| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(PPAR) Potentiates, whereas PPAR
Attenuates, Glucose-Stimulated Insulin Secretion in Pancreatic ß-Cells
Department of Biochemistry and Molecular Biology, University of Southern Denmark (K.R., M.B., J.K.L., T.N., J.F., S.M.), 5230 Odense M, Denmark; and Department of Cell Physiology and Metabolism, University Medical Center (B.R., P.M.), CH-1211 Geneva 4, Switzerland
Address all correspondence and requests for reprints to: Dr. Susanne Mandrup, Department of Biochemistry and Molecular Biology University of Southern Denmark, Campusvej 55, 5230 Odense M, Denmark. E-mail: s.mandrup{at}bmb.sdu.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and retinoid X receptor (RXR) in INS-1E ß-cells synergistically and in a dose- and ligand-dependent manner increases the expression of known PPAR target genes and enhances FA uptake and ß-oxidation. In contrast, ectopic expression of PPAR
/RXR increases FA uptake and deposition as triacylglycerides. Although the expression of PPAR/RXR leads to the induction of UCP2 mRNA and protein, this is not accompanied by reduced hyperpolarization of the mitochondrial membrane, indicating that under these conditions, increased UCP2 expression is insufficient for dissipation of the mitochondrial proton gradient. Importantly, whereas expression of PPAR/RXR attenuates GSIS, the expression of PPAR/RXR potentiates GSIS in rat islets and INS-1E cells without affecting the mitochondrial membrane potential. These results show a strong subtype specificity of the two PPAR subtypes and on lipid partitioning and insulin secretion when systematically compared in a ß-cell context. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, -, and -
) bind to PPAR-responsive elements as heterodimers with retinoid X receptor (RXR), another member of the nuclear receptor superfamily (2). The different PPAR subtypes show very limited subtype specificity in transient transfections and in the in vitro binding to PPAR-responsive elements, but it is evident from a number of in vivo and ex vivo studies that the different PPAR subtypes do have specialized functions when acting on endogenous genes. Thus, PPAR activates primarily genes encoding proteins involved in FA oxidation during fasting (3), whereas PPAR activates genes directly involved in lipogenic pathways and insulin signaling (4). The different biological actions of PPAR subtypes are undoubtedly due in part to the differential expression patterns of the PPAR subtypes; however, there are also indications that the PPARs are biochemically different and behave differently when ectopically expressed in the same cell (5).
All members of the PPAR family have been reported to be expressed in pancreatic ß-cells (6); however, several recent reports challenge the current dogma on the functions of the different subtypes. It was recently shown that PPAR ectopically expressed in INS-1 cells could induce lipid accumulation along with a modest increase in ß-oxidation (7). Likewise, others deduce that PPAR has lipooxidative potential promoting FA disposal in pancreatic ß-cells (8). This again disagrees with the conventional lipogenic role of PPAR, but is supported by reports indicating that PPAR ligands induce delipidation of pancreatic islets (9, 10). Thus, when related to findings in other tissues, the confusion about the direct function of PPAR subtypes in pancreatic ß-cell lipid partitioning is remarkable.
FAs are known to have profound effect on the ability of ß-cells to perform glucose-stimulated insulin secretion (GSIS) (1). It is well established that FAs acutely potentiate insulin secretion by glucose and other secretagogues in a manner dependent on both chain length and degree of saturation (11, 12). This effect is mediated by increased levels of intracellular long-chain FA-coenzyme A (FA-CoA) esters (13) as well as through direct activation of the G protein-coupled receptor 40 (14). Because FA oxidation decreases cytoplasmic levels of long-chain acyl-CoA esters, conditions that increase FA oxidation have been suggested to counteract the potentiating effect of FAs on insulin secretion (15).
In contrast to the acute effects of FAs on insulin secretion, exposure of ß-cells to high levels of FAs over a longer time (>24 h) leads to elevated basal insulin secretion and compromised GSIS, a phenomenon termed lipotoxicity. The molecular mechanisms underlying lipotoxicity are unclear, but they have been suggested to involve modulation of the glucose flux (16, 17), induction of uncoupling protein-2 (UCP2) (18, 19), activation of UCP2 by increased levels of reactive oxygen species (ROS) (20, 21), and facilitation of ceramide synthesis (22, 23).
Particular attention has recently been given to the role of UCP2 in modulating GSIS. It is well established that UCP2 activity in ß-cell mitochondria and GSIS capacity are negatively correlated in certain contexts (24, 25), and evidence was recently presented that this uncoupling activity is dependent on the generation of ROS (26). However, the physiological role of UCP2 in ß-cells explaining the evolutionary conservation remains less evident from these studies. Other reports indicate that UCP2 may facilitate FA oxidation. It was found that ectopic expression of UCP2 in either the INS-1 insulinoma cell line or Zucker diabetic fatty (ZDF) rat islets significantly accelerated FA oxidation (27, 28). This may be due to increased mitochondrial recycling of FAs (29, 30) or rapid neutralization of oxygen radicals (20, 31). Interestingly, both PPAR (7, 32) and PPAR (33, 34) have been reported to induce UCP2 expression in ß-cells and other cell types.
Hence, models positioning both PPAR and - as mediators of the adverse effects of FAs on ß-cell function have been suggested (7, 34). A significant reduction in GSIS by both ectopic PPAR expression and the PPAR agonist clofibrate led to the conclusion that PPAR activity could indeed cause ß-cell dysfunction, possibly through an induction of UCP2 (7). However, these results are at odds with the generally protective and restorative effects of PPAR activation in islets experiencing increased FA challenge (6, 35, 36). A net contribution of PPAR activation to ß-cell dysfunction also disagrees with the glucolipotoxicity hypothesis describing the additive toxic effect of elevated levels of FAs and glucose on ß-cell function as the ability of glucose to repress PPAR expression and, accordingly, FA oxidation (37, 38). Similarly, the finding that PPAR agonists counteract FA toxicity on pancreatic ß-cells (10, 36, 39) does not agree with PPAR being the mediator of FA induction of UCP2 and attenuation of GSIS (34)
The inconsistent observations on the effects of PPAR and PPAR on lipid partitioning and ß-cell function prompted us to reinvestigate and directly compare the roles of these PPAR subtypes in ß-cell function. In the present work we show that acute ectopic expression of PPAR and RXR synergistically and in a PPAR subtype-specific manner increases FA uptake and oxidation capacity and induces UCP2 expression in the INS-1E rat ß-cell line. Interestingly, despite this increase in FA oxidation and UCP2 expression, transduction with PPAR/RXR potentiates GSIS of both INS-1E cells and isolated rat islets. In contrast, acute ectopic expression of PPAR and RXR does not affect FA oxidation, but induces massive accumulation of triglycerides and impairs GSIS. These results confirm the roles of PPAR as a catabolic and PPAR as a lipogenic transcription factor also when ectopically expressed in pancreatic ß-cells. Importantly, our results show that acute activation of PPAR potentiates, whereas acute activation of PPAR compromises, GSIS.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Adenovirus generation and transduction
Recombinant adenoviruses containing full-length mouse PPAR, PPAR2, and RXR were generated using the AdEasy cloning system from Stratagene (La Jolla, CA). The three genes were initially cloned into pShuttle-CMV using SalI and EcoRV restriction sites and were moved to AdEasy-1 by homologous recombination. The linearized plasmids were transfected into 293-HEK cells, and the virus was amplified and purified using CsCl gradients. Viruses were initially titrated, and titers were estimated by a plaque assay-based approach. Subsequently, relative titers of functional viruses were equalized based on quantification of the adenoviral transcript AdE4 by real-time PCR. Relative titers 1, 2, 5, and 25 correspond to 4, 8, 20, and 100 plaque-forming units/cell respectively.
After medium removal, cells were washed once in PBS and transduced for 90 min in RPMI 1640 supplemented as described above. Viruses were removed, and new medium containing specific agonists was added for an additional 24- or 48-h incubation. In all experiments, equal total amounts of AdVector were used as a control, and dimethylsulfoxide (DMSO) was added to the incubation medium. All experiments for RNA and protein extraction were performed in duplicate.
[1-14C]Oleate uptake
Transduced INS-1E cells (relative titer, 1, 2, and 5; 24-h incubation) were washed twice in Hanks’ buffered salt solution (HBSS) and incubated for 30 min in HBSS and 10 µM oleate (3 nCi/ml [1-14C]oleate; Pharmacia Biotech, Uppsala, Sweden). Oleate uptake was stopped by the addition of ice-cold stop solution (HBSS and 2% BSA), and cells were washed three times herein and once in ice-cold PBS. The 14C content was quantified and normalized to total DNA per well. Dishes without cells were treated in parallel, and values were used for background subtraction. Experiments were performed in triplicate.
[1-14C]Oleate oxidation
Oleate oxidation over 4 h of transduced INS-1E cells (relative titer, 1, 2, 5, and 25; 24-h incubation) was measured by quantification of collected 14CO2 essentially as described by Berge et al. (41). Data were normalized to total cellular protein. Dishes without cells were treated in parallel and used for background subtraction. Experiments were performed in triplicate.
[1-14C]Glucose oxidation
Glucose oxidation of transduced INS-1E cells (relative titer, 5) was measured by quantification of 14CO2 produced from oxidation of [1-14C]glucose. Cells were harvested in supplemented RPMI 1640 with 5 mM glucose and collected in sealed glass tubes. Glucose was raised to 16.7 mM (0.4 µCi [1-14C]glucose/tube), and cells were incubated for 30 min at 37 C. Collection, quantification, and normalization of 14CO2 were performed as described above (40). Experiments were performed in triplicate.
Lipid accumulation assay
INS-1E cells were seeded on Falcon chamber slides (BD Biosciences, Franklin Lakes, NJ), transduced with viruses (relative titer, 2), and incubated for 48 h with DMSO or specific agonists as described. Cells were fixed in ice-cold PBS/4% paraformaldehyde, washed in 70% ethanol, and stained with Nile Red (1 ng/µl; Sigma-Aldrich Corp., St. Louis, MO). Stained cells were washed three times in ice-cold 70% ethanol. Lipid accumulation was visualized by confocal laser scanning microscopy (excitation, 543 nm) using the LSM 510 Meta confocal microscope (Carl Zeiss, Inc., New York, NY).
Protein analysis by Western blotting and ECL detection
Transduced INS-1E cells (relative titer, 1, 2, and 5; 24-h incubation) were harvested in hypotonic lysis buffer containing sodium dodecyl sulfate and separated by SDS-PAGE. Proteins were blotted onto polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA) and probed with specific antibodies. All primary antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA): anti-PPAR (SC-7273), anti-RXR (SC-553), antitranscription factor-IIB (anti-TFIIB; SC-225), and anti-UCP2 (SC-20). Secondary horseradish peroxidase-coupled anti-Fc antibodies, antimouse (P0447) and antirabbit (P0399), were obtained from DakoCytomation (Carpinteria, CA).
RNA isolation and cDNA synthesis
Transduced INS-1E cells (relative titer, 1, 2, and 5; 24-h incubation) were harvested in guanidium thiocyanate, and RNA was isolated according to a modified Chomczynski-Sacchi protocol (42). cDNA was prepared after deoxyribonuclease I treatment (Invitrogen Life Technologies, Inc.) by RT (Invitrogen Life Technologies, Inc., First-Strand Kit) of the isolated RNA primed by random hexamers (dexoy-NTP6).
Real-time PCR
Quantitative three-step real-time PCR was performed on the ABI-7700 PRISM real-time PCR instrument (Applied Biosystems, Foster City, CA) using 2x SYBR Green Master Mix (Sigma-Aldrich Corp.) and Sigma passive reference according to instructions from the manufacturer. PCRs were performed in duplicate. Primers for real-time PCR were designed using Primer Express 2.0 (Applied Biosystems), and specificity and efficacy were validated before use. All quantifications were performed with TFIIB as the internal standard and are presented as the fold increase over the control value.
Primer sequences (forward and reverse, respectively) are as follows: RXR, CCTGCCGAGACAACAAGGA and ACCGGTTCCGCTGTCTCTT; PPAR, GTACCACTACGGAGTTCACGCAT and CGCCGAAAGAAGCCCTTAC; PPAR, CACAATGCCATCAGGTTTGG and CAGCTTCTCCTTCTCGGCCT; carnitine palmitoyl transferase-1a (CPT-1a), CTGGTGGGCCACAAATTACG and AGGTAGATATATTCTTCCCACCAGTCA; CPT-1b, TCCAAACATCACTGCCCAAG and GAATTGTGGCTGGCACACTG; UCP2, CCTACAGCGCCAGATGAGCT and GAGTCGTAGAGGCCAATGCG; medium-chain acyl-CoA dehydrogenase (MCAD), AGCTGCTAGTGGAGCACCAAG and TCGCCATTTCTGCGAGC; long-chain acyl-CoA dehydrogenase (LCAD), TCACCAATCGTGAAGCTCGA and CCAAAAAGAGGCTAATGCCATG; acyl-coenzyme A oxidase (ACO), CAGATAATTGGCACCTACGCC and AAGATGAGTTCCGTGGCCC; fatty acid transport protein-1 (FATP1), GTGTACCCCATCCGTCTGGT and CAGTGGCTCCATCGTGTCCT; FATP2, TGGACGCGATCATCTCTGG and TGTGGCGAAGGTTTCAAGGT; catalase, CGTGGGAAACAACACCCCTA and GGAAACAACATGGCATCCCT; perilipin, GGGACCTGTGAGTGCTTCCA and GTGGGCTTCTTTGGTGCTGT; pyruvate dehydrogenase kinase-4 (PDK4), CACGCTGGTCAAAGTTCGAA and GCCATCGTAGGGACCACATT; CD36, TTCTGCTGCACGAGGAGGA and AATGAGCCCACAGTTCCGAT; TFIIB, GTTCTGCTCCAACCTTTGCCT and TGTGTAGCTGCCATCTGCACTT; and AdE4, CTCCGGAACCACCACAGAAA and GCAGACATGTTTGAGAGAAAAATGG.
Mitochondrial membrane potential (
m)
The m of transduced INS-1E cells (relative titer, 5; 24-h incubation) was measured using the fluorescent probe, rhodamine 123 (Molecular Probes, Eugene, OR) as described previously (40). Cells were maintained for 2 h in 2.5 mM glucose medium at 37 C before loading with 10 µg/ml rhodamine 123 for 20 min at 37 C in Krebs ringer bicarbonate HEPES. The m was monitored with excitation and emission filters set at 485 and 520 nm, respectively. Glucose (additions on top of basal, 2.5 mM), and then the protonophore carbonyl cyanide p-trifluoromethoxyphenylhydrazone was added to each well.
Insulin secretion
Insulin secretion from transduced INS-1E cells and rat islets (relative titer, 5; 24-h incubation) over a period of 30 min was measured by RIA using rat insulin as standard as previously described (40), and values were normalized to total insulin content.
Statistical analysis
Statistical evaluation of the data was performed using Student’s two-tailed t test on paired data. Data (fold increase over control value) are presented as the mean ± SD (n 
3) or the mean ± range (n = 2).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and RXR in the INS-1E ß-cell line (6, 37). To investigate the acute effect of increased PPAR activity on target gene expression in pancreatic ß-cells, we transduced INS-1E cells with adenoviral vectors expressing PPAR. To investigate whether the endogenous level of RXR would become limiting for gene activation, we cotransduced with adenoviral vectors expressing RXR. INS-1E cells were transduced with the indicated viruses and cultured for 24 h with DMSO, the PPAR agonist WY14643, and/or the RXR agonist LG100268 as indicated. AdVector was used as a control. The viruses were all used with the relative titer 2 based on Ad4E expression (see Materials and Methods). Real-time PCR was used to quantify expression levels of the two classical PPAR target genes, CPT-1b and ACO (Fig. 1
), which are rate-limiting for mitochondrial and peroxisomal FA oxidations, respectively. The expression of both genes was significantly induced by the specific agonists (P < 0.05–0.01) activating endogenous PPAR and RXR, and this induction was further potentiated by ectopic expression of the nuclear receptors. Interestingly, coexpression of AdPPAR and AdRXR synergistically increased the expression of the target genes in both the absence and the presence of exogenous ligands. Transactivation of the two target genes by PPAR/RXR in the absence of exogenous ligands indicates that either this transactivation is ligand independent, or significant levels of endogenous ligands are present in the INS-1E cells. However, the addition of receptor-specific, high affinity exogenous ligands potentiates the transaction of target genes by AdPPAR/AdRXR to levels 5- and 6-fold above basal levels (P < 0.002). The synergy between AdPPAR and AdRXR reveals that the endogenous level of RXR is not sufficient to fully support transcriptional activation by ectopically expressed PPAR, and that sequestration of RXR from other nuclear receptors could be an undesirable side effect of overexpressing PPAR alone. Thus, we chose to coexpress RXR with PPAR in subsequent experiments.
|
or PPAR together with RXR in INS-1E cells subtype, we transduced INS-1E cells with increasing titers of adenoviruses expressing either PPAR or PPAR in combination with AdRXR. To ensure comparable virus titers, we performed both plaque assays and routinely monitored the expression of the adenoviral E4 transcript (AdE4). After transduction, INS-1E cells were cultured 24 h with DMSO or the specific agonist, WY14643 or BRL49653 together with LG100268. Quantitative real-time PCR was used to quantify the ectopic expression levels of the PPARs and RXR (Fig. 2A). Protein expression was determined by Western blotting (Fig. 2B). At the level of both mRNA and protein, ectopically expressed PPAR and PPAR were equally abundant. Also, RXR levels were comparable in the two sets of samples. Even very low doses of ectopically expressed PPAR or PPAR together with RXR produced a significant and highly subtype-specific response at the level of target gene expression. The expressions of ACO and perilipin were dose- and subtype-dependently increased by AdPPAR and AdPPAR transduction, respectively (P < 0.05–0.02; Fig. 2C). Only a modest effect of AdPPAR was seen on ACO expression levels, whereas AdPPAR had no effect on perilipin expression. Adenoviral expression of PPAR is, therefore, a valuable control for PPAR-specific effects in INS-1E cells and vice versa.
|
or PPAR together with RXR and - expression in ß-cells, INS-1E cells were transduced with the indicated viruses using relative titers 1, 2, and 5 and were subsequently cultured 24 h with DMSO or the specific agonist, WY14643 or BRL49653 together with LG100268. Total RNA was isolated, and expression levels of the FA transporters CD36, FATP1, and FATP2 were quantified by real-time PCR (Fig. 3A). Both AdPPAR and AdPPAR increased CD36 expression dramatically, whereas FATP1 expression was increased mainly by AdPPAR, and FATP2 was increased mainly by AdPPAR. FA uptake was measured by incubating the transduced cells with 14C-labeled oleate and subsequently quantifying the 14C content. In keeping with the increased expression of genes encoding FA transporters, both AdPPAR and AdPPAR increased FA uptake in the cells in a dose-dependent manner up to 2-fold (P < 0.05–0.001; Fig. 3B).
|
together with RXR on FA oxidation genes in INS-1E cells on the metabolism of FAs taken up by the ß-cell, the expression levels of genes associated with FA oxidation was determined (Figs. 4 and 2C
). All genes presented, PDK4, catalase, liver and muscle forms of CPT-1 (CPT-1a and -b), MCAD, LCAD, and ACO, were dose-dependently up-regulated by AdPPAR (P < 0.05–0.001), whereas AdPPAR only modestly affected expression levels. Thus, ectopic expression of PPAR/RXR in INS-1E cells induces genes involved in both mitochondrial and peroxisomal FA oxidation in a PPAR subtype-specific manner. Neither the expression of cytochrome c nor that of nuclear respiratory factor-1 was affected by AdPPAR (data not shown), indicating that mitochondrial biogenesis was not stimulated by PPAR in this system.
|
/RXR or PPAR/RXR would be expected to increase the FA oxidation capacity of the ß-cell. To address this, INS-1E cells were transduced with the indicated viruses and respective ligands. Mitochondrial ß-oxidation was determined in the presence or absence of etomoxir, an inhibitor of CPT-1 activity (Fig. 5A). AdPPAR/AdRXR dose-dependently increased oleate oxidation 4- to 12-fold (P < 0.05–0.02) in a manner completely blocked by etomoxir. In contrast, AdPPAR/AdRXR had no effect on ß-cell FA oxidation capacity, clearly demonstrating that increased ß-oxidation is a PPAR-specific effect. Oleate oxidation was also increased 2-fold (P < 0.05) by activation of endogenous PPAR with WY14643 (Fig. 5B) and was synergistically potentiated by concomitant RXR activation (5-fold; P < 0.005). By contrast, glucose oxidation was unaffected by ectopic expression of either PPAR subtype together with RXR (Fig. 5C).
|
or PPAR alone or together with RXR/AdRXR increased FA uptake, but not oxidation, an increased storage of FAs as triacylglycerides would be expected. To investigate this, INS-1E cells were fixed and stained with the lipophilic fluorescent dye, Nile Red, 48 h after transduction with the respective viruses (relative titer, 2; Fig. 6). No lipid accumulation was seen in the cells transduced with AdVector, AdPPAR, or AdPPAR/AdRXR. In contrast, comparable expression of AdPPAR facilitated massive accumulation of lipid droplets forming ring-like structures in the cytoplasm. This accumulation was potentiated by PPAR activation by the agonist BRL49653or by cotransduction with AdRXR. Neither of the ligands alone or in combination induced lipid accumulation (data not shown). Hence, ectopically expressed PPAR and PPAR subtypes specifically drive FAs taken up by the pancreatic ß-cell along two opposite pathways, leading to FA oxidation and accumulation as neutral lipids, respectively.
|
m in INS-1E cells transduced with PPAR or PPAR together with RXRa activity. We therefore investigated whether ectopic expression of PPAR/RXR would also increase UCP2 levels in this system. As shown in Fig. 7, A and B, AdPPAR/AdRXR markedly increased both the gene expression (P < 0.05–0.02) and the protein level of UCP2 in a receptor dose-dependent manner. Similar to other PPAR target genes (Fig. 1), UCP2 expression was significantly induced by exogenous ligands without ectopically expressed receptors and by coexpression of AdPPAR/AdRXR in the absence of exogenous ligands (data not shown). In contrast, transduction with AdPPAR/AdRXR resulted in a lower and dose-independent induction of UCP2 expression. The induction of UCP2 by PPAR was independent of FA oxidation, because etomoxir slightly potentiated, rather than blocked, the induction.
|
/RXR-induced increase in FA oxidation capacity and/or the increased UCP2 levels would affect the basal mitochondrial potential or the glucose-induced hyperpolarization of the mitochondrial membrane, the m was measured. INS-1E cells were transduced with AdPPAR/AdRXR or AdVector at a relative titer of 5 and were incubated 24 h with DMSO or the specific agonists. Cells were subsequently loaded with rhodamine-123 and incubated at low glucose (2.5 mM) before stimulation with 15 mM glucose as indicated (Fig. 7C). Despite increased UCP2 levels, AdPPAR/AdRXR affected neither basal m nor glucose-induced mitochondrial membrane hyperpolarization. In both the absence and the presence of ectopically expressed receptors, ligands slightly blunted the glucose-induced mitochondrial membrane hyperpolarization. This effect may be ascribed to a direct, PPAR-independent effect on the mitochondrial membrane during the 24-h incubation as reported by others (21, 43).
GSIS from INS-1E cells and rat islets adenovirally transduced with PPAR or PPAR together with RXR
The above results clearly demonstrate that ectopic AdPPAR/AdRXR expression increases FA uptake and oxidation and results in up-regulation of UCP2 expression without affecting m, whereas AdPPAR/AdRXR expression results in FA uptake and lipid accumulation. To investigate the overall effect on ß-cell GSIS, we transduced INS-1E cells with either AdPPAR/AdRXR or AdVector and cultured 24 h with DMSO (control) or the respective agonists. Insulin secretion was measured over a 30-min period of stimulation. AdPPAR/AdRXR significantly potentiated GSIS (+54%; P < 0.02), whereas basal insulin secretion was unaffected (Fig. 8A). In clear contrast, AdPPAR/AdRXR markedly impaired GSIS (–65%; P < 0.01) implicating subtype-specific effects of PPAR and PPAR on GSIS (Fig. 8B). Blocking FA oxidation with etomoxir did not compromise the potentiating effect of PPAR (Fig. 8C), suggesting that the potentiation of GSIS is caused by mechanisms other than increased ATP generation from FA oxidation. The cellular insulin content was not affected by the treatments (data not shown). In keeping with the effects of the ligands on m (Fig. 7B), the ligands caused a decrease in GSIS (data not shown), and ligands slightly decreased, rather than potentiated, the effect of PPAR/RXR on GSIS (Fig. 8A).
|
/AdRXR would also potentiate GSIS in primary ß-cells. We therefore transduced isolated rat islets with AdPPAR/AdRXR or AdVector and cultured for 24 h. As shown in Fig. 8D, AdPPAR/ AdRXR specifically potentiated GSIS (+48%; P < 0.005) in primary rat ß-cells. In conclusion, ectopic expression of PPAR together with its heterodimerization partner, RXR, potentiated GSIS from both clonal INS-1E ß-cells and rat islets despite increased UCP2 levels and unaffected mitochondrial energy generation. In contrast, ectopic expression of PPAR and RXR significantly decreased GSIS. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
in leptin signaling and functional restoration of islets from diabetic ZDF rats was pointed out by Unger’s group (6, 44) and later confirmed in a whole-body analysis comparing wild-type and PPAR-null animals (45). The abnormal FA metabolism and defects in insulin secretion of ZDF rat islets could partly be ascribed to subnormal levels of PPAR. This linkage between PPAR expression and normal ß-cell function was also supported by studies of islets exposed to hyperglycemic conditions (37, 38). In these studies it was shown that high glucose levels significantly repressed PPAR expression and FA oxidation, and it was suggested that this effect of glucose would be central to the concept of glucolipotoxicity.
Also administration of PPAR agonists has been reported to be beneficial for ß-cell function under conditions with lipid excess. Both in human and murine islets, thiazolidinedione compounds known as potent PPAR agonists have been shown to neutralize the adverse effects of FAs on ß-cell function (10, 36, 39). Whether these effect are mediated by PPAR dependent mechanisms is currently not known. However, in other studies ectopic expression of PPAR and acute administration of thiazolidinedione compounds to normal islets unexpectedly increased islet lipid oxidation and attenuated the GSIS (8). Moreover, counteracting PPAR function by chronic administration of a PPAR antagonist totally attenuates FA induction of UCP2 and restored GSIS (34), suggesting that PPAR mediates, rather than protects against, FA toxicity within the time frame of investigation.
Recently, the beneficial physiological effects of PPAR activity on islet cell function were questioned (7). Upon adenoviral transduction of INS-1 insulinoma cells with PPAR or treatment with high doses of the hypolipidemic drug clofibric acid, a suppression of basal and glucose-stimulated insulin secretion was seen. This impairment of ß-cell function was hypothesized to be caused by increased proton leakage from mitochondria due to elevated levels of UCP2, by depletion of cytosolic free FA pools due to increased FA oxidation, or by mitochondrial desensitization. Studies of islets isolated form PPAR knockout mice are ambiguous. Some investigators reported an unaltered glucose response in PPAR–/– islets (46), whereas others found that cultured islets from fasted PPAR–/– mice hypersecrete insulin over a broad range of glucose concentrations compared with islets from wild-type mice (47).
The conflicting reports on the roles of PPAR and - in ß-cells prompted us to reinvestigate the effects of these nuclear receptors on ß-cells by acutely and ectopically expressing either of the PPAR subtypes in balance with their common heterodimerization partner RXR in the highly differentiated rat ß-cell line INS-1E and in rat islets. In experiments using more long-term or chronic activation (or deactivation) of the PPARs, it is difficult to distinguish between primary and secondary effects of the PPAR subtypes. By contrast, the strategy employed in this study, i.e. acute ectopic expression of the PPARs by adenoviral vectors, allowed us to determine the direct effects of elevated PPAR or - activity in pancreatic ß-cells. In this first parallel analysis of the potencies of the PPAR and - subtypes, each subtype served as a valuable control for the specificity of the effects the other subtype. In addition, to determine whether endogenous PPARs play a role similar to that suggested by ectopic expression, we investigated the effects of acute treatment with PPAR ligands on ß-cell gene expression and function.
The synergy between PPAR and RXR revealed that the endogenous level of RXR is insufficient to fully support transcriptional activation by ectopically expressed PPAR. Sequestration of endogenous RXR from its other dimerization partners could therefore be an undesirable side effect of overexpressing the PPARs alone. Thus, RXR was coexpressed in subsequent experiments. We showed that coexpression of PPAR and RXR led to significant induction of target gene expression even in the absence of exogenous ligands. This finding indicates that either considerable amounts of endogenous ligands are present in INS-1E cells or these receptors are able to activate endogenous target genes in the absence of ligands.
We found that both PPAR and PPAR dose-dependently increased the FA uptake of INS-1E cells. However, the two PPARs directed the incoming FAs along different metabolic pathways. Whereas both activation of the endogenous PPAR and ectopically expressed PPAR induced ß-oxidation of the FAs, ectopic expression of PPAR facilitated lipid accumulation in the ß-cell. This subtype specificity in lipid partitioning was reflected in the strict subtype-specific induction of target genes, which is conserved even when the PPARs are ectopically expressed. This finding is important for the interpretation of our data, because it indicates that the specific effects elicited by the ectopically expressed PPARs to a very large extent reflect the potencies of endogenous PPARs. A change in the relative levels of the PPAR subtypes thus is acutely translated into different metabolic profiles of the ß-cell. Under normal physiological conditions, ß-cell expression of PPAR exceeds that of PPAR (6, 38). This is in line with our present findings that endogenous PPAR levels are sufficient to stimulate target gene expression and FA oxidation, whereas agonist activation of the endogenous PPAR affects neither target genes (results not shown) nor lipid accumulation (Fig. 6). It can be speculated that under hyperglycemic or hyperlipidemic conditions, the picture would change, mimicked by PPAR overexpression, and lipid accumulation would be facilitated.
In agreement with previous reports (7, 32), PPAR induced UCP2 expression. This increase was a direct effect of PPAR, rather than secondary to the increased FA oxidation, and was also reflected in increased UCP2 protein levels. Interestingly, neither basal m nor glucose-induced mitochondrial hyperpolarization was affected by increased PPAR activity and the resulting UCP2 induction. Thus, the moderate induction of UCP2 per se appeared not to uncouple the mitochondrial membrane. On the contrary, ligands of PPAR, PPAR, and RXR partially compromised the glucose-induced hyperpolarization of the mitochondrial membrane in a receptor-independent manner. This observation is probably due to the effects of these lipophilic compounds on mitochondrial respiratory chain function, as has been reported by others (21, 43).
Importantly, GSIS was significantly potentiated by PPAR/RXR expression in both INS-1E cells and rat islets despite the ability of PPAR/RXR to increase UCP2 levels. PPAR/RXR, in contrast, only increased UCP2 mRNA levels 1.5- to 2-fold, but severely reduced both basal insulin secretion and GSIS by approximately 55%. To understand the discrepancy between UCP2 induction and effects on insulin secretion, the activity of UCP2 should be considered. In recent reports, superoxide was pointed out as the endogenous activator of UCP2 uncoupling activity (21, 26) in a potentially protective feedback loop. It is also well known that ROS production is induced not only by mitochondrial oxidation processes through reduction of respiratory complex I (21), but also by complex lipids, such as ceramide, in various cell types (48, 49, 50). In ß-cells, ceramide generation is part of the lipotoxic/cytotoxic effect associated with exposure to saturated FAs and deposition of these rather than oxidation (22, 51, 52, 53). Thus, PPAR expression might increase ROS production through increased lipid uptake in cells not metabolically adjusted to handle this challenge. In contrast, PPAR turns on a complete program of lipid uptake, lipid oxidation, neutralization of ROS, and adaptation of the citric acid cycle. This potentially equilibrates the FA flux at a higher level without increasing ROS generation or compromising the discrete compliant pools of FAs in the cytoplasm predicted to contribute to the full insulin response (13, 54, 55).
Another important aspect in understanding the observed differences between the two PPAR subtypes on GSIS may be the PPAR-stimulated PDK4 expression and expected augmented anaplerotic flow into the citric acid cycle. Anaplerosis is critical to the insulin response, and inhibition of pyruvate dehydrogenase by PDK4 could sensitize the ß-cell to glucose (56, 57), potentially through increased formation of mitochondrial coupling factors (58, 59). Such a mechanism would be in agreement with the observed potentiation of GSIS in this study and our finding that elimination of the ATP contribution from FA oxidation during GSIS did not interfere with the positive effect of PPAR. In a recent review, Sugden and Holness (60) proposed a concept, for PPAR action on insulin secretion, also consistent with our present observations. Hence, increased PPAR activity could play a role in liporegulation through improved ß-cell function and compensatory insulin secretion in the hyperlipidemic state in vivo. The recent suggestion that PPAR activation by FAs is responsible for ß-cell adaptation to fasting conditions by reducing the pool of cytoplasmatic acyl-CoA esters critical for GSIS (47) is an interesting proposal and does not directly conflict with the present gain of function data. Our results show that acute activation of PPAR has a net potentiating effect on GSIS, but do not exclude that PPAR under some conditions could promote exhaustion of the cytoplasmic acyl-CoA pool and blunt GSIS. However, under our conditions this is either not rate limiting or is counteracted by other positive effects of PPAR.
The finding that acute ectopic expression of PPAR causes massive triglyceride accumulation in INS-1E cells is in keeping with the current view that PPAR is a lipogenic transcription factor. However, this observation is at odds with the findings of Parton et al. (8), who showed that ectopic expression of PPAR increases FA oxidation and decreases lipid accumulation in rat islets. It is possible that these discrepancies are due to differences in ectopic expression levels, the duration of the experiment (24 vs. 48 h), or the fact that Parton et al. (8) ectopically expressed PPAR1, whereas we expressed PPAR2. However, to date, significant functional differences between PPAR isoforms 1 and 2 have not been described. In our system, glucose oxidation was not significantly affected, and it appears more likely that the PPAR-induced attenuation of GSIS is due to the massive lipid accumulation and increased ceramide synthesis reported to result from such accumulation. However, in the chronic hyperlipidemic state, PPAR could be important for compensatory ß-cell hyperplasia, as suggested by islet studies of ß-cell-specific PPAR knockout mice (61).
In conclusion, a significant subtype specificity is observed when comparing PPAR and PPAR in pancreatic ß-cells. Acute ectopic expression of PPAR/RXR increases the FA oxidation capacity in INS-1E cells and results in a net potentiation of GSIS in both INS-1E cells and primary rat islets. The increased expression of UCP2 protein does not per se ablate glucose-stimulated hyperpolarization of the mitochondrial membrane, an observation in agreement with the idea that uncoupling by UCP2 requires activation by reactive oxygen species. The exact molecular mechanisms underlying the potentiation of GSIS by PPAR are unclear at this point, but may involve enhanced anaplerotic feeding of the citric acid cycle and generation of mitochondrial coupling factors. This would be in keeping with the activation of PDK4 expression by PPAR and the observation that the potentiation of GSIS is not dependent on ongoing FA oxidation or accompanied by an increase in m. In contrast, acute ectopic expression of PPAR/RXR increases triglyceride accumulation and leads to attenuation of GSIS. Of note, agonist activation of endogenous PPAR, but not PPAR, affects gene expression and lipid metabolism similarly to ectopic expression of the nuclear receptor. Hence, under the present normoglycemic and normolipidemic conditions, PPAR appears to be the functionally more important subtype.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online May 5, 2005
Abbreviations: ACO, Acyl-coenzyme A oxidase; CoA, coenzyme A; CPT-1, carnitine palmitoyl transferase; DMSO, dimethylsulfoxide; FA, fatty acid; FATP, fatty acid transport protein; HBSS, Hanks’ buffered salt solution; LCAD, long-chain acyl-coenzyme A dehydrogenase; m, mitochondrial membrane potential; MCAD, medium-chain acyl-coenzyme A dehydrogenase; PDK4, pyruvate dehydrogenase kinase-4; PPAR, peroxisome proliferator-activated receptor; ROS, reactive oxygen species; RXR, retinoid X receptor; TBP, TATA box-binding protein; TFIIB, transcription factor-IIB; UCP2, uncoupling protein-2; ZDF, Zucker diabetic fatty.
Received November 1, 2004.
Accepted for publication April 5, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
mediates the adaptive response to fasting. J Clin Invest 103:1489–1498[Medline]
isoforms. J Biol Chem 277:41925–41930 in disease of pancreatic ß cells. Proc Natl Acad Sci USA 95:8898–8903 suppresses insulin secretion and induces UCP2 in insulinoma cells. J Lipid Res 43:936–943 overexpression and activation on pancreatic islet gene expression profile analyzed with oligonucleotide microarrays. Am J Physiol 287:E390–E404
, P < 0.02 (significantly different from AdVector). All results presented are representative of at least three independent experiments.
, P < 0.01; #, P < 0.005; 
, P < 0.002; ||, P < 0.001). The results presented are representative for at least three independent experiments.
, 2.5 mM glucose; 
, 15 mM glucose. C: