies to address how this role intersects with thebetter-known function of LAL in hydrolysing lipid fromLDL [17, 18].

Although the contribution of the amplification pathway toGSIS is now well established, the underlying mechanismsremain obscure. One possibility focuses on the turnover ofintracellular lipid stores [10, 39]. Evidence for this comes fromstudies in which lipase activity is inhibited either pharmaco-logically [14] or in HSL [11, 40] and ATGL [13] knockoutmice, which show aberrant insulin secretion. Thus, while thereis a key role for neutral lipases in GSIS, our study provides thefirst evidence that lysosomal lipid metabolism contributes aswell, but in a very different manner. First, and most notably,inhibition of LAL potentiates GSIS rather than inhibiting it.Second, whereas neutral lipase inhibitors such as orlistat blocksecretion acutely, Lalistat required >8 h to exert its effects. ThusLAL does not appear to generate lipid signalling moleculesitself, unlike neutral lipases. Third, the increment in GSIS dueto chronic exposure to Lalistat was itself blocked by acutetreatment with orlistat, suggesting that LAL acts upstream ofthe neutral lipases. The simplest explanation for these collectivefindings is that inhibition of LAL leads to slow accumulation ofa neutral lipid signalling pool, which is subsequently mobilisedby orlistat-sensitive lipases in response to glucose stimulation.

ControlsiRNA

Atg7siRNA

*

0

2

4

6

8

10f

Lipi

d dr

ople

ts(n

umbe

r pe

r ce

ll)

Control Lalistat

*

0

2

4

6

8c

Lipi

d dr

ople

ts(n

umbe

r pe

r ce

ll)

b

d

a

g h

DAPI Nile Red (LD)

DAPI LAL (lysosome) BODIPY (LD)

e

Fig. 7 Lalistat and Atg7 siRNAincrease lipid droplets. MIN6cells were stained with Nile Redfollowing either control (a), 24 hLalistat (b), 72 h control siRNA(d) or 72 h Atg7 siRNA (e) tovisualise lipid droplets (redstaining); nuclei were stained withDAPI (blue). (c , f) Quantificationof lipid droplets after control orLalistat (c) (n =3 independentexperiments, visualising 73–99cells per condition) or followingsiRNA treatment (f) (n=4independent experiments,visualising 58–66 cells percondition; *p <0.05, Student’sunpaired t test). (g , h) MIN6 cellswere also stained with BODIPY493/503 to visualise neutral lipid(green) and LAL to visualiselysosomes (red) to determineco-localisation between neutrallipid storage and lysosomes inuntreated (g) and Lalistat treated(h) cells. Scale bar, 10 μm

Lalistat - + - + - +

*

3MA - + - + - +

*****

LalistatNo Treatment

3MANo Treatment

*

1,500

0

0

10

20

30

40

20

0

40

60

80

100

100

200

300

400

500a b

c d

1,000

500

0

Insu

lin s

ecre

tion

(pg

isle

t-1 h

-1)

Insu

lin s

ecre

tion

(pg

isle

t-1 h

-1)

Insu

lin (

ng/n

g D

NA

)In

sulin

(ng

/ng

DN

A)

Fig. 8 Lalistat and inhibition of autophagy increase insulin secretion inprimary mouse islets. (a) Mouse islets were isolated and cultured for 48 hin RPMI or RPMI + Lalistat. Insulin secretion was measured over 1 h inthe presence of either 2 mmol/l (white bars) or 20 mmol/l (black bars)glucose, or 20 mmol/l glucose + 0.2 mmol/l orlistat (grey bars) (n =9–12;*p <0.05 Student’s paired t test). (b) Insulin content of islets followingtreatment for 48 h with Lalistat. (n =9–12). (c) Mouse islets were isolatedand cultured in the presence or absence of 5 mmol/l 3MA for 24 h. Insulinsecretion was measured over 1 h in the presence of either 2 mmol/l (whitebars) or 20 mmol/l (black bars) glucose, or 20 mmol/l glucose +0.2 mmol/l orlistat (grey bars) (n =6; *p <0.05, ****p <0.0001, one-way ANOVA, Bonferroni’s post hoc test). (d) Insulin content of isletstreated with 3MA (n =6); *p <0.05, Student’s unpaired t test)

136 Diabetologia (2014) 57:129–139

We also provide some evidence of potential crosstalk be-tween LAL inhibition and autophagy. Chronic Lalistat treat-ment inhibited autophagic fluxmeasured in the presence of CQ(Fig. 6d). This could help explain how inhibition of a lyso-somal enzyme, LAL, might lead to a build up of neutral lipidstores, which we observed in non-lysosomal sites (presumablycytosol) where they can be accessed by neutral lipases duringGSIS. However, there are other possible explanations, withvery recent studies pointing to a complex inter-regulation ofmore traditional lipid metabolism with LAL/lipophagy medi-ated by transcription factors, such as TFEB/HLH-30 andMXL-3 [41, 42]. The exact mechanism for interaction betweenLAL and neutral lipid droplets in beta cells will require exten-sive further studies to resolve.

Although we provide evidence supporting a pivotal role forlipid metabolism and signalling in regulating GSIS, the exactidentity of the signalling metabolite(s) responsible for thisremains obscure. Candidates include long-chain fatty acylCoAs, perhaps via protein acylation or modulation of ionchannels [43], and diacylglycerol, albeit most probably actingvia mechanisms independent of protein kinase C [44, 45].These metabolites can be derived from breakdown of TG.However, our results suggest that turnover of CE should alsobe investigated in future studies, since prior augmentation ofthis pool correlated with increased insulin secretion (Fig. 4c).Moreover, the fact that both inhibition of LAL and autophagyhave very similar effects in terms of potentiating GSIS andaccumulation of lipid droplets points to a major role for LALin mediating lipophagy. This is strongly underscored by theobservation that potentiation of GSIS due to prior disruptionof either LAL (by Lalistat) or autophagy (by 3MA) wassensitive to inhibition by orlistat, suggestive of a commonupstream mechanism. However, this sensitivity was only par-tial in the case of 3MA, pointing either to LAL-independenteffects of autophagy that contribute to GSIS, or LAL workingthrough other mechanisms in addition to lipophagy. The latter

would be consistent with the role of LAL in mobilising lipidsbound to lipoprotein [17, 18].

The beneficial effects of inhibiting autophagy observed inour study (enhanced GSIS) contrasts with some recent studiesutilising beta cell-specific ATG7 knockout mice, in which lossof autophagy impairs glucose tolerance [46, 47]. It must bestressed that our system employs a relatively mild andtransient inhibition of autophagy, using siRNA, in contrastto the life-long and complete deletion of Atg7 , which ischaracterised by reduced beta cell mass [46, 47]. Therefore,long-term inhibition of autophagy is detrimental to generalbeta cell function due to changes in protein, lipid and organ-elle degradation. In our study, we highlight the importance ofspecifically targeting one arm of these autophagic processes—lipid degradation. Also, there is some prior support forautophagy as a (short-term) negative regulator of beta cellresponses, in that activation of autophagy by rapamycinresulted in decreased insulin secretion [48].

In conclusion, this study provides the first evidence thatlipophagy/lysosomal lipid degradation contributes to neutrallipid metabolism in beta cells, as it does in other cell types(Fig. 9a). More importantly, our results suggest that lipophagyand LAL (lysosomal lipid degradation) serve as negativeregulators of insulin secretion, and under normal conditionslimit the size of the lipid signalling pool. Although it remainsto be determined whether dysregulation of these processescontributes to beta cell failure in any way, their inhibitionmight be a beneficial means for pharmacologically augmentinginsulin secretion (Fig. 9b).

Acknowledgements We thank M. Cahova (Institute for Clinical andExperimental Medicine, Prague, Czech Republic) for suggesting the useof Lalistat, and L. Marshall (Garvan Institute of Medical Research,Darlinghurst, NSW, Australia) for help in setting up the Nile Red stainingprotocol. We are grateful to L. O’Reilly and R. Laybutt (Garvan InstituteofMedical Research, Darlinghurst, NSW,Australia) for islet isolation andfor critical review of the manuscript, respectively.

Fig. 9 Proposed model forlipophagy in beta cells. (a)Lipophagy influences insulinsecretion by altering lipidmetabolism in beta cells, forrequirements during cellularprocesses. (b) Inhibition ofautophagy or LAL causes adecreased turnover of neutral lipidpools, which are then acted uponacutely by neutral lipases (such asHSL or ATGL) upon glucosestimulation, leading to enhancedinsulin release

Diabetologia (2014) 57:129–139 137

Funding This work was supported by a project grant (TJB), researchfellowships (TJB and PJM) and a post-graduate scholarship (GLP) fromthe National Medical Research Council of Australia.

Duality of interest The authors declare that there is no duality ofinterest associated with this manuscript.

Contribution statement GLP, NM, KYC, PJM, JC and AD performedexperiments, analysed data, contributed to the acquisition of data, and tothe revision of the manuscript. CCC, PB and PH contributed to theacquisition of data, and the revision of the manuscript. TJB and PJMcontributed to the revision of the manuscript. GLP wrote the manuscriptand designed experiments. TJB conceived and designed experiments andwrote the manuscript. All authors approved the final version of thisarticle.

References

1. Prentki M, Tornheim K, Corkey BE (1997) Signal transductionmechanisms in nutrient-induced insulin secretion. Diabetologia40(suppl 2):S32–S41

2. Ashcroft F, Rorsman P (1990) ATP-sensitive K+ channels: a linkbetweenβ-cell metabolism and insulin secretion. Biochem Soc Trans18:109–111

3. Henquin JC (2000) Triggering and amplifying pathways of regula-tion of insulin secretion by glucose. Diabetes 49:1751–1760

4. Henquin JC, Dufrane D, Nenquin M (2006) Nutrient control ofinsulin secretion in isolated normal human islets. Diabetes 55:3470–3477

5. PrentkiM, Vischer S, GlennonMC, Regazzi R, Deeney JT, Corkey BE(1992) Malonyl-CoA and long chain acyl-CoA esters as metaboliccoupling factors in nutrient-induced insulin secretion. J Biol Chem267:5802–5810

6. Nolan CJ, Prentki M (2008) The islet beta-cell: fuel responsive andvulnerable. Trends in Endocrinol Metab 19:285–291

7. MacDonald MJ, Fahien LA, Brown LJ, Hasan NM, Buss JD,Kendrick MA (2005) Perspective: emerging evidence for signalingroles of mitochondrial anaplerotic products in insulin secretion.Am J Physiol Endocrinol Metab 288:E1–E15

8. Jensen MV, Joseph JW, Ronnebaum SM, Burgess SC, Sherry AD,Newgard CB (2008) Metabolic cycling in control of glucose-stimulated insulin secretion. Am J Physiol Endocrinol Metab 295:E1287–E1297

9. Prentki M, Madiraju SRM (2012) Glycerolipid/free fatty acid cycle andisletβ-cell function in health, obesity and diabetes.Mol Cell Endocrinol353:88–100

10. Fex M, Mulder H (2008) Lipases in the pancreatic beta-cell: impli-cations for insulin secretion. Biochem Soc Trans 36:885–890

11. Fex M, Haemmerle G, Wierup N et al (2009) A beta cell-specificknockout of hormone-sensitive lipase in mice results in hyperglycaemiaand disruption of exocytosis. Diabetologia 52:271–280

12. Rl R, Masiello P, Wang SP, Li H, Mitchell GA, Prentki M (2001) Arole for hormone-sensitive lipase in glucose-stimulated insulin secre-tion. Diabetes 50:1970–1975

13. Peyot M-L, Guay C, Latour MG et al (2009) Adipose triglyceridelipase is implicated in fuel- and non-fuel-stimulated insulin secretion.J Biol Chem 284:16848–16859

14. Mulder H, Yang S, Winzell MS, Holm C, Ahren B (2004) Inhibitionof lipase activity and lipolysis in rat islets reduces insulin secretion.Diabetes 53:122–128

15. Cantley J, Biden TJ (2010) Targeting triglyceride/fatty acid cycling inbeta-cells as a therapy for augmenting glucose-stimulated insulinsecretion. Islets 2:127–129

16. Cantley J, Burchfield JG, Pearson GL, Schmitz-Peiffer C, Leitges M,Biden TJ (2009) Deletion of PKCε selectively enhances the ampli-fying pathways of glucose-stimulated insulin secretion via increasedlipolysis in mouse beta-cells. Diabetes 58:1826–1834

17. Jelinek D, Patrick SM, Kitt KN, Chan T, Francis GA, Garver WS(2009) Physiological and coordinate downregulation of the NPC1and NPC2 genes are associated with the sequestration of LDL-derived cholesterol within endocytic compartments. J Cell Biochem108:1102–1116

18. Miller WL, Bose HS (2011) Early steps in steroidogenesis: intracel-lular cholesterol trafficking. J Lipid Res 52:2111–2135

19. Singh R, Kaushik S, Wang Y et al (2009) Autophagy regulates lipidmetabolism. Nature 458:1131–1135

20. Stromhaug PE, Klionsky DJ (2001) Approaching the molecularmechanism of autophagy. Traffic 2:524–531

21. Glick D, Barth S, Macleod KF (2010) Autophagy: cellular andmolecular mechanisms. J Pathol 221:3–12

22. Singh R, Cuervo Ana M (2011) Autophagy in the cellular energeticbalance. Cell Metab 13:495–504

23. OuimetM, Franklin V,Mak E, LiaoX, Tabas I,Marcel Yves L (2011)Autophagy regulates cholesterol efflux from macrophage foam cellsvia lysosomal acid lipase. Cell Metab 13:655–667

24. Cahova M, Dankova H, Palenickova E, Papackova Z, Kazdova L(2010) The autophagy-lysosomal pathway is involved in TAGdegradation in the liver: the effect of high-sucrose and high-fat diet.Folia Biol 56:173–182

25. Cahova M, Dankova H, Palenickova E et al (2012) The increasedactivity of liver lysosomal lipase in nonalcoholic fatty liverdisease contributes to the development of hepatic insulin resistance.Biochem Res Int 2012:11

26. Miyazaki J-I, Araki K, Yamato E et al (1990) Establishment of apancreatic beta cell line that retains glucose-inducible insulin secre-tion: special reference to expression of glucose transporter isoforms.Endocrinology 127:126–132

27. Busch AK, Cordery D, Denyer GS, Biden TJ (2002) Expressionprofiling of palmitate- and oleate-regulated genes provides novelinsights into the effects of chronic lipid exposure on pancreaticbeta-cell function. Diabetes 51:977–987

28. Boslem E, MacIntosh G, Preston AM et al (2011) A lipidomic screenof palmitate-treated MIN6 β-cells links sphingolipid metaboliteswith endoplasmic reticulum (ER) stress and impaired protein traf-ficking. Biochem J 435:267–276

29. Vázquez CL, Colombo MI (2009) Assays to assess autophagy in-duction and fusion of autophagic vacuoles with a degradative com-partment, using monodansylcadaverine (MDC) and DQ-BSA. In:Daniel JK (ed) Methods in enzymology. Academic, Waltham, pp85–95

30. Meikle PJ, Wong G, Tsorotes D et al (2011) Plasma lipidomic analysis ofstable and unstable coronary artery disease. Arterioscler ThrombVasc Biol31:2723–2732

31. Rosenbaum AI, Cosner CC, Mariani CJ, Maxfield FR, Wiest O,Helquist P (2010) Thiadiazole carbamates: potent inhibitors of lyso-somal acid lipase and potential Niemann–Pick type C disease thera-peutics. J Med Chem 53:5281–5289

32. Rosenbaum AI, Rujoi M, Huang AY, Du H, Grabowski GA,Maxfield FR (2009) Chemical screen to reduce sterol accumulationin Niemann–Pick C disease cells identifies novel lysosomal acidlipase inhibitors. Biochim Biophys Acta 1791:1155–1165

33. Cnop M, Hannaert JC, Hoorens A, Eizirik DL, Pipeleers DG (2001)Inverse relationship between cytotoxicity of free fatty acids in pan-creatic islet cells and cellular triglyceride accumulation. Diabetes 50:1771–1777

34. Busch AK, Gurisik E, Cordery DV et al (2005) Increased fatty aciddesaturation and enhanced expression of stearoyl coenzyme Adesaturase protects pancreatic beta-cells from lipoapoptosis.Diabetes 54:2917–2924

138 Diabetologia (2014) 57:129–139

35. Larsson S, Resja S, Gomez MF, James P, Holm C (2012)Characterization of the lipid droplet proteome of a clonal insulin-producingβ-cell line (INS-1 832/13). J Proteome Res 11:1264–1273

36. Kaushik S, Rodriguez-Navarro Jose A, Arias E et al (2011)Autophagy in hypothalamic AgRP neurons regulates food intakeand energy balance. Cell Metab 14:173–183

37. Kovsan J, Bashan N, Greenberg AS, Rudich A (2009)Potential role of autophagy in modulation of lipid metabolism.Am J Physiol Endocrinol Metab 298:E1–E7

38. Sieber MH, Thummel CS (2012) Coordination of triacylglycerol andcholesterol homeostasis byDHR96 and theDrosophila LipA homologmagro. Cell Metab 15:122–127

39. Nolan CJ,MadirajuMS, Delghingaro-Augusto V, PeyotML, PrentkiM(2006) Fatty acid signaling in the beta-cell and insulin secretion.Diabetes 55(suppl 2):S16–S23

40. Peyot M-L, Nolan CJ, Soni K et al (2004) Hormone-sensitive lipasehas a role in lipid signaling for insulin secretion but is nonessential forthe incretin action of glucagon-like peptide 1. Diabetes 53:1733–1742

41. O’Rourke EJ, Ruvkun G (2013) MXL-3 and HLH-30 transcription-ally link lipolysis and autophagy to nutrient availability. Nat Cell Biol16:668–676

42. Settembre C, de Cegli R, Mansueto G et al (2013) TFEB controlscellular lipid metabolism through a starvation-induced autoregulatoryloop. Nat Cell Biol 16:647–658

43. Herrero L, Rubi B, Sebastian D et al (2005) Alteration of themalonyl-CoA/carnitine palmitoyltransferase I interaction in thebeta-cell impairs glucose-induced insulin secretion. Diabetes 54:462–471

44. Kwan EP, Xie L, Sheu L et al (2006) Munc13-1 deficiency reducesinsulin secretion and causes abnormal glucose tolerance. Diabetes 55:1421–1429

45. Schmitz-Peiffer C, Biden TJ (2008) Protein kinase C function inmuscle, liver, and beta-cells and its therapeutic implications for type2 diabetes. Diabetes 57:1774–1783

46. Ebato C, Uchida T, Arakawa M et al (2008) Autophagy is importantin islet homeostasis and compensatory increase of beta cell mass inresponse to high-fat diet. Cell Metab 8:325–332

47. Jung HS, Chung KW, Won Kim J et al (2008) Loss of autophagydiminishes pancreatic beta cell mass and function with resultanthyperglycemia. Cell Metab 8:318–324

48. Tanemura M, Ohmura Y, Deguchi T et al (2012) Rapamycin causesupregulation of autophagy and impairs islets function both in vitroand in vivo. Am J Transplant 12:102–114

Diabetologia (2014) 57:129–139 139