β-Cell Adaptation to Insulin Resistance

The β-cell biochemical mechanisms that account for the compensatory hyperfunction with insulin resistance (so-calledβ-cell adaptation) are unknown. We investigated glucose metabolism in isolated islets from 10–12-week-old Zucker fatty (ZF) and Zucker lean (ZL) rats (results expressed per mg/islet of protein). ZF rats were obese, hyperlipidemic, and normoglycemic. They had a 3.8-fold increased β-cell mass along with 3–10-fold increases in insulin secretion to various stimuli during pancreas perfusion despite insulin content per milligram of β-cells being only one-third that of ZL rats. Islet glucose metabolism (utilization and oxidation) was 1.5–2-fold increased in the ZF islets despite pyruvate dehydrogenase activity being 30% lowered compared with the ZL islets. The reason was increased flux through pyruvate carboxylase (PC) and the malate-pyruvate and citrate-pyruvate shuttles based on the following observations (% ZL islets): increasedV max of PC (160%), malate dehydrogenase (170%), and malic enzyme (275%); elevated concentrations of oxaloacetate (150%), malate (250%), citrate (140%), and pyruvate (250%); and 2-fold increased release of malate from isolated mitochondria. Inhibition of PC by 5 mm phenylacetic acid markedly lowered glucose-induced insulin secretion in ZF and ZL islets. Thus, our results suggest that PC and the pyruvate shuttles are increased in ZF islets, and this accounts for glucose mitochondrial metabolism being increased when pyruvate dehydrogenase activity is reduced. As the anaplerosis pathways are implicated in glucose-induced insulin secretion and the synthesis of glucose-derived lipid and amino acids, our results highlight the potential importance of PC and the anaplerosis pathways in the enhanced insulin secretion and β-cell growth that characterize β-cell adaptation to insulin resistance.

Insulin resistance is tissue insensitivity to the regulatory effects of insulin on glucose and fatty acid metabolism. Insulin resistance is a risk factor for type 2 diabetes (1). However, most affected individuals do not develop diabetes because of a compensatory increase in insulin secretion (2). The mechanism of the ␤-cell adaptation is poorly understood. Particularly unclear is the dichotomy that insulin resistance is typically accompa-nied by elevated blood and tissue triglyceride and fatty acid (FA) 1 levels (3) when multiple studies of isolated islets and ␤-cell lines cultured with FA have shown detrimental effects on ␤-cell function and viability (4 -7).
A well known mechanism for altered cellular function from excess FA is the glucose-fatty acid cycle of Randle, part of which is impaired activation of pyruvate dehydrogenase (PDH) (8). This effect is reported to occur in FA-cultured islets (9). PDH supplies pyruvate-derived acetyl-CoA to the citrate cycle so that in most tissues excess FA results in lowered ATP production. Intact ␤-cell mitochondrial function is required for glucose-induced insulin secretion (10,11), leading to speculation that this is a mechanism of FA-induced impaired ␤-cell function (9). However, lowering and raising PDH activity in clonal ␤-cells through adenoviral overexpression of PDH kinase and the catalytic subunit of PDH phosphatase, respectively, had little effect on glucose-induced insulin secretion (12). This may reflect a second pathway for mitochondrial pyruvate metabolism in ␤-cells through pyruvate carboxylase (PC). Downstream of PC are pyruvate shuttles (13,14) and malonyl-CoA (15,16) which are proposed to provide coupling factors for glucose-induced insulin secretion. The effect of FA on these pathways is uncertain.
We investigated rat islets cultured for 4 days with 0.25 mM oleate/5.5 mM glucose (17). PDH activity was 35% lowered, but glucose oxidation and insulin secretion were normal to supernormal depending on the studied glucose concentration. We proposed the reason was the observed unaffected PC activity and enhanced flux through the malate-pyruvate shuttle; the latter is proposed to be an effector of insulin secretion through the malic enzyme production of cytosolic NADPH (13,18,19). Support for our conclusion is the lowered PC expression resulting in defective malate-pyruvate shuttle activity reported in FA-cultured islets with impaired glucose-induced insulin secretion (20). Also, PC expression and activity are lowered in islets of diabetic rodents with defective glucose-induced insulin secretion (21,22).
These results focus on PC and the malate-pyruvate shuttle as components of the ␤-cell adaptation in states of excess FA such as insulin resistance. However, all of the cited studies were performed in vitro. The current study investigated Zucker fa/fa fatty rats. These rats are obese, insulin resistant, and hyperlipidemic because of mutated leptin receptors (23). Moreover, they are not diabetes-prone (24,25), making them an excellent model to investigate ␤-cell adaptation to insulin resistance.

MATERIALS AND METHODS
Animals and Isolated Islets-Zucker fatty (fa/fa, ZF)and Zucker lean (fa/Ϫ or Ϫ/Ϫ, ZL) rats were studied at 10 -12 weeks of age (Harlan, Indianapolis, IN) except for ␤-cell mass measurements that were performed in 10-week-old ZF and ZL rats. Tail vein blood was obtained from normally fed rats at 9:00 a.m.; glucose concentration was measured using Glucose Analyzer II (Beckman, Fullerton, CA), and plasma insulin, triglycerides, and FA were measured by commercial kits (Alpco; Winham, NH, Sigma; Wako; Richmond, VA, respectively). Islets were isolated by an adaptation of the Gotoh method (26): pancreas duct infiltration with collagenase, histopaque gradient separation, and hand picking. DNA was measured by the Labarca method (27), protein by a commercial kit that used BSA as standard (Bio-Rad, Hercules, CA), and insulin content post acid ethanol extraction using an insulin RIA (28). Freshly isolated islets were used after 30 min of incubation in KRB, 5.5 mM glucose, and 0.07% BSA unless stated otherwise. Because of the larger cell number of the islets from ZF versus ZL rats, methods utilized half the number of islets from obese versus lean rats and results are expressed per milligram of islet protein.
Pancreas ␤-Cell Mass-Rats were euthanized, and pancreata dissected and immersion-fixed in 4.0% paraformaldehyde/100 mM phosphate-buffered saline. Following paraffin embedding, 5-m sections throughout the entire pancreas were mounted as ribbons on microscope slides to facilitate section counting. At 200-m intervals, (ϳ20 slides/ pancreas), sections were rehydrated and immunostained with guinea pig anti-insulin IgG (Linco, St. Charles, MO) followed by donkey antiguinea pig IgG-peroxidase (Jackson ImmunoResearch, West Grove, PA). Following development with diaminobenzidine/H 2 O 2 , sections were counterstained with hematoxylin, cleared, and mounted in Permount. The proportion of ␤-cell surface area versus surface area of the whole pancreas was determined by digitally imaging at least three non-overlapping 3.89 mm 2 fields per section on a Zeiss Universal microscope coupled to a Spot RT color charge-coupled device camera (Diagnostics Instruments, Sterling Heights, MI). ␤-Cell mass was calculated for each animal as the average proportional ␤-cell surface area multiplied by pancreatic weight.
In Situ Perfused Pancreas and Pancreas Insulin Content-The perfusion technique has been described elsewhere (29). The perfusate was oxygenated KRB buffer, pH 7.4, that contained 4% dextran T70 and 0.2% BSA fraction V (Sigma). The perfusion protocol is shown at the top of Fig. 1. Following 17 min of equilibration at 7.8 mM glucose, 1-min samples were collected in tubes on ice that contained 8 mg of EDTA and stored at Ϫ20°C pending analysis by insulin RIA. Immediately after the perfusion, the pancreas was excised, blotted, and stored at Ϫ20°C in acid ethanol pending homogenization and insulin RIA.
Islet Glucose Metabolism and Insulin Secretion-Glucose utilization and oxidation were determined with D-[5-3 H]glucose and [U-14 C]glucose as described (30). Insulin secretion was measured in duplicate vials of five islets following a 60-min incubation in KRB, 10 mM HEPES, 0.5% BSA, glucose (2.8, 8.3, 27.7 mM) at 37°C in a shaking water bath. Islets were sedimented by gentle centrifugation and assayed for insulin content post sonication and extraction in acid ethanol. Insulin in the medium was measured by insulin RIA and expressed as percent of islet insulin content. Experiments were performed with and without 5 mM phenylacetic acid (PAA), which is an inhibitor of PC activity (14).
Islet Metabolites-Freshly isolated islets were incubated 60 min at 37°C in prewarmed and oxygenated KRB that contained the stated glucose concentration in the text or figure followed by biochemical analysis. Citrate concentration was measured as described (31). Islets (200 ZL, 100 ZF) were lysed with trichloroacetic acid and placed on ice followed by centrifugation to remove precipitated proteins. Supernatant was neutralized by ether extraction 5ϫ, lyophilized, and resolubilized in 100 l of H 2 O. Extract (20 l) or citrate standard (0.1-2 nmol) was added to 0.1 M Tris-HCl buffer, pH 7.6, 40 M ZnCl 2 , 3 M NADH, 0.4 g/ml malate dehydrogenase, final volume 0.5 ml. Fluorescence was determined at 340 nm excitation and 465 nm emission, then repeated 5 min after adding 10 l of citrate lyase solution with the ⌬ value representing the citrate content.
G6P concentration was measured as described (32). Islets underwent rapid lysis after the 60-min incubation using 10 l of 40 mM NaOH and were placed on ice for 10 min, and then 3 l of 0.15 M HCl was added with incubation at 75°C for 20 min to destroy cellular enzymes to ensure stability of the G6P. Extract (20 ZL islets, 10 ZF islets) was added to 15 l of reaction buffer (0.15 M Tris/HCl, pH 8.1, 20 M NADP, 0.02 units/ml grade 1 glucose-6-phosphate dehydrogenase from yeast; Roche Molecular Biochemicals) and incubated 30 min at 28°C in a shaking water bath followed by addition of 4 l of 1 M NaOH and incubation at 75°C for 20 min to destroy any remaining NADP. The formed NADPH was amplified by the cycling method by adding 100 l of reagent (9 units/ml of type II glutamate dehydrogenase, 5 mM ␣-ketoglutaric acid, 1 mM G6P, 6 units/ml glucose-6-phosphate dehydrogenase) for 60 min at 38°C, followed by heating to 100°C for 10 min to stop the reaction. The sample (100 l) was transferred to a UV cuvette containing 1 ml of 0.006 units/ml   10 -12-week-old Zucker fatty rats and Zucker lean rats Blood or tissue was obtained at 9:00 a.m. from normally fed 10 -12week-old Zucker fatty and Zucker lean rats. ␤-Cell mass was measured in 10-week-old rats. Data points are mean Ϯ S.E. values of the number of studied rats shown in the parentheses. NS, not significant.
for 30 min. The formed 6-phosphogluconate was measured by a fluorometer at 340 nm excitation and 420 nm emission, and islet G6P concentration was determined from G6P standards (1-20 pmol). Malate and pyruvate were measured as described (17). Islets (300 ZL, 150 ZF) were lysed in 150 l of 2 M perchloric acid on ice for 20 min and then centrifuged 10 min at 12,000 ϫ g. Supernatant was neutralized with 3 M KHCO 3 and recentrifuged at 12,000 ϫ g. Malate and pyruvate standards were prepared in perchloric acid. Malate: 50 l of extract or malate standard (0.1-1 nmol) was added to 250 l of reaction buffer (20 mM 2-amino-2-methyl propanol, pH 9.9, 2 mM glutamate, 50 M NAD, 1 g of aspartate aminotransferase, 2.5 g of malate dehydrogenase) for 20 min at 30°C. Nonmetabolized NAD was removed by addition of 0.25 ml of potassium phosphate, pH 11.9, and incubation at 60°C for 15 min, followed by addition of 12 l of 1 M imidazole and another 15 min at 60°C. Fluorescence was measured at 340 nm excitation and 465 nm emission. Pyruvate: 30 l of extract or pyruvate standard (20 -200 pmol of sodium pyruvate) was added to 100 l of reaction buffer (50 mM imidazole, pH 7, 0.6 mM ascorbate, 0.2 mg/ml BSA, 6 M NADH, 0.125 units/ml lactate dehydrogenase) at room temperature for 20 min. Nonmetabolized NADH was removed by addition of 20 l of 2 N HCl at room temperature for 30 min followed by 1 ml of 6 N NaOH for 10 min at 60°C. Fluorescence was measured at 360 nm excitation and 460 nm emission. Oxaloacetate was measured as described (17). Islets (50 ZL, 25 ZF) were lysed 20 min in 40 l of 0.25 M perchloric acid at Ϫ20°C followed by sonication and addition of 20 l of 0.94 M KOH. Extract (30 l) or oxaloacetate standard in perchloric acid (0.2-2.0 pmol) was added to 200 l of reaction buffer (75 mM K 2 PO 4 , pH 7.4, 80 nM acetyl-[ 3 H]CoA, 50 g/ml citrate synthase) at room temperature for 60 min. The reaction was stopped by 600 l of charcoal mixture (8 g of charcoal, 38 g of citric acid monohydrate, 120 ml of 95% ethanol) followed by centrifugation at 12,000 ϫ g and liquid scintillation counting of the supernatant.

FIG. 2. Glucose utilization (A) and glucose oxidation (B) in isolated islets of 10 -12-week-old ZL and ZF rats determined with D-[5-3 H]glucose and
[U-14 C] glucose, respectively. Experiments were conducted with and without 5 mM PAA, which inhibits pyruvate carboxylase activity (14). Data are mean Ϯ S.E. Fluorescence was measured 20 min at 340 nm excitation and 420 nm emission.
Mitochondrial glycerol-3-phosphate dehydrogenase (mGPD) was measured as described (17). Islets (100 ZL, 50 ZF) were sonicated in 0.1 ml of 5 mM HEPES, pH 7.5, 230 mM mannitol, and 70 mM sucrose. Extract (10 l) was added to 200 l of reaction buffer (50 mM Bicine buffer, pH 8, 1 mM KCN, 50 mM L-glycerol-3-phosphate, 4 mM iodonitrotetrazolium violet) for 10 min at 37°C. The reaction was stopped by adding 1 ml of ethyl acetate followed by centrifugation and absorbance determination of the ethyl acetate layer at 490 nm. Results were compared with the standard curve made with 0 -10 nmol of iodonitrotetrazolium violet in reaction mixture with or without (as blank) islet extract.
Malate Release from Isolated Islets-The method is previously described (17). Mitochondrial isolation: islets (400 ZL, 200 ZF) were homogenized in 0.4 ml of 5 mM potassium HEPES, pH 7.5, 230 mM mannitol, and 70 mM sucrose and centrifuged at 600 ϫ g for 5 min to sediment the nuclear and cell debris, followed by recentrifugation of the supernatant at 5,500 ϫ g for 10 min to sediment the mitochondria. The sedimented mitochondria were resuspended in 120 l of ice cold buffer (5 mM potassium HEPES, pH 7.3, 5 mM K 2 PO 4 , 5 mM KHCO 3 , 2 mM Na 2 ADP, 230 mM mannitol, 70 mM sucrose, with or without 10 mM pyruvate) and kept on ice. Malate release was measured using the method of MacDonald (13). The mixture was placed at 37°C, and 50-l samples were taken at 0 and 10 min. Samples were centrifuged at 14,000 rpm for 2 min followed by addition of 15 l of 0.92 M perchloric acid to the supernatant, return of pH to 7.0 with 1 M KOH, and recentrifugation at 14,000 rpm for 2 min. Malate was measured using the method of Sener et al. (33). Supernatant (40 l) or malate standard (0 -30 pmol) was added to 200 l of reaction buffer (100 mM Tris/KCl, pH 8.0, 1 mM NAD, 0.2 mM [ 3 H]acetyl-CoA, 20 g/ml malate dehydrogenase from pig heart, 60 g/ml citrate synthase from pig heart) at room temperature for 60 min. The final product of [ 3 H]citrate was separated with 600 l of charcoal mixture (120 ml of 95% ethanol, 8 g of charcoal, 38 g of citrate acid monohydrate) and centrifuged at 14,000 rpm for 5 min followed by liquid scintillation counting of the supernatant.
Data Presentation-All data are expressed as mean Ϯ S.E. For protocols that used isolated islets from a single rat, the n values are the number of rats studied. If pooled islets from more than one rat were needed, the n values are the number of experiments performed. Statistical significance was determined by the unpaired Student's t test.

General Characteristics of ZF and ZL Rats and Isolated
Islets-The 10 -12 week-old ZF rats were obese, normoglycemic, hyperinsulinemic, and hyperlipidemic versus the lean controls (Table I). ZF rats had a 3.8-fold increased mass of islet ␤-cells. Pancreas insulin content was 150% of the ZL rats so that insulin content per mg of ␤-cells in the ZF rats was only 35% of the ZL rats (0.9 Ϯ 0.1 g/mg ␤-cells ZF versus 2.6 Ϯ 0.1 g/mg ␤-cells ZL, p Ͻ 0.001). Isolated islets were similar: ZF islets were hypercellular compared with ZL islets (3-fold increased DNA and protein contents), and insulin content per islet protein was 40% of the ZL islets (23 Ϯ 3 ng of insulin/g of protein ZF versus 56 Ϯ 9 ng of insulin/g of protein ZL, p Ͻ 0.001). Pancreas perfusions carried out at 7.8, 16.7, and 16.7 mM glucose plus 10 mM arginine (Fig. 1) showed a 10-fold increase of insulin secretion at 7.8 mM glucose, and 3-fold increases at both high glucose conditions in the ZF rats (all p Ͻ 0.001).
Islet Glucose Metabolism-Glucose utilization ( Fig. 2A) and oxidation (Fig. 2B) per mg of islet protein were 1.5-2-fold increased in the ZF islets at 2.8 and 16.7 mM glucose (utilization 2.8 mM glucose p Ͻ 0.025, 16.7 mM glucose p Ͻ 0.001; oxidation 2.8 mM glucose p Ͻ 0.015, 16.7 mM glucose p Ͻ 0.005). We next investigated potential mechanisms for the increased glucose metabolism.
We previously studied rat islets cultured for 1 day with palmitate and noted increased basal glucose metabolism and insulin secretion (30), possibly because of a lowered concentration of G6P that diminished end-product inhibition of hexokinase. We proposed that the mechanism was up-regulation of PFK activity through a dual effect, increased PFK expression and lowered production of the PFK allosteric inhibitor citrate by citrate synthase, so that flux through glycolysis was increased (31,32). However, we found no difference in G6P level between the ZF and ZL islets after 60 min of incubation at 8.3 mM glucose (0.99 Ϯ 0.08 nmol/mg protein ZL versus 1.26 Ϯ 0.14 nmol/mg protein ZF, n ϭ 4, p ϭ NS), V max values for PFK (52 Ϯ 2 nmol/min/mg protein ZL versus 68 Ϯ 14 nmol/min/mg protein ZF, n ϭ 6, p ϭ NS), or citrate synthase (44 Ϯ 3 nmol/min/mg protein ZL versus 43 Ϯ 5 nmol/min/mg protein ZF, n ϭ 4, p ϭ NS). Rather, the citrate concentration after 60 min of incubation at 8.3 mM glucose was increased in the ZF islets (Fig. 3A).
Islet Pyruvate Metabolism-An alternate mechanism based on our study of 4-day oleate-cultured islets (17) was impaired PDH activity with diversion of pyruvate metabolism through PC. That study noted increased flux through the malate-pyruvate shuttle, which entails pyruvate transport into the mitochondria, conversion to oxaloacetate by PC, malate production by MDH, and export of malate to the cytoplasm where it is converted back to pyruvate by malic enzyme. We proposed that the observed normal glucose oxidation in the FA-cultured islets stemmed from the pyruvate concentration being twice normal,

␤-Cell Adaptation in Zucker Fatty Rats
overcoming the observed 35% reduction in PDH activity through a mass action effect.
Islet Glycerol Phosphate Shuttle-The glycerol phosphate shuttle is another pathway in ␤-cells that links cytosolic glucose metabolism to mitochondrial regulation of insulin secretion (34). We investigated this pathway by studying activity of the rate-limiting enzyme, mGPD. mGPD V max was 50% lowered in the ZF islets (34.1 Ϯ 0.8 nmol/min/mg protein ZL versus 14.6 Ϯ 1.5 nmol/min/mg protein ZF, n ϭ 7, p ϭ 0.001).
Inhibition of PC Activity by Phenylacetic Acid-We studied the effect of inhibition of PC activity by 5 mM PAA (14) on islet glucose metabolism and insulin secretion. In both the ZL and ZF islets, glucose utilization and oxidation were unaffected by PAA at 2.8 mM glucose and lowered 30% at 16.7 mM glucose (Fig. 2). Also, PAA markedly lowered glucose-induced insulin secretion in the ZL and ZF islets (Fig. 6). DISCUSSION The glucose homeostasis system to insulin resistance entails compensatory increases in ␤-cell mass and function so that glycemia is unaltered (2). A confounding issue is how the excess FA that accompanies insulin resistance affects the ␤-cell adaptation as multiple detrimental effects on ␤-cell mass and function have been reported based on in vitro studies (4 -7). We and others have reported lowered PDH activity in FA-cultured islets (9,17) and now have made the same observation in ZF rats. This inhibitory effect of excess FA on PDH activity is well known to impair glucose oxidation and ATP production in cardiac muscle and insulin-sensitive tissues (8). Also, a necessary role for ␤-cell mitochondrial metabolism has been proven for glucose-induced insulin secretion (10,11). Surprisingly, neither glucose oxidation nor glucose-induced insulin secretion was inhibited in the ZF islets or in our FA-cultured islets (17). The question addressed in this study is what confers the paradoxical effect of increased glucose oxidation and preserved glucose-induced insulin secretion in the ZF islets when PDH activity is suppressed?
␤-cell glucose metabolism is complex, with highly active pathways in addition to pyruvate decarboxylation by PDH that link cytosolic and mitochondrial metabolism. One that has attracted considerable attention is metabolism of pyruvate through PC, the anaplerosis pathway (13)(14)(15)(16)35). PC is unusually active in ␤-cells so that half the pyruvate is normally metabolized through PC, the other half through PDH (36 -38). The reason for the atypically high PC activity is under investigation, with regulatory roles proposed for glucose-induced insulin secretion through downstream coupling factors: reducing equivalents produced by the malate-pyruvate (13,18) or citrate-pyruvate (14,39) shuttles, glutamate (40), and malonyl-CoA that inhibits fatty acid oxidation so that cytoplasmic levels of long chain acyl-CoAs and complex lipids rise (15,16). The importance of this pathway is supported by studies that show rises and falls in ␤-cell PC activity closely correlate with glucose-induced insulin secretion (13, 20 -22, 41) and also by the observation in this study and by others (14,18) that glucoseinduced insulin secretion is impaired by the PC inhibitor, PAA.
In contrast to the lowered PDH activity in the ZF islets was increased PC activity. Also, our results are compatible with enhanced flux through the malate-pyruvate shuttle based on increased V max values for MDH and malic enzyme, raised concentrations of pyruvate, malate, and oxaloacetate, and 2-fold increased release of malate from isolated mitochondria. A second shuttle, the citrate-pyruvate shuttle, is similar except that mitochondrial oxaloacetate proceeds to citrate by citrate synthase, and the citrate is exported to the cytoplasm for conversion to oxaloactetate by citrate lysase, then to malate by cytoplasmic MDH with conversion of NADH to NAD ϩ , and then to pyruvate by malic enzyme (14). Thus, this shuttle not only produces NADPH through the malic enzyme reaction as occurs in the malate-pyruvate shuttle but also maintains a high cytoplasmic NAD ϩ to NADH ratio. Our finding an increased citrate concentration in the ZF islets may suggest increased flux through this shuttle; this would be particularly advantageous to the ZF islets due to the lowered mGPD activity possibly causing impairment of the glycerol phosphate shuttle, which is normally an important source in ␤-cells for the cytoplasmic NAD ϩ that is required for glycolysis (42). Enhanced trafficking through these shuttles likely explains how glucose oxidation is not impaired in the ZF islets by providing a loop for pyruvate to rise to supernormal levels that overcome the lowered PDH activity by a mass action effect; the pyruvate concentration was nearly 3-fold increased in the ZF islets.
What potential insight do these findings provide into the ␤-cell adaptation of the ZF rats? By showing that there is a mitochondrial pyruvate metabolism pathway in ␤-cells that is not impaired by FA, they provide a mechanism for ␤-cells to escape the glucose dysmetabolism that occurs with excess FA in other tissues. A more speculative possibility is that the FIG. 6. Glucose-induced insulin secretion from isolated islets of 10 -12week-old ZL (A) and ZF (B) rats in the absence and presence of 5 mM phenylacetic acid, which inhibits pyruvate carboxylase activity (14). Data are mean Ϯ S.E. and are expressed as insulin secretion as a proportion of insulin content. diversion of pyruvate metabolism through PC plays a direct role in the ␤-cell adaptive hyperfunction and/or expansion. This possibility is based on the growing evidence for downstream pathways from PC directly influencing insulin secretion, in particular the pyruvate shuttles (13,14,18). Furthermore, a facilitative role in the compensatory increase in ␤-cell mass may be played by providing glucose-derived amino acids and lipids for the needed structural macromolecules (38). How then to reconcile the pancreas perfusion and static incubation findings in this study that find that insulin secretion was not increased in the ZF rats after correcting for their larger ␤-cell mass? This result must be viewed in the context of the markedly reduced ␤-cell insulin stores in the ZF rats that is reported to stem from impaired proinsulin biosynthesis (43). We and others have shown a regulatory role for pancreas insulin content over insulin secretory responses in rats (44,45). With that understanding, the observed insulin secretion in the ZF rats is in considerable excess to what would be expected for their level of stored insulin, which is more compatible with the observed increase in islet glucose metabolism.
An unexpected finding in the ZF islets was the lowered activity of mGPD, which is the rate-limiting enzyme of the glycerol phosphate shuttle. mGPD is 50 times more active in ␤-cells than other tissues (46). Interest in this enzyme was spurred by the observation that its activity is lowered in islets of diabetic rodents (21,22), leading to the suggestion that it played a causative role in the defective glucose-induced insulin secretion that characterizes diabetic states. However, null mice for mGPD were found to have no defect in insulin secretion unless combined with pharmacological disruption of the malate-aspartate shuttle (34). Furthermore, overexpressing mGPD in islets of diabetic GK rats failed to restore glucoseinduced insulin secretion (47). Thus, defective activity of the glycerol phosphate shuttle alone is not sufficient for impaired glucose-induced insulin secretion, which is consistent with our findings in the ZF islets. An alternate possibility is a metabolic or functional benefit from an impaired glycerol phosphate shuttle in the ZF rats, perhaps by shifting reoxidization of cytoplasmic NAD ϩ to the citrate-pyruvate shuttle so that production of NADPH through the malic enzyme reaction also is enhanced.
In summary, the well known effect of excess FA to impair PDH activity appears to be present in islets of insulin-resistant ZF rats. Also, mGPD activity, and by inference, the glycerol phosphate shuttle are defective. However, ␤-cell mitochondrial metabolism and glucose-induced insulin secretion are preserved, which we propose stems from diversion of pyruvate metabolism through PC and downstream pathways. Our results provide a mechanism for how ␤-cells escape the mitochondrial dysmetabolism that occurs with excess FA in other tissues. Also, as the anaplerosis pathways are implicated in glucose-induced insulin secretion and the synthesis of glucosederived lipid and amino acids, greater than normal flux through these pathways may be key components of the ␤-cell hyperfunction and increased ␤-cell mass that make up the ␤-cell adaptation of the ZF rats. Assigning such importance to PC and the anaplerosis pathways focuses interest on studies showing lowered islet PC activity in diabetic rodents as a potential mechanism of failed ␤-cell adaptation (21,22).