| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
activation reverses adverse effects induced by high-saturated-fat feeding on pancreatic 
-cell function in late pregnancyCentre for Diabetes and Metabolic Medicine, Institute of Cell and Molecular Science, St. Bartholomew's and the Royal London School of Medicine and Dentistry, Queen Mary, University of London, London, United Kingdom
Submitted 27 July 2006 ; accepted in final form 8 December 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
activation or substitution of 7% of dietary fatty acids (FAs) with long-chain 
-3 FA, manipulations that improve insulin action in the nonpregnant state. Glucose tolerance at day 19 of pregnancy in the rat was impaired by high-saturated-fat feeding throughout pregnancy. Despite modestly enhanced glucose-stimulated insulin secretion (GSIS) in vivo, islet perifusions revealed an increased glucose threshold and decreased glucose responsiveness of GSIS in the saturated-fat-fed pregnant group. Thus, insulin resistance evoked by dietary saturated fat is partially countered by augmented insulin secretion, but compensation is compromised by impaired islet function. Substitution of 7% of saturated FA with long-chain -3 FA suppressed GSIS in vivo but did not modify the effect of saturated-fat feeding to impair GSIS by perifused islets. PPAR activation (24 h) rescued impaired islet function that was identified using perifused islets, but GSIS in vivo was suppressed such that glucose tolerance was not improved, suggesting modification of the feedback loop between insulin action and secretion.
islet function; peroxisome proliferator-activated receptor-; fatty acids; insulin resistance
(PPAR) expression and suppresses expression levels of PPAR target genes (32). In INS-1 cells, this causes elevated rates of insulin secretion at low glucose (32). This mimics in vitro one aspect of the islet adaptation to pregnancy, namely a lowering of the glucose threshold. Increasing insulin resistance, together with failure of the pancreatic 
-cell to meet the increased demand for insulin secretion, will lead to glucose intolerance and gestational diabetes.
Glucose intolerance and gestational diabetes occurs in 2–6% of pregnancies in European women, and gestational diabetes is associated with an increased rate of complications during the second half of pregnancy (22) together with an increased risk of subsequent development of type 2 diabetes (21). The excessive development of insulin resistance during pregnancy most likely underlies the development of gestational diabetes (1, 9; reviewed in Ref. 9a). The nongenetic factors that may precipitate glucose intolerance and gestational diabetes are incompletely elucidated. In the nonpregnant state, dietary saturated fat causes insulin resistance (13, 28, 29, 34), which can be accompanied by altered islet function permitting compensatory hypersecretion of insulin (11, 14, 17). However, insulin resistance induced by dietary saturated fat in nonpregnant rats can be alleviated by PPAR activation in vivo (14, 45), which, probably by promoting lipid catabolism, prevents the excess accumulation of intracellular lipid that is associated with impaired insulin signaling (reviewed in Refs. 30 and 40). Effects of PPAR activation on GSIS occur within 24 h of exposure in vivo and can be demonstrated in the intact animal and after islet isolation (14, 15, 37–39). Improved insulin action obviates the requirement for islet compensatory hypersecretion of insulin, and GSIS in vivo is accordingly reduced (14, 39). PPAR-null mice are protected from high-fat diet-induced insulin resistance, and so the global absence of PPAR signaling also obviates any requirement for islet compensation for insulin resistance (10).
In nonpregnant rats, substitution of long-chain -3 fatty acids (FAs) for a small percentage (6–7%) of saturated FAs in the diet can also oppose the development of insulin resistance in response to high-saturated-fat (HF) feeding (11, 35, 36). This intervention can also reverse compensatory insulin secretion in vivo and with perifused islets (11). The impact of long-chain -3 FA on insulin sensitivity may be related to PPAR function, since PPAR-dependent genes in liver are upregulated by high-fat diets containing fish oils compared with other high-fat diets, including lard-based diets (8).
The present study investigated whether HF feeding throughout pregnancy further impairs maternal insulin sensitivity during late pregnancy and, if so, whether GSIS can be augmented to an extent adequate to compensate for the augmented insulin resistance. In addition, it examined whether either PPAR activation or substitution of a small percentage of the dietary saturated FAs with long-chain -3 FA can ameliorate HF-induced insulin resistance during pregnancy and how these manipulations impact on GSIS in vivo and in islet perifusions. The former intervention provides insight into the mechanism of action involved in modulating insulin secretion during late pregnancy, whereas the latter provides a potential nonpharmacological intervention to ameliorate impaired insulin secretion and/or action in late pregnancy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Animals.
All procedures performed in this study were approved by the Local Ethics Review Committee of Queen Mary and were carried out under the authority of the appropriate Home Office (UK) personal and project licences in accordance with the regulations of the Home Office Animals (Scientific Procedures) Act 1986. Female rats (200–250 g), maintained at 21 ± 2°C and subjected to a 12:12-h light-dark cycle, were time mated. Pregnancy was confirmed by the appearance of sperm plugs. Pregnant rats with less than eight fetuses were excluded from the study. Analyses were undertaken on day 19 of pregnancy (term = 22–23 days). Energy intake was monitored by weighing on a daily basis food contained in specially designed food hoppers. Rats were allowed access ad libitum to water throughout. One group of pregnant rats (P group) was given free access to standard, low-fat/high-carbohydrate rodent diet (367 kcal/100 g) containing 52% carbohydrate, 15% protein, 3% lipid, and 30% nondigestible residue (by weight). A second group of pregnant rats were maintained on a HF diet from the onset of pregnancy (P-HF group). The semisynthetic HF diet (419 kcal/100 g diet) contained 34% carbohydrate, 19% protein, and 22% lipid [lard as the major source of lipid, together with corn oil (1.9 g/100 g diet), to prevent essential FA deficiency] by weight percentage. (17). The lipid component of the HF diet comprised 16% saturated FA (mainly stearic), 16% monounsaturated FA (mainly oleic), and 7% polyunsaturated FA (mainly linoleic) by energy. Despite increased energy intakes in the P-HF group compared with the P group (a mean increase of 35%), body weight gain during pregnancy was not further increased by HF feeding (data not shown). A third group of rats was provided with free access to long-chain -3 FA-enriched HF diet on confirmation of pregnancy (P--3HF group). The long-chain -3 FA-enriched HF (419 kcal/100 g diet) was isocaloric with the HF diet and was also lard/corn oil based, but 
7% of the dietary saturated FAs were replaced with long-chain -3 FA from marine oil. By gas-liquid chromatography analysis, 49% of the long-chain -3 FA was eicosapentaenoic acid (20:5) and 33% docosahexaenoic acid (22:6). Mean fetal numbers per dam were similar in each of the groups: 14 ± 1 (n = 6) for the P group, 13 ± 1 (n = 10) for the P-HF group, and 13 ± 1 (n = 6) for the P--3HF group. Holding the total lipid content constant, substituting 7% of the dietary saturated fat with long-chain -3 FAs (P--3HF group) did not significantly affect energy intake during pregnancy compared with the P-HF group; however, body weight gain was significantly (P < 0.05) lower from day 13 (by 8%) to day 19 (by 21%) of pregnancy. In addition, although fetal number was unchanged, mean fetal weights at day 19 of pregnancy tended to be lower for the P--3HF group [2.3 ± 0.1 (n = 6) for the P group, 2.2 ± 0.1 (n = 10) for the P-HF group, 2.3 ± 0.1 (n = 8) for the P-HF-WY group, and 1.8 ± 0.2 (n = 6) for the P--3HF group]. There were no obvious differences in physical activity (assessed by visual observation) between the pregnant groups (results not shown). Because the decline in fetal mass in the P--3HF group (mean of 5.2 g) makes only a minor contribution to the difference in total body weight (mean of 76.8 g) between the P-HF and P--3HF rats at day 19 of pregnancy, the failure to gain weight in P--3HF rats between days 13 and 19 of pregnancy is consistent with increased energy dissipation. A subgroup of pregnant rats fed the HF diet from day 1 of pregnancy was transferred to the -3HF diet on day 18 of pregnancy (P-acute -3HF group) and studied on day 19 of pregnancy. WY14,643 was administered to 18-day pregnant rats as a single intraperitoneal injection (50 mg/kg body wt), and rats were sampled after an additional 24 h (14, 15, 37–39). This period of exposure to the PPAR agonist is adequate to elicit markedly enhanced protein expression of the PPAR-linked gene pyruvate dehydrogenase kinase-4 in islets of nonpregnant rats (37). We selected this period of exposure to WY14,643 since previous studies show that, as the period of PPAR activation increases beyond 2 days, pregnant (but not unmated) rats become refractory to its action (33). In addition, more prolonged PPAR activation suppresses fetal growth and increases circulating triglycerides, most likely as a consequence of increased adipocyte lipolysis (33). Control nonpregnant and 19-day pregnant rats were injected with vehicle. WY14,643 administration from day 18 to day 19 of pregnancy did not affect food intake, maternal body weight gain, fetal number, or fetal weight (results not shown).
Intravenous glucose challenge.
Glucose was administered as an intravenous bolus (0.5 g glucose/kg body wt, 150 µl/100 g body wt) to conscious, unrestrained rats via a chronic indwelling jugular cannula that was inserted 1 wk before the experiment (see Ref. 17). The indwelling cannula was flushed with saline after the injection of glucose to remove residual glucose. Blood samples (100 µl) were withdrawn from the indwelling cannula before and at 2, 5, 10, 15, and 30 min after intravenous glucose administration. Samples of whole blood (50 µl) were deproteinized with ZnSO4-Ba(OH)2 and centrifuged (10,000 g) at 4°C, and the supernatant was retained for subsequent assay of blood glucose. The remaining blood sample was immediately centrifuged (10,000 g) at 4°C, and plasma was stored at –20°C until assayed for insulin. Insulin and glucose responses during the glucose tolerance test were used for calculation of the acute insulin response (AIR; mean of suprabasal 2- and 5-min plasma insulin values). Areas under the insulin and glucose curves were calculated as incremental plasma insulin values above values at 0 min integrated over the 30-min period after the injection of glucose (
I0–30) and the corresponding incremental integrated blood glucose values above values at 0 min integrated over the 30-min period after intravenous injection of glucose (G0–30). The rate of glucose disappearance (k) was calculated from the slope of the regression line obtained with log-transformed glucose values from 2 to 15 min after intravenous glucose administration. Homeostasis model assessment (HOMA) of insulin sensitivity scores were calculated as [fasting insulin (µU/ml) x fasting glucose (µmol/ml)]/22.5 (19, 41; reviewed in Ref. 43).
Hyperglycemic clamps.
An intravenous glucose bolus was administered to conscious, unrestrained, postabsorptive rats via an indwelling cannula, and 25% glucose solution was subsequently infused to maintain blood glucose concentrations at a target concentration of 10 mmol/l. Blood samples were obtained before the glucose bolus and at 5-min intervals thereafter for 40 min. Samples of whole blood (50 µl) were treated as described for the intravenous glucose tolerance tests. Insulin responses were used for calculation of the incremental plasma insulin values integrated over the 40-min period of glucose infusion (I0–40).
Islet perifusions. Rats were anesthetized by injection of pentobarbital sodium (60 mg/ml in saline, 1 ml/kg body wt ip). Once locomotor activity had ceased, pancreases were excised and the rats were killed. Islets isolated by collagenase digestion (24) were collected under a dissecting microscope into HEPES-buffered Hanks' balanced salt solution containing 5% BSA. Insulin release from freshly isolated islets (of approximately equal size) was measured in a perifusion system as described previously (39). In this system, 50 islets were housed in small chambers on Millicell culture inserts. Islets were perifused in basal medium (Krebs-Ringer, containing 20 mM HEPES, pH 7.4, 5 mg/ml BSA, and 2 mM glucose) for 60 min at a flow rate of 1 ml/min at 37°C prior to collection of fractions. Perifusate glucose concentrations were modified to generate rises from 2 mM (basal) to the mid-physiological range (8 mM), then to the high physiological range (16 mM), before being switched back to 2 mM glucose. Fractions (2 ml) were collected at 2-min intervals and stored at –20°C prior to assay for insulin and glucose.
Statistical analysis. Data are presented as means ± SE, with the number of individual observations (rats or islet preparations from individual rats) in parentheses. Statistical analysis was performed by ANOVA followed by Fisher's post hoc tests for individual comparisons or Student's t-test as appropriate (Statview; Abacus Concepts, Berkeley, CA). A P value of <0.05 was considered to be statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
G0–30) was significantly increased (by 33%, P < 0.05; Fig. 1D), and rates of glucose disappearance (k values) were significantly suppressed (by 44%, P < 0.001; Fig. 1E). HF feeding during pregnancy caused a 45% increase in HOMA scores (Fig. 2F) compared with the P group; although this did not achieve statistical significance, this trend is consistent with augmented maternal insulin resistance.
|
|
I0–30) (45%; Fig. 2E) in P-HF compared with P rats.
Insulin secretion during hyperglycemic clamps is not impaired by HF feeding during pregnancy.
To test whether impaired glucose tolerance after HF feeding during late pregnancy reflected a limitation of the capacity for insulin secretion, perhaps as a consequence of a reduced insulin content of the islets, we analyzed insulin responses in P and P-HF rats in vivo during hyperglycemia elicited by sustained glucose infusion. Glycemia was maintained at 10 mmol/l for 40 min. HF feeding during pregnancy resulted in a 22% lower mean rate of glucose infusion required to maintain glycemia at the target concentration of 10 mmol/l, reinforcing data obtained by HOMA scores indicating exaggerated development of insulin resistance during pregnancy due to this dietary intervention. Steady-state plasma insulin levels in vivo during steady-state hyperglycemia did not differ significantly between P-HF and P groups (Fig. 2G). Similarly, the area under the insulin curves (I0–40) was unaffected by HF feeding during pregnancy (results not shown). Thus, HF feeding during pregnancy does not limit the capacity for insulin secretion in response to sustained hyperglycemia during late pregnancy.
HF feeding throughout pregnancy greatly impairs GSIS by perifused islets, decreasing the response to submaximal glucose and the maximal glucose responsiveness of GSIS.
The inability of GSIS during late pregnancy to respond adequately to dietary HF to maintain glucose tolerance, despite an ability to maintain high sustained insulin concentrations during a hyperglycemic clamp, suggested that islet function could be impaired such that GSIS cannot be augmented sufficiently to maintain glucose tolerance. To investigate this directly, we examined the impact of HF feeding throughout pregnancy on GSIS ex vivo using stepwise glucose perifusions of isolated islets. High-fat feeding throughout pregnancy elicited a stable impairment in insulin secretion such that GSIS ex vivo with isolated perifused islets was severely compromised (Fig. 3A), as reflected by a 45% decline (P < 0.01) in the area under plasma insulin curves during islet perifusion (I60–140; Fig. 3C).
|
1 (Fig. 3D) and 2 (Fig. 3E) refer to periods during which the perifusate glucose concentration was increased from 2 to 8 mM and from 8 to 16 mM, respectively. Perifused islets from the P-HF group exhibited a markedly suppressed response to a rise in glucose in the low physiological range, as reflected by a 71% decline (P < 0.05) in the value for 1 compared with perifused islets from the P group (Fig. 3D). GSIS by perifused islets from P-HF rats relative to P rats was also decreased, at higher perifusate glucose concentrations (see Fig. 3A), as demonstrated by a 53% decrease in 2 values (Fig. 3E).
PPAR activation opposes the effect of HF feeding during pregnancy and reverses the impaired response to submaximal glucose and maximal glucose responsiveness of GSIS by perifused islets.
Because GSIS during hyperglycemic clamps was unimpaired by high-fat feeding, we hypothesized that the inability to maintain glucose tolerance in vivo in the P-HF group and the lowered insulin secretion observed with perifused islets from the P-HF group were unlikely to reflect inadequate islet insulin content but rather altered control of insulin secretion as a consequence of increased lipid delivery. Previous studies (4) have demonstrated that PPAR agonists can prevent FA-induced impairment of GSIS (25) and that PPAR may be required to ensure appropriate insulin secretion in certain circumstances through its ability to maintain islet lipid homeostasis. We therefore examined whether upregulation of FA oxidation by activation of PPAR in vivo could rectify impaired insulin secretion by subsequently isolated islets. We administered the PPAR agonist WY14,643 in vivo at 24 h prior to sampling, a procedure shown previously (7, 37, 38) to enhance expression of genes involved in lipid catabolism. Others (46) have shown that this period of exposure to a PPAR ligand (bezafibrate) does not affect islet insulin content, although more prolonged (48 h) exposure lowers islet insulin content. On assessment of islet function with perifused islets, antecedent PPAR activation in vivo was found to reverse the adverse effects of HF feeding to impair GSIS by perifused islets (compare 
in Fig. 3A with 
in Fig. 3B). The I60–140 value was increased by 63% (P < 0.05) in the P-HF-WY group compared with the P-HF group such that the I60–140 value was 89% of the value in the P group (Fig. 3C). PPAR activation by WY14,643 partially reversed effects of HF feeding to impair insulin secretion at submaximal glucose concentrations, as indicated by a 44% reversal of the decline in 1 (Fig. 3D) and 66% of the decline in 2 (Fig. 3E) induced by HF feeding. In addition, as shown in Fig. 3B (), PPAR activation largely reversed the effects of HF feeding on glucose responsiveness.
Impact of dietary long-chain -3 FA supplementation on the effect of HF feeding during pregnancy to impair GSIS by perifused islets.
A recent study (8) systematically comparing the metabolic and molecular effects of high-fat diets with differing FA compositions demonstrated that PPAR-dependent genes in liver were predominantly upregulated by high-fat diets containing fish oils compared with other high-fat diets, including lard-based diets. We therefore examined the effects of substitution of 7% of dietary saturated FA in the HF diet with long-chain -3 FA from fish oil (P--3HF group) with those achieved by PPAR activation in vivo using the pharmacological agonist WY14,643 (P-HF-WY group). Contrasting with effects of pharmacological PPAR activation to enhance GSIS by perifused islets from HF-fed rats, substitution of 7% of dietary saturated FA with long-chain -3 FA throughout gestation failed to reverse the effects of HF feeding on islet function ex vivo (Fig. 3, B–E). We observed that dietary long-chain -3 FA supplementation of the HF diet during pregnancy was accompanied by compromised maternal body weight gain together with a corresponding, but nonsignificant, decline in mean fetal weight (18%) (see MATERIALS AND METHODS). We therefore considered that a potential explanation for the discrepancy between the data obtained with acute PPAR activation by WY14,643 and the administration of dietary long-chain -3 FA throughout gestation could be that chronic activation of PPAR for 19 days of gestation with long-chain -3 FA is detrimental during pregnancy, whereas acute activation (for 1 day with WY 14,643) is not. To investigate this possibility, we examined the effects of more acute (24 h) substitution of a small amount of dietary lipid with long-chain -3 FA. A subgroup of pregnant rats fed the HF diet from day 1 of pregnancy was transferred to the -3-HF diet on day 18 of pregnancy and studied on day 19 of pregnancy (P-acute--3HF group). Energy intakes and body weight gain over this 24-h period in the P-acute--3HF group were unchanged compared with the P-HF group. Although the period of -3HF feeding was only 24 h, the effect of long-chain -3 FAs on GSIS by perifused islets was almost identical to that elicited in response to long-term -3HF feeding. I60–140 values were similar for the P--3HF and the P-acute--3HF groups [P--3HF, 1.83 ± 0.28 mU·ml–1·min–1 (n = 5); P-acute--3HF, 2.00 ± 0.14 mU·ml–1·min–1 (n = 5)], as were values for 1 [P--3HF, 44 ± 18 µU·ml–1·min–1 (n = 5); P-acute--3HF, 61 ± 20 µU·ml–1·min–1 (n = 5)] and 2 [P--3HF, 430 ± 81 µU·ml–1·min–1 (n = 5); P-acute--3HF, 481 ± 73 µU·ml–1·min–1 (n = 5)] for islets from P--3HF and P-acute--3HF groups. Thus, effects of acute PPAR activation in HF-fed pregnant rats to partially restore insulin secretion at submaximal glucose concentrations and completely normalize the glucose responsiveness for GSIS are not mimicked by acute or chronic long-chain -3 FA supplementation of the HF-fat diet during pregnancy.
Comparison of effects of PPAR activation and long-chain -3 FA supplementation of the HF diet on effects of HF feeding during pregnancy to modulate insulin sensitivity, glucose tolerance, and GSIS in vivo.
Because of the dramatic effects of PPAR activation in late pregnant rats to restore GSIS with perifused islets from HF-fed rats, we examined the effects of PPAR activation during late pregnancy on insulin sensitivity, glucose tolerance, and insulin secretion in vivo, making comparisons with the effects of long-chain -3 FA supplementation of the HF diet. PPAR activation during late pregnancy in late pregnant rats fed a control (low-fat) diet (P-WY group) resulted in higher blood glucose concentrations at 10 min (42%, P < 0.05) and 15 min (by 69%, P < 0.05) after the glucose load (Fig. 1B) and a 28% decline (P < 0.05) in rates of glucose disappearance (k values; Fig. 1E), although it failed to affect insulin sensitivity (as assessed by HOMA scores; Fig. 2F). Insulin concentrations at 10 min (50%, P < 0.05), 15 min (44%, P < 0.05), and 30 min (62%, P < 0.05) after the glucose load were lower in the P-WY group (Fig. 2B); there were associated (although nonsignificant) trends toward lower AIR and I0–30 values compared with the P group (Fig. 2). Neither blood glucose profiles after an intavenous glucose bolus (Fig. 1B), G0–30 values (Fig. 1D), nor k values (Fig. 1E) were significantly affected by PPAR activation in HF-fed late-pregnant rats. However, significant effects of PPAR activation on GSIS in vivo were observed in the high-fat fed rats, with a 53% decrease (P < 0.05) in AIR and a 48% decrease in I0–30 values compared with the P-HF group (Fig. 2, D and E, respectively). Since glucose tolerance is similar in the P-HF and P-HF-WY groups, despite markedly lower rates of insulin secretion in the latter group, it is implied that PPAR activation in vivo alleviates the induction of insulin resistance induced by HF feeding during pregnancy (a conclusion supported by attenuated increases in HOMA scores in the P-HF-WY group compared with the P-HF group; Fig. 2F) but precludes insulin secretion of a magnitude adequate to normalize glucose tolerance to that observed in pregnant rats fed a low-fat diet. Supplementation of the HF diet during pregnancy with -3 FAs resulted in a marked increase in insulin sensitivity, as reflected by a 64% decline (P < 0.05) in HOMA scores compared with the P-HF group (Fig. 2F). Thus, -3-FA supplementation of the HF diet during pregnancy markedly enhanced insulin sensitivity, as is observed in nonpregnant rats (11, 36). Consistent with enhanced insulin sensitivity, blood glucose concentrations after the intravenous glucose load and G0–30 and k values were unaffected by -3-FA supplementation of the HF diet (Fig. 1, D and E) despite substantial decreases in both the AIR and I0–30 values compared with the P-HF and P groups (P < 0.001; Fig. 2, D and E). The acute (24 h) substitution of a small amount of dietary lipid with long-chain -3 FAs also modestly enhanced insulin sensitivity (reducing HOMA scores by 27%); insulin concentrations after an intravenous glucose load were also attenuated but in this case were associated with high G0–30 and low k values that were not restored to values observed in the P group.
In additional experiments, we examined effects of acute PPAR activation in the P--3HF group. Interestingly, acute PPAR activation in the P--3HF group was also associated with enhanced insulin sensitivity compared with the P-HF-WY group (assessed by a 34% decline in HOMA scores). Importantly, in addition, insulin concentrations after an intravenous glucose load were significantly higher compared with the P--3HF group (Fig. 2C) such that AIR and I0–30 values were not significantly different from those observed in the P group (Fig. 2, D and E). Furthermore, blood glucose concentrations after an intravenous glucose load were lower compared with values observed in the P--3HF group, with normalization of G0–30 values (Fig. 1D) and partial restoration of k values (Fig. 1E) to those observed in the P group.
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
activation. Perifusions of isolated islets were used to establish whether modulation of dietary lipid composition and/or PPAR activation directly affected the characteristics of islet GSIS ex vivo. Finally, we examined whether effects of PPAR activation were mimicked by substitution of a small amount (7%) of dietary saturated FA with long-chain -3 FAs.
The existence of an in vivo functional feedback loop whereby impaired insulin action evoked by dietary saturated fat during pregnancy may be partially countered by augmented insulin secretion is indicated in late pregnancy. Impaired glucose tolerance observed in response to high-saturated-fat feeding during pregnancy is not associated with decreased GSIS in response to acute glucose challenge in vivo. Furthermore, high-saturated-fat feeding during pregnancy did not limit the capacity for insulin secretion in response to sustained hyperglycemia during late pregnancy. These observations suggest that glucose intolerance may reflect inadequate compensatory hypersecretion of insulin in response to the development of impaired insulin action. The inadequate insulin secretion in vivo may reflect direct alterations in islet function and/or indirect effects of circulating nutrients or hormones or altered neuronal signals to the islet. Ex vivo islet perifusions that are not influenced by these factors demonstrated that insulin secretion by islets isolated from pregnant rats fed a high-fat diet was greatly impaired. Thus, the modest insulin hypersecretion observed in the P-HF group in vivo reflects a partial compensatory response of insulin secretion to peripheral insulin resistance that is limited by an intrinsic impairment in the capacity for islet insulin secretion. In the nonpregnant state, long-term administration of PPAR agonists are reported to improve islet function and oppose the development of diabetes. Chronic fenofibrate administration (for 28 wk) to diabetes-prone Otsuka Long-Evans Tokushima fatty rats increases FA oxidation, lowers islet triglyceride content, and prevents morphological changes in the islets contributing to the prevention of the development of diabetes (23). Similarly, chronic PPAR treatment increases the islets/total pancreatic surface area and pancreatic insulin content in Zucker diabetic fatty rats (2). Similarly, in the nonpregnant state, PPAR agonist administration can prevent FA-induced impairment of GSIS (25), and it has been suggested that PPAR may ensure appropriate insulin secretion in certain circumstances through maintaining islet lipid homeostasis (4, 47). In the present study, effects of acute (24 h) PPAR activation in vivo to rescue the impairment in islet GSIS elicited by high-saturated-fat feeding during pregnancy are identified using isolated perifused islets in vitro. PPAR deficiency has been reported to blunt the response of insulin mRNA expression to 24-h culture in the presence of high glucose, leading to the suggestion that PPAR is needed for induction of insulin mRNA (4). However, since exposure to PPAR agonists over this time scale used in the present study does not affect islet insulin content (and may decrease islet insulin content if exposure is more prolonged) (46), our data suggest that PPAR activation rectifies the impairment in islet function that occurs in response to high-saturated-fat feeding during pregnancy. Although PPAR-dependent genes in liver are upregulated by high-fat diets containing fish oils compared with lard-based high-fat diets (8), the effects of PPAR activation on islet GSIS are not reproduced solely by supplementation of the high-saturated-fat diet with long-chain -3 FAs. However, the combination of dietary supplementation of the high-saturated-fat diet with -3 FA and acute PPAR activation reproduces in vivo the restoration of GSIS that is observed with perifused islets from high-saturated-fat-fed pregnant rats treated with WY14,643 but which cannot be achieved by dietary supplementation alone.
High-saturated-fat feeding throughout pregnancy moderately increased GSIS in vivo after an acute intravenous glucose challenge at day 19 of pregnancy compared with control. Glucose intolerance nevertheless ensued. The increase in GSIS observed in vivo in response to an intravenous glucose bolus in P-HF compared with P group reflects factors (e.g., circulating nutrients or hormones or altered neuronal signals to the islet) that operate in the intact organism that signal the deterioration of peripheral insulin sensitivity. However, the function of the normal compensatory feedback loop between peripheral insulin resistance and altered islet function permitting compensatory hypersecretion of insulin is compromised; impaired glucose tolerance indicates that pancreatic -cell compensatory hypersecretion of insulin is inadequate for the degree of insulin resistance. It is possible that the consumption of excess dietary saturated FAs, by compromising -cell function, could precipitate the development of gestational diabetes, because the compensatory feedback loop between insulin resistance and islet compensation also appears to be defective in women with a predisposition towards the development of gestational diabetes (see Ref. 3). The inability to maintain glucose tolerance after high-saturated-fat feeding throughout pregnancy contrasts with unimpaired glucose tolerance after a slightly longer (28-day vs. 19-day) period of administration of the same high-saturated-fat diet to female nonpregnant rats of a similar age (14, 17). Similarly, the enhancement of GSIS elicited in response to high-saturated-fat feeding during pregnancy is less than that elicited in response to 28 days of high-saturated-fat feeding in female nonpregnant rats (14, 17). Islet perifusions showed markedly impaired -cell function in pregnant rats maintained on high-saturated-fat diet compared with those maintained on normal (low-fat) diet, as reflected by impaired insulin secretion at submaximal glucose concentrations and a decrease in glucose responsiveness (assessed as insulin secretion at perifusate glucose concentrations in the high physiological range) for GSIS by isolated perifused islets. Thus, pregnancy renders the pancreatic -cell susceptible to adverse effects of increased saturated fat delivery that are not observed in the nonpregnant state. Although accumulation of intracellular triglyceride in islets precedes islet dysfunction in the development of type 2 diabetes (5, 27, 31, 42), previous studies by Winzell et al. (44) failed to observe altered islet triglyceride content in mice fed low- or high-fat diets supplemented with a combination of dietary -3 polyunsaturated FAs and conjugated linoleic acids.
We anticipated that impaired glucose tolerance evoked by high-saturated-fat feeding in late pregnancy would be ameliorated by pharmacological PPAR activation, as occurs in the nonpregnant state (14), but this was not the case. Thus, glucose intolerance in late pregnant rats maintained on high-saturated-fat diet is refractory to PPAR activation. However, importantly, the adverse effects of high-saturated-fat feeding on -cell function identified ex vivo using islet perifusions were completely attenuated by PPAR activation despite the continued provision of high-saturated-fat diet. There is thus a selective action of PPAR activation on the islet itself. The inability of insulin secretion in vivo to increase sufficiently to restore glucose tolerance following WY14,643 treatment, despite the improved -cell function that was demonstrated in islet perifusions, suggests that factors in vivo, humoral or neuronal, suppress GSIS.
In the nonpregnant state, substitution of a small percentage of saturated FAs with long-chain -3 FAs opposes high-saturated-fat-induced insulin resistance (11, 35, 36) and reverses compensatory insulin secretion (11). On the basis of studies in the nonpregnant state, we predicted that impaired glucose tolerance evoked by high-saturated-fat feeding in pregnancy would also be ameliorated by replacement of 7% of the dietary saturated fat with long-chain -3 FAs. This did not occur. However, whole body insulin sensitivity was improved in late pregnancy by enrichment of a high-saturated-fat diet with long-chain -3 FAs (as assessed by a decline in HOMA scores), as is observed in nonpregnant rats (11, 35, 36). Lowered GSIS despite retained glucose intolerance suggests that the insulin response to glucose was suppressed to a greater extent than whole body insulin sensitivity was enhanced by enrichment of the high-saturated-fat diet by long-chain -3 FA. Our present studies with perifused islets revealed that substitution of 7% of dietary lipid with long-chain -3 FAs failed to oppose the effect of high-saturated-fat feeding during pregnancy to increase insulin secretion at submaximal glucose concentrations and lower glucose responsiveness for GSIS in perifused islets. A confounding factor in the interpretation of the impact of enrichment of a high-saturated-fat diet with long-chain -3 FA on insulin secretion and action in late pregnancy is that the pregnant rats fed the high-saturated-fat diet supplemented with long-chain -3 FAs failed to gain weight during the later stages of gestation. Reduced adipose tissue mass has been reported to be a consequence of dietary supplementation with a combination of dietary -3 polyunsaturated FAs and conjugated linoleic acids in mice (44) and was associated with impaired insulin secretion, leading to glucose intolerance. Body composition was not measured in the present study. However, both triacylglycerol and nonesterified FA concentrations were lower in the P-HF, P--3HF, and P-HF-WY groups compared with the P group (results not shown), possibly reflecting increased uptake and augmented lipid metabolism. We addressed the discrepancy between effects of acute (24 h) PPAR activation by WY14,643 compared with dietary long-chain -3 FA supplementation throughout gestation. To test whether this reflected detrimental effects of the latter chronic intervention, we examined the effects of acute (24-h) substitution of a small amount of dietary lipid with long-chain -3 FAs. Although -3HF feeding was limited to 24 h, the period of exposure to WY14,643, its impact on GSIS by perifused islets was almost identical to that elicited in response to the longer-term (19 days) -3HF feeding. Thus, effects of acute pharmacological PPAR activation by WY14,643 in high-saturated-fat-fed pregnant rats are not mimicked by either acute or chronic long-chain -3 FA supplementation of the high-saturated-fat fat diet during pregnancy.
Interestingly, acute PPAR activation in combination with dietary long-chain -3 FA supplementation throughout gestation was associated with enhanced insulin sensitivity, elevated insulin concentrations after an intravenous glucose load, and normalization of glucose tolerance. Thus, the combination of dietary supplementation of the high-saturated-fat diet with -3 FA and acute PPAR activation reproduces in vivo the restoration of GSIS that is observed in vitro with perifused islets from high-saturated-fat-fed pregnant rats treated with WY14,643 but which cannot be achieved by dietary -3 FA supplementation alone. Thus, both HF and -3HF elicit stable adaptations that impair -cell function and limit the capacity for insulin secretion to compensate for insulin resistance. These effects can be reversed by WY14643; however, this is associated only with restoration of glucose tolerance in vivo when in combination with enhancement of insulin sensitivity elicited by dietary long-chain -3 FA supplementation of the high-saturated-fat diet throughout gestation.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
/
ragaglitazar eliminates fatty liver and enhances insulin action in fat-fed rats in the absence of hepatomegaly. Am J Physiol Endocrinol Metab 284: E531–E540, 2003.
| ||||||||||||||||||||||||
Statistically significant effects of WY14,643 treatment, P < 0.05.
