Chronic leucine supplementation improves glycemic control in etiologically distinct mouse models of obesity and diabetes mellitus

Seven to eight-week-old male NONcNZO10/LtJ (RCS10) and B6.Cg-A ^y /J (A ^y ) mice were purchased from Jackson Laboratories. Animal protocols were in compliance with the accepted standards of animal care, and were approved by the Columbia University Institutional Animal Care and Use Committee. Mice were maintained at 22°C on a 12:12 light-dark cycle (0700-1900), and had ad libitum access to the breeder chow (Purina Mouse Diet 20 5058, 21.8%, 21.6% and 56.6% calories from protein, fat, and carbohydrates, respectively). The breeder chow diet, which contains twice as much fat calories as the regular chow, increases the rate of weight gain in A ^y mice and is necessary for the development of overt diabetes with high frequency in male RCS10 mice (Jackson Laboratory website). Leucine was supplemented via the drinking water containing 1.5% (wt vol^-1) L-leucine (Sigma, St. Louis, MO) as previously described [7, 16]. Since rodents consume the majority of water with meals during the dark cycle, supplementation via drink water should achieve similar effects as supplementation through food. This method of supplementation was chosen mainly out of the consideration of convenience. The controls were sex- and age-matched mice fed the same diet with regular tap water as drinking water.

Body composition was determined using a Minispec TD-NMR Spectrometer (Bruker Optics, TX). HbA1c levels were determined using DCA 2000 Hemoglobin A1c (HbA1c) Reagent Kit (Bayer, HealthCare, LLC, Elkhart, IN). Plasma amino acid concentrations were determined by the Hormone Analytic Core of the Mouse Metabolic Phenotyping Center at the Vanderbilt University. Blood glucose levels were measured in tail-vein blood using Glucometer Elite (Bayer, Elkhart, IN). Plasma insulin levels were determined using an ELISA kit (Mercodia Inc, Winston Salem, NC). Because plasma glucose and insulin levels are regulated by different mechanisms in different feeding states, we measured them in three feeding states to assess the effect of leucine treatment on these mechanisms. The three feeding states are defined as: the fed state - measurements were taken at the 4^th hour of the dark cycle; the basal state - measurements were taken at the 7^th hour of the light cycle after 5 hour food and leucine deprivation; the fast state - measurements were taken after 24 hour food and leucine deprivation. During the food and leucine deprivation, all mice were supplied with the regular tap water to prevent them from dehydration. In order to minimize the effects of feeding manipulation and handling (collecting ~50 ul blood for insulin analysis) on their metabolism, the mice were tested in the three feeding states in the following order, the fed state, the basal state, and the fast state during the testing period after 2, 4 or 8 month leucine treatment. A 2-day resting period was given between testing, and the animals were allowed to recover from the 24 hour fast for at least one week. The interpretations of the data obtained in the different feeding states are described in the text where they are relevant. To further compare the regulation of glucose-insulin homeostasis in response to acute meal ingestion in leucine-treated and control mice, we also conducted a fasting-refeeding experiment in RCS10 mice at the end of 8 month leucine treatment. The mice in both leucine and control groups were deprived of food and leucine (for the leucine group) and supplied with the regular tap water for 24 hours (11 am-11 am). Food was then re-introduced with the regular tap water and 1.5% leucine solution for the control and leucine groups, respectively. The mice were allowed to feed ad lib for 3 hours before bloods were collected for determination of plasma glucose and insulin. Food and water intake during the 3 hour refeeding period were recorded.

Daily food intake and water (or leucine solution) consumption were determined twice a week in individually-housed mice during the first two months of leucine treatment in both RCS10 and A ^y mice. Indirect calorimetry was performed in A ^y mice at the end of 4 month leucine-treatment (LabMaster, TSE Systems, Inc. Chesterfield MO). Oxygen consumption, locomotive activity, respiration exchange ratio (RER, V_CO2/V_O2), and food intake were measured continuously during the same 12:12 light-dark cycles. The same diet and leucine regimens were continued during indirect calorimetry. Data were collected over a 3 day period following 2 days of adaptation to the metabolic cage.

Quantitative real-time RT-PCR was used to determine mRNA expression as previously described [17]. Briefly, total RNA was isolated using RNAeasy mini-columns (Qiagen, CA), and reverse-transcribed into single-stranded cDNA using random hexamers and M-MLV Reverse Transcriptase (Invitrogen, CA). Quantitative amplification by polymerase chain reaction (PCR) was carried out using Bio-Rad iQ SYBR Green Supermix (Bio-Rad Laboratories, CA). All RT-PCR reactions were done in duplicates. Cyclophilin A mRNA level was used to normalize total RNA input. Cycle numbers at a set fluorescence threshold within the exponential amplification range (Ct) were used to calculate expression levels of the genes of interest (relative to cyclophilin A). Changes in gene expression in leucine-treated mice were expressed relatively the control values. Following gene-specific primers were used for mRNA quantification in this study: cyclophilin A, 5'-atggcactggcggcaggtcc-3' (forward), 5'-ttgccattcctggacccaaa-3' (reverse); UCP3, 5'-caagggagcggaccactcc-3' (forward) 5'-ctctctcctccagttcccatg-3' (reverse); CrAT, 5'- tgccagttgagttcctggga-3', 5'- tctgatctgaggtgagcggt-3' (reverse); CPT-1B, 5'-ggcaacagttggttccaactact-3' (forward), 5'-caggaagcttaggcatgtacgtt-3' (reverse); PPAR-alpha, 5'-tcatacatgacatggagaccttg-3' (forward), 5'-actggcagcagtggaagaatc-3' (reverse); NRF-1, 5'-ggagcacttactggagtcc-3' (forward), 5'-ctgtccgatatcctggtggt-3' (reverse); NRF-2, 5'-ggggaacagaacaggaaaca-3' (forward), 5'-ccgtaatgcacggctaagtt-3' (reverse); TBP, 5'-ggcctctcagaagcatcacta-3' (forward), 5'-gccaagccctgagcataa-3' (reverse); aP2, 5'-gacgacaggaaggtgaagagc-3' (forward), 5'-gcctttcataacacattccacc-3' (reverse); leptin, 5'-tgacaccaaaaccctcatca-3' (forward), 5'-agcccaggaatgaagtcca-3' (reverse); MCP-1, 5'-ctgaagccagctctctcttcct-3' (forward), 5'-tccttcttggggtcagcacaga-3' (reverse); TNF-alpha, 5'-ccaccacgctcttctgtcta-3' (forward), 5'-agctgctcctccacttggtg-3' (reverse); F4/80, 5'-ctttggctatgggcttccagtc-3' (forward), 5'-gcaaggaggacagagtttatcgtg-3' (reverse).

Adipose tissue was fixed in 4% paraformaldehyde for 3 days and then paraffin-embedded according to the standard procedure. Tissue sections (7 um) were stained with anti mouse F4/80 antibody (Cat # MF8000, Invitrogen, CA) and then counter-stained with hematoxylin as previously described [18]. The level of macrophage infiltration was assessed by visual examination of the positive staining under the light microscopy.

All data were expressed as mean ± SEM. Differences between leucine-treated and control mice were assessed using t-tests (STATISTICA V6, StatSoft, Tulsa, OK). A 2-tailed p < 0.05 was considered statistically significant. Correlation analysis was used to assess relationships between HbA1C and plasma glucose and insulin levels.

Leucine supplementation was started in 8-week-old RCS10 mice and lasted for 8 months. The average water consumption measured during the first two months was not significantly different between the control (7.8 ± 0.2 ml/day) and leucine group (7.6 ± 0.1 ml/day). The average leucine intake via drinking water was 114.5 ± 15 mg/day, 1.9-fold the daily leucine intake from the chow during this period. Food intake during the first 2 months of treatment was significantly lower in leucine-treated mice (3.80 ± 0.14 g/day), relative to the control mice (4.19 ± 0.16 g/day) (p < 0.05) (Fig 1A). Weight gain during this period was also reduced in leucine-treated mice (10.1 ± 0.66 g vs. 13.1 ± 0.41 g, p < 0.01) (Fig 1B). However, no significant difference in body weight or adiposity was observed between the leucine-treated and control mice at the end of 4 and 8 month treatment (Fig 1B-1C).

HbA1c levels were significantly lower in leucine-treated RCS10 mice relative to the control mice at the end of 2, 4, and 8 month treatment (p < 0.05 at all points) (Fig 1D). At the end of 8-month study period, more than 50% of the control mice developed overt diabetes with HbA1c levels greater than 9%, while none of the leucine-treated mice had HbA1c levels greater than 7.8%. The average HbA1c levels at 8 month were 8.68 ± 0.60% (ranging from 6.2%-10.4%) in the control groups and 6.67 ± 0.41% in the leucine group (ranging from 4.7%-7.8%). Blood glucose levels were also significantly lower or trended lower in leucine-treated RCS10 mice, compared to the control mice in all of the feeding states (Fig 2A). Basal and fast insulin levels were not significantly different at the end of 8 month-study (Fig 2B), but insulin secretion in response to refeeding was significantly more robust in leucine-treated mice than in control mice (Fig 2B). Three-hour refeeding following a 24 hr fast resulted in an average 15.8-fold increase (over the fast level) in plasma insulin levels in leucine-treated mice (ranging from 6.1 to 29.6 fold), but only 6.5-fold increase in control mice (ranging from 2.8 to 10.0 fold) (p < 0.05). The absolute plasma insulin levels after the 3 hour refeeding were also significantly higher in leucine-treated mice (29.5 ± 5.5 vs. 11.0 ± 2.1 ng/ml, p < 0.01). Food intake during the refeeding period was not significantly different between the groups (1.23 ± 0.05 vs. 1.37 ± 0.15 g). Plasma leucine concentration was 38.6% (p < 0.05) higher in leucine-treated mice than in the control mice after the 3 hour refeeding; the concentrations of other amino acids examined were not significantly different between the two groups (Fig 2C). To further examine the relationship between glycemic control and beta cell function in RCS10 mice at the end of 8 month leucine treatment, HbA1c levels were related to plasma glucose and insulin levels after the 3 hour refeeding. HbA1c levels were positively correlated with the fed plasma glucose levels (r = 0.78, p < 0.001) (Fig 2D). The plotting of HbA1c levels against plasma insulin levels (Fig 2E) revealed that beta-cell decompensation was most conspicuous in the mice with HbA1c levels >9.0% (circled data points in Fig 2D and 2E), all of which were in the control group. These results suggest that the improvement in glycemic control in leucine-treated RCS10 mice may be attributable in part to the increased insulin response to feeding and decreased postprandial plasma glucose levels.

Metabolic effects of leucine supplementation were examined in two age groups of A ^y mice. In the first group, the treatment was started at 8 weeks of age and lasted for 4 months. The average food intake (3.99 ± 0.10 vs. 4.21 ± 0.09 g/day) and water consumption (7.3 ± 0.6 vs. 7.5 ± 0.7 ml/day) during the first two months of treatment were not significantly different between leucine-treated A ^y mice and their controls. Body weight was not significantly different between the control and leucine groups after 2 and 4 months of treatment (Fig 3A), nor was the body composition (data not shown). Again, HbA1c levels were significantly lower in leucine-treated mice than in control mice after 2 months (4.8 ± 0.3% vs. 5.7 ± 0.5%, p < 0.05) and 4 months (6.3 ± 0.2% vs. 7.2 ± 0.3%, p < 0.05) (Fig 3B). Effects of leucine supplementation on body weight and HbA1c levels were similar in the second group of A ^y mice (old A ^y ), in which the treatment was started at 5 months of age and lasted for 10 weeks. At the end of the 10-week treatment, body weight was not significantly different between the control and leucine groups (Fig 3C), but the HbA1c level was significantly lower in leucine-treated mice, relative to the control mice (p < 0.05) (Fig 3D). While the HbA1c level rose from 4.2% to 6.4% in the control group during this period, it was virtually unchanged in the leucine-treated group (4.3% to 4.6%).

Fed plasma glucose levels were also significantly lower in leucine-treated A ^y mice than in control mice at the end of 4 month treatment (p < 0.01) (Fig 4A), and were positively correlated with the HbA1c levels in these mice (r = 0.73, p < 0.001) (Fig 4B). Although basal and fast blood glucose levels were not significantly different between the control and leucine groups, the corresponding plasma insulin levels were 38.5% and 33.9% lower, respectively, in leucine-treated mice, relative to the control mice (both p < 0.05) (Fig 4C-4D), suggesting that leucine supplementation may have improved insulin sensitivity in the basal and fast states in these mice. Plasma insulin levels in the fed state did not differ significantly between the control and leucine groups (0.80 ± 0.11 vs. 1.11 ± 0.21 μg ml^-1, p = 0.16) (Fig 4E).

In order to determine the effect of long term leucine supplementation on energy metabolism, indirect calorimetry was performed in the A ^y mice at the end of 4 month leucine-treatment. Rates of oxygen consumption were ~10.7% and 8.9% higher, respectively, in the light and dark cycles in leucine-treated mice, relative to control mice (both p < 0.05) (Fig 5A). No increase in locomotive activity in leucine-treated mice was observed (Fig 5B). The respiratory exchange ratio (RER, V_CO2/V_O2) was slightly lower and food intake was slight higher in leucine-treated mice, compared to the control mice, although the differences were not statistically significant (Fig 5C-5D).

In order to further understand the effect of leucine supplementation on energy metabolism at the molecular level, we next examined the expression level of key genes involved in fatty acid metabolism and mitochondrial function in the skeletal muscle of A^y mice. The mRNA levels for uncoupling protein 3 (UCP3), carnitine acetyltransferase (CrAT), peroxisome proliferators-activated receptor alpha (PPAR-alpha), and nuclear respiratory factor 1 (NRF-1) were significantly higher in the soleus muscle of leucine-treated mice than in control mice (Fig 6). The mRNA levels for carnitine palmitoyltransferase-1B (CPT-1B) and nuclear respiratory factor 2 (NRF-2) were also increased although the differences were not statistically significant. The expression level of TATA-binding protein (TBP) was unaltered. These results suggest that long term leucine treatment may increase energy expenditure by selectively stimulating the expression of a set of key genes involved in fatty acid metabolism and mitochondrial biogenesis in the skeletal muscle.

Obesity and insulin resistance have been associated with adipose tissue inflammation in both humans and rodents [18‐22]. In order to assess the effect of leucine supplementation on adipose tissue inflammation, we examined the expression of several proinflammatory markers and macrophage infiltration in leucine-treated and control A ^y mice. Messenger RNA levels of monocyte chemoattractant protein-1 (MCP-1), tumor necrosis factor-alpha (TNF-alpha), and F4/80, a specific marker of macrophage, were significantly decreased in the epididymal adipose tissue of leucine-treated A ^y mice, relative to the control mice (Fig 7A). In contrast, mRNA levels of two adipocyte-specific genes, aP2 and leptin, were not significantly different between the two groups (Fig 7A). Consistent with the decreased mRNA expression of F4/80 gene, immunostaining of the epididymal adipose tissue with anti-mouse F4/80 antibody confirmed that the degree of macrophage infiltration in the epididymal adipose tissue was also decreased in leucine-treated A ^y mice, relative to the control mice (Fig 7B). These results suggest that long term leucine supplementation may attenuate adipose tissue inflammation associated with obesity.