Histidine Augments the Suppression of Hepatic Glucose Production by Central Insulin Action

Experiments using mice were conducted in accordance with the guidelines for the care and use of laboratory animals of Kanazawa University. Male mice (8–10 weeks old) were housed under a 12-h light-dark cycle with free access to food and water. Wild-type C57BL/6J Slc mice were purchased from Japan SLC (Shizuoka, Japan), and histamine H₁ receptor knockout (Hrh1KO) (16) mice were from Oriental Bio Service (Kyoto, Japan). We generated LST3KO mice by crossing albumin-cre and floxed-STAT3 mice (5), while NIRKO mice were generated by crossing Nestin-cre and floxed–insulin receptor mice (17). Floxed-STAT3 and floxed–insulin receptor mice were used as the corresponding controls.

Chicken meat extract was prepared as a protein source for gavage administration. Shredded chicken meat was cooked in a pressure cooker for 270 min and strained, the lipids were removed by centrifugation, and the protein concentration was adjusted. Snakehead fish meat was processed in the same way and used as a low-histidine (Low-His) protein source that did not increase blood histidine levels. After 16 h of fasting, the mice were orally administered protein at 2 g/kg body wt via a feeding tube. Histidine or each amino acid was administered at 0.5 mmol/kg body wt i.p. or intravenously at the concentrations shown in Figs. 2A and 5G after a 16-h fast.

A glucose, pyruvate, or fructose tolerance test was performed after a 16-h fast. For the glucose tolerance test, histidine was injected intraperitoneally at 5 min before injection of 2 g/kg body wt glucose (Otsuka Pharmaceutical, Tokushima, Japan). For the pyruvate or fructose tolerance test, histidine was injected twice at 120 and 5 min before 1 g/kg body wt sodium pyruvate or fructose, respectively (Sigma-Aldrich, St. Louis, MO). Blood glucose levels were measured using a GLUCOCARD G+ Meter (Arkray, Kyoto, Japan). Blood insulin and IL-6 levels were measured using a Mouse Insulin ELISA kit (Shibayagi, Gunma, Japan) or a Quantikine Mouse IL-6 ELISA kit (R&D Systems, Minneapolis, MN). Plasma levels of amino acids were quantified using high-performance liquid chromatography. Hypothalamic histamine levels were measured using a Histamine EIA kit (Bertin Pharma, York, U.K.). Hepatic glycogen levels were measured by the anthrone–sulfuric acid method.

Hyperinsulinemic-euglycemic clamping was performed by injecting the awake and unrestrained mice with human insulin (Eli Lilly, Indianapolis, IN) after a 16-h fast. Blood glucose levels were measured every 10 min and maintained between 90 and 120 mg/dL after intravenous insulin injections were started. Further, between 90 and 120 min after starting insulin administration, we stabilized blood glucose volatility with 20 mg/dL insulin (Supplementary Table 1). We measured the glucose infusion rate (GIR) and plasma [³H]glucose specific activity every 10 min. The R_d under steady-state conditions for plasma glucose concentration was determined from the rate of [³H]glucose infusion divided by the plasma [³H]glucose specific activity. The rate of HGP was obtained from the difference between R_d and GIR. An internal cannula (Plastics One, Roanoke, VA) was inserted into the lateral ventricle, followed by intravenous cannulation 7–10 days later. After 4–6 days of recovery and habituation, a hyperinsulinemic-euglycemic clamp or pancreatic clamp was performed with ICV administration. After a 16-h fast, the mice received an ICV administration of human insulin, histidine, or pyrilamine with an intravenous injection of human insulin and somatostatin, as shown in Fig. 7A. Blood glucose levels were maintained between 90 and 120 mg/dL. The mice received 10 nmol/mouse LY294002 (Calbiochem, San Diego, CA), 50 μg/mouse pyrilamine (Wako, Osaka, Japan), 50 nmol/mouse thioperamide (Sigma), pyridylethylamine (TOCRIS Bioscience, Ellisville, MO), or artificial cerebrospinal fluid via the lateral ventricle. For the hyperinsulinemic-euglycemic clamp, pyrilamine was administered to the lateral ventricle (Fig. 5G).

An IL-6 neutralizing antibody and control IgG (R&D Systems) were injected as previously described (9). Clodronate (LKT Laboratories, St. Paul, MN) was encapsulated into empty liposomes (NOF, Tokyo, Japan), and clodronate-containing liposomes (CLDs) were injected into the mice via the tail vein for two consecutive days to deplete Kupffer cells in the liver. The control mice received empty liposomes. Immunostaining with an anti-Mac-2 antibody (Cedarlane Laboratories, Burlington, Ontario, Canada) was performed to visualize the Kupffer cells.

Mouse hepatocytes were isolated as previously described (6). Immunoblotting was performed using anti–phosphorylated (phospho)-STAT3 (Tyr705), anti–phospho-Akt (Thr308), anti–phospho-p70 S6 kinase (Thr389), anti–p70 S6 kinase, anti–phospho-S6 ribosomal protein (Ser235/236), and anti-S6 ribosomal protein antibodies from Cell Signaling Technology (Danvers, MA); anti-STAT3, anti-Akt, anti–G6Pase-α, anti-PEPCK, and anti–insulin receptor β antibodies from Santa Cruz Biotechnology (Santa Cruz, CA); and an anti–β-actin antibody from Sigma. Immunoblot images were representative of at least three independent immunoblot analyses and quantified by densitometry on an LAS-3000 Imager (Fujifilm, Tokyo, Japan).

The results of quantitative PCR were analyzed using the 36B4 gene as an internal control. The primer sequences used in this study are available upon request.

Statistical analysis was performed using Student t test and one- or two-way ANOVA followed by post hoc tests, and differences were considered significant for P values <0.05.

Phosphorylation of hepatic Akt and STAT3 was increased at 120 min after the oral administration of the protein extract (Fig. 1A). As in our previous study (9), glucose loading activated hepatic Akt and STAT3, suggesting that both the hepatic and central actions of insulin were induced by glucose loading. In addition, protein administration also augmented the activation of these molecules by glucose loading (Fig. 1A). The intake of proteins is known to facilitate the secretion of insulin (18); in fact, the continuous administration of somatostatin, which suppresses the secretion of endogenous insulin, kept Akt phosphorylation at a virtually undetectable level after ingestion of the protein extract (Fig. 1B), whereas protein intake increased STAT3 phosphorylation even in the presence of somatostatin (Fig. 1B). These findings suggest that hepatic STAT3 activation associated with protein intake is independent of the secretion of endogenous insulin. Because blood amino acid levels are increased after the oral administration of proteins (19), we administered individual amino acids intraperitoneally to investigate the effects of blood amino acid levels on hepatic STAT3 activation. We found that the hepatic STAT3 phosphorylation was increased by the administration of histidine and histidine-related amino acids and dipeptides, such as 1-methyl-histidine, β-alanyl-histidine (carnosine), and β-alanyl-1-methyl-histidine (anserine) (Fig. 1C). Therefore, we investigated the importance of histidine and histidine-containing dipeptides in the activation of hepatic STAT3 using a protein and Low-His protein extract. Coincident with the plasma concentration of histidine and other histidine-containing dipeptides (Supplementary Table 2), a milder increase in STAT3 phosphorylation was observed in the Low-His group compared with the protein group (Fig. 1D).

Because histidine administration activates hepatic STAT3, we then conducted a glucose tolerance test to investigate the role of histidine in glucose homeostasis. With the intraperitoneal administration of 500 μmol/kg body wt histidine, which maintained the blood concentration of histidine close to that obtained after the oral administration of 2 g/kg body wt protein (Supplementary Table 3), blood glucose levels were significantly decreased at 120 min after glucose loading (Fig. 1E). Conversely, there was no change in plasma insulin levels after histidine administration (Fig. 1F). Phosphorylation of hepatic STAT3 increased at 120 min after histidine administration, which was further increased by glucose loading (Fig. 1G). While the intraperitoneal administration of histidine alone did not significantly change blood glucose levels, it significantly downregulated the expression of the G6pc gene, which encodes G6Pase, with and without glucose loading (Fig. 1H). However, the expression of the Pck1 gene, which encodes PEPCK, did not change significantly after the intraperitoneal administration of histidine alone or with glucose. The administration of fructose and pyruvate is known to increase blood glucose levels mildly through gluconeogenesis (20,21); however, histidine injection attenuated the increase in blood glucose levels after the administration of fructose (Fig. 1I) and pyruvate (Fig. 1J).

We used a hyperinsulinemic-euglycemic clamp technique to investigate the effect of histidine on HGP (Fig. 2A). The GIR and the suppression rate of HGP obtained after insulin administration were increased by histidine administration (Fig. 2B). The suppression of hepatic gene expression of Pck1 and G6pc by insulin was augmented by histidine (Fig. 2C), even though histidine did not affect plasma insulin levels in the clamp tests (Supplementary Fig. 1). While PEPCK protein expression did not change significantly during the 120-min hyperinsulinemic-euglycemic clamping period, insulin administration reduced the levels of G6Pase protein, and this reduction was augmented by histidine administration (Fig. 2D). Phosphorylation of STAT3, but not Akt, was observed at 120 min after insulin administration (Fig. 2E), and histidine enhanced the phosphorylation of STAT3 caused by hyperinsulinemia (Fig. 2E). To evaluate the histidine effect of HGP with fasting levels of insulin, we performed euglycemic-pancreatic clamp tests (Fig. 2F). No significant changes in GIR or HGP resulted from the presence or absence of histidine injection (Fig. 2G).

We investigated the significance of hepatic STAT3 in the suppression of HGP after histidine administration using 8- to 10-week-old LST3KO mice, which had not yet developed insulin resistance (5). In a glucose tolerance test with histidine administration, blood glucose levels were significantly decreased in the control, but not LST3KO, mice at 120 min after glucose loading (Fig. 3A), even though no significant difference in plasma insulin levels was observed between the groups (Fig. 3B). While G6pc gene expression was downregulated at 120 min after the intraperitoneal administration of histidine in the control mice, no such downregulation was observed in the LST3KO mice (Fig. 3C). In the hyperinsulinemic-euglycemic clamp tests, histidine increased the GIR and the suppression of HGP caused by insulin administration in the control animals but not in the LST3KO mice (Fig. 3D). Insulin activates hepatic STAT3 via the central nervous system and the subsequent upregulation of Il6 gene expression in the liver (9). Indeed, not only insulin but also histidine failed to activate STAT3 in primary hepatocytes (Fig. 3E). Histidine also had no effect on the Akt and S6 kinase pathway (Fig. 3E). Therefore, we used an IL-6 neutralizing antibody to investigate the role of IL-6 in the histidine-mediated activation of hepatic STAT3. The administration of the anti–IL-6 antibody prior to histidine administration attenuated the histidine-induced phosphorylation of STAT3 (Fig. 3F) as well as the suppression of G6pc gene expression (Fig. 3G). Histidine administration did not significantly change the plasma levels of IL-6 (Fig. 3H) but significantly upregulated Il6 gene expression in the liver (Fig. 3I). These results suggest that, similar to insulin, histidine activates hepatic STAT3 through the upregulation of hepatic IL-6.

In the liver, Kupffer cells play an important role in the production of IL-6 (22,23); therefore, we depleted Kupffer cells by the intravenous injection of CLD and investigated the function of histidine (22). CLD administration resulted in the disappearance of Mac-2–positive cells (Fig. 4A) and significantly downregulated the expression of the Emr1 gene (Fig. 4B), which is specific to Kupffer cells in the liver. In addition, the depletion of Kupffer cells diminished the expression of hepatic Il6 (Fig. 4C). Histidine administration, which significantly downregulates the expression of the hepatic G6pc gene, did not change its expression after CLD administration (Fig. 4D). In the glucose tolerance test, histidine administration significantly reduced blood glucose levels at 120 min after glucose loading; however, this reduction was diminished after CLD administration (Fig. 4E). In a hyperinsulinemic-euglycemic clamp experiment, histidine augmented the effects of insulin on Pck1 and G6pc gene expression, HGP, and the GIR, but the augmentation of insulin action by histidine was diminished by CLD administration (Fig. 4F and G). Although histidine augmented the STAT3 phosphorylation caused by hyperinsulinemia, this augmentation was attenuated by CLD administration (Fig. 4H). These results indicate that hepatic IL-6 expression in Kupffer cells is essential for the histidine-mediated regulation of hepatic glucose metabolism.

Histidine is a precursor of histamine; therefore, the levels of histamine in the hypothalamus are reportedly increased after the systemic administration of histidine (24,25). Thus, we investigated the involvement of central histamine in the histidine-mediated activation of hepatic STAT3 and regulation of HGP. As reported previously, histidine administration increased the levels of histamine in the hypothalamus (Fig. 5A). In addition, the ICV administration of histamine induced the phosphorylation of hepatic STAT3 (Fig. 5B) and upregulated the expression of Il6 in the liver (Fig. 5C). The lateral ventricular administration of the histamine H₁ receptor antagonist pyrilamine prior to systemic histidine administration attenuated the phosphorylation of hepatic STAT3 (Fig. 5D), whereas the ICV administration of the histamine H3 receptor antagonist thioperamide increased STAT3 phosphorylation (Fig. 5E). Conversely, the ICV administration of the histamine H₁ receptor agonist pyridylethylamine activated hepatic STAT3 (Fig. 5F). To elucidate the role of central histamine/histamine H₁ receptor in the histidine-mediated regulation of HGP, we performed a hyperinsulinemic-euglycemic clamp study with the concurrent ICV administration of pyrilamine and intravenous administration of histidine (Fig. 5G). Although histidine administration augmented the action of insulin on Pck1 and G6pc gene expression, HGP, and the GIR, the ICV administration of pyrilamine blocked this augmentation (Fig. 5H and I).

We administered histidine into the ventricle to investigate the involvement of the central-mediated regulation of HGP by histidine. As with intraperitoneal injection, the ICV administration of histidine increased the hypothalamic histamine content (Fig. 6A). In addition, histidine ICV administration upregulated the phosphorylation of hepatic STAT3 (Fig. 6B) and suppressed G6pc gene expression in the liver (Fig. 6C). We then investigated the effect of histidine ICV administration on the insulin-dependent suppression of HGP (Fig. 6D). In hyperinsulinemic-euglycemic clamping, the ICV administration of histidine increased the GIR and augmented the suppression of HGP (Fig. 6E). The ICV administration of histidine suppressed the expression of the G6pc gene and augmented the suppression of gene expression by the systemic administration of insulin (Fig. 6F), whereas Pck1 gene expression tended to decrease, albeit insignificantly, after histidine administration (Fig. 6F). Furthermore, histidine ICV administration increased the phosphorylation of hepatic STAT3 and additively enhanced the hepatic phosphorylation of STAT3 associated with hyperinsulinemia (Fig. 6G). These results indicate that the suppression of HGP by the systemic administration of insulin is augmented by the central administration of histidine.

We then injected insulin and histidine into the ventricle to investigate the effect of central histidine on the regulation of HGP by central insulin action (Fig. 7A). The ICV administration of insulin was performed with the concurrent administration of somatostatin to block the secretion of endogenous insulin caused by insulin ICV administration (Fig. 7A). As reported previously, the ICV administration of insulin suppressed HGP, while the ICV administration of histidine augmented this suppression (Fig. 7B). In addition, with the ICV administration of insulin the expression of the Pck1 gene decreased, albeit insignificantly, whereas the expression of the G6pc gene was significantly downregulated (Fig. 7C). Histidine administration augmented the suppression of G6pc gene expression (Fig. 7C). Hepatic STAT3 was phosphorylated after the ICV administration of insulin or histidine, and their coadministration enhanced its phosphorylation (Fig. 7D). These results indicate that the central administration of histidine augments the suppression of HGP by central insulin action.

Next, we investigated the cross-talk between central histidine and central insulin action, both of which activate hepatic STAT3. As pyrilamine administration inhibited hepatic STAT3 phosphorylation mediated by the systemic administration of histidine (Fig. 5D), the phosphorylation of hepatic STAT3 was attenuated by the administration of pyrilamine prior to the ICV administration of histidine (Fig. 8A). Conversely, the phosphorylation of hepatic STAT3 induced by the ICV administration of insulin was not altered by the preadministration of pyrilamine (Fig. 8B). Furthermore, in a study using Hrh1KO mice (16) the phosphorylation of hepatic STAT3 was significantly attenuated after the ICV administration of histidine (Fig. 8C) in the same manner as that after pyrilamine administration. However, no significant change in hepatic STAT3 phosphorylation was induced by the ICV administration of insulin to the Hrh1KO mice (Fig. 8D).

Insulin-dependent phosphorylation of hypothalamic Akt was attenuated by the ICV administration of the PI3-K inhibitor LY294002 (Fig. 8E). We therefore investigated the effect of the ICV administration of histidine or insulin, with the concurrent administration of LY294002, on the phosphorylation of hepatic STAT3. LY294002 inhibited the insulin-induced phosphorylation of hepatic STAT3 but did not affect its phosphorylation by histidine (Fig. 8F). We also examined mice lacking central insulin receptors. In the NIRKO mice, as in the control mice, the ICV administration of histidine induced hepatic STAT3 phosphorylation (Fig. 8G). These results suggest that central histidine action mediated by the histamine H₁ receptor is independent of central insulin action but that they function additively to suppress HGP.

The central nervous system monitors hormonal and nutritional changes that regulate energy metabolism in peripheral tissues. In this study, we revealed that an increase in blood histidine levels acts on the central nervous system, activates hepatic IL-6/STAT3 signaling, suppresses the expression of gluconeogenic genes in the liver, and augments the insulin-dependent suppression of HGP. These findings suggest that histidine can be a target molecule for the treatment of type 2 diabetes. Recently, plasma histidine levels were reported to show a significant inverse association with fasting and 2-h blood glucose levels during glucose tolerance tests in individuals without diabetes or newly diagnosed male patients with type 2 diabetes (14). Histidine was also shown to improve the hyperglycemia induced by the central administration of 2-deoxyglucose in rodents (26); however, the mechanism by which histidine lowers blood glucose levels is not clear. In the current study, the intraperitoneal administration of histidine induced the phosphorylation of hepatic STAT3 and reduced blood glucose levels associated with glucose loading. Consistent with a previous study showing that histidine is not involved in the secretion of insulin (27), histidine did not significantly affect plasma insulin levels in the current study. In the hyperinsulinemic-euglycemic clamp experiment, the intravenous administration of histidine increased GIR and the suppression of HGP. Such histidine action was not observed in LST3KO mice or Kupffer cell–depleted mice, suggesting that the activation of hepatic IL-6/STAT3 signaling plays an important role in the histidine-mediated suppression of HGP. The gene expression of G6pc and Pck1 was downregulated by histidine administration in the hyperinsulinemic-euglycemic clamp experiment, with a clear reduction in the protein levels of G6Pase but not PEPCK. These findings suggest that the downregulation of G6Pase expression causes the suppression of HGP by histidine. However, the activation of hepatic STAT3 by histidine alone is relatively weak, resulting in the mild suppression of G6pc gene expression but not of Pck1. Furthermore, euglycemic-pancreatic clamp tests with histidine did not reveal a significant change in HGP. These findings may explain why histidine administration alone is not enough to affect the homeostasis of blood glucose levels.

In the brain, histidine is converted into histamine by histidine decarboxylase in the tuberomammillary nucleus of the posterior hypothalamus (24,25,28). Consistent with a previous study (25), hypothalamic histamine content was increased by histidine administration in the current study. A previous study using rats showed that enhanced central histamine action due to histidine administration suppresses food intake (25). In contrast, both Hrh1KO mice (29) and histidine decarboxylase knockout mice (30) present with insulin resistance as well as obesity. In addition, the development of obesity and diabetes caused by the antipsychotic drug olanzapine in humans is reportedly linked to inhibition of the histamine H₁ receptor (31). Such phenotypes may indicate that the action of histamine affects not only food intake but also glucose metabolism. In this study, the central administration of histamine H₁ receptor antagonist attenuated the activation of hepatic STAT3 and the suppression of HGP induced by the systemic administration of histidine. Furthermore, the central administration of histidine also activated hepatic STAT3 while suppressing HGP. These findings indicate that the histidine-induced suppression of HGP, as well as the suppression of food intake, is mediated by central histamine action. Because the histamine-containing neurons in the tuberomammillary nucleus project into most of the brain (28), the precise mechanism in the brain by which central histamine activates hepatic STAT3 remains unclear. This histidine action is additive to the insulin-induced suppression of HGP. The activation of hepatic STAT3 by central insulin action was not blocked by the central administration of a histamine H₁ receptor antagonist or in Hrh1KO mice. In contrast, the activation of hepatic STAT3 by central histidine was not blocked by the central administration of a PI3-K inhibitor or in NIRKO mice, suggesting that the mechanism underlying the action of central histidine is independent of insulin action.

This study demonstrates that the activation of hepatic IL-6/STAT3 signaling pathways and downregulation of hepatic gluconeogenic gene expression by histidine are mediated by central histamine action involving the histamine H₁ receptor. Central histidine action, which has a central nervous system mechanism that is independent of insulin, augments the insulin-dependent suppression of glucose production in the liver. This study indicates that the central mechanism by which histidine regulates hepatic glucose metabolism and activates hepatic STAT3 is a potential target for the treatment of type 2 diabetic patients with elevated gluconeogenesis.

This work was supported by the Program for Promotion of Basic and Applied Research for Innovations in Bio-oriented Industry to H.I.; Ministry of Education, Culture, Sports, Science and Technology MEXT KAKENHI Grant 23126509 to H.I.; and Japan Society for the Promotion of Science KAKENHI grants 23300274 to H.I. and 24790919 to K.K.

ThinkSCIENCE, Inc. (Tokyo, Japan), assisted with manuscript preparation. No other potential conflicts of interest relevant to this article were reported.

K.K., Y.N., Y.I., S.As., T.M., H.W., A.M., and F.I. produced the data. M.M., T.O., and H.N. produced the data and contributed to the discussion. Y.K., C.M., K.T., S.Ak., and S.K. contributed to the discussion. M.K. contributed to the discussion and reviewed and edited the manuscript. H.I. researched the data, designed the study, contributed to the discussion, and wrote the manuscript. H.I. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank C.R. Kahn (Harvard Medical School) for NIRKO mice; H. Nakamura (Kanazawa University) for expert advice on the statistical analysis; and Y. Anraku, M. Nishio, and C. Asahi (Kanazawa University) for technical assistance.