Protein Phosphatase 1 (PP-1)-Dependent Inhibition of Insulin Secretion by Leptin in INS-1 Pancreatic...

Protein Phosphatase 1 (PP-1)-Dependent Inhibition ofInsulin Secretion by Leptin in INS-1 Pancreatic �-Cellsand Human Pancreatic Islets

Peter Kuehnen, Katharina Laubner, Klemens Raile, Christof Schofl, Franz Jakob,Ingo Pilz, Gunter Path, and Jochen Seufert

Institute of Experimental Pediatric Endocrinology (P.K., K.R.), Charite Universitatsmedizin Berlin, 13353Berlin, Germany; The Experimental and Clinical Research Center (K.R.), Berlin-Buch, 13125 Berlin,Germany; Division of Endocrinology and Diabetology (K.L., I.P., G.P., J.S.), Department of InternalMedicine II, University Hospital of Freiburg, 79095 Freiburg, Germany; Division of Endocrinology andDiabetology (C.S.), Department of Internal Medicine I, University Hospital of Erlangen, 91052 Erlangen,Germany; and Center for Musculoskeletal Research (F.J.), Orthopedic Department, University ofWuerzburg, 97070 Wuerzburg, Germany

Leptin inhibits insulin secretion from pancreatic �-cells, and in turn, insulin stimulates leptin biosyn-thesis and secretion from adipose tissue. Dysfunction of this adipoinsular feedback loop has beenproposed to be involved in the development of hyperinsulinemia and type 2 diabetes mellitus. At themolecular level, leptin acts through various pathways, which in combination confer inhibitory effectson insulin biosynthesis and secretion. The aim of this study was to identify molecular mechanisms ofleptin action on insulin secretion in pancreatic �-cells. To identify novel leptin-regulated genes, weperformed subtraction PCR in INS-1 �-cells. Regulated expression of identified genes was confirmed byRT-PCR and Northern and Western blotting. Furthermore, functional impact on �-cell function wascharacterized by insulin-secretion assays, intracellular Ca2� concentration measurements, and enzymeactivity assays. PP-1�, the catalytic subunit of protein phosphatase 1 (PP-1), was identified as a novelgene down-regulated by leptin in INS-1 pancreatic �-cells. Expression of PP-1� was verified in humanpancreatic sections.PP-1�mRNAandproteinexpression isdown-regulatedby leptin,whichculminatesin reduction of PP-1 enzyme activity in �-cells. In addition, glucose-induced insulin secretion was in-hibited by nuclear inhibitor of PP-1 and calyculin A, which was in part mediated by a reduction ofPP-1-dependent calcium influx into INS-1 �-cells. These results identify a novel molecular pathway bywhich leptin confers inhibitory action on insulin secretion, and impaired PP-1 inhibition by leptin maybe involved in dysfunction of the adipoinsular axis during the development of hyperinsulinemia andtype 2 diabetes mellitus. (Endocrinology 152: 1800–1808, 2011)

It is hypothesized that the adipocyte hormone leptin playsa central role in the development of type 2 diabetes mel-

litus, due to the formation of leptin resistance at the pan-creatic islet level, a mechanism that has been proposed inhumans and clearly demonstrated in animal models (1–3).

Leptin encoded by the ob gene (4) is secreted from whiteadipose tissue and regulates food intake and energy ex-penditure within the hypothalamus (5), but next to thesecentral effects, leptin also inhibits insulin biosynthesis

and, independently, insulin secretion at the pancreatic isletlevel (6–8). In detail, leptin reduces preproinsulin mRNAexpression and inhibits transcriptional activity of the in-sulin promoter. Furthermore, leptin has been demon-strated to interfere with ATP-dependent potassium chan-nel activity (9), and glucagon-like-peptide 1 signaling (10)in pancreatic �-cells. As part of a negative feedback loop,insulin stimulates leptin secretion in adipose tissue creat-ing the so-called adipoinsular axis (11, 12).

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in U.S.A.Copyright © 2011 by The Endocrine Societydoi: 10.1210/en.2010-1094 Received September 29, 2010. Accepted February 28, 2011.First Published Online March 22, 2011

Abbreviations: AMPK, AMP-activated protein kinase; [Ca2�]i, intracellular Ca2� concen-tration; Cal A, calyculin A; NIPP1, nuclear inhibitor of PP-1; PKA, protein kinase A; PP-1,protein phosphatase.

D I A B E T E S - I N S U L I N - G L U C A G O N - G A S T R O I N T E S T I N A L

1800 endo.endojournals.org Endocrinology, May 2011, 152(5):1800–1808

Rodents (ob/ob mice) with homozygous defects in theleptin gene or in the leptin receptor gene (db/db mice)develop massive obesity, hyperinsulinemia, hyperphagia,hypertriglyceridemia, and type 2 diabetes mellitus (13,14). Treatment of leptin-deficient ob/ob mice with leptinnormalizes hyperinsulinemia and metabolic abnormali-ties, indicating a mandatory role of central and peripheralleptin signaling for appropriate �-cell function (15). Theimportance of the peripheral leptin action has been spec-ified in �-cell-specific leptin receptor knockout mice (1, 2),which are suffering from hyperinsulinemia, obesity, andinsulin resistance, indicating the importance of �-cell-specific leptin signaling for the regulation of insulin se-cretion. Moreover, in these mouse models, hyperinsu-linemia precedes the development of insulin resistance(3), which points out that impaired leptin signaling inpancreatic �-cells, and thereby, disturbances of the adi-poinsular axis may play a central role early in the de-velopment of type 2 diabetes mellitus. Current resultsindicate that leptin acts through multiple cellular mech-anisms in pancreatic �-cells.

To identify novel cellular mechanisms of leptin-depen-dent inhibition of insulin secretion, we performed a sub-traction hybridization assay in INS-1 �-cells. We foundthat expression of the gene encoding the catalytic subunitof the serine/threonine protein phosphatase 1 (PP-1) wasstrongly inhibited by leptin, leading to reduced enzymeactivity, intracellular calcium concentrations ([Ca2�]i),and consequently blunted insulin secretion.

The serine/threonine phosphatase PP-1 (16) is highlyconserved in different species. PP-1 is subdivided intothree isoforms, PP-1�, PP-1�, and PP-1�. It is expressed indifferent organs such as the central nervous system, heart,liver, and pancreas and modifies a broad range of differentsubstrates by dephosphorylation. PP-1 isoforms are in-volved in different processes, such as glycogen synthesis,cell cycle regulation, muscle contraction, and ion channelregulation (17–19). PP-1 is a multiprotein complex con-sisting of the PP-1 catalytic subunit and different regula-tory subunits (20, 21). Several hormones like insulin andthyroid hormones regulate PP-1 activity, which has beenproposed to be mediated by protein kinase A (PKA) (22,23). In adipocytes, studies reveal a potential regulation ofPP-1 by leptin via PKA and MAPK (24).

Moreover, PP-1 has been suggested to play an impor-tant role in the regulation of insulin secretion in pancreatic�-cells. PP-1 inhibition by okadaic acid or calyculin A (CalA) leads to decreased Ca2� influx into insulin-producingpancreatic �-cells (25, 26). PP-1 inhibition decreases glu-cose-stimulated insulin secretion and does not affect non-glucose-induced secretion (27–29). However, the inhibi-tion of Ca2� entry may not fully explain the influence on

insulinsecretion.Furthermore, studies indicate thatPP-1reg-ulates vesicle fusion and exocytosis in pancreatic �-cells (30)andregulatesgeneralprotein translation inpancreatic�-cells(31). In addition to these findings, expression of the endog-enous PP-1 inhibitors dopamine and cAMP regulated phos-phoprotein-32 and inhibitor 1 (32), which are activated byPKA-mediated phosphorylation, has been demonstrated inpancreatic�-cells.Recently,PP-1�hasbeendemonstratedtoinhibit AMP-activated protein kinase (AMPK) activity inMIN6 �-cells and thereby may regulate insulin secretion (33).

These observations support a role for PP-1in modula-tion of insulin secretion and vesicle mobilization in pan-creatic �-cells. In this study, we further characterized theinvolvement of PP-1 activity in the leptin-mediated effectson insulin biosynthesis and secretion in pancreatic �-cellsmore in detail.

Materials and Methods

Recombinant mouse leptin was obtained from R&D Systems(Minneapolis, MN), insulin from Deltapharma (Pfullingen, Ger-many), and nuclear inhibitor of PP-1 (NIPP1) from Calbiochem(Merck, Darmstadt, Germany). Polyclonal antiserum directedagainst rat PP-1� was received from Exalpha Biologicals (May-nard, MA) and insulin antiserum from Invitrogen (Carlsbad,CA). Subtraction hybridization was performed with Clontech(Palo Alto, CA) PCR-select cDNA subtraction kit. Fura-2/AMwas purchased from Molecular Probes (Eugene, OR). All otherproducts were analytical grade and obtained from Sigma (Mu-nich, Germany). ELISA was performed with the DRG high-rangerat insulin ELISA (DRG International, Mountainside, NJ).

Cells and cell cultureINS-1 rat pancreatic �-cells were a kind gift from Dr. Claes

Wollheim (University of Geneva, Geneva, Switzerland). Mediumfor cell culture contained RPMI 1640 medium (GIBCO BRL,Carlsbad, CA) with fetal bovine serum (10%), HEPES (10 mM),L-glutamine (2 mM), penicillin (100 U/ml), streptomycin (100�g/ml), sodium pyruvate (1 mM) and �-mercaptoethanol (50�M), as reported previously. HepG2 cell medium containsDMEM/F12 medium (GIBCO BRL) supplemented with L-glu-tamine and pyridoxine HCl, fetal bovine serum (10%), HEPES(10 mM), penicillin (100 U/ml), and streptomycin (100 �g/ml).For stimulation experiments, cells were incubated in a serum-freemedium with BSA (0.1%) and trasylol (1%). Human pancreaticislets were obtained from organ donors as part of the humanpancreatic islet transplantation program at the University Hos-pital of Freiburg. Islets (hand-picked) were placed 10 per well (ina 12-well culture plate) with 1 ml/well 5.5 mM glucose RPMImedium containing 0.5% human serum albumin (Sigma Chem-ical Co., St. Louis, MO) and 1% aprotinin (Miles, Kankakee, IL).After incubation for 1 h, the medium was carefully removed andreplaced with fresh medium at 11.1 mM glucose with 100 mM

leptin, 40 nM Cal A, or vehicle control. After 1 h incubation, themedium was removed, centrifuged (15,000 rpm, 5 min), trans-ferred to a fresh Eppendorf tube, and processed for subsequentinsulin RIA (human insulin ultrasensitive ELISA kit; Mercodia,

Endocrinology, May 2011, 152(5):1800–1808 endo.endojournals.org 1801

Uppsala, Sweden). Human pancreatic islets, which were used inthe insulin ELISA experiments, were obtained from organ do-nors as part of the islet transplantation program of the UniversityHospital of Giessen, Germany.

Subtraction PCRTo identify genes regulated by leptin in the INS-1 cell line,

poly(A)� RNA was prepared from leptin-stimulated and un-stimulated INS-1 cells and used for PCR-selected subtractionanalysis. The subtraction PCR (Clontech PCR select cDNA sub-traction kit) was performed according to the manufacturer’s pro-tocol. INS-1 cells were stimulated with either leptin (100 nM) orvehicle control (15 mM HCl and 7.5 mM NaOH) for 24 h. mRNAwas isolated with micro-fast track kit (Invitrogen). After sub-traction hybridization and gel electrophoresis the obtainedbands were extracted and subcloned with Topo TA cloning (In-vitrogen). Thereafter, DNA was extracted and purified with HiS-peed Plasmid Mini (QIAGEN, Hilden, Germany) or RPM 4AG(Invitrogen). ABI Prism Big Dye Terminator cycle sequencingready reaction kit (Applied Biosystems, Foster City, CA) wasused for sequencing analysis in an ABI Prism 310 geneticanalyzer.

Competitive semiquantitative RT PCRTotal RNA from INS-1 cells was extracted by single-step gua-

nidinium isothiocyanate method according to the Trizol protocol(Invitrogen). RNA was quantitated by OD 260/280 measurement,and quality was checked by gel electrophoresis. Equal amounts of2 �g isolated RNA were used for reverse transcription (Super ScriptII; Invitrogen).RT-PCRwasperformed(n�5) forPP-1�withsenseprimer 5�-ATGATGAATGCAAGAGACGATACAACA-3� andantisense primer 5�-TGGGGCAGAGGGCAGCACA-3� (600-bpPCR product, GenBank accession number NM_031527) and 18Swith Taq polymerase (TaKaRa, Takara Shuzo Co., Japan) on aGene Amp thermal cycler (Applied Biosystems) according to thefollowing conditions: 1 min at 94 C (one cycle); 30 sec at 94 C, 30sec at 55 C, and 45 sec at 72 C (36 cycles); and 72 C for 5 min (onecycle).

Northern blot analysisRNA was extracted from INS-1 cells, which were stimulated

by 100 nM leptin or vehicle for up to 48 h in serum-free incuba-tion medium. The RNA was separated on a formaldehyde gel,and samples were transferred to a nylon membrane and fixed at80 C for 1 h. Doubled-stranded DNA probe was generated bycloning the RT-PCR product of PP-1�. After digestion and ex-traction of the PP-1� fragment, the purified sample was labeledwith [32P]dCTP (Hartmann Analytics, Braunschweig, Germany)using Klenow polymerase. This probe was hybridized to the ny-lon membrane at 74 C overnight. Washing steps were performedat room temperature and 42 C for 40 min. Samples were exposedto x-ray films. The experiments were repeated three times.

Western immunoblottingTotal protein lysates from INS-1 cells (15 �g), which were

stimulated with 100 nM leptin or vehicle control for 48 h, wereresolved by 7.5% SDS-PAGE and transferred to an ImmobilonP polyvinylidene difluoride membrane (Sigma, St. Louis, MO).The membrane was blocked with 4 g nonfat dry milk in 40 ml0.04% PBS-Tween and sequentially incubated with a specificantirat PP-1� antiserum and the appropriate horseradish per-

oxidase-conjugated, secondary antirabbit antibody (dilution1:10,000). Bands were visualized by enhanced chemilumines-cence (ECL; Amersham Inc., Buckinghamshire, UK). Westernblotting experiments were repeated three times.

Pancreas processing for immunofluorescenceThe human pancreas specimen was obtained through the Islet

Transplantation Program of the University Hospital of Freiburg.The specimen was fixed in 2% formaldehyde (Sigma, Germany)overnight and processed immediately. After rehydration, antigenretrieval of 5-�m sections was carried out with 0.1% pepsin/0.01N HCl (Sigma, Germany) for 5 min at room temperature. Cut-tings were stained sequentially for 1 h at room temperature withantibody dilutions of 1:100 in 0.1% Tween 20/PBS (Sigma, Ger-many) containing 3% normal goat serum (Vector Laboratories,Burlingame, CA). Staining for human PP-1� was carried out witha rabbit monoclonal antibody (Epitomics, Burlingame, CA) fol-lowed by Cy3-conjugated secondary donkey antirabbit antibody(Jackson ImmunoResearch, West Grove, PA) staining. Insulindetection was carried out with a polyclonal guinea pig antiswineinsulin antibody (DakoCytomation, Hamburg, Germany) fol-lowed by Cy2-conjugated secondary goat anti-guinea pig anti-body (Jackson ImmunoResearch) staining. Nuclei were stainedwith 1 �g/ml 4�,6-diamidino-2-phenylindole (Sigma, Germany)for 10 min at room temperature. Processed slides were mountedwith fluorescent mounting medium (DakoCytomation).

Images were obtained via confocal microscopy (Leica TCSSP2 AOBS) and analyzed by Zeiss LSM Image Browser software.

PP-1 enzyme activityPP-1 enzyme activity was measured according to the pub-

lished protocol (34).In brief, INS-1 cells cultivated at 5.6 mM glucose were homog-

enized in 0.5 ml phosphatase extraction/assay buffer [20 mM imi-dazole HCl (pH 7.2), 2 mM EDTA, �-mercaptoethanol (0.2%), 2mg/ml glycogen (rabbit liver, type III), 1 mM benzamidine, 1 mM

phenylmethylsulfonyl fluoride, 10 �g/ml aprotinin, leupeptin, an-tipain, soy trypsin inhibitor (STI), and pepstatin] by sonicating for10 sec with 20% power. Cell extract was centrifuged at 2000 � gfor 5 min, and the supernatant, the cytosolic fraction, was centri-fuged at 100,000 � g for 40 min. The pellet was dissolved in phos-phatase extraction/assay buffer (particulate fraction). [P32]ATP-la-beled phosphorylase A was used as a substrate to measure PP-1enzyme activity. The released radioactivity was detected, and thespecific substrate activity was calculated as follows: specific activityof the substrate � (total cpm/reaction tube cpm blank)/0.62 nM

phosphate, where cpm is counts per minute.

Measurement of cytosolic calcium concentrationINS-1 cells cultured on coverslips were loaded with 5 �M fura-

2/AM for 30 min at 37 C. The loading buffer was as follows: 130mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 1.2 mM MgSO4, 1.5 mM

CaCl2, 10 mM glucose 20 mM HEPES, 2% BSA (wt/vol), and 0.1%pluronic acid (wt/vol), gassed with 100% (vol/vol) O2 (pH 7.4).After loading, the coverslips were washed, mounted in a tempera-ture-controlledsuperfusionchamber(37C),andplacedonthestageof a Zeiss Axiovert IM 135 equipped with a �40/1.3 Achrostigmatoil immersion objective (Zeiss, Gottingen, Germany). The chamberwas superfused with the same buffer as used for fura-2 loading with0.1% BSA (wt/vol) and 0.5 mM glucose and without pluronic acidat a flow rate of 0.75–2 ml/min. [Ca2�]i was measured in cells of

1802 Kuehnen et al. PP-1 Mediated Inhibition of Insulin Secretion Endocrinology, May 2011, 152(5):1800–1808

average sizeandhealthyappearance (round in shape,nomembraneblebs). After 5 min at 0.5 mM glucose in the presence or absence of100 nM leptin or 10 nM Cal A, glucose was raised to 15 mM for 15min.Aftera5-minwashwith0.5mMglucose,cellsweredepolarizedwith 45 mM potassium chloride to check for functional integrity ofthecells.Leptin-treatedcellshadbeenpretreatedwith100nM leptinfor 12 h in normal growth medium. Fura-2 fluorescence from asingle cell was recorded with a dual-excitation spectrofluorom-eter system (Deltascan 4000; Photon Technology Instruments,Wedel, Germany). [Ca2�]i was calculated according to the for-mula [Ca2�]i � KD � B � (R � Rmin)/(Rmax � R), where KD(dissociation constant) � 224 nM, R is the ratio of fluorescenceintensities excited at 340 and 380 nm, and Rmax, Rmin, and B areconstants that were determined in the superfusion chamber fromsolutions containing fura-2 (1 �M) and various concentrations offree Ca2� (data not shown).

Insulin secretion measured by ELISATo analyze insulin secretion of �-cells, ELISA was performed

according to the DRG high-range rat insulin ELISA protocol (DRG

International). A total of 300,000 INS-1 cells per wellwere cultivated in six-well plates. INS-1 cells were pre-treated overnight with leptin. Thereafter, the sampleswere treated with leptin for 3 h or incubated togetherwith leptin and NIPP1 for 3 h before these sampleswere stimulated with glucose for 1 h (2.8 vs. 28 mM).The samples that were stimulated only by NIPP1 werenot treated by leptin overnight. This group was onlytreated for 3 h with NIPP1 before stimulation withglucose for1h.Thecontrol groupwasonly stimulatedwith glucose for 1 h (2.8 vs. 28 mM). Moreover, glu-cose-stimulated (11 mM) insulin secretion was ana-lyzed in human pancreatic islets after stimulation with100 mM leptin or vehicle control and leptin pretreat-ment overnight. INS-1 cells and human pancreatic is-lets were treated in a second independent experimentwith 40 nM Cal A or vehicle control for 1 h in mediumcontaining 11.1 mM glucose. The glucose stimulationexperiments were performed with either 11.1 mM glu-cose or 28 mM glucose concentration because of thedifferences of glucose sensitivity in INS-1 cells and hu-man pancreatic islets.

Statistical analysisAll values are expressed as the means � SD. Differ-

ences between groups were compared by Student’s ttest, and a P value � 0.05 was considered statisticallysignificant.

Results

Identification of PP-1� as aleptin-regulated gene in INS-1 cells

One of the detected leptin-regulated candi-date genes in the PCR-selected subtraction anal-ysis was the gene encoding the catalytic subunitof PP-1, PP-1� (Fig. 1A). The identity of the reg-ulated gene product was confirmed by subclon-ingandsequencing(datanotshown).Toconfirm

theeffectof leptinonPP-1�mRNAexpression inINS-1cells,we performed semiquantitave RT-PCR and Northern blotanalysis. Administration of 100 nM leptin decreased PP-1�

mRNA in INS-1 �-cells (Fig. 1, B and C) in a time-dependentmanner. After 48 h, a clear suppression of PP-1� mRNAlevels was observed in the semiquantitative RT-PCR as wellas in the Northern blot analysis (P � 0.05). We also con-firmed that leptin reducesPP-1� protein expressionbyWest-ern blot analysis time dependently (Fig. 1D) (P � 0.05). Theprotein levels of PP-1� are reduced after 48 h of leptin stim-ulation in INS-1 cells, as a consequence of transcriptionalregulation of PP-1� by leptin in INS-1 cells.

PP-1� protein expression in INS-1 cells and inhuman pancreas

Immunohistochemistry of PP-1� in INS-1 cells demon-strated homogeneous protein expression in the cells (Fig.

600 bp

600 bp

18s

28s

**

*

A B

C

D

FIG. 1. Leptin inhibits PP-1� mRNA and protein expression in INS-1 cells. A, SubtractionPCR, showing weight marker (WM), nonsubtracted cDNA (Non-SUB), and subtractedPCR (SUB) with the PP-1� band; B, representative semiquantitative RT-PCR for PP-1� inINS-1 cells after treatment with leptin (100 nM) or vehicle for up to 48 h (n � 5) (P �0.05); C, representative Northern blot for rat PP-1� in INS-1 cells treated for up to 48 hwith leptin (100 nM) (n � 3) (P � 0.05); D, representative Western blot for PP-1� in INS-1 cells after incubation with 100 nM leptin. PP-1� protein expression is inhibited in atime-dependent manner, whereas CREB (cAMP response element-binding) expression(control) was unchanged (n � 3). Bar graphs indicate the individual densitometric scanof the representative gel. *, P � 0.05.


2A). Immunohistochemical analysis of PP-1� revealed abroad distribution of PP-1� within the whole humanpancreatic tissue, in the pancreatic islets as well as in theexocrine pancreatic tissue (Fig. 2B). However, PP-1�

expression was seemingly enriched in pancreatic isletsand coexpressed with insulin in �-cells. PP-1� seems tobe expressed in the cytoplasm and in the nucleus. Thisis corresponding to former publications (35) where it isdemonstrated that PP-1� is present in nuclei as well asin the cytoplasm during interphase in HeLa cells. Inconclusion, we could demonstrate PP-1� expressionboth in the INS-1 �-cell line (Fig. 2A) and in humanpancreatic islets (Fig. 2B).

Leptin inhibits PP-1 enzyme activity in INS-1 cellsTo examine the effect of leptin on PP-1 enzymatic ac-

tivity in pancreatic �-cells extracts isolated from INS-1cells stimulated with 100 nM leptin, vehicle, or 10 nM CalA for 1 h were assayed for PP-1 enzymatic activity. Leptin

significantly inhibited PP-1 enzyme activity (P � 0.05)compared with vehicle control. The same inhibitory effectwas obtained when INS-1 cells were treated with Cal A, aknown inhibitor of PP-1 (Fig 3). When INS-1 cells wereincubated with leptin and Cal A together, there was noadditive effect, suggesting that both act through the samesignaling pathway. As an internal control for the PP-1assay, HepG2 cells were incubated with 100 nM insulin for2 h, and increase of PP-1 enzyme activity (P � 0.05) wasobserved as has been reported before (36).

Leptin and PP-1 inhibitor Cal A affect cytosoliccalcium concentrations in INS-1 cells

As shown in the enzyme activity studies, PP-1 activityis reduced by leptin exposure. Because PP-1 has also beendemonstrated to be involved in transmembrane ion chan-nel regulation (25), [Ca2�]i, a pivotal signal upstream ofinsulin secretion, was studied after leptin or Cal A treat-ment. To assess whether PP-1 interacts with Ca2� in INS-1cells, [Ca2�]i was measured by the fura-2 method.

INS-1 cells were preincubated with 100 nM leptin for12 h before leptin treatment or stimulation with 10 nM CalA. Glucose-stimulated intracellular Ca2� concentrationswere determined as the difference between average pre-and poststimulation levels of [Ca2�]i and were signifi-cantly reduced by both leptin and Cal A (39 and 49%,

FIG. 2. Immunohistochemical detection of PP-1� in INS-1 cells andhuman pancreatic islets. A, Nuclei of INS-1 cells were visualized by4�,6-diamidino-2-phenylindole (DAPI) staining. Anti-PP-1� polyclonalantibody (red) was used for immunofluorescence staining of PP-1�; B,staining for human PP-1� in human pancreatic frozen sections by arabbit monoclonal antibody, followed by Cy3-conjugated secondarydonkey antirabbit antibody staining (red). Insulin detection was carriedout with a polyclonal guinea pig antiswine insulin antibody followed byCy2-conjugated secondary goat anti-guinea pig antibody (green).Nuclei were stained with 4�,6-diamidino-2-phenylindole (blue). Thebackground panels are visualizing the grade of nonspecific staining bythe antibodies. PC, Phase contrast.

FIG. 3. Leptin inhibits PP-1 enzyme activity. In INS-1 cells, leptin (100nM) as well as Cal A (10 nM) reduces PP-1 enzyme activity (P � 0.05)after 1 h of incubation each. This inhibition was also found when INS-1cells were incubated with leptin and Cal A (P � 0.05) (n � 8). Incontrol experiments with HepG2 cells, the PP-1 enzyme activity wasstimulated after 60 min by insulin (100 nM) (inset). *, P � 0.05.


respectively, P � 0,05) compared withvehicle control-treated INS-1 cells (Fig.4A). Leptin reduced [Ca2�]i to a similarextent as Cal A, a PP-1 inhibitor (Fig.4A). This was confirmed by looking at[Ca2�]i traces of individual cells (rep-resentative images, Fig. 4B), where glu-cose-induced [Ca2�]i spiking was com-parably reduced by leptin and Cal A.

These results suggest that leptin inhib-its insulin secretion through negative ac-tion on PP-1 enzyme activity that conse-quently blunts Ca2� influx as a pivotalprerequisite for insulin exocytosis.

Effects of leptin, Cal A, and NIPP1on insulin secretion of INS-1 cellsand human pancreatic islets

When INS-1 cells were stimulatedwith glucose, we found a slight increasein insulin secretion as has been docu-mented before (37). These results indi-cate the INS-1 cells could be used forglucose-stimulated insulin secretion ex-periments. Although the basal insulinsecretion in INS-1 cells could not be sig-nificantly reduced by leptin, we found asignificant reduction in the glucose-stimulated insulin secretion by leptin,as has been published previously (7). Toinvestigate whether an inhibition ofPP-1� activity regulates insulin secre-tion, INS-1 cells were treated with twodifferent inhibitors of PP-1, Cal A andNIPP1. In additional confirmatory ex-periments, the effect of Cal A on insulinsecretion was also analyzed in humanpancreatic islets. NIPP1 is one of fourancient regulators of PP-1 and is ubiq-uitously expressed in multicellular eu-karyotes (38).

Both the basal as well as the glucose-stimulated insulin secretion is signifi-cantly reduced by NIPP1 (P � 0.05).Leptin and NIPP1 combined alsoshowed a significant suppression of thebasal and glucose-induced insulin se-cretion, but there was no additive effect(Fig. 5A). Furthermore, insulin secre-tion was analyzed in human pancreaticislets. Leptin inhibits the glucose-stimu-lated insulin secretion (Fig. 5B). More-

FIG. 4. Leptin and inhibition of PP-1� reduces [Ca2�]i in INS-1 cells. A, Average changesin [Ca2�]i levels in INS-1 pancreatic �-cells (n � 12 for each treatment) in response toelevation of extracellular glucose from 0.5 to 15 mM for 15 min in the presence orabsence of 100 nM leptin or 10 nM Cal A. Leptin-treated cells had been pretreated with100 nM leptin for 12 h in normal growth medium. B, Representative traces of [Ca2�]ireadings in individual cells in response to elevation of extracellular glucose from 0.5 to 15mM for 15 min in the absence (upper panel) or presence of 100 nM leptin (middle panel)or 10 nM Cal A (lower panel). KCl (45 mM) was used at the end of the readings as acontrol for maintained depolarization capacity in each cell. *, P � 0.05.


over, also PP-1 inhibition by Cal A leads to a reducedinsulin secretion in INS-1 cells and human pancreatic islets(Fig. 5C). This is compatible with the results of the PP-1enzyme activity studies. In summary, these results indicatethat PP-1 plays a role in the regulation of glucose-inducedinsulin release in INS-1 cells and may mediate inhibition ofinsulin secretion by leptin.

Discussion

Humoral interactions of adipose tissue with cells and or-gans regulating glucose homeostasis have been increas-ingly demonstrated to be involved in the pathogenesis oftype 2 diabetes mellitus (39).

The adipose tissue hormone leptin, within an adipoin-sular feedback loop, has been linked to a complex systemmodulating glucose-dependent insulin secretion in pan-

creatic islets (1, 2, 40). In addition to leptin receptor-me-diated inhibition of preproinsulin mRNA expression (10),leptin-mediated opening of ATP-dependent potassiumchannels and activation of phosphodiesterase 3B andthereby reduction in glucose transport had been identified(41). Tissue-specific leptin receptor knockout of either thewhole pancreas or combined in hypothalamus and �-cellsindicates the existence of an adipose tissue-islet endocrinefeedback loop (1, 2). During the development of type 2diabetes, hyperinsulinemia has been interpreted after themanifestation of obesity as a result of insulin resistance.However, in the tissue-specific leptin receptor knockoutmouse model, hyperinsulinemia was observed before thedevelopment of insulin resistance (3). Therefore, distur-bances of physiological leptin-mediated inhibition of in-sulin secretion due to the development of leptin resistancein obesity may be important for the development of type2 diabetes mellitus.

Because leptin seems to act at multiple cellular levels inpancreatic �-cells (40), we sought to gain deeper insightinto leptin signaling in pancreatic �-cells. We performed aPCR-based subtraction hybridization assay and identifiedPP-1�, in addition to insulin (not shown), to be stronglyregulated by leptin in INS-1 cells. Leptin inhibits PP-1�

mRNA and protein expression in a time-dependent man-ner. Moreover, leptin reduced PP-1 enzyme activity inINS-1 �-cells to an extent that could not be further bluntedby the PP-1 inhibitor Cal A. Furthermore, in both INS-1cells and human pancreatic islets, the PP-1 inhibitor Cal Aand, in INS-1 cells, the more specific PP-1 inhibitor NIPP1mimicked leptin-dependent suppression of glucose-in-duced insulin secretion. The necessity of longer NIPP1treatment with a 3-h incubation period to reduce insulinsecretion compared with 1 h in the Caln A incubationexperiment might be due to the more indirect effect ofNIPP1 on PP-1 (42). Finally, leptin action was associatedwith a reduction of [Ca2�]i similar to those that wereachieved by inhibition of PP-1 by Cal A. These resultstogether indicate short-term as well as long-term effects ofleptin on the pancreatic �-cell as documented before dur-ing the leptin-mediated inhibition of preproinsulin expres-sion (7). Long-term effects might involve a different sig-naling pathway until gene expression of PP-1 is reducedafter 24–48 h. Furthermore, leptin acts in a short periodof time potentially directly on target proteins and therebymay lead to a reduced PP-1 enzyme activity within only afew hours of stimulation. Together these results identifyleptin-dependent inhibition of PP-1� expression and con-sequently reduced PP-1 enzyme activity in pancreatic�-cells as a novel and independent cellular mechanism inthe cross talk between adipose tissue and endocrine pan-creas to regulate glucose homeostasis.

C

Control Cal A Control Cal A

Rel

. 1 h

Insu

lin S

ecre

tion

INS-1 Cells Human Isletsn = 4

11,1 mM Glucose

0,0

0,2

0,4

0,6

0,8

1,0

*

Leptin 0

50

100

150

200

*'

' **

Insu

lin (%

2,8

mM

con

trol)

A

Control NIPP1 Leptin+NIPP1

B

Rel

. 1 h

Insu

lin S

ecre

tion

Human Isletsn = 3

11,1 mM Glucose

0,0

0,2

0,4

0,6

0,8

1,0

Control Leptin

*

Glucose2,8 mM28 mM

FIG. 5. Insulin secretion is reduced by leptin and inhibition of PP-1� byNIPP1 and Cal A. A, The 1-h insulin secretion in INS-1 cells in responseto 28 mM glucose stimulation (black bars) after preincubation at 2.8mM glucose (white bars) in the absence (control) or presence of 100nM leptin or the specific PP-1 inhibitor NIPP-1 or both (L � N). *, P �0.05 (n � 4 per incubation condition; �, significant difference (P �0.05) from control at 2.8 mM glucose concentration. B, The 1-h insulinsecretion in human pancreatic islets in response to 11.1 mM glucosestimulation in the absence (control) or presence of 100 mM leptin (P �0.05). C, Insulin secretion of INS-1 cells (P � 0.1) and humanpancreatic islets (P � 0.05) before and after treatment with 40 nM CalA for 1 h at a glucose concentration of 11.1 mM.


PP-1 � has been suggested previously to play a role ininsulin secretion in pancreatic �-cells. Inhibition of PP-1 invitro by okadaic acid or Cal A decreased glucose-inducedinsulin secretion of pancreatic �-cells (25, 27, 28). Thiswas due to PP-1-mediated down-regulation of Ca2� influxinto pancreatic �-cells (25) and impaired regulation ofvesicular fusion of secretory vesicles (30, 43). Further-more, PP-1 has been shown to regulate overall translationin pancreatic �-cells (31). These findings indicate thattightly regulated PP-1 action represents a mechanism tomodulate insulin secretion at a vesicular and translationallevel in pancreatic �-cells.

The fact that we identified PP-1� as a leptin down-regulated gene in pancreatic �-cells provides strong evi-dence for the notion that leptin may mediate its inhibitoryeffects on insulin secretion at least in part through thisPP-1-dependent cellular signaling system. In addition topreviously characterized molecular effects of leptin in pan-creatic �-cells, this is a novel cellular mechanism of leptinaction that, to our knowledge, has been identified for thefirst time in insulin-producing cells.

Currently, it is unclear, however, how leptin inhibitsPP-1� expression in pancreatic �-cells at the molecularlevel. It also remains to be demonstrated whether this lep-tin effect is dependent on functional leptin receptor sig-naling or whether this is a leptin receptor-independentfunction. We have previously shown that leptin inhibitsproinsulin gene expression at the transcriptional level byinduction of the signal transducer and activator of tran-scription inhibitory molecule suppressor of cytokine sig-naling 3 in pancreatic �-cells (44). Therefore, one mayspeculate that inhibition of expression of PP-1� by leptinmay be mediated by similar transcriptional mechanisms.

PP-1 has also been demonstrated to be involved in in-sulin signaling (23) and phosphoinositol-3 kinase activityhas been proposed to integrate leptin and insulin signalingin the hypothalamus (45, 46). Because both insulin andleptin signaling are active in pancreatic �-cells, PP-1 maybe involved in a signal transduction network that inte-grates hormonal signals to fine-tune pancreatic �-cellfunction. An interesting candidate as a mediator betweenthese different signaling cascades seems to be AMPK. Re-cently, it has been demonstrated that a PP-1-R6 complexleads to an inactivation of AMPK in MIN6 �-cells (33). Itis tempting to speculate that the leptin-mediated regula-tion of PP-1, and thereby AMPK leads to an interfacebetween the signals of both hormones insulin and leptin onthe pancreatic �-cell, which is finally combined to a singleresponse of the insulin-producing �-cell.

Taken together, we have identified PP-1� expressionand PP-1 enzyme activity to be inhibited by leptin in pan-creatic �-cells. PP-1-dependent leptin action negatively

modulates insulin secretion in a calcium-dependent man-ner. Because intact leptin signaling is pivotal for the phys-iological function of pancreatic �-cells, leptin resistance ininsulin-producing cells may also affect normal PP-1 func-tion and thereby contribute to �-cell dysregulation withinthe adipoinsular axis (11, 40).

Thus, PP-1 may represent a molecular target for thedevelopment of strategies to improve or maintain pancre-atic �-cell function in adipogenic diabetes mellitus.

Acknowledgments

We thank J. Roller and S. Royer for expert technical assistance.We thank M. D. Brendel and R. G. Bretzel from the Islet Trans-plantation Center in Giessen, Germany, for human pancreaticislets.

Address all correspondence and requests for reprints to: Dr.Peter Kuehnen, Institute of Experimental Pediatric Endocrinol-ogy, Charite Children�s Hospital, Universitatsmedizin Berlin,Augustenburger Platz 1, 13353 Berlin, Germany. E-mail: [email protected].

J.S. has received support from the German Diabetes Associ-ation (Deutsche Diabetes-Gesellschaft), from the Juvenile Dia-betes Research Foundation (U.S.), the German Ministry of Ed-ucation and Research within the “Competence NetworkDiabetes Mellitus,” the DANONE Foundation for Nutrition,and the German Research Council (Deutsche Forschungsge-meinschaft). K.L. was supported by the Research Commission(Forschungskommission) of the Faculty of Medicine at the Uni-versity of Freiburg, Germany.

Disclosure Summary: The authors have nothing to disclose.

References

1. Morioka T, Asilmaz E, Hu J, Dishinger JF, Kurpad AJ, Elias CF, LiH, Elmquist JK, Kennedy RT, Kulkarni RN 2007 Disruption ofleptin receptor expression in the pancreas directly affects �-cellgrowth and function in mice. J Clin Invest 117:2860–2868

2. Covey SD, Wideman RD, McDonald C, Unniappan S, Huynh F,Asadi A, Speck M, Webber T, Chua SC, Kieffer TJ 2006 The pan-creatic �-cell is a key site for mediating the effects of leptin on glucosehomeostasis. Cell Metab 4:291–302

3. Gray SL, Donald C, Jetha A, Covey SD, Kieffer TJ 2010 Hyperin-sulinemia precedes insulin resistance in mice lacking pancreatic�-cell leptin signaling. Endocrinology 151:4178–4186

4. Zhang Y, Proenca R, Maffei M, Barone M, Leopold L, Friedman JM1994 Positional cloning of the mouse obese gene and its humanhomologue. Nature 372:425–432

5. Flier JS 1997 Leptin expression and action: new experimental par-adigms. Proc Natl Acad Sci USA 94:4242–4245

6. Cases JA, Gabriely I, Ma XH, Yang XM, Michaeli T, Fleischer N,Rossetti L, Barzilai N 2001 Physiological increase in plasma leptinmarkedly inhibits insulin secretion in vivo. Diabetes 50:348–352

7. Seufert J, Kieffer TJ, Leech CA, Holz GG, Moritz W, Ricordi C,Habener JF 1999 Leptin suppression of insulin secretion and geneexpression in human pancreatic islets: implications for the develop-


ment of adipogenic diabetes mellitus. J Clin Endocrinol Metab 84:670–676

8. Fehmann HC, Berghofer P, Brandhorst D, Brandhorst H, Hering B,Bretzel RG, Goke B 1997 Leptin inhibition of insulin secretion fromisolated human islets. Acta Diabetol 34:249–252

9. Kieffer TJ, Heller RS, Leech CA, Holz GG, Habener JF 1997 Leptinsuppression of insulin secretion by the activation of ATP-sensitiveK� channels in pancreatic �-cells. Diabetes 46:1087–1093

10. Seufert J, Kieffer TJ, Habener JF 1999 Leptin inhibits insulin genetranscription and reverses hyperinsulinemia in leptin-deficientob/ob mice. Proc Natl Acad Sci USA 96:674–679

11. Kieffer TJ, Habener JF 2000 The adipoinsular axis: effects of leptinon pancreatic �-cells. Am J Physiol Endocrinol Metab 278:E1–E14

12. Holz GG, Habener JF 1992 Signal transduction crosstalk in theendocrine system: pancreatic �-cells and the glucose competenceconcept. Trends Biochem Sci 17:388–393

13. Coleman DL 1978 Obese and diabetes: two mutant genes causingdiabetes-obesity syndromes in mice. Diabetologia 14:141–148

14. Coleman DL 1982 Diabetes-obesity syndromes in mice. Diabetes31:1–6

15. Pelleymounter MA, Cullen MJ, Baker MB, Hecht R, Winters D,Boone T, Collins F 1995 Effects of the obese gene product on bodyweight regulation in ob/ob mice. Science 269:540–543

16. Lin Q, Buckler 4th ES, Muse SV, Walker JC 1999 Molecular evo-lution of type 1 serine/threonine protein phosphatases. Mol Phylo-genet Evol 12:57–66

17. Ceulemans H, Bollen M 2004 Functional diversity of protein phos-phatase-1, a cellular economizer and reset button. Physiol Rev 84:1–39

18. Herzig S, Neumann J 2000 Effects of serine/threonine protein phos-phatases on ion channels in excitable membranes. Physiol Rev 80:173–210

19. Wera S, Hemmings BA 1995 Serine/threonine protein phosphatases.Biochem J 311(Pt 1):17–29

20. Terrak M, Kerff F, Langsetmo K, Tao T, Dominguez R 2004 Struc-tural basis of protein phosphatase 1 regulation. Nature 429:780–784

21. Barford D, Das AK, Egloff MP 1998 The structure and mechanismof protein phosphatases: insights into catalysis and regulation. AnnuRev Biophys Biomol Struct 27:133–164

22. Bollen M, Stalmans W 1992 The structure, role, and regulation oftype 1 protein phosphatases. Crit Rev Biochem Mol Biol 27:227–281

23. Brady MJ, Saltiel AR 2001 The role of protein phosphatase-1 ininsulin action. Recent Prog Horm Res 56:157–173

24. Mehebik N, Jaubert AM, Sabourault D, Giudicelli Y, Ribiere C2005 Leptin-induced nitric oxide production in white adipocytes ismediated through PKA and MAP kinase activation. Am J PhysiolCell Physiol 289:C379–C387

25. Sato Y, Mariot P, Detimary P, Gilon P, Henquin JC 1998 Okadaicacid-induced decrease in the magnitude and efficacy of the Ca2�

signal in pancreatic �-cells and inhibition of insulin secretion. Br JPharmacol 123:97–105

26. Haby C, Larsson O, Islam MS, Aunis D, Berggren PO, Zwiller J1994 Inhibition of serine/threonine protein phosphatases promotesopening of voltage-activated L-type Ca2� channels in insulin-secret-ing cells. Biochem J 298(Pt 2):341–346

27. Tamagawa T, Iguchi A, Uemura K, Miura H, Nonogaki K, IshiguroT, Sakamoto N 1992 Effects of the protein phosphatase inhibitorsokadaic acid and calyculin A on insulin release from rat pancreaticislets. Endocrinol Jpn 39:325–329

28. Ratcliff H, Jones PM 1993 Effects of okadaic acid on insulin secre-tion from rat islets of Langerhans. Biochim Biophys Acta 1175:188–191

29. Murphy LI, Jones PM 1996 Phospho-serine/threonine phosphatasesin rat islets of Langerhans: identification and effect on insulin se-cretion. Mol Cell Endocrinol 117:195–202

30. Sim AT, Baldwin ML, Rostas JA, Holst J, Ludowyke RI 2003 Therole of serine/threonine protein phosphatases in exocytosis.Biochem J 373:641–659

31. Vander Mierde D, Scheuner D, Quintens R, Patel R, Song B, Tsu-kamoto K, Beullens M, Kaufman RJ, Bollen M, Schuit FC 2007Glucose activates a protein phosphatase-1-mediated signaling path-way to enhance overall translation in pancreatic �-cells. Endocri-nology 148:609–617

32. Lilja L, Meister B, Berggren PO, Bark C 2005 DARPP-32 and in-hibitor-1 are expressed in pancreatic �-cells. Biochem Biophys ResCommun 329:673–677

33. Garcia-Haro L, Garcia-Gimeno MA, Neumann D, Beullens M, Bol-len M, Sanz P 2010 The PP1–R6 protein phosphatase holoenzymeis involved in the glucose-induced dephosphorylation and inactiva-tion of AMP-activated protein kinase, a key regulator of insulinsecretion, in MIN6 �-cells. FASEB J 24:5080–5091

34. Ragolia L, Begum N 1998 The relationship between insulin signal-ing and protein phosphatase 1 activation. Methods Mol Biol 93:157–167

35. Andreassen PR, Lacroix FB, Villa-Moruzzi E, Margolis RL 1998Differential subcellular localization of protein phosphatase-1 �, �1,and � isoforms during both interphase and mitosis in mammaliancells. J Cell Biol 141:1207–1215

36. Syed NA, Khandelwal RL 2000 Reciprocal regulation of glycogenphosphorylase and glycogen synthase by insulin involving phospha-tidylinositol-3 kinase and protein phosphatase-1 in HepG2 cells.Mol Cell Biochem 211:123–136

37. Sekine N, Fasolato C, Pralong WF, Theler JM, Wollheim CB 1997Glucose-induced insulin secretion in INS-1 cells depends on factorspresent in fetal calf serum and rat islet-conditioned medium. Dia-betes 46:1424–1433

38. Ceulemans H, Stalmans W, Bollen M 2002 Regulator-driven func-tional diversification of protein phosphatase-1 in eukaryotic evolu-tion. Bioessays 24:371–381

39. Rasouli N, Kern PA 2008 Adipocytokines and the metabolic com-plications of obesity. J Clin Endocrinol Metab 93:S64–S73

40. Seufert J 2004 Leptin effects on pancreatic �-cell gene expressionand function. Diabetes 53(Suppl 1):S152–S158

41. Lam NT, Cheung AT, Riedel MJ, Light PE, Cheeseman CI, KiefferTJ 2004 Leptin reduces glucose transport and cellular ATP levels inINS-1 �-cells. J Mol Endocrinol 32:415–424

42. Boudrez A, Beullens M, Groenen P, Van Eynde A, Vulsteke V, Jag-iello I, Murray M, Krainer AR, Stalmans W, Bollen M 2000 NIPP1-mediated interaction of protein phosphatase-1 with CDC5L, a reg-ulator of pre-mRNA splicing and mitotic entry. J Biol Chem 275:25411–25417

43. Peters C, Andrews PD, Stark MJ, Cesaro-Tadic S, Glatz A, Podtele-jnikov A, Mann M, Mayer A 1999 Control of the terminal step ofintracellular membrane fusion by protein phosphatase 1. Science285:1084–1087

44. Laubner K, Kieffer TJ, Lam NT, Niu X, Jakob F, Seufert J 2005Inhibition of preproinsulin gene expression by leptin induction ofsuppressor of cytokine signaling 3 in pancreatic �-cells. Diabetes54:3410–3417

45. Hill JW, Williams KW, Ye C, Luo J, Balthasar N, Coppari R, CowleyMA, Cantley LC, Lowell BB, Elmquist JK 2008 Acute effects ofleptin require PI3K signaling in hypothalamic proopiomelanocortinneurons in mice. J Clin Invest 118:1796–1805

46. Xu AW, Kaelin CB, Takeda K, Akira S, Schwartz MW, Barsh GS2005 PI3K integrates the action of insulin and leptin on hypotha-lamic neurons. J Clin Invest 115:951–958


Protein Phosphatase 1 (PP-1)-Dependent Inhibition of Insulin Secretion by Leptin in INS-1 Pancreatic...

Documents

Transcript of Protein Phosphatase 1 (PP-1)-Dependent Inhibition of Insulin Secretion by Leptin in INS-1 Pancreatic...