Endocrinology and Metabolism

Leucine in food mediates some of the postprandial rise in plasma leptin concentrations

Christopher J. Lynch, Beth Gern, Carolyn Lloyd, Susan M. Hutson, Rachel Eicher, Thomas C. Vary


In vitro, leptin secretion is regulated at the level of mRNA translation by the rapamycin-sensitive mammalian target of rapamycin (mTOR) and its agonist leucine (Leu). Studies were conducted on meal-trained rats to evaluate the potential physiological relevance of these in vitro findings and the role of Leu in affecting rises in plasma leptin observed after a meal. In the first study, we correlated changes in plasma insulin and Leu to mTOR-signaling pathway activation and plasma leptin at different times during meal feeding. Rapid rises in plasma insulin and Leu, along with mTOR signaling (phosphorylation of eIF4G, S6K1, rpS6, and 4E-BP1) in adipose tissue were observed during the 3-h meal and declined thereafter. Plasma leptin rose more slowly, peaking at 3 h, and was inhibited by rapamycin (0.75 mg/kg) pretreatment. In another experiment, oral Leu or norleucine was provided instead of a meal. Leu and norleucine stimulated a rise in plasma leptin; however, the magnitude was less than the response to a complete meal. In a third study, rats were provided a meal that lacked Leu, branched-chain amino acids, or all amino acids. Stimulation of leptin secretion was reduced ∼40% in animals provided the Leu-deficient meal. Further reductions were not observed by removing the other amino acids. Thus Leu appears to regulate most of the effects of dietary amino acids on the postprandial rise in plasma leptin but is responsible only for part of the leptin response to meal feeding.

  • protein synthesis
  • translation initiation
  • obesity
  • rats
  • meal feeding
  • meal trained

leptin is an 16-kda hormone encoded by the obese (ob) gene and synthesized predominantly in adipose tissue. Leptin regulates body weight by regulating hunger and food consumption, as well as lipolysis, energy consumption, and body temperature. These actions of the hormone are brought about mainly through its actions on hypothalamic leptin receptors and to a lesser extent by activating leptin receptors directly in peripheral tissues.

Even though injection of leptin has rather immediate effects, it is frequently described as being important for long-term regulation of body weight or an indicator of fat cell stores (10). This model suggests that there is constitutive synthesis and release of leptin. Indeed, larger adipocytes appear to synthesize more leptin per cell, and leptin is secreted via a constitutive secretory pathway. Also supporting this long-term regulation idea is the observation that the concentrations of leptin in the blood are generally positively correlated with adipose tissue mass in both lean and obese animals and humans.

Nevertheless, there is considerable evidence that the synthesis of leptin is robustly regulated at the transcriptional and translational levels by nutrition and hormones (8, 16, 17, 21, 3135, 37, 43). Importantly, plasma leptin concentrations increase with feeding (11, 17, 20, 22, 37, 45). Leptin synthesis is stimulated by the nutrient- and insulin-regulated mammalian target of rapamycin (mTOR)-signaling pathway at the level of protein translation in isolated-fat cell studies (7, 42), and this translational regulation of leptin synthesis is stimulated by Leu (42). These effects are blocked by the drug rapamycin, implicating the involvement of the mTOR complex 1 (TORC1), which is notably activated by nutrients and is sensitive to rapamycin (23).

The reports of in vitro effects of Leu on leptin are of interest because previous reports suggest that Leu is a nutrient signal that regulates adipose mTOR and protein synthesis (24, 2628). When added to the diet or provided by gavage, Leu activates the mTOR-signaling pathway in adipose tissue. The effect of Leu on mTOR signaling is not mediated by insulin, because norleucine, which is not an insulin secretagogue, also activates mTOR signaling in fat cells (28). In contrast, the effects of rises in glucose and free fatty acids on leptin appear to be mediated by the rise in plasma insulin (21, 32, 34, 35).

Activation of mTOR by Leu or insulin rapidly affects initiation and elongation steps in protein synthesis (for reviews see Refs. 18 and 39). These short-term effects can have long-term consequences. For example, two substrates of mTOR include the ribosomal protein S6 kinase-1 (S6K1) and the translational repressor eukaryotic initiation factor (eIF)4E-binding protein-1 (4E-BP1). S6K1 is phosphorylated on multiple sites; Thr389 is associated with activation of the kinase. Hyperphosphorylation of ribosomal protein S6 by activated S6K1 is associated with increased translation of mRNAs with a polypyrimidine tract in their 5′-untranslated region (5′-UTR). Many of these mRNAs code for proteins involved in protein synthesis. Thus mTOR activation may promote ribosome biogenesis, leading to increased protein synthetic capacity. Similarly, multisite phosphorylation of the translational repressor family of proteins exemplified by 4E-BP1 (aka PHAS-I, first discovered in fat) allows eIF4E to assemble with eIF4G and other factors into the eIF4F complex. eIF4E is required for a rate-limiting step in translation, recognition of the mRNA cap structure. This step may be particularly important for the translation of messages with significant secondary structure in their 5′-UTR. COOH-terminal phosphorylation sites have been identified on eIF4G1 (Ser1108, Ser1148, and Ser1192) that are regulated by PI 3-kinase/mTOR-signaling pathways. Phosphorylation of eIF4G has been positively correlated with protein synthesis and activation of translation initiation (6, 13, 47) as well as translational regulation via an internal ribosome entry site (IRES) (40). In summary, mTOR regulates global protein synthesis but also preferential translation of certain mRNAs. By leading to the preferential translation of such messages, activation of mTOR may affect cell function and, in the case of leptin, which has effects on reproduction, feeding, and metabolism, could affect the behavior and function of the organism.

To determine whether the in vitro effects of Leu on leptin are physiologically relevant in the intact animal in vivo, we examined the effect of a meal on the phosphorylation states of key proteins involved in translation initiation in adipose tissue and correlated these data with changes in plasma Leu and leptin as well as other potential nutrient regulators of leptin synthesis and secretion. The role of Leu was examined by measuring plasma leptin concentrations in meal-trained animals provided with Leu or norleucine instead of a meal. Finally, we determined the contribution of Leu in food to the rise in plasma leptin that occurs in response to eating a meal and the contribution of rapamycin-sensitive cell signaling to that meal response.


Animals and meal feeding.

The animal facilities and protocol were reviewed and approved by the Institutional Animal Care and Use Committee of The Pennsylvania State University College of Medicine. Narrow-weight range male Sprague-Dawley rats (∼220–225 g) were purchased from Charles River Laboratories (Cambridge, MA). After quarantine, animals were maintained on a reverse 12:12-h light-dark cycle (the dark cycle began at 7 AM and the light cycle at 7 PM) with water provided ad libitum. Animals were caged in pairs rather than singly to reduce anxiety-induced changes in food intake and were trained over a period of ≥12 days to consume their food in a 3-h time window. Rat chow or purified diets, as indicated, were provided in two metal food cups (several times more than could be consumed) for 3 h beginning 30 min after the beginning of the dark cycle. At the conclusion of the meal, all food was collected out of the cage, including any food that the animals clutched. Initial experiments used Teklad Diet 8604, which has a protein concentration of 24.48% (2.04% leucine). Other purified test diets were obtained from Dyets (Bethlehem, PA).

Experimental protocol 1.

The first series of experiments examined the time course of nutrient and endocrine changes in response to feeding in meal-trained animals. After 12 or more days of meal training, animals were weighed and euthanized at six different times relative to the start of the meal consisting of Teklad 8604 (rat chow) or other purified amino acid diets as indicated. One group of animals was euthanized 0.5 h before the meal (t = −0.5 h). The meal was provided at t = 0 h and was removed at t = +3 h. Animals were euthanized at t = +0.5, 1, 3, 6, and 9 h. Truncal blood was collected with EDTA for plasma preparation, and plasma prepared from the truncal blood was snap-frozen in liquid nitrogen for subsequent analyses. Epididymal adipose tissues were frozen between clamps cooled to the temperature of liquid nitrogen for subsequent examination of cell signaling end points.

Experimental protocol 2.

In the second set of experiments, the effect of providing oral Leu or norleucine instead of a meal on plasma leptin concentration was examined. Animals were meal trained for ≥12 days using Teklad 8604 chow and then allocated to one of three treatment groups: saline (control), Leu, or norleucine. Because it was not practical to make all of the measurements in 1 day, the number of animals in the control group was doubled, and on 1 day control and Leu groups were studied. On the subsequent day, control and norleucine animals were studied. As the control data were not significantly different on the 2 days, the data from both days were pooled.

On the day of an experiment, blood was drawn from the tail of the meal-trained animals at −0.5 h before mealtime. Blood samples were collected in Microvette 500 potassium-EDTA-coated sample tubes (Sarstedt, Newton, NC) following a tail snip using a sterile scalpel and kept on ice for plasma preparation. A bandage covered the snipped tail between blood samplings; thus only one cut was necessary. Samples were centrifuged at 1,800 g for 10 min at 4°C; plasma was collected and frozen at −80°C for further analysis. No meal was provided at t = 0 h. After blood sampling, a gavage solution/suspension was provided at t = 0 and +1.5 h according to the designated treatment group: saline (0.155 mol NaCl/l), Leu (54.0 g/l l-leucine), norleucine (54.0 g/l l-norleucine). The dose volume was 2.5 ml/100 g body wt. The two doses of Leu are equivalent to twice the amount of leucine consumed by rats of this age and strain during 24 h of free access to a commercial rodent diet (2). Additional tail blood was taken at +3 h. Plasma prepared from the blood was stored frozen at −84°C until subsequent assays.

Experimental protocol 3.

In the third set of experiments, we examined the effect of removing amino acids from the meal on plasma leptin concentrations. In pilot studies described in the text, animals were meal trained on a commercially available purified amino acid diet (“Commercial Purified”) described by Anthony et al. (5). For the reasons outlined in results, subsequent studies utilized a custom-purified amino acid diet containing 2.04% Leu (“Custom All amino acids”) and other concentrations of amino acids found in the rat chow. These two diets and other diets used for this experiment are compared in Table 1 (columns 1 and 2). Each of the diets in that table has a caloric content of 3.8 kcal/g. After meal training on the control purified diet for ≥12 days, animals were weighed and allocated to one of two groups: control or amino acid-deficient diet groups. Animals assigned to the amino acid-deficient diet group were studied on that day. Tail blood was collected from those animals at t = −0.5 h. At t = 0 h, those animals received a diet deficient in one or more amino acids, as indicated. On that same day, control animals received their usual 3-h meal; no blood was taken from them. At t = +3 h, food was removed and weighed, and blood was taken at t = +3 h from the deficient-diet animals as described in Experimental protocol 2. Plasma was prepared from the blood samples for plasma amino acid and leptin determination.

View this table:
Table 1.

Components of test diets and rat chow

On the next day, the control animals were tested. Tail blood was taken from those animals at t = −0.5 h, and at t = 0 they were provided with a weighed portion of control diet corresponding to the average amount of food consumed by the animals who had been presented the deficient diet on the previous day. At t = +3 h, additional tail blood was sampled for plasma leptin determination.

Experimental protocol 4.

This experiment measured the effect of a single dose of rapamycin on the meal-associated rise in plasma leptin and ribosomal protein S6 phosphorylation as a measure of the mTOR activation, using the protocol of Anthony et al. (1). After meal training on the control diet from protocol 1 for at ≥12 days, animals were weighed and allocated to one of two groups: vehicle control (307 ± 8 g body wt) or rapamycin (320 ± 8 g) groups. Tail blood was collected from animals at t = −0.5 h before the meal. Animals assigned to the rapamycin diet group were studied on the first day. At t = 0 h, the animals (n = 11) received an intraperitoneal injection of rapamycin (0.75 mg/kg) along with the control diet. On that same day, animals in the vehicle control group received their usual 3-h meal; no blood was taken from them. At t = +3 h, food was removed and weighed for both groups (average of 16.5 g consumed per rat). Blood was taken at t = +3 h from the rapamycin group, and the animals were then anesthetized with Nembutal (100 mg/kg) for laparotomy removal of epididymal adipose tissues. Tissues were frozen between clamps cooled to the temperature of liquid nitrogen and stored at −84°C for subsequent examination of cell-signaling end points. Plasma was prepared and stored frozen with the tissues at −84°C for subsequent plasma leptin and insulin determinations.

On the day that the control animals were tested, some animals were anesthetized for adipose tissue removal at t = −0.5 h (i.e., before the meal). At t = 0 h, the remaining animals were provided with a weighed portion of control diet corresponding to the average amount of food consumed by the animals that had been injected by rapamycin on the previous day. At t = +3 h, blood plasma and adipose tissues were obtained and stored frozen as above. Some meal-trained animals (n = 8) were put into a separate group for examination of S6 phosphorylation in adipose tissue at t = −0.5 h before the meal and treatments.

Plasma measurements.

Active ghrelin was measured in freshly thawed samples close to the time of sampling using an RIA from Linco Research (St. Charles, MO). Plasma leptin was also assayed by RIA (Linco). Insulin concentrations were calculated using a commercial ELISA kit for rat insulin (American Laboratory Products, Windham, NH). Glucose and triglycerides were determined using the Vitros Chemistry System (DT60 II and DTSc II modules; Ortho-Clinical Diagnostics, Rochester, NY). Amino acids were determined by HPLC as previously described (27, 28).

Phosphorylation of substrates in the mTOR-signaling pathway.

Frozen, powdered adipose tissue was homogenized on ice using a Polytron in 3 volumes of a phospho-preserving homogenization buffer. The components of that buffer (in mM) were: 20 HEPES, pH 7.4 (buffer pH adjusted with 1 or 10 mM NaOH); 2 EGTA, pH 7.4; 50 NaF; 100 KCl; 0.2 EDTA, pH 7.4; 50 β-glycerophosphate; 0.1 4-(2-aminoethyl)benzenesulfonyl fluoride HCL (AEBSF), 1 benzamidine, 0.5 sodium vanadate, and 1 M microcystin LR, 0.4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 1% Triton X-100. The homogenate was centrifuged at 10,000 g for 10 min at 4°C. An aliquot of the infranatant was reserved for protein assay, and the rest was added to an equal volume of 2× Laemmli sodium dodecyl sulfate (SDS) sample buffer. The mixtures were boiled for 4 min and centrifuged at 16,000 g for 4 min.

The extent of phosphorylation of the translational repressor 4E-BP1 was determined in Western blots by examining changes in electrophoretic mobility during SDS-PAGE of adipose tissues lysates, as previously described (12), using antibodies from Bethyl Laboratories (Montgomery, TX). The amount of 4E-BP1 in the most phosphorylated form (γ-form) relative to the total immunoreactivity was then determined.

Phosphorylation of eIF4G1 was measured as described previously (47). Briefly, lysates prepared as above were stabilized in SDS-PAGE sample buffer and separated on 7.5% Tris·HCl gels. Western blots from the gels were probed with a 1:1,000 dilution of antibody from Cell Signaling Technology (Beverly, MA) that recognizes eIF4G when it is phosphorylated at Ser1108. Total eIF4G was also determined on parallel blots using a rabbit antibody that recognizes phosphorylated and unphosphorylated forms of eIF4G1 from Bethyl Laboratories as a gel loading control.

Phosphorylation of S6K on Thr389 was examined in immunoprecipitates, as previously described (26), using an antibody that recognizes the Thr389 phosphorylated form of p70S6K according to the manufacturer's protocol (Cell Signaling Technology). Hyperphosphorylation of the 40S ribosomal 31-kDA protein S6 was examined in Western blots from 10% Tris-glycine gels using two antibodies simultaneously, phospho-S6 ribosomal protein (Ser235/236) antibody from Cell Signaling Technology and Ser240/244 antibody from Santa Cruz Biotechnology.

Phosphorylation of eIF2α.

Lysates from adipose tissue were prepared in phospho-preserving buffer, as described above. The lysates were clarified by centrifugation and the infranatant proteins separated by SDS-PAGE using 10% Tris·HCl gels. After transfer to PVDF, eIF2α phosphorylation was assessed as described by Anthony et al. (5).

Statistical analysis.

Data are expressed as means ± SE. To calculate statistical significance (P < 0.05), a two-tailed Student's t-test or one-way ANOVA with or without Sidak posttests was used. All statistical analyses and data manipulations were made using GraphPad Prism (GraphPad Software, San Diego, CA) or Statistica 6.0 software (Statistica, Tulsa, OK).


Effects of meal feeding on plasma leptin concentrations.

Because plasma Leu, alanine, and glucose have been reported to regulate leptin secretion in vitro (e.g., Refs. 9, 21, 32, 33, 46), the first step was to help determine the time course and extent to which these nutrients rise in the peripheral circulation after a meal. Animals were trained for ≥12 days to eat their food as a 3-h meal. Observation of the trained animals under diminished indirect lighting (dark cycle) during the last 15 min of the feeding period indicated that, on the whole, animals were generally consuming food until the end of the meal. On the day of the experiment, animals were weighed (averaging 304 ± 3.9 g) and euthanized either 30 min before the normal meal start time or at various times during and after the 3-h meal.

The plasma concentration of nutrients was measured in plasma before (t = −0.5 h), during (t = +0.5, +1, and +3 h), and after meal feeding (t = +6 and +9 h). The means ± SE of the Ala concentrations were as follows: −0.5 h = 277 ± 59, +0.5 h = 402 ± 93, +1 h = 373 ± 8, +3 h = 349 ± 80, and +9 h = 302 ± 37 μM. ANOVA analysis of these results indicated that there was not a statistically significant change in Ala with respect to sampling time (P = 0.09).

In contrast to lack of changes in Ala, Leu rose rapidly during the meal and stayed elevated as the animals continued meal feeding (Fig. 1). Plasma Leu concentrations measured before the meal (∼120–140 μM) tended to be lower than concentrations in overnight food-deprived animals (∼150–180 mM; Refs. 26 and 27). Leu concentrations declined after removal of the food, reaching prefeeding levels ∼9 h after the start of the meal.

Fig. 1.

Plasma nutrient concentrations in meal-fed animals. Truncal blood was collected from meal-trained animals before, during, and after a meal of rat chow. Plasma from the blood was used to measure nutrient concentrations. Results are means ± SE from 8 determinations. The meal period is indicated by the gray bar. Compared with the −0.5 time point, Leu concentrations were significantly different (P < 0.05, repeated-measures ANOVA with Bonferroni posttest) at t = +0.5, 1, 3, and 6 h after the meal (P < 0.05). Triglycerides were significantly elevated at t = 6 and 9 h relative to the −0.5-h time point.

Figure 1 also shows that plasma glucose concentrations remained relatively constant in response to meal feeding. The glucose concentration was moderated by the rise in plasma insulin that occurred with meal feeding (Fig. 2). Insulin concentrations increased and remained elevated during the meal and then fell. In contrast to leptin and insulin, so-called “active ghrelin” did not change significantly (data not shown). Triglyceride concentrations did not rise significantly until 3 h after the meal and continued to rise out to 9 h after the end of the meal. The delayed rise in triglyceride was not as robust as the changes in leucine (Fig. 1).

Fig. 2.

Plasma insulin concentrations during a meal. Rat plasma insulin concentrations were determined by ELISA. Results are means ± SE from 8 determinations. Gray bar indicates feeding period. Compared with the −0.5-h time point, results were significantly different (P < 0.05, repeated-measures ANOVA with Bonferroni posttest) at t = +0.5, 1, 3, and 6 h after the meal.

Because mTOR is a nutrient-regulated kinase that has been implicated in leptin secretion in vitro (7, 42), a second objective was to measure mTOR-mediated signaling events in adipose tissue in response to meal feeding. Activation of the mTOR-signaling pathway in adipose tissue occurs when Leu concentrations are elevated (26, 28). For example, in food-deprived animals, adipose tissue mTOR is maximally elevated at Leu concentrations at or above ∼550 μM; half-maximal is ∼270 μM (26). Because Leu concentrations rose in the range of ∼400–600 μM during meal feeding, we measured the phosphorylation state of proteins associated with mTOR activation in adipose tissue from the meal-fed animals. Phosphorylation of the translational repressor 4E-BP1 rose at the first time point in the meal and remained elevated during feeding (Fig. 3, top). Phosphorylation of eIF4G followed the same time course as 4E-BP1 (Fig. 3, top), declining with a time course similar to the decline in Leu concentration after the meal. S6K1 phosphorylation also increased to a maximal level 30 min after initiation of meal feeding (Fig. 3, bottom). Multisite phosphorylation of the S6K1 substrate, ribosomal protein S6, was also increased with meal feeding. As expected, the time course of the decline in S6 phosphorylation was slower than that of S6K1.

Fig. 3.

Phosphorylation (p) of substrates of the mammalian target of rapamycin (mTOR) cell-signaling pathway in adipose tissue from meal-trained animals. Adipose tissues lysates were prepared from animals at indicated times before or after initiation of meal feeding. Top: eukaryotic initiation factor (eIF)-binding protein-1 (4E-BP1) and eIF4G phosphorylation. Phosphorylation of 4E-BP1 was measured by determining %total 4E-BP1 in the γ-form. Phosphorylation of eIF4G on Ser1108 was measured after electrophoresis of adipose tissue lysates on 5% Tris-glycine gels and corrected for loading using eIF4G anti-peptide antibody. Bottom: phosphorylation of S6K1 and ribosomal protein (rp)S6. Phosphorylation of S6K1 on Thr389, the site associated with activation, was determined in Western blots using phospho-Thr389-specific antisera and corrected for the loading of total rpS6 kinase-1 (S6K1) using S6K1 COOH-terminal peptide antibody. Multisite phosphorylation of the S6K1 substrate, rpS6, was detected in Western blots using phospho-rpS6 (Ser235/236) antibody. Results are means ± SE and were statistically different (P < 0.05, repeated-measures ANOVA with Bonferroni posttest) from the −0.5 time point at t = +0.5, 1.0, and 3 h.

Last, we sought to find the best time to observe the maximal leptin responses to meal feeding. Figure 4 shows that, compared with the signaling events and plasma changes in insulin and Leu concentrations, the plasma leptin concentration followed a slower time course in response to meal feeding than plasma Leu and phosphorylation of signaling proteins. Thus the rise in the leptin concentration peaked near the end of the 3-h meal, decreasing after the meal. Replicate studies produced similar results. On the basis of these results, it seemed that the 3-h time point was optimal for measuring the leptin response to a meal in subsequent experiments.

Fig. 4.

Time-dependent changes in plasma leptin concentrations associated with meal feeding. Truncal blood was collected from meal-trained animals before, during, and after a meal of rat chow. Plasma from the blood was used to measure leptin concentrations by. Results are means ± SE from 4 determinations. Meal period is indicated by the gray bar. Leptin concentrations were significantly different at t = +3 and +6 h relative to t = −0.5 h (P < 0.05).

Leu and norleucine effects on plasma leptin.

The aim of the second set of experiments was to examine the effect of providing pure Leu or norleucine instead of a meal on leptin secretion. Norleucine administration was studied because it activates mTOR signaling without raising plasma insulin concentration (28). In one experiment, the animals were split into two groups, saline (n = 12) or Leu. In the other experiment, the groups were saline and norleucine (n = 12). All groups received two boluses via oral gavage at t = 0 and +1.5 h. On the basis of previous reports defining the Leu kinetics after a gavage, we were concerned that a single dose was not sufficient to provide a sustained increase in plasma Leu. Thus the dosing scheme was designed to ensure a sustained elevation of plasma Leu that was higher than the values measured during the meal. In these experiments, blood was taken at t = −0.5 and +3.0 h for plasma leptin determinations. Because the data from the saline animals were not significantly different between experiments, they were pooled. Animals that received saline (n = 24) had a trend to a higher 3-h plasma leptin compared with their −0.5-h time point; however, this was not statistically significant (data not shown).

Figure 5 shows that plasma leptin concentrations were significantly greater at the 3-h time point in both the Leu and norleucine groups compared with the saline group. Although plasma leptin rose in response to Leu or norleucine gavage, the observed rises in leptin were not as great as the response to a meal. That is, a complete meal was more efficacious than a Leu gavage at raising the plasma leptin concentration. A possible explanation could be that the Leu gavage did not sufficiently raise the plasma Leu concentration compared with a meal (i.e., not above 600 μM). To evaluate this possibility we measured the plasma Leu and norleucine concentrations at the 3-h time point in the gavaged animals. Leu concentrations were 134 ± 22, 2,144 ± 122, and 98 ± 24 μM for the saline, Leu, and norleucine groups, respectively. The norleucine concentration was 6,778 ± 377 μM in the norleucine group and undetectable in the other groups. Therefore, the smaller rise in the plasma leptin in response to a Leu gavage compared with a meal could not be explained by an insufficient rise in the plasma Leu concentration. In contrast to leptin, insulin concentrations were not elevated by the Leu or norleucine gavage (saline control group, 0.21 ± 0.05 ng/ml; Leu gavage, 0.25 ± 0.05 ng/ml; norleucine gavage, 0.28 ± 0.05 ng/ml).

Fig. 5.

Oral Leu and norleucine administration stimulates leptin secretion in meal-trained animals. After 12 days of meal training using rat chow, animals were provided oral solutions containing saline, Leu suspension, or norleucine suspension instead of their meal. Animals received 2.5 ml/100 g body wt of saline (0.155 mol NaCl/l, n = 24), Leu (54.0 g/l l-leucine, n = 12), or norleucine (54.0 g/l l-norleucine, n = 12) at t = 0 and +1.5 h (relative to the usual meal time). Results are means ± SE of plasma leptin concentration at t = +3 h. *Significantly different compared with saline control.

Effect of purified diets deficient in amino acids and branched-chain amino acids (BCAA).

To further understand the contribution of Leu in food to the rise in plasma leptin concentrations following a meal, we examined the effect of purified amino acid-containing diets. Using purified amino acid diets allowed us to examine the effect of diets deficient in certain amino acids. The strategy was to meal train animals to eat amino acid-purified diets and then to replace those diets with a deficient diet, missing one or more amino acids, for one meal.

In preliminary experiments we trained animals to eat a commercial amino acid diet (Teklad, TD 86529; Table 1). However, this diet did not produce the same magnitude of leptin responses and activation of the mTOR-signaling pathway observed in response to a meal of rat chow (data not shown). We hypothesized that this might be due to the relatively low concentration of Leu (1.11%) and other amino acids in the commercial diet compared with our rat chow (2.04% Leu). Therefore, a custom-purified diet was obtained that had amino acid concentrations approximating those in our rat chow (Table 1). This diet showed comparable rises in plasma leptin concentration at the +3-h time point (data not shown) and was therefore used as a control diet for the deficient diet experiments.

Three separate experiments were conducted to examine the effect of eating a deficient diet on plasma leptin and insulin concentrations. Figure 6 shows the first results of the experiment with the Leu-free diet. The plasma leptin concentrations before the meal were not significantly different between the control and Leu-free diet groups (Fig. 6). However, at t = +3 h, the plasma leptin was significantly lower in the animals given the Leu-deficient diet compared with the control diet. The difference in the leptin concentration between the −0.5- and the +3-h time points was called the meal-associated rise in plasma leptin concentration.

Fig. 6.

Removal of Leu from food decreases the meal-stimulated rise in plasma leptin. Animals were meal-trained on a purified amino acid diet containing 2.04% Leu for ≥12 days. On day 1 of the experiment, animals were assigned to 2 groups: control diet (open bars) or Leu-deficient diet (gray bars) groups. Tail blood was collected from the Leu-deficient group at t = −0.5 and at +3 h, after animals had consumed the Leu-deficient diet. Amount of deficient diet consumed was measured. The next day, blood was collected from control animals at the same times before and after being provided with a measured amount of control diet, equivalent to the average amount of food that the deficient animals had eaten. Leptin concentrations were measured in plasma prepared from the blood sample. Bars represent means ± SE; n = 12. Body weights: control, 311 ± 3 g body wt; Leu-deficient diet, 315 ± 5 g body wt. There was no significant difference between −0.5-h leptin concentrations. *Significant difference between −0.5- and +3-h time points within a group (ANOVA, Bonferroni posttest, P < 0.001); §difference (P < 0.05) between control and Leu-deficient group leptin concentrations at t = 3 h.

The meal-associated rise in plasma leptin for this experiment and the experiments with the other deficient diets is shown in Fig. 7. All of the amino acid-deficient diets reduced plasma leptin. However, it is noted that, compared with the Leu-deficient diet, removing other amino acids from the diets did not lead to a further reduction in the meal-associated rise in plasma leptin concentration compared with the reduction observed with the Leu-deficient diet.

Fig. 7.

Meal-stimulated rises in plasma leptin after complete and deficient diets. Meal-stimulated rises in plasma leptin were measured in a series of experiments such as those described in Fig. 6 legend. Results of those experiments are shown together here and include data derived from Fig 6. Bars show the difference between the leptin measured at t = +3 and −0.5 h. Test diets were deficient in Leu (data from Fig. 6), branched-chain amino acids (BCAA; Leu, Ile, and Val), or all amino acids (AA). Results are means ± SE; n = 12. *Difference between deficient diets compared with corresponding control (neighboring bar). Body weights: Leu-free group and control (see Fig. 6 legend), BCAA-free group (339 ± 6 g) and control (337 ± 4 g), and AA-free (374 ± 8 g) and control (370 ± 3 g) were not significantly different between control and experimental groups with which the deficient diets were compared.

Mammals are able to detect diets deficient in amino acids, and in rodents this can lead to a rapid reduction in intake of the deficient diet (14). Thus, in these experiments, rats fed the deficient diets ate less compared with when they were provided complete diets. For example, when provided with the Leu-free diet, there was a 21% reduction in the amount of food intake compared with a previous day (data not shown). This reduction was comparable to that observed in a previous study (5). Our experiments were designed to overcome this problem as described in materials and methods. That is, after determination of how much of the deficient diet the animals ate, on the next day the control animals received the same weight of food. The amount of deficient diet eaten was not significantly different between the different deficient diets (Leu-free diet: 4.9 ± 0.4 g/100 g body wt; BCAA-free diet: 4.1 ± 0.3 g/100 kg body wt; amino acid-free diet: 4.1 ± 0.3 g/100 kg body wt; P > 0.05).

Because insulin regulates leptin secretion, it was also measured. No differences in plasma insulin concentration were observed in animals fed the different diets at 3 h (Fig. 8). In parallel studies, adipose tissue was taken to measure eIF2α phosphorylation in adipose tissue from animals provided control and Leu-deficient diet; however, no differences in eIF2α phosphorylation were observed (data not shown).

Fig. 8.

Effects of purified AA diet and AA-free diets on plasma insulin concentrations. Insulin concentrations were measured in plasma collected at the +3-h time point in the experiments described in Figs. 6 and 7 legends. Bars indicate means ± SE. None of the results were significantly different.

We also examined the influence of a maximally effective dose of rapamycin on the meal-associated rise in plasma leptin. An intraperitoneal dose of 0.75 mg /kg body wt rapamycin has been previously shown to ablate nutrient activation of mTOR signaling in rats (4). To assess the effectiveness of this dose of rapamycin in our own experiment, we examined Ser235/236 phosphorylation of ribosomal protein S6 (Fig. 9). There was a significant increase seen in S6 phosphorylation associated with meal feeding in the control animals by comparing the −0.5- and +3-h time points. This result is in agreement with the data from our earlier experiment (Fig. 3, bottom). In agreement with the results from Anthony et al. (4), this dose of rapamycin resulted in a complete reduction of ribosomal protein S6 phosphorylation to prefeeding concentrations (Fig. 9).

Fig. 9.

Effect of rapamycin on phosphorylation of rpS6 in adipose tissue from meal-trained animals. Adipose tissue lysates were prepared from animals at indicated times before or after initiation of meal feeding. Rapamycin group was injected ip with 0.75 mg/kg rapamycin at t = 0 h before the meal. Control group animals were injected with vehicle and provided the amount of food eaten by the rapamycin group the day before. Multisite phosphorylation of the S6K1 substrate rpS6 was detected in Western blots of adipose tissue lysates after SDS-PAGE using phospho-rpS6 (Ser235/236) antibody. Top: representative Western blot. Bottom: results are means ± SE from densitometry of films made from Western Blots using ECL; n = 8 tissues/group. Results were statistically different (P < 0.05, ANOVA). Bonferroni posttest indicated that the 3-h control was significantly different from the −0.5-h group (P < 0.05) vs. significant difference between 3-h control and 3-h rapamycin groups (P < 0.05).

We were concerned that the rapamycin might affect food intake. Therefore, on the day when rapamycin animals were studied, we measured the control animals' food intake. The control animals ate 5.8 ± 0.2 g/100 g body wt that day, whereas the rapamycin group ate 5.3 ± 0.6 g/100 g body wt. Although this was not significantly different, it tended to be somewhat lower, so we pair-fed the vehicle control animals the amount of food eaten by the rapamycin group to compare the meal effect on plasma leptin. The meal-associated rise in plasma leptin was significantly reduced by ∼54% in rapamycin-treated compared with vehicle-injected animals (Fig. 10).

Fig. 10.

Effect of rapamycin on meal-stimulated rises in plasma leptin. Plasma was obtained in animals at t = −0.5 h before a meal and at t = +3 h after a meal. Meal-stimulated rises in plasma leptin were calculated by subtracting the concentration at +3-h from the −0.5-h concentration in the same animal. Rapamycin-treated animals received ip injection of rapamycin (0.75 mg/kg) at t = 0 h. Vehicle controls received vehicle only by ip injection at t = 0 h and were pair fed to the rapamycin group. Results are means ± SE expressed as %mean control meal-stimulated rises in plasma leptin (3.4 ± 0.25 ng/ml); n = 11/group. *Statistical difference between control and rapamycin groups, P < 0.05.

The 3-h insulin concentrations were not significantly affected by rapamycin (vehicle control, 4.7 ± 0.3; rapamycin, 5.7 ± 0.7), in agreement with the previous findings (4).


In this report, we have shown that orally administered Leu and norleucine can increase the plasma concentrations of leptin in meal-trained animals and provided evidence that Leu in food influences the rise in the plasma concentration of leptin during a meal. Studying meal-trained animals provided information on the time course of the changes in leptin associated with eating a meal and demonstrated that meal feeding is associated with a measurable and significant increase in the plasma leptin concentration. Our results are consistent with previous studies showing that plasma concentrations of leptin increase after food intake in humans and mice (11, 20, 44). The effects of a meal on leptin secretion are likely due to a combination of hormonal and nutrient regulation of leptin at both the level of ob gene transcription and mRNA translation initiation. Another component may be secretion of preexisting stores of leptin, which appears to be subject to metabolic regulation (e.g., 43).

Plasma insulin concentrations, which are known to regulate leptin secretion (7, 9, 19, 34, 44), rose with meal feeding as expected. Insulin is capable of upregulating both ob mRNA transcription and translation (for review see Ref. 49). However, when short- and long-term regulation of leptin by insulin is examined, it can be concluded that the shorter-term effects of insulin (i.e., hours, as might be expected in response to a meal) are mainly on ob mRNA translation, not transcription (21, 31).

Previous in vitro studies have also posited a role for glucose in leptin synthesis and secretion. It is difficult to completely dissect the independent effects of glucose, because insulin (which clearly exerts effects) is needed for glucose uptake. The effect of glucose on leptin may be related to consequent increases in tissue levels of UDP-N-acetylglucosamine (48); however, alternate mechanisms involving mitochondrial metabolism have also been proposed (9, 33). Despite the interest from in vitro studies in glucose regulation of leptin secretion, in our studies plasma glucose concentration did not change appreciably in response to a meal. Therefore, although the mechanism whereby insulin regulates leptin could conceivably involve glucose, it seems unlikely that glucose is a direct physiological nutrient signal to fat that is independent of insulin, because plasma glucose concentrations do not change appreciably in the peripheral circulation after a meal.

In vitro studies by Roh et al. (42) indicated a role for Leu in leptin synthesis mediated by mTOR. Earlier studies also implicated mTOR in leptin regulation (7). However, adding Leu and norleucine to food in chronic experiments (27) showed a trend for elevated leptin concentrations that were not statistically significant (27). Although those studies were statistically powered for the end points measured, the pilot studies performed for the present report suggested that more animals are needed for leptin measurements (data not shown). Plasma leptin concentrations tend to be more variable from animal to animal than other end points we measured.

To address the potential for Leu to regulate leptin in vivo, we first examined the effect of oral administration of Leu or the Leu mimetic norleucine on leptin secretion in meal-trained rats. The time course from the meal study was used to choose the best time for examining changes in plasma leptin, 3 h. Although norleucine is not as potent as Leu, we used it because it has an advantage over Leu in vivo: norleucine does not affect insulin secretion (36). This might seem unnecessary, as we did not see a rise in Leu in the 3-h blood. Although Leu is a secretagogue, its effects in food-deprived animals are difficult to observe because they are transient (3). Thus, even though a change in insulin is not observed at a certain time point, that does not mean that it cannnot have an influence on insulin signaling during the experiment. Therefore, norleucine can be used to delineate between the primary effects of Leu vs. secondary effects of elevated insulin secondary to increased Leu concentrations. Both amino acids increased plasma leptin; however, the magnitude of the response to leucine gavage was not as large as that observed with a complete meal. We wondered whether the small responses could be explained by a problem in achieving sufficient plasma concentrations of the amino acids. However, we found that the concentration of Leu in the blood of animals provided the Leu gavage was actually higher than that observed during the consumption of a rat chow meal. Thus the small leptin response to a Leu gavage does not seem to be due to inadequate plasma concentrations of Leu. This observation, taken together with results from the deficient diet and rapamycin experiment, discussed below, supports the conclusion that other factors associated with meal feeding must be responsible for the rest of the leptin response to a meal. Nevertheless, these results show that Leu and norleucine can significantly affect plasma leptin concentrations in vivo, consistent with in vitro studies showing that Leu regulates leptin synthesis at the level of translation initiation (42).

Several factors make Leu in an ideal direct nutrient signal to adipose tissue. First, in contrast to plasma glucose and triglyceride concentrations, which did not change appreciably with feeding, Leu concentrations increased severalfold in the peripheral circulation during a meal. Second, Leu regulates the mTOR-signaling pathway in adipocytes in vitro and in vivo (12, 15, 2430, 38, 42). Third, Leu is more efficacious than other amino acids, including valine and isoleucine, at regulating mTOR signaling (25). Indeed, previous studies from our laboratory have indicated that the other BCAA are not likely to be physiological regulators of mTOR signaling in fat. Fourth, Leu is the most abundant amino acid in many dietary proteins. Fifth, in contrast to glucose, the other putative nutrient signal for leptin, Leu cannot be synthesized by mammals. Last, the enzymes for Leu metabolism are highly expressed in fat, where Leu is regulatory (27). This provides a mechanism to terminate the signal from this nutrient.

We showed that, during a meal, plasma Leu achieved concentrations that have previously been shown extensively to activate mTOR signaling in adipose tissue in vivo. For example, maximal stimulation of the mTOR-signaling pathway by Leu in vivo occurs at a mean plasma Leu concentration of ∼540 μM, whereas half-maximal activation occurs at ∼260 μM (22). Indeed, components of the mTOR-signaling pathway were activated in adipose tissue during meal feeding. Although this is consistent with mTOR-mediated translational regulation of leptin, both Leu and insulin probably contribute to this response. Indeed, although removal of Leu caused an ∼40% decrease in the meal-associated rise in plasma leptin concentration, rapamycin caused a 54% decrease. The difference may be attributable to other leptin-regulatory factors that are altered by meal feeding, such as insulin.

To further explore the potential nutrient regulation of leptin by Leu, we examined the effect of amino acid-deficient diets using the experimental paradigm of Anthony et al. (5). Studies by Gietzen's group have shown that when animals are presented with a deficient diet they exhibit a compensatory decrease in food intake associated with eIF2α phosphorylation in the piriform cortex (14). This phenomenon is most readily observed when the animals are adapted to a low-protein diet and can be reduced and overcome by 1) having a diet that is not protein/amino acid deficient, 2) the motivation brought about by meal feeding/training, and 3) correcting for the decrease in eating by pair-feeding the control animals with the amount of food consumed by the deficient diet group.

The leptin response in the different diets showed that the amino acid concentrations achieved in the plasma influenced the leptin response. Thus animals eating the rat chow diet had higher leptin responses than animals eating a commercial, purified amino acid diet (5). By designing a balanced purified diet with higher amino acid concentrations (Custom All Amino Acid diet), leptin responses were similar to those observed with rat chow. Phosphorylation of eIF2α did not seem to be a factor in the different leptin responses.

The results also suggest that Leu is responsible for ∼40% of the meal effect on leptin and may be responsible for nearly all of the effects of nutrient amino acids on leptin secretion in response to meal feeding. This is consistent with our observation that the amplitude of the rise in leptin concentration in response to the Leu gavage is ∼50–60% smaller than the response to a complete meal. The rest of this response to the meal is expected to be due to the meal-associated rises in insulin or other nutrients that may affect leptin secretion.

Compared with a complete meal, a Leu-deficient meal led to a decrease in the meal-associated rise in plasma leptin concentrations. Removing either BCAA or all amino acids from the food did not bring about a further decrease. Insulin concentrations were not significantly different when deficient and complete diets were compared. Thus insulin does not seem to be a confounding factor in our interpretation that Leu seems to be responsible for most of the effects of amino acids on the meal-associated rise in plasma leptin concentration.

Consistent with the hypothesis that Leu is responsible for only a portion of the rapamycin-sensitive response of plasma leptin to a meal, rapamycin treatment caused a greater reduction in the meal-associated rise in the plasma leptin (∼54%) compared with when Leu was removed from the food (∼40%). The effectiveness of the rapamycin treatment was reflected in the ablation of ribosomal protein S6 phosphorylation at 3 h. Thus other factors in the meal, such as insulin, must be responsible for the remainder of the rapamycin-sensitive portion of the rise in plasma leptin in response to a meal. The rapamycin-insensitive portion of the meal-associated rise in plasma leptin must be brought about by other mechanisms such as, but not limited to, release of preexisting stores (41) or rapamycin-insensitive translation.

In conclusion, we have shown that Leu or norleucine can increase the plasma concentration of leptin in vivo. During the consumption of a meal of rat chow, plasma Leu levels rise to concentrations that have been previously shown to activate mTOR in adipose tissue. Indeed, downstream targets of this pathway were affected by meal feeding and blocked by rapamycin, as previously reported (3). The amount of inhibition in response to rapamycin (∼54%) is somewhat greater than the inhibition that occurs when Leu is removed from the food (∼40%). This implies that Leu is not responsible for all of the rapamycin-sensitive effects of a meal on plasma leptin. However, the difference is small, implying that Leu is probably responsible for most of the rapamycin-sensitive response. Our results also indicate that Leu appears to be responsible for most of the effects of nutrient amino acids in the meal on leptin secretion. However, again, amino acids appear to be responsible for only a portion of the meal effect on plasma leptin concentrations. Further studies are required to determine the factors responsible for the rest of this response, but the nutrient-stimulated rise in plasma insulin is posited to be the most likely candidate. The findings reported here, in concert with previous in vitro and in vivo data, continue to support a role of Leu as a direct-acting nutrient signal in adipose tissue.


This project was supported by the following grants from the National Institutes of Health: DK-053843 (C. J. Lynch), DK-062880 (C. J. Lynch and S. M. Hutson), GM-39277 (T. C. Vary), and AA-12814 (T. C. Vary).


  • 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.


View Abstract