Physiological Increase in Plasma Leptin Markedly Inhibits Insulin Secretion In Vivo

Animals. Three-month-old male Sprague-Dawley rats (n = 28)(Charles River Laboratories, Wilmington, MA) were used for this study. Rats were housed in individual cages and were subjected to a standard light (6:00 A.M. to 6:00 P.M.) and dark (6:00 P.M. to 6:00 A.M.) cycle. All rats were fed ad libitum using regular rat diet that consisted of 64% carbohydrate, 30%protein, and 6% fat with a physiological fuel value of 3.3-kcal/g food. One week before the in vivo study, rats were anesthetized by inhalation of methoxyflurane, and indwelling catheters were inserted in the right internal jugular vein and in the left carotid artery. This method of anesthesia allows fast recovery and normal food consumption after 1 day. The venous catheter extended to the level of the right atrium, and the arterial catheter was advanced to the level of the aortic arch. Recovery was continued until body weight was within 3% of the preoperative weight (∼4-6 days). Studies were performed in awake, unstressed, chronically catheterized rats(19,20).

Hyperglycemic clamp study. To demonstrate the in vivo effect of leptin on insulin secretion, all rats were subjected to 4-h moderate hyperglycemia (∼11 mmol/l) ∼6 h postprandially. Briefly, 25% glucose was infused intravenously to raise the plasma glucose concentration acutely to∼11 mmol/l. The glucose infusion rate was then varied to maintain the plasma glucose concentration at this level for 240 min. The first 120 min of the clamp study was maintained before the infusion of either saline or leptin to avoid the confounding effects of the acute increases in insulin levels during the first phase of insulin secretion and achieve similar levels of glucose-stimulated plasma insulin levels.

Leptin study. At minute 120, rats were assigned to receive either saline (1 ml/h, n = 5, control) or leptin infusion for an additional 2 h. Leptin was infused as primed (0.5, 2.5, and 25 μg ·kg^-1 · min for 2 min) continuous infusion at 0.1 (L0.1; n = 4), 0.5 (L0.5; n = 4), and 5 μg ·kg^-1 · min (L5; n = 4), respectively, to determine the dose-response effect on plasma insulin levels.

Milrinone study. At minute 120 to 240, either milrinone (1, 3, and 28 μg · kg^-1 · min, n = 7) or milrinone(28 μg · kg^-1 · min) + leptin (5 μg ·kg^-1 · min, L5) (n = 4) was infused. Plasma samples for insulin were obtained at 10-min intervals throughout the study. Samples were also obtained for determination of plasma leptin and free fatty acid(FFA) concentrations. A solution (1:1 vol/vol) of ∼3.0 ml fresh blood(obtained by heart puncture from a littermate of the test animal) and heparinized saline (10 U/ml) was infused at a constant rate throughout the study to prevent volume depletion and anemia. At the end of the clamp study,rats were killed using 60 mg pentobarbital sodium/kg body wt i.v.

The study protocol was reviewed and approved by the Animal Care and Use Committee of the Albert Einstein College of Medicine.

Analytical procedures. Plasma glucose was measured by the glucose oxidase method (Glucose Analyzer II; Beckman Instruments, Palo Alto, CA) and plasma insulin by radioimmunoassay using rat insulin standards. Plasma leptin was assayed using the Linco leptin assay kit (Linco Research, St. Charles,MO). Plasma nonesterified fatty acid concentrations were determined by an enzymatic method with an automated kit according to the manufacturer's specification (Waco Pure Chemical Industries, Osaka, Japan).

Terminology and calculations. To express insulin sensitivity, metabolic clearance rate (MCR) of glucose was determined to account for glucose uptake in a variety of nonmaximal stimulations of insulin levels. Thus, the MCR of glucose was calculated as follows: MCR = glucose infusion rate (GIR) divided by 2 plasma insulin.

Statistical analysis. The significance of group differences was evaluated by the two-sample t test. Pearson correlation coefficients were calculated to estimate the linear relationship between variables. All values are presented as means ± SE. All statistical analyses were performed using SPSS.

Leptin studies. Before the study, body weights (305 ± 10 g),plasma glucose (8 ± 0.2 mmol/l), insulin (2.4 ± 0.4 ng/ml),leptin (1.5 ± 0.1 ng/ml), and FFA (0.6 ± 0.04 meq/l) levels were similar in all groups. During the clamp study, glucose levels were maintained at similar levels (∼11 mmol/l), and insulin levels continued to increase throughout the saline (control) study and tended to reach a plateau after minute 160 of hyperglycemia (Fig. 1), as previously demonstrated(21). With variable infusion rates of leptin (0.1, 0.5, and 5 μg · kg^-1 · min),a dose response could be assessed because averaged plasma leptin levels during the clamps were 6 ± 1, 29 ± 5, and 261 ± 4 ng/ml(P < 0.01 between all). These levels were demonstrated 10 min after the primed infusion of leptin. Plasma insulin levels (averaged for the final 120 min of the clamps) were abruptly and significantly decreased by leptin and an approximately twofold decrease was sustained until the termination of the study (14.8 ± 5.8 vs. 34.8 ± 2.6 ng/ml, with leptin and saline infusion, respectively, P < 0.001)(Table 1 and Fig. 1). Moreover, decrease in plasma insulin levels occurred earlier with L5 than with L0.1 (insulin levels at 10 min of leptin infusion: 12 ± 3 vs. 18 ± 1 ng/ml with L5 vs. saline compared with insulin levels at 20 min 17 ± 2 vs. 22± 2 ng/ml with L0.1 vs. saline, respectively, P < 0.05)(Fig. 1). Intermediate results were obtained at plasma leptin levels of ∼29 ng/ml(Table 1). Figure 2 further demonstrates leptin's dose-response effect and the significant inverse correlation between leptin and plasma insulin levels (r = -0.731, P < 0.01). Of note, ∼70% of leptin inhibition occurred with the lowest infusion rate that increased leptin levels by only approximately threefold. Leptin infusion significantly decreased GIR by ∼30% in all groups compared with saline(P < 0.05). However, when GIR was further expressed in terms of the MCR of glucose (GIR/plasma insulin), this remained unchanged in all groups compared with saline (Table 1). These data indicate no change in insulin sensitivity during this short experimental period.

Milrinone study. Plasma insulin levels increased promptly upon infusion of milrinone during the first 10 min (insulin: 31 ± 6 vs. 18± 1 ng/ml, with milrinone and saline respectively, P <0.01) (Fig. 3). Surprisingly,the stimulatory effect of milrinone lasted only for 30 min and was subsequently followed by a decrease in plasma insulin levels. This may be explained by a direct effect of milrinone on insulin secretion, glucose uptake, or other unrelated pathway by which milrinone may influence glucose homeostasis. Co-infusion of milrinone with L5 also resulted in an abrupt increase in insulin secretion during the first 10 min (insulin: 27 ± 6 ng/m vs. saline; P < 0.01)(Fig. 3), suggesting an initial milrinone action. However, there was an acute decrease in plasma insulin levels during the next 10 min, eventually reaching levels similar to those noted with leptin infusion alone (insulin: 17 ± 3 vs. 13 ± 4 ng/ml, with milrinone + L5 and L5, respectively)(Fig. 3).

This study demonstrates that acute increases in plasma leptin that are in a physiologically relevant range suppress glucose-stimulated insulin secretion in a dose-dependent manner.

Many in vitro studies on islets, perfused pancreas, and cell lines demonstrated that leptin had an inhibitory effect on glucose-stimulated insulin secretion(14,15,16,17),presumably through the ObRb receptor. However, leptin also has an important role in the central regulation of food intake and energy expenditure and possibly peripherally through autonomic output from the hypothalamus(14,24). Thus, the effects of leptin on β-cell secretion may also work centrally. Also, leptin may have other indirect effects on the function and action of other hormones associated with the control of insulin secretion(14).

Thus, in vivo studies may be most appropriate to examine the overall effect of leptin on insulin secretion. This study was designed to optimize the in vivo model of moderate hyperglycemia that will allow a valid interpretation of the short-term in vivo effect of physiological increases in plasma leptin levels on insulin secretion. Two hours of sustained hyperglycemia were allotted before the infusion of leptin or saline to avoid the confounding effects of the acute rise in insulin levels during the first phase of insulin secretion and to obtain similar levels of insulin release in individual rats before the infusion of leptin or saline at minute 120. Thereafter, the clamp study was performed for an additional 2 h with either saline or leptin to examine the ability of leptin to suppress insulin secretion. A 4-h study was conceived to be adequate and appropriate since glucose-stimulated insulin secretion was previously shown to continue to rise during most of this time period (21).

This study also demonstrated that the inhibitory effect of leptin on insulin secretion was immediate (∼5-10 min after infusion) and dose dependent (Fig. 2). We have previously determined that leptin has no effect on the metabolic clearance rate of insulin (4). This abrupt inhibition of insulin secretion with leptin treatment is highly suggestive that this effect is solely due to leptin action. The physiological plasma leptin levels achieved to significantly suppress insulin secretion are at levels found in mild obesity(25) and with impaired glucose tolerance (26). Whereas Kieffer and Habener (27)proposed that the chronic hyperleptinemic state of obesity may lead to a relative insulin-deficient state (the adipoinsular axis hypothesis), this short-term study cannot make valid implications regarding the effects of chronic hyperleptinemia on insulin secretion. Hyperleptinemia for 14 days in vivo and 3 days in vitro in normal and fa/fa rats suppressed insulin secretion due to its lipopenic effect on islet cells(28,29). These observations (with leptin administration for several days) were different from our study in which leptin effects had a rapid onset (within minutes), and without changes in markers of lipid homeostasis.

Clues to direct cellular mechanisms by which leptin suppresses insulin release come from earlier studies from ob/ob and db/db mouse islets because both phenotypes result from the absence of leptin signaling. One of the leading proposed mechanisms is through a reduction of cAMP by leptin. This proposed mechanism was further supported when milrinone, a selective inhibitor of PDE3, completely blocked the inhibitory effect of leptin on glucose- or glucagon-like peptide 1-potentiated insulin secretion in vitro (13). In an effort to examine whether leptin action on glucose-induced insulin secretion is through the activation of PDE3, we infused high doses of milrinone alone and milrinone with high doses of leptin. The interpretation of these data is limited,because milrinone increased insulin secretion initially, but with time insulin secretion decreased (Fig. 3). This could have been due to a downregulation of its own effect on insulin secretion through the cAMP pathway. However, if we limit our discussion to the first 60 min of milrinone infusion, leptin inhibited milrinone-induced insulin secretion, an effect not readily explained by leptin activation of PDE3(13).

Our study does not clearly demonstrate the site of leptin action on insulin secretion. Although it was previously assumed from in vitro studies that the metabolic effects of leptin are mediated locally through its tissue receptors,such a notion had been challenged. In fact, many of leptin's effects were demonstrated both centrally and peripherally. For example, it was demonstrated that marked and acute hyperleptinemia modulated hepatic gene expression of the gluconeogenic enzyme PEPCK and the rate of gluconeogenesis in vivo by intracerebroventricular and intravenous routes(4). The same results were demonstrated in liver cell line(30), suggesting that the efferent pathway of leptin from the hypothalamus can influence peripheral metabolism. Combined with negative results from examining the PDE3 system,this example cautions us from concluding that the major effect of leptin on insulin secretion is directly through its receptor on the β-cells.

Finally, chronic hyperleptinemia in obesity is proposed to uncouple leptin action on its receptor in the hypothalamus, thereby attenuating signal transduction pathways that exert resistance to the hormone on satiety and energy expenditure and metabolism(2). If leptin resistance extends to the effect of leptin on insulin secretion, then high leptin levels may not contribute to the modulation of insulin secretion. However, this hypothesis can be supported only by further studies on obese leptin-resistant rodents compared with lean controls.

In summary, this is the first study designed to demonstrate the acute effects of physiological increases in plasma leptin levels on the inhibition of glucose-stimulated insulin secretion in vivo. Although our findings link hyperleptinemia and decreased insulin secretion, the role of chronic effects of hyperleptinemia with obesity leading to the transition to diabetes requires further studies.

This work was supported by grants from the National Institutes of Health(KO8-AG00639 and R29-AG15003 to N.B. and R01-DK 45024 and ROI-DK48321 to L.R),the American Diabetes Association, and by the Core laboratories of the Albert Einstein Diabetes Research and Training Center (DK 20541). N.B. is a recipient of the Paul Beeson Physician Faculty Scholar in Aging Award.

The authors wish to thank Bing Liu, Robin Squeglia, and Manju Suranja for expert technical assistance.