Pancreatic (cid:1) -Cell Lipoprotein Lipase Independently Regulates Islet Glucose Metabolism and Normal Insulin Secretion*

Lipid and glucose metabolism are adversely affected by diabetes, a disease characterized by pancreatic (cid:1) -cell dysfunction. To clarify the role of lipids in insulin secretion, we generated mice with (cid:1) -cell-specific overexpression ( (cid:1) LPL-TG) or inactivation ( (cid:1) LPL-KO) of lipoprotein lipase (LPL), a physiologic provider of fatty acids. LPL enzyme activity and triglyceride content were increased in (cid:1) LPL-TG islets; decreased LPL enzyme activity in (cid:1) LPL-KO islets did not affect islet triglyceride content. Surprisingly, both (cid:1) LPL-TG and (cid:1) LPL-KO mice were strikingly hyperglycemic during glucose tolerance testing. Impaired glucose tolerance in (cid:1) LPL-KO mice was present at one month of age, whereas (cid:1) LPL-TG mice did not develop defective glucose homeostasis until ap-proximately five months of age. Glucose-simulated insulin secretion was impaired in islets isolated from both mouse models. Glucose oxidation, critical for ATP production and

Normal secretion of insulin, the anabolic hormone central to glucose metabolism, is necessary for normal lipid metabolism. The converse may also be true; normal lipid metabolism may be necessary for normal insulin secretion. Acute exposure to fatty acids potentiates glucose-stimulated insulin secretion, whereas chronic exposure causes a muted insulin secretory response to glucose (1,2). Pharmacological depletion of plasma fatty acids after a prolonged fast inhibits glucose-stimulated insulin secretion (3)(4)(5). The excessive delivery of lipids to the ␤-cell is thought to contribute to insulin secretory failure and the development of type 2 diabetes, part of a general process commonly called lipotoxicity (6).
How lipids affect insulin secretion is unknown. In most tissues, glucose and fatty acids compete for respiration. Increased provision of lipids might be expected to decrease insulin secretion by decreasing pyruvate dehydrogenase activity and glucose oxidation, known as the Randle effect (7). Reports addressing this issue are conflicting (8,9), but it is likely that any effects on pyruvate dehydrogenase are blunted because of the high activity of pyruvate carboxylase in islets, which ultimately prevents effects on glucose oxidation by preserving cytoplasmic pyruvate (10). Fatty acids have been postulated to interfere with insulin secretion by disrupting normal mitochondrial pyruvate metabolism (11), inducing apoptosis (12), and by depleting ATP levels through the induction of uncoupling protein-2 (13,14). Fatty acids have been postulated to promote insulin secretion through the effects of long chain acyl-CoA molecules on a variety of potential targets (15), and more recently, through direct effects on the G-protein-coupled receptor GPR40 (16).
The source of lipids that modulate insulin secretion is also unknown. Fatty acids delivered to ␤-cell regulatory pools could be derived from circulating non-esterified fatty acids (mostly bound to albumin), intracellular triglycerides (the substrate for the intracellular enzyme hormone-sensitive lipase) (14), or circulating triglycerides traveling in lipoproteins. The relative contribution of each of these pools in ␤-cells to insulin secretion is unknown. In other tissues, hydrolysis of circulating lipoprotein-associated triglycerides is the principal source of fatty acids destined for metabolism (17). This process is accomplished by the extracellular enzyme lipoprotein lipase (LPL). 1 LPL is regulated by hormones and nutrients in a tissuespecific manner (18). In adipose tissue, where fatty acids are stored, LPL activity is induced by feeding and suppressed by fasting. In muscle, which relies on fatty acids for energy production in the postabsorptive state, LPL activity is induced by fasting and suppressed by feeding. Insulin stimulates LPL in adipocytes and adipose tissue (19) but may inhibit LPL in heart (20). We have shown that LPL is expressed in human and rodent islets (21) where its regulation resembles that of adipocytes, because ␤-cell LPL expression is induced by glucose and insulin (22).
Here we describe two novel mouse models generated to define how LPL, a physiologic provider of fatty acids, affects insulin secretion. Both overexpression and inactivation of ␤-cell LPL activity cause diabetes phenotypes despite opposite effects on islet glucose metabolism. Both also cause defects in the amplifying pathway of insulin secretion, suggesting that optimal ␤-cell function requires maintenance of a discrete range of triglyceride-derived fatty acids by LPL.

EXPERIMENTAL PROCEDURES
Animals-The Washington University Animal Studies Committee approved these studies. Mice were housed in a specific pathogen-free barrier facility with unrestricted access to water and standard mouse chow containing 6% fat.
For mice with ␤-cell overexpression of LPL (␤LPL-TG), a human LPL cDNA engineered to encode a hemagglutinin (HA) tag at the carboxyl terminus of the protein was inserted at the EcoRI site of the vector pRIP I (a gift from Dr. Bess Marshall), which contains an intron and polyadenylation signal and is known to direct expression of transgenes in mouse ␤-cells (23). The recombinant plasmid was sequentially treated with SacI and XhoI to liberate the LPL transgene and then the fragment was separated from vector DNA, and C57BL/6 x CBA hybrid embryos were injected at the Washington University Mouse Genetics Core. Transgenic animals, identified by PCR genotyping using human LPL/HA-specific primers, were born in expected Mendelian frequencies, and all experiments were performed with littermates.
For mice with ␤-cell-specific inactivation of LPL (␤LPL-KO), Ins2Cre mice (The Jackson Laboratory, number 003573) were crossed with mice carrying LPL alleles flanked by loxP recombination sites (LPL lox/lox ) (24). First generation animals hemizygous for the RIP-Cre gene and bearing one "floxed" LPL allele (LPL lox/wt Cre ϩ ) were crossed with LPL lox/lox animals to generate ␤-cell LPL-deficient (LPL lox/lox Cre ϩ ) and ␤-cell LPL wild-type (LPL lox/lox Cre Ϫ ) littermates that were used for experiments. The following primers (schematically represented in Fig.  2) were used to document LPL gene rearrangement: Primer A, 5Ј-GTA  GGT TTG GAA TGG TCA TTT GGC ATG TCC-3Ј; Primer B, 5Ј-TTT  CCA CTG CAC AGC TGT TTA AGT GAC TGG-3Ј; and Primer C, 5Ј-CCT AGT CTT CTC TAG GCA GAG AGC AGC AGA-3Ј. Amplification of non-rearranged DNA by primers A and B yields a product of ϳ700 bp, whereas the amplification of appropriately rearranged DNA with primers A and C yields a product of ϳ400 bp.
Serum glucose, cholesterol, triglyceride, and non-esterified fatty acids were measured using reagents from Sigma. Islet triglycerides were determined after extraction of lipids with 2:1 (v/v) chloroform and methanol followed by sonication of the extract in reagent buffer using a Branson 250-probe-tip Sonifier. Serum insulin was assessed by enzyme-linked immunosorbent assay (Crystal Chem Inc., Downer's Grove, IL). Islets in phosphate-buffered saline containing 0.1% BSA were sonicated, and the total insulin content was determined by radioimmunoassay in the Washington University Diabetes Research and Training Center.
For Western blotting, tissues were Dounce-homogenized, solubilized in sample buffer under reducing conditions, boiled for 10 min, and then proteins were separated by 10% SDS-polyacrylamide gel electrophoresis followed by transfer to nitrocellulose. Detection of human LPL protein produced by transgenic mice was performed using an antihemagglutinin antibody (MMS-101P; Covance).
ATP levels were determined within the linear response of a luciferinbased assay. Freshly isolated islets were homogenized in a nucleotidereleasing mixture and then frozen. Standard curves were generated with serial dilutions of ATP, and bioluminescence was detected using a luminometer.
Static Insulin Secretion-Islets were rinsed with CMRL 1066 medium containing 3 mM glucose and 0.1% BSA, placed in microcentrifuge tubes in the same buffer, and incubated at 37°C in a 5% CO 2 /95% air incubator for 30 min. The buffer was then replaced with CMRL 1066 medium containing 3 or 20 mM glucose and 0.1% BSA, and the samples were incubated for 30 (␤LPL-TG) or 20 min (␤LPL-KO). Insulin secreted into the medium was measured by radioimmunoassay.
Glucose and Insulin Tolerance Tests-Mice in clean cages with free access to water were fasted for 4 h and then weighed, and baseline blood glucose was determined using a glucose oxidase analyzer (HemoCue Inc., Lake Forest, CA). The animals were injected intraperitoneally with 50% (w/v) dextrose at a dose of 2 mg/g body weight, and blood glucose was measured at 30, 60, and 120 min. Insulin tolerance tests were conducted in an analogous manner, except that mice received an intraperitoneal injection of human regular insulin (Lilly, Indianapolis, Indiana) at a dose of 0.5 units/kg body weight.
Glucose Oxidation-Islets were washed and then incubated in Krebs-Ringer bicarbonate buffer (25 mM Hepes, 115 mM NaCl, 24 mM NaHCO 3 , 5 mM KCl, 1 mM MgCl 2 , 2.5 mM CaCl 2 , pH 7.4) without glucose or BSA, gassed with 5% CO 2 /95% air, and incubated at 37°C for 30 min. Krebs-Ringer bicarbonate buffer containing 3 or 16.5 mM glucose in the presence of D-[U-14 C]glucose was added, and the tubes were transferred to scintillation vials containing filter paper and sealed with septum caps (26,27). The vials were incubated for 2 h in a shaking water bath at 37°C, and then 200 l of 2 N NaOH was added to the filter paper through the stopper cap using a needle and syringe. Two hundred l of 6 N HCl was added to the tubes containing islets, and the vials were incubated without shaking at room temperature overnight. The tubes were then removed, scintillation fluid was added to the vials containing filter paper, and radioactivity representing 14 CO 2 production determined in a scintillation counter.
Statistical Methods-Results are presented as mean Ϯ S.E. Data were evaluated by unpaired, two-tailed Student's t test or by analysis of variance with appropriate post-hoc tests. Significance levels are described in the individual figure legends.

Generation of ␤LPL-TG and ␤LPL-KO
Mice-For ␤LPL-TG mice, two independent founders carrying the transgene encoding an HA-tagged LPL molecule, shown schematically in Fig.  1A, were characterized and found to share the same phenotype. An HA-tagged protein of the predicted size was detected in islets from ␤LPL-TG animals but not in control islets (Fig. 1B). Certain regions of the brain, such as the hypothalamus, transiently activate the insulin promoter during embryogenesis (28). No HA-tagged LPL was detectable in brain extracts from ␤LPL-TG or control mice (Fig. 1B), and there was no difference in brain LPL enzyme activity between genotypes (data not shown). Compared with non-transgenic islets, ␤LPL-TG islets had three-fold more LPL activity (Fig. 1C) in the presence of 3 mM glucose (p ϭ 0.0019) and nearly five-fold more LPL activity in the presence of 20 mM glucose (p ϭ 0.0004). Islet triglyceride content was modestly but significantly increased in the transgenic animals (Fig. 1D). Isolated islets in culture respond to elevated glucose concentrations by increasing LPL activity (22). The increased intracellular triglyceride accumulation ( Fig. 1D) suggests that the increased LPL activity measured in our in vitro assay (Fig. 1C) leads to increased fatty acid flux and subsequent metabolic adaptations in vivo.
For ␤LPL-KO mice, the presence of the Cre recombinase in ␤-cells was predicted to remove exon 1 of the LPL gene ( Fig.  2A). PCR reactions, including Primers A, B, and C as shown in Fig. 2A, showed the expected rearrangement of the LPL gene (as indicated by the presence of a 400 bp band) in islet DNA from Cre ϩ mice (Fig. 2B). Detection of the 700-bp wild-type band in the KO lane is expected, as Cre is driven by a ␤-cellspecific promoter, and DNA from non-␤-cells in islets should not be rearranged. That the conditional rearrangement of the LPL gene occurred only in the ␤-cell was shown by the presence of the wild-type product and the absence of the rearranged product in DNA from the following tissues of RIP-Cre ϩ mice: brain, heart, adipose tissue, spleen, and liver (data not shown).
LPL enzyme activity (Fig. 2C) was ϳ40% lower in ␤LPL-KO as compared with control islets in the presence of 20 mM glucose (p ϭ 0.0307). Residual LPL activity detected in islets from ␤LPL-KO mice is likely because of the presence of non-␤-cells. Although we observed a reduction in glucose-stimulated LPL activity in ␤LPL-KO islets (Fig. 2C), this change in LPL activity did not alter the islet content of triglycerides in vivo (Fig. 2D).
Glucose Intolerance and Defective Insulin Secretion with Perturbations of ␤-cell LPL-Neither genetic manipulation of LPL expression affected fasting serum levels of cholesterol, triglycerides, free fatty acids, or insulin (data not shown). There was also no effect of increased or decreased islet LPL activity on body mass. Weights for ␤LPL-TG mice at the age of 12 weeks were: TG males, 32.8 Ϯ 1.8 g; WT males, 31.2 Ϯ 0.9 g; TG females, 24.5 Ϯ 4.9 g; and WT females, 25.5 Ϯ 5.0 g. Weights for ␤LPL-KO mice at the age of 12 weeks were: KO males, Insulin sensitivity as assessed by insulin tolerance testing was unaffected in both ␤LPL-TG and ␤LPL-KO mice (Fig. 3, B  and D). However, ␤LPL-TG and ␤LPL-KO animals became hyperglycemic compared with their respective wild-type littermates during glucose tolerance testing (Fig. 3, A and C). The data of Fig. 3 represent males, but the same results were seen with females. To characterize in vivo insulin secretory responses, insulin was measured during glucose tolerance tests in separate cohorts. Compared with the basal state, plasma insulin nearly doubled 30 min after glucose administration in both the ␤LPL-TG and ␤LPL-KO animal models (Table I).
Although ␤LPL-TG and ␤LPL-KO mice displayed significantly higher blood glucose levels 30 min after challenge, there were no significant differences in circulating insulin concentrations between these animals and their respective wild-type litter-

␤-Cell LPL and Insulin Secretion
mates. The failure to detect higher insulin levels in the setting of hyperglycemia suggests the presence of defective insulin secretion, which was confirmed using isolated islets (Fig. 4). ␤LPL-TG (Fig. 4A) and ␤LPL-KO (Fig. 4C) islets secreted less insulin than controls when stimulated with 20 mM glucose. Islet insulin content was unaffected in both models (Fig. 4, B  and D).
Although both ␤LPL-TG and ␤LPL-KO mice display impaired glucose tolerance relative to their relevant wild-type littermates, the age at which this occurs differs between the two models. ␤LPL-TG mice and their wild-type siblings have normal glucose tolerance up to four months of age, then manifest glucose intolerance by the age of five months (data not shown). This delayed onset of glucose intolerance suggests that these animals successfully compensate for the increased delivery of fatty acids for a period of time. In contrast, ␤LPL-KO mice as young as one month of age, the earliest age examined, exhibit impaired glucose tolerance (data not shown), suggesting that these animals have insufficient levels of a fatty acidderived signal required for normal insulin secretion. Because these two animal models developed impaired glucose tolerance at different ages, ␤LPL-TG animals were studied at 5-6 months of age, whereas ␤LPL-KO mice were studied at 2-3 months of age.
Opposite Effects on Islet Glucose Oxidation and Defects in the Amplifying Pathway of Insulin Secretion-These genetic models of altered islet LPL expression had opposite effects on islet glucose oxidation. Glucose oxidation was decreased at 16.5 mM glucose in ␤LPL-TG islets relative to islets from wild-type siblings (Fig. 5A), but this decrease was unlikely to be due to lipotoxicity-induced mitochondrial dysfunction, because islet ATP content was increased (Fig. 5B). Previous studies have suggested that increased availability of fatty acids in islets can increase expression of uncoupling protein-2 (14), known to decrease insulin secretion presumably through effects on ATP generation (29). Consistent with the ATP data of Fig. 5B, there was no effect of LPL expression on islet uncoupling protein-2 mRNA levels (not shown). There was also no effect of LPL modulation on islet mRNA levels for glucokinase (not shown), a key mediator of ␤-cell glucose metabolism.

TABLE I Plasma insulin and glucose concentrations before and after glucose challenge
Blood samples were obtained from mice after fasting 4 h (time 0) or after subsequently being challenged with 50% dextrose at a dose of 2 mg/g body weight (time 30). Plasma insulin was determined by an enzyme-linked immunosorbent assay using mouse insulin for the standard curve. ␤LPL-TG mice and their respective littermates were 5-6 months old when tested, whereas ␤LPL-KO mice and wild-type siblings were 2-3 months of age when examined.

␤-Cell LPL and Insulin Secretion
Glucose oxidation was increased at both 3 and 16.5 mM glucose in ␤LPL-KO islets as compared with islets from wildtype littermates (Fig. 5C). Qualitatively, the increased glucose oxidation as a consequence of lowered ␤-cell LPL expression is the opposite to that of ␤LPL-TG islets, which have reduced glucose oxidation. However, differences in the ages of animals studied and in their genetic backgrounds limit the validity of direct quantitative comparisons of glucose oxidation between these two models of altered LPL expression. There was no effect of ␤-cell LPL knockdown on islet ATP content (Fig. 5D).
Despite differing effects on glucose oxidation, ␤LPL-TG and ␤LPL-KO islets had similar defects in the amplifying pathway of insulin secretion (Fig. 6). Treatment with 250 M diazoxide and 30 mM KCl, conditions inducing insulin secretion independent of the K ATP channel, revealed defects in insulin secretion at 3 and 20 mM glucose in ␤LPL-TG islets (Fig. 6A) and at 20 mM glucose in ␤LPL-KO islets (Fig. 6B). DISCUSSION Diabetes, a disease projected to affect 366 million people worldwide by the year 2030 (30), is characterized by two salient pathophysiologies: insulin resistance and the inability of the The same results for these panels were seen in 2-5 independent experiments. For the ␤LPL-TG experiments, islets were obtained from 5-6-month-old ␤LPL-TG animals and littermates. For the ␤LPL-KO model, islets were isolated from 2-3-month-old ␤LPL-KO mice and wild-type siblings. *, p Ͻ 0.005 and **, p Ͻ 0.01 versus wild-type littermates under the same condition.

␤-Cell LPL and Insulin Secretion
␤-cell to secrete sufficient insulin to maintain normal levels of blood glucose. Both may be affected by abnormal lipid metabolism. By modulating ␤-cell expression of LPL, which is known to be altered in diabetes and critical for the physiologic delivery of fatty acids to tissues, we have identified three important roles for this enzyme in ␤-cell function. First, normal levels of ␤-cell LPL are required to maintain glucose homeostasis in an intact physiologic system, because an increase in islet LPL activity (␤LPL-TG mice, Fig. 3A) and a decrease in islet LPL activity (␤LPL-KO mice, Fig. 3C) resulted in systemic glucose intolerance. Second, LPL activity inversely regulates islet glucose metabolism. Glucose oxidation was decreased in islets with increased LPL activity (Fig. 5A) and increased in islets with deficient LPL activity (Fig. 5C). Third, LPL appears to independently regulate insulin secretion at a level distal to events associated with membrane depolarization. Islets from ␤LPL-TG and ␤LPL-KO mice secreted less insulin under conditions causing calcium-dependent insulin release independent of the K ATP channel (Fig. 6).
The current work emphasizes the importance of LPL, and hence free fatty acid availability, for glucose metabolism and insulin secretion, because too much or too little LPL has negative effects on ␤-cell function. The intracellular accumulation of fatty acids in triglycerides is just one possible fate for nonesterified fatty acids liberated by LPL, and in the case of the ␤LPL-TG model, it served as a relevant indicator of the increased provision of lipids to the ␤-cell (Fig. 1D). However, in the case of the ␤LPL-KO model, the triglyceride content was unaffected. This result is not surprising given the young age of these animals and the very low levels of triglycerides, near the threshold of detection, in both wild-type and knock-out islets (Fig. 2D). Too much LPL impairing insulin secretion is consistent with the concept of lipotoxicity (31). Increased triglycerides in ␤LPL-TG islets (Fig. 1D) may represent storage of potentially toxic fatty acids as inert neutral lipid (32,33). Increased LPL activity would be expected to accelerate loss of this potentially protective mechanism as the capacity for lipid storage is exceeded, and we interpret the delayed onset of glucose intolerance in the transgenic model as representing the eventual negative consequences of exceeding this capacity in the ␤-cell. Too little LPL, also impairing insulin secretion, is consistent with the concept that a lipid-derived coupling factor is required for insulin secretion (34,35). Circulating free fatty acids are required for glucose-stimulated insulin secretion following prolonged fasting (18 -24 h in rodents, 24 -48 h in humans) and treatment with nicotinic acid (3,4). ␤LPL-KO mice were fasted for 4 h, and there was no effect on serum free fatty acid levels. The fact that circulating free fatty acids were available to islets in the current work and yet did not prevent a hyperglycemic phenotype in ␤LPL-KO mice (animals with a reduction in lipid delivery limited to those fatty acids derived from the activity of ␤-cell LPL) strongly suggests that ␤-cell LPL controls a discrete pool of fatty acids required for normal insulin secretion. These results suggest that the ␤-cell detects triglycerides in the blood, via the action of LPL, and that the ␤-cell is capable of distinguishing between free fatty acids supplied by LPL versus those present as fatty acid-albumin complexes.
Effects on glucose oxidation in our mouse models contribute to a growing body of evidence that LPL participates in the regulation of glucose metabolism. Haplotype analysis in humans indicates that variations in the LPL gene impact glucose metabolism (36). Two independent groups have shown that glucose metabolism is impaired in mice with skeletal muscle overexpression of LPL (37,38). Cardiac-specific knockdown of LPL activity increases myocardial glucose metabolism (24). Now we show that ␤-cell-specific overexpression of LPL decreases and ␤-cell-specific knockdown of LPL increases islet glucose metabolism (Fig. 5).
Surprisingly, these effects on glucose metabolism did not appear to account for defective insulin secretion in our models. Glucose oxidation was decreased in ␤LPL-TG islets, but ATP levels were increased (Fig. 5), and insulin secretion was decreased in the presence of KCl and diazoxide (Fig. 6), implicating an effect on events distal to membrane depolarization. Glucose oxidation was increased in ␤LPL-KO islets (Fig. 5), which might be expected to increase insulin secretion by increasing ATP production and decreasing K ATP activity, but instead, glucose-stimulated insulin secretion was decreased in the presence and absence of KCl/diazoxide (Fig. 6). Studies using insulinoma cells have shown that the inhibition of glucose-stimulated insulin secretion caused by the chronic exposure of free fatty acids occurs independently from the effects of those fatty acids on glucose metabolism (39,40). Our animal models of altered ␤-cell LPL expression suggest that the dissociation of the effects of fatty acids on glucose oxidation and insulin secretion also occurs in vivo. Other groups have provided evidence that a lipid-coupling factor mediates a K ATP channel-independent pathway of insulin secretion (34,35). Our data demonstrate that ␤-cell LPL activity is a key physiologic generator of the lipids involved in this effect.
We did not address the in vivo effects of individual fatty acids on ␤-cell function in this study. However, the in vitro effects of different free fatty acid species on islet function have been studied extensively. No difference between saturated and un-saturated fatty acids, in terms of acute potentiating effects on insulin secretion, was observed by Warnotte et al. (41). Chronic stimulation by saturated fatty acids, such as palmitate, provokes ␤-cell apoptosis, whereas monounsaturated fatty acids do not cause apoptosis and appear to protect against palmitateinduced apoptosis by redirecting this saturated fatty acid into the triglyceride esterification pathway (32,33). A slight preference of rat LPL for saturated fatty acid species over unsaturated ones was detected using an in vitro assay (42), but whether this difference affects LPL substrate selection in vivo is unclear.
A downstream event likely to be affected by LPL expression is the exocytosis of vesicles containing insulin (43). Docking and fusion of vesicles with the plasma membrane requires the assembly of a complex between tSNAREs, located on the plasma membrane, and vSNAREs, located on secretory vesicles. The tSNARE protein SNAP-25, important for insulin release (44), is palmitoylated (45), a modification that allows proteins to be associated with membranes. Palmitoylation is poorly understood (46). If the process is enzymatic, it is likely to be dependent on an optimal concentration of fatty acids. There is precedence for this notion; serine palmitoyltransferase, a key enzyme in the ceramide synthesis pathway, is functional within a narrow range of fatty acid concentrations (47). It is thus plausible that lipid modification of vesicles would be impaired both with LPL deficiency and LPL excess when LPL activity controls a pool of fatty acids available to modify tSNAREs. If palmitoylation is non-enzymatic (substratedriven), the low level of lipids could be insufficient for vesicle anchoring, and high levels could lead to excessive palmitoylation and self-association of vesicles that would interfere with docking. Other components of the exocytotic machinery may also be involved. Munc-18 is a tSNARE binding protein expressed in ␤-cells that is released to allow assembly of the active SNARE complex (48). Skeletal muscle overexpression of LPL is associated with increased Munc-18 expression (49).
In summary, overexpression or inactivation of LPL in the ␤-cell disrupts glucose metabolism in mice. LPL appears to deliver triglyceride-derived fatty acids to the ␤-cell within a discrete concentration range required for optimal glucose-stimulated insulin secretion. ␤-cell LPL thus represents an important translator of the cross-talk between glucose and lipid metabolism involved in the development of diabetes.