Remodeling of Hepatic Metabolism and Hyperaminoacidemia in Mice Deficient in Proglucagon-Derived Peptides

The establishment of GCGKO mice and breeding conditions has previously been described in detail (9). All the mice used in the experiments had a genetic background of C57Bl6 J through backcrossing for at least eight generations, except mice used for the microarray analysis, which were offspring of animals backcrossed for three generations. All the animal experimental procedures were performed in accordance with the Nagoya University institutional guidelines for animal care, which conform to the National Institutes of Health animal care guidelines.

Total RNA was extracted from the livers. Microarray analyses were performed using OpArray Mouse V4.0.1 (Operon Biotechnology, Tokyo, Japan) and a GenePix 4000B microarray scanner (Axon Instruments, Union City, CA). To obtain more rigorous quantitative data, gene expression levels were analyzed by quantitative PCR. The sequences of the primers used for the analyses are available upon request. The details of the procedure have been described previously (10).

Liver samples were excised from 12–16-week-old male mice, added to five volumes of methanol, and snap frozen in liquid nitrogen. After homogenization, the water-extractable phase was subjected to capillary electrophoresis electrospray ionization mass spectrometry using Agilent CE-TOFMS systems (Agilent Technologies Japan, Tokyo, Japan). The detected peaks were quantified, and each peak was annotated based on the database of Human Metabolome Technologies (Yamagata, Japan), as previously described in detail (11).

Serum blood chemistry was analyzed using a Biomajesty JCA-BM2250 automatic analyzer (JEOL, Tokyo, Japan). Heparinized blood samples were obtained from 12-week-old male mice after a 5-h starvation. The plasma concentrations of amino acids were determined by liquid chromatography/mass spectrometry (12). The nicotinamide N-methyltransferase (NNMT) activity in the liver homogenates was measured as previously described in detail (13) and expressed as nanomoles of methylnicotinamide produced per hour per gram of liver.

For the pyruvate tolerance test, 1.5 g/kg of sodium pyruvate was administered intraperitoneally to 4-month-old male mice after 18-h fasting. For the glucagon challenge test, 30 μg/kg of glucagon was administered intraperitoneally to mice fed ad libitum. Tail blood was taken at specified times and the glucose levels were determined using a Medisafe glucose meter (Terumo, Tokyo, Japan).

Oxygen consumption, CO₂ production, food intake, and locomotor activity were measured using a Comprehensive Laboratory Animal Monitoring System (CLAMS; Columbus Instruments, Columbus, OH). The mice were allowed to acclimatize to the monitoring environment for 24 h, followed by real-time monitoring of the variables under free access to food for 48 h. The mice were also monitored for 24 h without access to food.

Serum insulin levels were measured in duplicate using High Sensitivity “PLUS” insulin kit (Morinaga-Seikagaku Co. Ltd., Yokohama, Japan) according to the manufacturer’s instructions.

Antibodies against Akt, phospho-Akt, and β-actin were purchased from Cell Signaling Technology Japan (Tokyo, Japan). The secondary antibodies were purchased from Jackson ImmunoResearch (West Grove, PA). The livers were homogenized in ice-cold radioimmunoprecipitation assay buffer (Pierce, Rockford, IL). Aliquots containing 40 μg of protein were separated on 8% SDS-PAGE gels and transferred onto Immobilon-P polyvinylidene fluoride membranes (Millipore, Bedford, MA). Signals were visualized with the Supersignal West Pico Chemiluminescent Substrate (Pierce) and an LAS-1000 lumino-image analyzing system (Fujifilm, Tokyo, Japan).

Data are presented as the mean ± SEM. One-way ANOVA or a two-tailed Student t test was applied to all statistical analyses. P values of <0.05 were regarded as statistically significant.

To evaluate the metabolic status of GCGKO mice, we carried out gene expression and metabolome analyses using capillary electrophoresis mass spectrometry (11) in GCGKO and control livers. The gene expression profiles in the livers of mice starved overnight were analyzed using a two-colored microarray, thus allowing identification of genes differentially expressed between the GCGKO and control livers. The expression levels of such genes were further evaluated by quantitative PCR using cDNA prepared from the livers of GCGKO and control mice under various feeding conditions: fed ad libitum or starved for 5 or 24 h. Although genes encoding rate-limiting enzymes for gluconeogenesis (3) were not identified as differentially expressed genes by the microarray analysis, the expression levels of pyruvate carboxylase (PYCS), phosphoenolpyruvate carboxykinase (PEPCK), and glucose-6-phosphate phosphatase (G6PC) mRNA were also analyzed. As shown in Fig. 1A–C, no significant differences were observed between the control and GCGKO livers for the expression levels of PYCS, PEPCK, and G6PC mRNA. Since there were no differences between the GCGKO and control mice in the concentrations of pyruvate, phosphoenolpyruvate, and glucose-6-phosphate as the substrates for these enzymes (Fig. 1D), it is considered likely that gluconeogenesis from pyruvates is not severely attenuated in the GCGKO liver.

On the other hand, 10 genes expressed at significantly different levels in the GCGKO livers compared with the control livers were identified through the microarray and quantitative PCR analyses (Fig. 1E). These 10 genes comprised six genes involved in amino acid metabolism, namely soluble glutamate oxaloacetate transaminase (GOT1), serine dehydratase (SDS), alanine-glyoxylate aminotransferase (AGXT), argininosuccinate synthetase 1 (ASS1), cystathionine γ-lyase (CTH), and ornithine aminotransferase (OAT); three involved in fatty acid metabolism, namely fibroblast growth factor 21 (FGF21), acetyl-CoA carboxylase α (ACAC), and fatty acid–binding protein 5 (FABP5); and one involved in nicotinamide metabolism, namely NNMT.

Figure 2A–C depicts the expression levels of the genes encoding enzymes involved in amino acid metabolism, mostly the conversion of amino acids to substrates available for gluconeogenesis or the tricarboxylic acid (TCA) cycle. The interactions of the metabolic pathways mediated by these enzymes are summarized in Fig. 2D. SDS converts serine to pyruvate and ammonia (14). AGXT is also involved in the metabolism of pyruvate and serine (15). GOT1 is involved in the malate-aspartate shuttle and thereby in the production of oxaloacetate, which serve as substrates for gluconeogenesis via a PEPCK-mediated reaction (16). SDS, AGXT, and GOT1 mRNA were expressed at significantly lower levels in the GCGKO livers than in the control livers, especially when the mice were fed ad libitum. The expression levels of CTH and tyrosine aminotransferase (TAT) mRNA were also significantly lower in the GCGKO livers. CTH generates cysteine and α-ketobutyrate from cystathionine, which is synthesized from serine and homocysteine (17), and TAT is involved in the conversion of glutamate and α-ketoglutarate (18). ASS1 and arginase (ARG) are involved in the urea cycle, which generates fumarate and urea from l-aspartate and carbamoyl phosphate (19,20). ASS1 produces argininosuccinate from the substrate l-citrulline, and the produced argininosuccinate is then converted to l-arginine and fumarate. The expression levels of ASS1 mRNA were significantly decreased in the GCGKO livers. The expression levels of the mRNA encoding ARG, which converts arginine to ornithine and ammonia, were also lower in the GCGKO livers than in the control livers, although the difference did not reach statistical significance. The expression levels of the mRNA encoding OAT, which is involved in the conversion of α-keto acids and ornithine to l-glutamate 5-semialdehyde and l-amino acids (21), were also decreased in the GCGKO livers.

The above findings suggest that GCGKO mice have defects in the conversion of amino acids to metabolites available for gluconeogenesis. As shown in Fig. 3A, the relative concentrations of 12 of the 20 amino acids that comprise polypeptides were significantly higher in the GCGKO livers than in the control livers. These findings demonstrate that the altered expression levels of the genes depicted in Fig. 2D result in altered amino acid metabolism. The relative concentrations of arginine, ornithine, l-citrulline, and argininosuccinate were also increased in the GCGKO livers, although the difference in the citrulline levels did not reach statistical significance (Fig. 3B). The relative concentration of cystathionine, as the substrate of CTH, was also significantly increased in the GCGKO livers (Fig. 3C).

The plasma amino acid concentrations were subsequently evaluated by high-performance liquid chromatography. It was found that 18 of the 20 amino acids were significantly increased in GCGKO mice, the exceptions being Trp and Phe (Fig. 3D). The plasma concentrations of ornithine, citrulline, urea, and cystathionine were also increased in GCGKO mice (Fig. 3E). Collectively, these findings indicate that the altered metabolism in the GCGKO liver results in high concentrations of amino acids in the plasma. Accordingly, it is suggested that peptides derived from proglucagon are required for utilization of amino acids as a source of energy and for gluconeogenesis.

While the expression levels of genes encoding enzymes involved in amino acid metabolism were lower in the GCGKO livers than in the control livers, the expression levels of three genes involved in fatty acid metabolism, FGF21, ACAC, and FABP5, were higher in the GCGKO livers (see Fig. 1E). Figure 4A–C depicts the expression levels of the genes involved in fatty acid metabolism. Similar to the case for the genes involved in amino acid metabolism, the differences between the GCGKO and control livers in the expression of these genes were greatest in the fed state. ACAC is involved in the production of malonyl-CoA, thereby serving as the limiting step in lipogenesis (22). On the other hand, FGF21 is upregulated by starvation and stimulates lipolysis (23,24). FABP5 is involved in the intracellular transport of fatty acids, which also play an important role in lipolysis (25). These changes in gene expression in the GCGKO livers suggest that both the anabolism and catabolism of fatty acids are increased. Although the serum concentration of triglyceride was decreased, the serum concentrations of HDL, LDL, and total cholesterol were increased in the GCGKO mice (Fig. 4D). Therefore, peptides derived from proglucagon appear to exert both anabolic and catabolic effects on lipid metabolism.

Among the differentially expressed genes, the decreased expression of NNMT was most pronounced in the GCGKO livers (Fig. 5A). NNMT is involved in the methylation of nitrogen in nicotinamide, and thereby in the excretion and disposal of nicotinamide (26). The enzymatic activity of NNMT was significantly lower in the GCGKO livers than in the control livers (Fig. 5B). Figure 5C shows the relative concentrations of nicotinamide and its related metabolites. The mean concentration of N-methylnicotinamide in the GCGKO livers was only 17.6 ± 3.7% of that in the control livers, whereas the mean concentration of nicotinamide was 75.7 ± 7.7% of that in the control livers (Fig. 5C). Therefore, it is suggested that methylation of nicotinamide is decreased in the GCGKO liver. On the other hand, the concentrations of NAD⁺, NADP⁺, and NADPH were significantly increased, suggesting that the catabolism and excretion of nicotinamide are decreased in GCGKO mice.

Upon pyruvate administration, the blood glucose levels were significantly increased in both the GCGKO and control mice, although the increases at 30 and 60 min after the pyruvate injection were significantly greater in the control mice than in the GCGKO mice (Fig. 6A). The differences between the GCGKO and control mice are likely caused by the absence or presence of pyruvate-induced glucagon secretion (27). Nevertheless, since pyruvate also induces gluconeogenesis through mechanisms that are independent of glucagon action, these results indicate that glucagon-independent gluconeogenesis does take place in GCGKO mice.

The effects of glucagon administration were also examined (Fig. 6B). At 15 min after glucagon administration, the blood glucose levels in the GCGKO mice were significantly lower than those in the control mice. The expression levels of the genes depicted in Fig. 1E were not significantly increased by glucagon injection in either the GCGKO or control mice, whereas the expression levels of the PEPCK and G6PC genes were significantly increased in both the GCGKO and control mice (data not shown). These findings indicate that gluconeogenesis is partially attenuated in GCGKO mice compared with control mice, and suggest that both the lack of proglucagon-derived peptides and the altered gene expression pattern in the liver are involved in the attenuated gluconeogenesis.

Because the combined transcriptome/metabolome analyses suggested impaired utilization of amino acids and altered lipid metabolism, the aspects of whether oxygen consumption and the respiratory exchange ratio (RER) are altered in GCGKO mice were investigated. However, there were no differences in the RERs between the GCGKO and control mice, both with ad libitum feeding and after starvation for 24 h (Fig. 7A). Although the GCGKO mice were heavier than the control mice, no significant difference was observed in the oxygen consumption regardless of the difference in body weight (Fig. 7B) or gross oxygen consumption (data not shown). In addition, there were no significant differences in food intake or locomotor activity (Fig. 7C and D). Collectively, there was no substantial difference in energy consumption between GCGKO and control mice, despite the altered metabolic status of the former.

As blood glucose levels in the GCGKO mice are not significantly different despite lack of glucagon, we evaluated plasma insulin levels in GCGKO mice fed ad libitum. As shown in Fig. 8A, plasma insulin levels were significantly lower in GCGKO than in control mice. We also evaluated phosphorylation of Akt at Ser473 by immunoblot analyses. In concordance with lower serum insulin levels, phosphorylation of Akt was significantly lower in GCGKO than in control mice (Fig. 8B). These findings suggest that the absence of proglucagon-derived peptides is compensated for, at least in part, by lower serum insulin levels.

In the current study, we analyzed the metabolic status of the liver in GCGKO mice, which lack most, if not all, proglucagon-derived peptides, including glucagon and GLP-1. Although the expression of the rate-limiting enzymes for gluconeogenesis at the mRNA level was only marginally altered, the expression of genes involved in the metabolism of amino acids, lipids, and nicotinamides was dysregulated in GCGKO mice.

Amino acids serve as the major source for gluconeogenesis through the removal of amino bases. Oxaloacetate and α-ketoglutarate serve as recipients for amino bases in such reactions and are converted to aspartate and glutamate, respectively. On the other hand, SDS converts serine into pyruvate and ammonia, which is then converted into carbamoyl phosphate. Aspartate, together with carbamoyl phosphate, is converted into urea and fumarate through the urea cycle (see Fig. 2D). Our analyses of the gene expression patterns suggested that the expression of a number of enzymes involved in these procedures was suppressed in the GCGKO liver, and that utilization of amino acids as a source for gluconeogenesis and energy was attenuated. In accordance with our deductions based on the altered gene expression pattern, the concentrations of amino acids were increased in both the liver and plasma of GCGKO mice. Therefore, we concluded that proglucagon-derived peptides play pivotal roles in the regulation of amino acid metabolism. Administration of glucagon or exendin-4, a GLP-1 analog, did not result in increased expression of the genes involved in amino acid metabolism in either the control or GCGKO mice (data not shown). Accordingly, the remodeling of amino acid metabolism is a chronic, rather than an acute, effect of the absence of proglucagon-derived peptides, and mechanisms such as epigenetic regulation (28) are possibly involved in such remodeling.

Taken alone, the current study does not provide answers to the question of which peptide or peptides derived from proglucagon are responsible for the metabolic remodeling in GCGKO mice. However, several reports have supported the pivotal importance of glucagon among the proglucagon-derived peptides in the regulation of amino acid metabolism. Early studies demonstrated that injection of glucagon reduces the plasma amino acid concentrations in healthy humans, whereas suppression of glucagon secretion by injection of somatostatin increases these concentrations (29,30). It is also known that hypoaminoacidemia is often observed in glucagonoma syndrome (31), and that necrolytic migratory erythema can be successfully treated by infusion of amino acids combined with fatty acids (32,33). Therefore, it is reasonable to consider that the remodeling of amino acid metabolism in GCGKO mice is mainly caused by the absence of glucagon. While the current manuscript was in preparation, it was reported that the concentration of amino acids is increased in the liver of glucagon receptor knockout mice (34). Taken together, among the proglucagon-derived peptides, glucagon appears to play the most important role in the regulation of amino acid metabolism, and the combination of transcriptome and metabolome analyses in the current study has clarified the mechanism for the remodeling of amino acid metabolism evoked by the absence of glucagon.

Although amino acid metabolism was greatly altered in GCGKO mice, their glucose metabolism was only marginally altered. We previously showed that the fasting blood glucose levels in adult GCGKO mice are comparable to those in control mice (9). The mild glycemic phenotype of GCGKO mice is markedly different from those of Pcsk2^−/− and Gcgr^−/− mice, which display 20–40% lower blood glucose levels than control mice (4,5). The lower blood glucose levels in the latter models are most likely caused by increased GLP-1 production, which may lower these levels through both insulin-dependent and insulin-independent mechanisms (35). Concordant with the blood glucose levels, the expression levels of genes encoding rate-limiting enzymes for gluconeogenesis did not differ significantly between the GCGKO and control mice (Fig. 1A–D). Analyses of such genes in Pcsk2^−/− and Gcgr^−/− mice should further clarify the differential glycemic phenotypes among the animal models deficient in glucagon action.

We previously showed that the serum insulin concentrations in GCGKO mice fed ad libitum are significantly lower than those in control mice (9), and this was reproduced in the current study (Fig. 8A). Recently it has been reported that gluconeogenesis during prolonged fasting is sustained by glucagon-independent mechanisms that involve SIRT1 protein deacetylase and Foxo1 transcription (3,36). For such mechanisms, suppression of insulin signaling due to lower serum insulin levels, but not activation of glucagon signaling, might be required. Therefore, the expression of gluconeogenetic genes in the GCGKO liver may be sustained by lower serum insulin levels. In concordance with this assumption, the phosphorylation of Akt was decreased in the GCGKO livers compared with the control livers (Fig. 8B).

The regulatory mechanisms for amino acid metabolism have been far less explored than those for glucose metabolism. On the other hand, the roles of amino acids in the regulation of metabolism, as well as the intracellular signaling pathways involved in such regulation, have recently been investigated. Complexes containing mammalian target of rapamycin respond to diverse extracellular and intracellular cues for promoting anabolic, and inhibiting catabolic, cellular processes. Nutrients, including amino acids, are one of these cues, and amino acids have been shown to regulate the phosphorylation status of p70S6K, S6K1, and eukaryotic initiation factor 4E binding protein through activation of mammalian target of rapamycin complexes 1 and 2 (37–40). Therefore, in GCGKO mice, the increased amino acid concentrations in the liver and plasma, as a consequence of the attenuated utilization for gluconeogenesis and energy production, might promote anabolic processes in GCGKO mice. Although it remains unclear whether protein synthesis is increased in GCGKO mice, it is noteworthy that GCGKO mice display increased body weight, which is proportional to their increased body length (9).

In the current study, we have clarified through combined transcriptome and metabolome analyses that amino acid, but not glucose, metabolism is remodeled in the GCGKO liver. We also found that the expression of genes involved in the metabolism of fatty acids and nicotinamide is altered in the GCGKO liver. However, in the current study, the physiological relevance of the changes in these metabolic pathways could not be explored to the same extent as the changes in amino acid metabolism. It has not been clarified how the metabolism of fatty acids is altered in the GCGKO liver, because the genes involved in fatty acid oxidization and fatty acid synthesis were both expressed at higher levels. Because it has recently been reported that the oxidization of fatty acids is attenuated, whereas their synthesis is increased, in Gcgr^−/− mice (41), GCGKO mice might harbor similar alterations in fatty acid metabolism. Although no significant differences in the serum concentrations of triglyceride were observed, the serum concentration of cholesterol was increased in GCGKO mice. This increase is most likely caused by reduced recruitment of LDL cholesterol receptor on the surface of hepatocytes (42).

Decreased expression of the NNMT gene was the most prominent change in gene expression in the GCGKO liver. Decreased enzymatic activity of NNMT and altered concentrations of related metabolites were also confirmed. Recently, the involvement of nicotinamide in the regulation of various metabolic pathways has been highlighted, because nicotinamide inhibits the enzymatic activity of sirtuins, protein deacetylases/ADP-ribosyltransferases involved in the regulation of many biological processes (43). Therefore, it is interesting to hypothesize that reduced NNMT activity and/or altered nicotinamide metabolism modifies metabolic status by regulating the activity of sirtuins. Accordingly, we analyzed the expression levels of SIRT1 in the liver, but did not detect any significant difference between GCGKO and control mice in either the mRNA or protein levels (data not shown). Interestingly, suppressed expression of NNMT mRNA has been observed in microarray data sets from the liver of mice fed a ketogenic (44) or high fat (45) diet. These findings suggest that NNMT expression is regulated by nutritional status. However, the metabolic status of GCGKO mice appears to be different from that of mice fed a ketogenic or high fat diet, because the RER in GCGKO mice did not differ from that in control mice (Fig. 7). Both the regulatory mechanism for NNMT mRNA expression and the effects of differential expression of NNMT mRNA on metabolic status remain to be elucidated.

In summary, our results have shown a pivotal role for proglucagon-derived peptides in the regulation of not only amino acid metabolism but also fatty acid and nicotinamide metabolism. The results further indicated that the effect of absent glucagon action on glucose metabolism may be compensated by lower serum insulin concentration in GCGKO mice. GCGKO mice should serve as a unique animal model for analyzing the interactions of various metabolic pathways, including those of glucose, amino acids, lipids, and nicotinamide.

This work was supported in part by grants-in-aid to Y.H. from the Takeda Science Foundation, Daiko Foundation, Suzuken Memorial Foundation, Novo Nordisk Growth Award, and Danone Institute for Nutrition and Health; a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (21659232); and a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology-Japan (23122507).

No potential conflicts of interest relevant to this article were reported.

C.W. researched data, contributed to discussion, and wrote the manuscript. Y.S., A.F., N.O., T.F., and K.S. researched data and contributed to discussion. H.M., M.Y., Y.T., and J.S. researched data. Y.O. and Y.M. contributed to discussion. Y.H. researched data, contributed to discussion, reviewed and edited the manuscript, and is the guarantor of this article.

The authors thank the staff of the Animal Care Unit in the Futuristic Environmental Simulation Center at the Research Institute of Environmental Medicine (Nagoya University) and the Technical Center of Nagoya University for their help with the animal experiments. The authors also thank the staff of the Japan Mouse Clinic at the BioResource Center (RIKEN Tsukuba, Ibaraki, Japan) and Human Metabolome Technologies for their help with analytical procedures.