20-HETE Induces Hyperglycemia through the cAMP/PKA-PhK-GP Pathway

Abstract

We previously generated cytochrome P450 4F2 (CYP4F2) transgenic mice and showed high 20-hydroxyeicosatetraenoic acid (20-HETE) production, which resulted in an elevation of blood pressure. However, it was unclear whether 20-HETE affected glucose metabolism. We measured fasting plasma glucose, insulin, hepatic CYP4F2 expression, and 20-HETE production by hepatic microsomes, and hepatic 20-HETE levels in transgenic mice. We also assessed glycogen phosphorylase (GP) activity and the cAMP/protein kinase A (PKA)-phosphorylase kinase (PhK)-GP pathway, as well as expressions of insulin receptor substrate 1 and glucose transporters in vivo and in vitro. The transgenic mice had overexpressed hepatic CYP4F2, high hepatic 20-HETE and fasting plasma glucose levels but normal insulin level. The GP activity was increased and the cAMP/PKA-PhK-GP pathway was activated in the transgenic mice compared with wild-type mice. Moreover, these alterations were eliminated with the addition of N-hydroxy-N′-(4-butyl-2 methylphenyl) formamidine, which is a selective 20-HETE inhibitor. The results were further validated in Bel7402 cells. In addition, the transgenic mice had functional insulin signaling, and 20-HETE had no effect on insulin signaling in Bel7402 cells, excluding that the observed hyperglycemia in CYP4F2 transgenic mice resulted from insulin dysfunction, because the target tissues were sensitive to insulin. Our study suggested that 20-HETE can induce hyperglycemia, at least in part, through the cAMP/PKA-PhK-GP pathway but not through the insulin-signaling pathway.

Hyperglycemia is common in essential hypertension and is an independent risk factor of cardiovascular diseases (1, 2). Thus, hypertension and hyperglycemia most likely share substantial overlap in the underlying pathophysiological mechanisms. Although it is generally recognized that insulin resistance plays a major role in the pathogenesis and clinical course of patients with hypertension and diabetes (3, 4), insulin resistance, per se, does not explain why fasting hyperglycemia develops, because the magnitude of insulin resistance varies dramatically in nondiabetic individuals over a narrow range of fasting plasma glucose concentrations. Therefore, an understanding of this process could provide information that may help explain some of the mechanistic links between an increase in blood pressure and fasting hyperglycemia.

The metabolite 20-hydroxyeicosatetraenoic acid (20-HETE) is a modulator of renal ion transport and vascular reactivity and plays an important role in the development of hypertension (5–7). However, some hypertensive animal models have been shown to have elevated 20-HETE levels with accompanying hyperglycemia or hyperinsulin, such as stroke-prone spontaneously hypertensive rats, which exhibit elevated 20-HETE levels with hyperglycemia, and spontaneously hypertensive rats, which exhibit elevated 20-HETE levels with hyperinsulin (5, 8–10). In addition, some diabetic animal models show alterations in 20-HETE. For example, OVE26 type 1 diabetic mice have increased CYP4A-dependent 20-HETE in glomeruli (11), and streptozotocin-induced diabetic rats have in creased CYP4A expression and 20-HETE production in the renal and cardiac microsomes but not in the renal vasculature (12). By contrast, other studies have reported that the expression of CYP4A and production of 20-HETE are decreased in the glomeruli and renal microvessels of streptozotocin-induced diabetic rats (13, 14). In a human study, Tsai et al. (15) demonstrated that plasma and urinary 20-HETE levels were significantly elevated in metabolic syndrome (including hypertension, hyperglycemia, hyperlipemia, and obesity) patients. Therefore, we hypothesized that 20-HETE might be a candidate regulator that participates in the common pathogenesis of hypertension and hyperglycemia.

The human cytochrome P450 4F2 (CYP4F2) gene encodes an ω-hydroxylase that converts arachidonic acid to 20-HETE and is responsible for the majority of 20-HETE synthesis in the human kidney and liver (16). In a previous study, we reported that a gain-in-function mutation of the CYP4F2 gene is associated with elevated urinary 20-HETE levels and hypertension in a Chinese population (17). We also generated a novel CYP4F2 transgenic mouse model that exhibits increased expression and catalytic activity of CYP4F2 and elevated 20-HETE production, which consequently leads to hypertension (18, 19). In this current study, we used the CYP4F2 transgenic model to clarify the relationship between 20-HETE-mediated hypertension and hyperglycemia. We investigated glucose metabolism from the points of hepatic glycogenolysis and insulin signaling in the transgenic mice and tested whether 20-HETE may affect hepatic glycogenolysis and insulin signaling in vivo and vitro.

Materials and Methods

Experimental animals

Experiments were performed on 12- to 16-wk-old male CYP4F2 transgenic mice that weighed 24–33 g. All mice were matched by sex, weight, and age with wild-type mice as controls. Mice were fed with standard mouse chow, provided water ad libitum, and bred under a 12-h light,12-h dark cycle system. All experiments conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication no. 85–23, revised 1996). For 20-HETE inhibition, mice underwent ip with either N-hydroxy-N′-(4-butyl-2 methylphenyl) formamidine (HET0016) (Cayman Chemical Co., Ann Arbor, MI) by 10 μg/g body weight daily or lecithin (Roche Applied Science, Basel, Switzerland) vehicle (10% weight per volume lecithin in saline) for 14 d.

Blood pressure measurements

Blood pressure of the mice was measured with the tail-cuff method using an IITC Life Science Model 1631 tail pulse detection system (IITC Life Science,Woodland Hills, CA). Mice were trained for 5 d, and acclimated mice were warmed at 33 C in a heating container for 5 min before the measurement was taken. One measurement session involved six repetitions, and at least three sessions were performed on each mouse to calculate an average.

Biochemical analysis

Most of the blood samples were gotten by enucleating eyeballs, whereas the partial samples (HET0016 treated 0 or 1 wk) were taken from the cavernous sinus with a capillary under anesthesia after mice were fasted overnight. Plasma glucose and insulin were respectively determined, respectively, by the oxidase-peroxidase method and ELISA (R&D Systems, Minneapolis, MN). Free fatty acid (Applygen Technologies, Inc., Beijing, China), triglyceride, and cholesterol (BHKT Clinical Reagent Co., Beijing, China) concentrations were measured as kits described. The cAMP levels in liver and cells were detected by ELISA (R&D Systems).

Hepatic 20-HETE analysis

Hepatic 20-HETE was measured by API 3200 Q-trap liquid chromatography-tandem mass spectrometry System (Applied Biosystems, Foster City, CA) as previously reported (19). Generally, liver samples were homogenated in methonal (0.1% formic acid), then added 2 ng 20-HETE-d6 (Cayman Chemical) as internal standard. Lipids were extracted with ethyl acetate, dried under nitrogen, and resuspended in methanol. Samples were separated on reversed-phase Symmetry C18 column (3.5 μm, 2.1 × 150 mm; Waters Associates, Milford, MA) at a flow rate of 0.2 ml/min using solvent A (water, 0.1% formic acid) and solvent B (acetonitrile-methanol 6:1, 0.1% formic acid) (0–2 min 25% B, 2–10 min 25–75% B, 10–18 min 75–95% B, 18–30 min 95% B, 30–30.5 min 95–25% B, 30.5–40min 25% B). The effluent was ionized using negative ion electrospray and quantified by multiple reaction monitoring. The ion abundance of 20-HETE in the peaks vs. that of 20-HETE-d6 was determined and compared with standard curves generated over a range from 0.2 to 10 ng.

Arachidonic acid (AA) hydroxylation assay of hepatic microsomes

Hepatic microsomes were prepared according to the method described previously (18). The conversion of AA to 20-HETE was assessed in a reaction mixture of 100 mm potassium phosphate buffer (pH 7.4) containing 3.3 mm MgCl₂, 80 μm AA (Cayman Chemical), 1 mm reduced nicotinamide adenine dinucleotide phosphate (Roche Applied Science, Basel, Switzerland) and 0.6 μg/μl mouse hepatic microsomes. After violent vortex and 5 min preincubation at 37 C, nicotinamide adenine dinucleotide phosphate was added to start the reaction at 37 C for 30 min. The reaction was stopped by acidification with formic acid to pH 3.5, after which the 20-HETE production was measured as described above.

Protein extraction

Cytoplasm and plasmalemma protein were extracted using a kit according to the manufacturer's instructions (Beyotime Institute of Technology, Jiangsu, China). Briefly, tissues or cells were first homogenized. Homogenates were then centrifuged at 20,000 × g for 30 min at 4 C. The supernatant was collected as the cytoplasm protein. The pellet was resuspended, mixed, and then centrifuged at 20,000 × g for 10 min at 4 C. This supernatant was collected as the fraction containing plasmalemma protein. Total protein was prepared by homogenizing the frozen tissues or cells in a lysis buffer containing protease and phosphatase inhibitors. Protein concentration was determined using the Bradford method.

Immunoprecipitation (IP) and Western blot (WB)

Total protein samples were preincubated with the respective antibody at 4 C overnight with rotation followed by the addition of 20 μl equivalents of protein A/G PLUS-Agarose (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and rotated for an additional 2 h. Protein A/G beads were collected and washed with lysis buffer three times. For WB analysis, denatured protein was separated by electrophoresis and transferred onto polyvinylidene fluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA) at 4 C. Membranes were subsequently incubated with primary antibody in 5% bovine specific albumin (BSA) at 4 C overnight followed by horseradish peroxidase-conjugated IgG (KangChen Bio-Tech, Hangzhou, China) as the secondary antibody at room temperature for 1 h. The final detection reaction was performed with an ECL detection kit (Bio-Rad). Densitometry scan was performed using ImageJ software, and the wild-type mice or control group were normalized as one. The antibodies used for IP and WB were as follows: anti-CYP4F2 (Fitzgerald, USA), anti-glycogen phosphorylase (GP), anti-phosphorylase kinase (PhK)β, anti-GLUT1, and anti-ATPase (Proteintech Group, Chicago, IL); anti-insulin receptor substrate (IRS)-1 and anti-phospho-serine/threonine (Cell Signaling Technology, Danvers, MA); anti-phospho-tyrosine (BD Biosciences, Palo Alto, CA); anti-GLUT4 (Santa, Cruz Biotechnology); and anti-glyceraldehyde 3-phosphate dehydrogenase (KangChen Bio-Tech).

GP, PKA, and adenylate cyclase (AC) activity analysis

The activity of GP and PKA was measured using a kit according to the manufacturer's instructions (GenMed, Plymouth, MN). In brief, total protein was added to a buffer containing substrate and then measured at an absorbance of 340 nm once every 60 sec for 5 min. The difference between the absorbance at the fifth minute and the immediate absorbance represented the activity. The AC activity measurement was according to the production of cAMP. Protein was added to a buffer containing 50 mm Tris-HCl, 2 mm MgSO₄, 12.5 mm theophylline, and 10 mm ATP for 5 min at 37 C, after which the reaction was immediately stopped at 100 C, and the cAMP concentration was measured.

Cell culture

The Bel7402 cell line was cultured in RPMI 1640 media using standard cell culture techniques. Once the cells reached 75% confluence in six-well plates, the cells were placed in fetal bovine serum-free media for 4 h and then treated with vehicle (0.1% ethanol) as control, 20-HETE at different concentrations (0.1, 0.5, 1, 2.5, or 10 μm for 1 h; Cayman Chemical), 20-HETE for different times (1 μm for 10, 30, 60, 120, and 240 min), or the PKA-selective inhibitor 10 μm N-[2-(p-Bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide·2 HCl hydrate (H-89) or 100 μm AC inhibitor 2′,3′-dideoxyadenosine (DDA) (Sigma-Aldrich, St. Louis, MO) 30 min before the addition of 20-HETE. The cells were then harvested as described above.

Statistical analysis

Data were expressed as means ± sd. All data were analyzed using SPSSv17.0 software (IBM Corp., Armonk, NY). Student's independent two-tailed test, a one-way ANOVA, or two-way ANOVA followed by S-N-K test were used for statistical analysis. Statistical significance was set at P < 0.05.

Results

CYP4F2 transgenic mice are both hypertensive and hyperglycemic

The systolic blood pressure of CYP4F2 transgenic mice was elevated compared with wild-type mice due to the overproduction of 20-HETE as previously reported (18, 19). As Table 1 showed, fasting plasma glucose was higher in transgenic mice than in wild-type mice (8.4 ± 1.86 vs. 4.9 ± 1.23 mmol/liter). However, there was no difference in plasma insulin between transgenic and wild-type mice (4.9 ± 0.25 vs. 4.8 ± 0.19 mU/liter). The free fatty acid, triglyceride, cholesterol, and body weight also had no difference in the two groups. Therefore, these data indicated that increased 20-HETE production resulted not only in hypertension, but also in hyperglycemia.

Table 1.

CYP4F2 transgenic (TG) mice exhibit both hypertension and hyperglycemia.

	WT	TG	P value
Glucose (mmol/liter)	4.9 ± 1.23	8.4 ± 1.86	1.38E-4a
Insulin (mU/liter)	4.8 ± 0.19	4.9 ± 0.25	0.135
FFA (mmol/liter)	0.4 ± 0.10	0.4 ± 0.11	0.848
Triglyceride (mg/dl)	182.8 ± 34.75	197.3 ± 43.46	0.420
Cholesterol (mg/dl)	164.9 ± 38.63	174.2 ± 49.20	0.643
Body weight (g)	27.2 ± 2.21	27.4 ± 2.23	0.812
SBP (mm Hg)	107.1 ± 7.02	130.2 ± 7.89	1.81E-6a

All parameters except the last two were measured in the plasma. Data were shown as means ±sd (n = 10). FFA, Free fatty acids; SBP, systolic blood pressure; TG, transgenic WT, wild-type.

^aP < 0.01 vs. WT mice.

Hepatic glycogen phosphorylase (GP) activity is increased in transgenic mice

Because the fasting plasma glucose level is mostly maintained by hepatic glycogenolysis in which GP plays a key role (20), we measured GP activity in the livers of the transgenic mice. As showed in Fig. 1, the hepatic GP activity increased by 3.5-fold in transgenic mice compared with wild-type mice. There was also a strong positive correlation between fasting plasma glucose and hepatic GP activity (r = 0.91). Furthermore, CYP4F2 was overexpressed in liver; hepatic 20-HETE level was elevated (10.4 ± 1.57 vs. 6.7 ± 1.37 ng/mg protein), and the production of 20-HETE by hepatic microsome was augmented (77.2 ± 17.50 vs. 52.7 ± 10.18 pg/min/mg mi crosome) in transgenic mice. Based on these findings, we inferred that hepatic GP activity was induced by the increased 20-HETE production in transgenic mice.

Fig. 1.

Hepatic glycogen GP activity is increased in transgenic (TG) mice. A, The hepatic GP activity in TG mice. B, The correlation between fasting plasma glucose and hepatic GP activity. C, Hepatic 20-HETE level. D, The CYP4F2 expression. E, The 20-HETE production by hepatic microsomes in TG mice compared with wild-type (WT) mice (n = 6). *, P < 0.05 vs. WT mice. GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

HET0016 decreases plasma glucose and hepatic GP activity in transgenic mice

To test the hypothesis that 20-HETE increases hepatic GP activity and results in hyperglycemia, for 2 wk we ip injected transgenic mice with HET0016, which is a selective inhibitor of 20-HETE synthesis. As expected, 20-HETE inhibition with HET0016 led to a reduction of plasma glucose, hepatic GP activity, and systolic blood pressure (reduced by 2.9 mmol/liter, 58.1% and 18.5 mmHg, respectively), whereas no significant alteration of plasma glucose, hepatic GP activity, and systolic blood pressure was observed in wild-type mice treated with HET0016 (Fig. 2). These data showed that inhibition of 20-HETE synthesis alleviated the hyperglycemia in CYP4F2 transgenic mice and supported the hypothesis that increased 20-HETE levels can elevate fasting plasma glucose by increasing hepatic GP activity.

Fig. 2.

HET0016 decreases the plasma glucose and hepatic GP activity in transgenic (TG) mice. A, Plasma fasting glucose. B, Hepatic GP activity. C, Hepatic 20-HETE level. D, Systolic blood pressure (SBP) after 2 wk of HET0016 treatment. WT, wild-type (n = 4). *, P < 0.05 vs. WT mice; ^#, P < 0.05 vs. control.

20-HETE activates hepatic GP through the cAMP/PKA-PhK-GP pathway in transgenic mice

It has been shown that GP is activated by phosphorylase kinase (PhK), which is controlled by protein kinase A (PKA) and cAMP. PhK remains inactive until α- or β-subunit is phosphorylated (21). To gain insight into the mechanism whereby 20-HETE activates hepatic GP, we measured AC activity, cAMP levels, PKA activity, and the phosphorylation status of PhKβ and GP in the livers of transgenic and wild-type mice. As shown in Fig. 3A, the AC activity was 1.8-fold increased in transgenic mice compared with wild-type mice, and reduced after HET0016 treatment. cAMP levels were higher in transgenic mice than in wild-type mice (15.9 ± 5.88 vs. 9.1 ± 3.56 pmol/mg protein), but decreased after treatment with HET0016 (11.0 ± 3.70 vs. 15.9 ± 5.88 pmol/mg protein) (Fig. 3B). The PKA activity was 2.2-fold higher in transgenic mice compared with wild-type mice, and reduced by 39.9% after HET0016 treatment (Fig. 3C). IP with anti-PhKβ or anti-GP antibodies followed by Western blot with an antibody specific for serine/threonine phosphorylation demonstrated that the phosphorylation levels of PhKβ and GP were markedly higher in transgenic mice than in wild-type mice (Fig. 3, D and E). Moreover, HET0016 treatment eliminated the increased PhKβ and GP phosphorylation in transgenic mice. These data indicated that 20-HETE can activate hepatic GP through the cAMP/PKA-PhK-GP pathway.

Fig. 3.

The cAMP/PKA-PhK-GP pathway is activated in transgenic (TG) mice and decreases after HET0016 treatment. A, AC activity. B, cAMP level. C, PKA activity. D and E, Serine/threonine phosphorylation (P-Ser) of GP and PhKβ in control and HET0016 treatment groups. WT, wild-type (n = 4). *, P < 0.05 vs. WT mice; ^#, P < 0.05 vs. control.

20-HETE signals through the cAMP/PKA-PhK-GP pathway in Bel7402 cells

We initially determined the optimal 20-HETE concentration (0.1, 0.5, 1, 2.5, and 10 μm for 1 h) and treatment time (10, 30, 60, 120, and 240 min at 1 μm) in Bel7402 cells and found that PKA reached maximum activity in Bel7402 cells with 1 μm of 20-HETE treatment for 1 h (as seen in Fig. 4). At this optimal condition, 20-HETE treatment increased AC activity (elevated by 38%), cAMP levels (37.8 ± 7.28 vs. 20.3 ± 3.98 pmol/mg protein) and GP activity (increased by nearly 50%). Moreover, both PhKβ and GP phosphorylation increased after 20-HETE treatment in Bel7402 cells as determined by IP and WB. The addition of a PKA or AC inhibitor, H-89 or DDA, for 30 min before 20-HETE incubation was found to block 20-HETE action on GP activity as well as PhKβ and GP phosphorylation. These findings confirmed that 20-HETE can activate GP through the cAMP/PKA-PhK-GP pathway in vitro.

Fig. 4.

The cAMP/PKA-PhK-GP pathway is activated by 20-HETE and decreases after H-89 treatment in Bel7402 cells. A and B, PKA activity after different 20-HETE concentration (0.1, 0.5, 1, 2.5, and 10 μm for 1 h) and treatment times (10, 30, 60, 120, and 240 min at 1 μm) were used in Bel7402 cells. C, GP activity in 20-HETE and H-89/DDA-treated cells. D, AC activity and cAMP levels in cells. E and F, Serine/threonine phosphorylation (P-Ser) of PhKβ and GP in 20-HETE and H-89/DDA-treated cells. *, P < 0.05 vs. control; ^#, P < 0.05 vs. 20-HETE treatment. All experiments were independently performed at least three times.

IRS-1 and GLUT expression in transgenic mice

To further address the pathophysiological mechanism of hyperglycemia in CYP4F2 transgenic mice with normal plasma insulin, we investigated insulin function in major tissues. Insulin plays a central role in glucose metabolism through a series of signaling cascades of glucose homeostasis wherein the phosphorylation of insulin receptor substrates (IRS) on tyrosine or serine/threonine residues enhances or inhibits the insulin effect through the translocation of glucose transporters (GLUT) (22–24). Therefore, we detected IRS-1, the most important isoform of IRS, and the insulin-sensitive GLUT4 expression in mouse muscle, fat, and liver tissues. Because GLUT1 is the major isoform of GLUT in liver, we also measured the hepatic GLUT1 expression. As shown in Fig. 5, transgenic mice had much higher levels of tyrosine phosphorylation and lower levels of serine/threonine phosphorylation of the protein in the three tissues tested. In addition, the expression of GLUT was higher in the cytoplasm and plasmalemma of cells from the transgenic mice than wild-type mice. Therefore, we concluded that the target tissues analyzed were responsive to insulin and therefore the hyperglycemia could not result from insulin dysfunction in transgenic mice.

Fig. 5.

The insulin-signaling pathway is not disrupted in transgenic (TG) mice. A, Tyrosine (P-Tyr) and serine/threonine phosphorylation (P-Ser) of IRS-1. B, Plasmalemma (PM) and cytoplasm (CP) GLUT4 expression in muscle, fat, and liver tissues. C, Hepatic GLUT1 expression in TG mice compared with wild-type (WT) mice. Na⁺-K⁺-ATPase α 1 (ATPase) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) are shown as internal control for PM and CP, respectively (n = 4). *, P < 0.05 vs. WT mice.

The effect of 20-HETE on IRS-1 phosphorylation and GLUT expression in Bel7402 cells

We next treated Bel7402 cells with 20-HETE to assess its effect on insulin signaling in vitro. No alteration of IRS-1 phosphorylation was observed after IP and WB analysis with phosphorylation-specific antibodies (Fig. 6). In addition, Western blot analysis showed that the expression of GLUT did not change after 20-HETE treatment. Therefore, the observed effects of 20-HETE in transgenic mice did not occur through insulin signaling in vitro.

Fig. 6.

20-HETE has no effect on insulin signaling in vitro. A, Tyrosine (P-Tyr) and serine/threonine phosphorylation (P-Ser) of IRS-1. B and C, GLUT1 and GLUT4 expression in Bel7402 cells after 20-HETE treatment. Na⁺-K⁺-ATPase α 1 (ATPase) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) are shown as internal control for plasmalemma (PM) and cytoplasm (CP), respectively. All experiments were independently performed at least three times.

Discussion

In this study, we found that an augmentation of 20-HETE production in CYP4F2 transgenic mice contributed not only to hypertension, but also to hyperglycemia. Importantly, we showed, for the first time, that 20-HETE can activate GP through the cAMP/PKA-PhK-GP pathway in vivo and in vitro and not through the insulin-signaling pathway, which consequently leads to hyperglycemia. This finding extends our understanding of the pathophysiological mechanism of hypertension and hyperglycemia comorbidity.

Several studies have indicated that 20-HETE participates in the development of hypertension by regulating vascular and renal tubular functions (5–7). Recently, we established a CYP4F2 transgenic mice model and found that it exhibited a wide CYP4F2 expression spectrum, including high renal expression and moderate extrarenal expression (18, 19). Of note, we showed that the overproduction of renal 20-HETE led to an elevation of systolic blood pressure. However, the function of extrarenal 20-HETE, and hepatic 20-HETE, in particular, was not clear in the transgenic mice model, because native CYP4F2 is predominantly expressed in the kidney and liver (16). Thus, we detected the hepatic expression of CYP4F2, production of 20-HETE by hepatic microsomes, and hepatic 20-HETE level in the transgenic mice. The data demonstrated that overexpressed CYP4F2 and overproduced 20-HETE occur in the liver of transgenic mice. It has been documented that hypertension and hyperglycemia are intertwined and share common pathophysiological features, including complications (1, 2). Therefore, we measured glucose, insulin, and lipid levels in CYP4F2 transgenic mice and found that an isolated fasting hyperglycemia occurred in addition to hypertension. Although no studies to date have shown that elevated 20-HETE production can cause hyperglycemia, an increase in 20-HETE has been observed in diabetic animals and patients with metabolic syndrome (11, 12, 15). To investigate the relationship between 20-HETE and hyperglycemia, a selective inhibitor of 20-HETE synthesis, HET0016, was administered to the CYP4F2 transgenic mice. We made two important observations: 1) that the hyperglycemia in the transgenic mice was resolved after inhibition of 20-HETE synthesis by HET0016, suggesting that 20-HETE may induce hyperglycemia; and 2) that hepatic GP activity was increased in transgenic mice and had a positive correlation with fasting plasma glucose, and the increased GP activity was eliminated by HET0016. Moreover, these in vivo results were confirmed in Bel7402 cells. It is known that hepatic GP plays a key role in glycogenolysis, which maintains fasting plasma glucose levels in a dominant manner (20). Therefore, this is the first finding that 20-HETE can activate hepatic GP, which consequently led to hyperglycemia.

To our knowledge, PhK is the only kinase that can convert GP from the inactive to the active form. PhK is a hexadecamer composed of four copies of four different subunits (αβγδ)_4, in which the γ-subunit provides catalytic activity and the α-, β-, and δ-subunits regulate the activity. PhK remains inactive until α or β is phosphorylated by PKA, which is controlled by cAMP (21, 25). Therefore, we assessed whether 20-HETE can activate hepatic GP through the cAMP/PKA-PhK-GP pathway. We found that compared with wild-type mice, the CYP4F2 transgenic mice had increased AC activity, cAMP levels, PKA activity, and serine/threonine phosphorylation of PhKβ and GP, which was in agreement with the increase in 20-HETE levels. Importantly, all of these parameters decreased after 20-HETE synthesis was inhibited by HET0016. In vitro experiments performed in Bel7402 cells further validated our finding that 20-HETE treatment increased the serine/threonine phosphorylation of PhKβ and GP as well as GP activity. Moreover, treatment with H-89 or DDA, a PKA or AC inhibitor, reversed the 20-HETE-induced increase in GP activity and phosphorylation level. Therefore, 20-HETE indeed activates the cAMP/PKA-PhK-GP pathway. 20-HETE has been considered to be a secondary messenger for the modulation of blood pressure and natriuresis through the activation of protein kinase C, MAPK, src-type tyrosine kinase, and rho kinase pathways (5, 26). To date, very few studies have shown that 20-HETE influences the cAMP/PKA pathway. Li et al. (27) reported that 20-HETE inhibits a cAMP-dependent pathway in the kidney of rats consuming a potassium-deficient diet. By contrast, we demonstrated that 20-HETE can activate the cAMP/PKA pathway in the liver of CYP4F2 transgenic mice as well as in Bel7402 cells. One possible explanation for this discrepancy is that 20-HETE may have differential effects on different tissues: 20-HETE has been reported to be both a dilator of bovine coronary arteries (28) and a potent constrictor of renal, cerebral, and mesenteric arteries (29). Even in the kidney, 20-HETE has been found to decrease cortical blood flow and increase medullary blood flow (30). Therefore, we conclude that 20-HETE can activate hepatic GP, resulting in hyperglycemia. at least in part, through the cAMP/PKA-PhK-GP pathway. However, whether 20-HETE directly/indirectly affects this pathway or via receptor is still unclear and will require additional studies.

Although we found that 20-HETE can elevate the fasting plasma glucose via activating glycogenolysis, we could not exclude its effect on gluconeogenesis pathway. In the fasting state, the plasma glucose is maintained by glycogenolysis and gluconeogenesis, but glycogenolysis accounts for more than 70%, especially fasted within 24 h (20). In addition, Jenssen et al. (31) reported that an increased supply of gluconeogenesis precursors cannot elevate the plasma glucose, because glycogenolysis is decreased in the human study. Similarly, inhibition of gluconeogenesis did not correct hyperglycemia in type 2 diabetic patients (32). In the animal study, when GP is inhibited, both glycogenolysis and gluconeogenesis are decreased and plasma glucose is reduced (33). A similar result was also acknowledged by Fosgerau et al. (34). In our study, we also found that the hyperglycemia in transgenic mice can be recovered when the GP activity was decreased after treatment with HET0016.

In addition, we did not observe any significant change in the plasma insulin in CYP4F2 transgenic mice compared with wild-type mice. It was clearly necessary to address the contribution of insulin action to hyperglycemia in these mice. Insulin action involves a series of signaling cascades that include the binding of insulin to its receptor, the phosphorylation of IRS, and the translocation of GLUT from the cytoplasm to plasmalemma through a protein phosphorylation cascade involving the phosphatidylinositol 3-kinase/AKT pathway (24). IRS-1 is required for the effects of insulin mainly through the translocation of the insulin-sensitive GLUT, which depends on IRS-1-positive tyrosine phosphorylation or -negative serine/threonine phosphorylation (23). We found that transgenic mice had a much higher level of tyrosine phosphorylation and lower level of serine/threonine phosphorylation of IRS-1 in muscle, fat, and liver tissues. Moreover, in the transgenic mice, GLUT4 expression was higher in cytoplasm and plasmalemma of muscle, fat, and liver tissues, and the hepatic GLUT1 expression was increased as well, indicating that the translocation of GLUT from cytoplasm to plasmalemma was sufficient for transporting glucose. These data suggested that the CYP4F2 transgenic mice were sensitive to insulin due to either 20-HETE action on insulin or the feedback of hyperglycemia. Further in vitro analyses showed that there was no difference in IRS-1 phosphorylation levels or GLUT expression in the cytoplasm and plasmalemma of Bel7402 cells treated with 20-HETE compared with controls, suggesting that 20-HETE does not directly enhance insulin signaling. Therefore, the feedback of hyperglycemia in CYP4F2 transgenic mice most likely triggered the insulin effect. Importantly, other studies have shown evidence that supports this hypothesis. Adochio et al. (35) reported that tyrosine phosphorylation of IRS-1 in human muscle is significantly increased in response to a 5-d high-carbohydrate diet, indicating that insulin signaling was elevated in response to a high-carbohydrate load. In addition, Laffer et al. (36) showed that 20-HETE levels were not significantly different between insulin-resistant and insulin-sensitive groups of patients experiencing essential hypertension with obesity, implying that 20-HETE does not have an effect on insulin signaling. However, the question of why hyperglycemia does not promote insulin secretion in CYP4F2 transgenic mice will require further exploration.

The field has recently focused on GP as a new target for diabetes therapy, because hepatic glycogenolysis is increased in diabetic patients and plasma glucose is decreased when a GP inhibitor is administered (20, 33–34). The evidence presented in this study supports the hypothesis that increased GP activity in the liver can result in hyperglycemia. Interestingly, 20-HETE may elevate fasting plasma glucose by increasing hepatic GP activity in addition to elevating blood pressure through alterations in kidney function. Therefore, we propose that 20-HETE may share the common mechanism of hypertension and hyperglycemia and may be a potential new target for hypertension and hyperglycemia therapy.