Wnt Signaling Regulates Blood Pressure by Downregulating a GSK-3β–Mediated Pathway to Enhance Insulin Signaling in the Central Nervous System

The nucleus tractus solitarii (NTS), which is located in the dorsal medulla of the brainstem, is the primary site of blood pressure (BP) and sympathetic nerve activity modulation. The NTS participates in cardiovascular, gastric, and gustatory regulation. Our previous studies revealed that several neuromodulators are involved in BP control of the NTS, including insulin (1,2) and nitric oxide (NO) (3).

Insulin is primarily produced by pancreatic β-cells. Margolis and Altszuler (4) reported that insulin can cross the blood-brain barrier, as demonstrated by sensitive and specific radioimmunoassay results. However, insulin mRNA and immunoreactive insulin are present in the rat central nervous system (5,6). The insulin receptor (IR) is a tetrameric glycoprotein that belongs to the receptor tyrosine kinase superfamily. Insulin can bind to the IR, leading to the autophosphorylation of the β-subunit of the IR at tyrosine residues, which induces the tyrosine phosphorylation of IR substrates (IRSs) (7). IRs/IRSs can activate the following two major signaling pathways: the phosphatidylinositol 3-kinase (PI3K)-Akt pathway, and the Ras-mitogen-activated protein kinase pathway. Insulin has also been implicated in the regulation of the baroreceptor reflex in the NTS (8,9). We previously found (1,2) that insulin is involved in the control of central BP via the PI3K-Akt-NO synthase (NOS) signaling pathway in the NTS.

Wnt proteins are a family of secreted glycoproteins that bind to the receptor Frizzled (Fzd) and to coreceptors referred to as LDL receptor–related protein (LRP) 5 and LRP6 (10). The canonical Wnt pathway involves the activation of Fzd receptors and LRP5/6, which results in the stabilization of β-catenin and importation into the nucleus, where it acts as a cofactor of T-cell factor (TCF)/lymphoid enhancer factor family transcription factors (11). Dickkopf (DKK) molecules have been revealed to function as specific antagonists of canonical Wnt signaling by directly binding to LRP5/6 (12). Wnt signaling regulates various biological activities, including cancer (13,14), stem cell maintenance (15), cardiac hypertrophy (16), and neural differentiation (17,18). In addition, recent clinical and in vitro studies revealed that aberrant Wnt signaling appears to play an important role in the onset of hypertension and diabetes (19–23). Nevertheless, the relationships among Wnt signaling and the insulin pathway and the modulation of BP in the central nervous system have yet to be established.

In the current study, we investigated whether Wnt stimulation promotes central insulin signaling in the NTS to modulate BP. We clarified whether aberrant Wnt signaling causes a neuronal insulin signaling defect that induces hypertension. In addition, we investigated which molecular mechanisms are essential for the fructose-induced hypertension-mediated dysfunction of the canonical Wnt signaling pathway in the NTS. Overall, our results suggest that this neuronal insulin signaling defect is a core mechanism that induces hypertension and that activation of the Wnt signaling pathway may alleviate this defect.

All experimental drugs were purchased from Sigma-Aldrich (St. Louis, MO), unless otherwise noted.

Twenty-week-old male WKY rats and spontaneously hypertensive rats (SHRs) were obtained from the National Science Council Animal Facility and were housed in the animal room of Kaohsiung Veterans General Hospital (Kaohsiung, Republic of China). Humane treatment was administered at all times. The rats were maintained in individual cages in a room in which the lighting was controlled (12 h on/12 h off), and the temperature was maintained between 23°C and 24°C. The rats were acclimatized to the housing conditions for 1 week and were then trained for 1 week to acclimate the animals to the procedure of indirect BP measurement. The rats were randomly assigned to nine groups of six rats per group, as follows: 1) the SHR group, in which the SHRs received an intracerebroventricular injection of artificial cerebrospinal fluid (aCSF) as a vehicle control; 2) the SHR + Wnt group, in which the SHRs received an intracerebroventricular injection of Wnt; 3) the SHR + Wnt + DKK1 group, in which the SHRs received an intracerebroventricular injection of Wnt and DKK1; 4) the WKY group, in which the WKY rats received an intracerebroventricular injection of aCSF as a vehicle control; 5) the WKY + Wnt group, in which the WKY rats received an intracerebroventricular injection of Wnt; 6) the WKY + Wnt + DKK1 group, in which the WKY rats received an intracerebroventricular injection of Wnt and DKK1; 7) the WKY group, in which the WKY rats received an intracerebroventricular injection of aCSF as a vehicle control; 8) the Fructose group, in which the WKY rats were provided with 10% fructose water for 2 weeks; and 9) the Fructose + Wnt group, in which the fructose-fed WKY rats received an intracerebroventricular injection of Wnt. The rats were provided with normal rat chow (Purina, St. Louis, MO) and tap water ad libitum. All animal research protocols were approved by the Research Animal Facility Committee of Kaohsiung Veterans General Hospital.

Intracerebroventricular infusion experiments were performed following a stabilization period of at least 30 min after the insertion of the microinjector into the ventricular-guided cannula. The BP was monitored for 3 days after drug infusion. As a vehicle control, the effect of intracerebroventricular injection of aCSF (142 mmol/L NaCl, 5 mmol/L KCl, 10 mmol/L glucose, and 10 mmol/L HEPES, pH 7.4) was analyzed. Wnt3a (0.9 pmol/day) and DKK1 (1 μg/day) were dissolved in aCSF; the GSK-3β inhibitor (TWS119, 0.173 μg/day) was initially dissolved in DMSO and then diluted in aCSF at a final concentration of 1% DMSO. The basal BP was examined prior to injection. The daily intracerebroventricular drug infusions were performed over a 2-min period and delivered as a single bolus with a final volume of 5 μL from day 0 to day 14. Wnt3a and the inhibitors were injected simultaneously.

Using a previously described tail-cuff method (Model MK-2000 Storage Pressure Meter; Muromachi Kikai, Tokyo, Japan), the systolic BP (SBP) and the heart rate were measured prior to Wnt3a or DKK1 treatment (day 0). The animals were placed in the chamber for 30 min. In this method, the reappearance of pulsation on a digital display of the BP cuff was detected using a pressure transducer and was amplified and recorded as the SBP. During the measurement, a series of 10 individual readings was rapidly obtained. The highest and lowest readings were dropped from consideration, and the remaining eight readings were averaged.

Total protein was prepared by homogenizing the NTS tissue in lysis buffer and was deproteinized using Microcon YM-30 centrifugal filter units (Millipore, Bedford, MA). The total amount of NO in the samples was determined using a modified procedure based on the purge system of a Sievers Nitric Oxide Analyzer (NOA 280i; Sievers Instruments, Boulder, CO), which involves the use of chemiluminescence (24). The samples (10 μL) were injected into a reflux column containing 0.1 mol/L VCl₃ in 1 mol/L HCl at 90°C to reduce any nitrates and nitrites to NO. Then, the NO reacted with the O₃ produced by the analyzer to form NO₂. The resulting emission from the excited NO₂ was detected using a photomultiplier tube and was digitally recorded (in millivolts). Then, the emission values were interpolated to a concurrently determined standard concentration curve of NaNO₃. The measurements were performed in triplicate for each sample. The measured NO levels were corrected for the volume of the examined rats. The NTS insulin levels were measured using an Ultrasensitive Rat Insulin ELISA kit (Mercodia, Uppsala, Sweden) and detected with a Biochrom Anthos Zenyth 200rt Microplate Reader (Biochrom, Cambridge, U.K.).

The cells were transfected with TOPflash or FOPflash Wnt reporter plasmids (Millipore) containing wild-type or mutant TCF DNA binding sites by using Lipofectamine 2000. The cells were also cotransfected with β-galactosidase reporter plasmids. The reporter activity was analyzed as described above.

Total protein was prepared by homogenizing the NTS tissue in lysis buffer containing a protease inhibitor cocktail and a phosphatase inhibitor cocktail and subsequently incubating the sample for 1 h at 4°C. The protein extracts (20 μg/sample; Pierce BCA Protein Assay; Life Technologies, Rockford, IL) were resolved on a 6% polyacrylamide gel and transferred to a polyvinylidene fluoride membrane (GE Healthcare, Buckinghamshire, U.K.). The membranes were incubated in the appropriate anti–phosphorylated (P)-LRP6^S1490 (2568), anti–P-Akt^S473 (4060), anti–P-ribosomal protein S6 kinase (RSK)^T359/S363 (9344), anti–P-extracellular signal–related kinase (ERK)1/2^T202/Y204 (4370), anti-LRP6 (2560), anti-Akt (9272), anti-RSK (9341), anti-ERK1/2 (137F5), anti–P-IRS1^S332 (2580) (Cell Signaling Technology, Beverly, MA); anti–P-GSK-3β^S9 (05–643), anti–GSK-3β (07–389), anti-IRS1 (05–784R), anti–unphosphorylated (Un-P)-β-catenin (05–665), anti–β-catenin (AB19022) (EMD Millipore, Billerica, MA); and anti–P-β-catenin^Ser33 (SC-16743) and anti-Dvl1 (Santa Cruz Biotechnology, Dallas, TX) antibodies. Then, the membranes were then incubated in an horseradish peroxidase (HRP)-labeled goat anti-rabbit secondary antibody at 1:10,000. The membranes were developed using the ECL Plus detection kit (GE Healthcare).

The sections were deparaffinized, quenched in 3% H₂O₂/methanol, heated (using a microwave) in citrate buffer (10 mmol/L, pH 6.0), blocked in 5% goat serum, and incubated in the anti–P-LRP6^S1490 antibody overnight at 4°C. Next, the sections were incubated in a biotinylated secondary antibody (1:200; Vector Laboratories, Burlingame, CA) for 1 h and in AB complex (1:100) for 30 min at room temperature. The sections were visualized using a DAB Substrate Kit (Vector Laboratories) and were counterstained with hematoxylin. Then, the sections were photographed using a microscope equipped with a charge-coupled device camera.

The Catch and Release Reversible Immunoprecipitation System (Millipore) was used according to the manufacturer instructions. The proteins were eluted in 70 μL elution buffer and subjected to immunoblotting analysis using the anti–GSK-3β and anti–P-IRS1^S332 antibodies.

For the in vitro kinase assay, coimmunoprecipitation using the anti–GSK-3β and anti-IRS1 antibodies was performed, and the GSK-3β-IRS1 complex was eluted using the Catch and Release Reversible Immunoprecipitation System. The kinase reaction was initiated by adding kinase buffer and the terminated by adding ×2 sample buffer and boiling for 10 min. The phosphorylation of IRS1 was determined via immunoblotting analysis using the anti–P-IRS1^S332 and anti-IRS1 antibodies.

The brainstem sections were incubated in the anti–GSK-3β (1:200) and anti-IRS1^S332 (1:50) antibodies. After washing with PBS, the sections were incubated in green fluorescent Alexa Fluor 488 anti-rabbit IgG (1:200; Invitrogen) at 37°C for 2 h. The sections were analyzed using a confocal microscope and LSM Image software (Carl Zeiss MicroImaging).

An unpaired Student t test was used to compare protein levels (SHR and WKY groups, WKY and fructose-treated groups, SHR and SHR + TWS119 groups, or fructose and fructose + TWS119 groups) and the BP measurements (SHR and SHR + TWS119 groups or fructose and fructose + TWS119 groups). One-way ANOVA followed by Scheffé post hoc analysis were applied to compare the differences between groups. Differences in which P < 0.05 were considered to be significant. All data are expressed as the mean ± SEM.

To assess whether Wnt signaling is defective in hypertension, we determined the expression patterns of several molecules involved in Wnt signaling in genetic (SHR) and dietary (fructose feeding–induced) models of hypertension. Our results revealed decreased LRP6 and GSK-3β phosphorylation levels in the NTS of SHRs and fructose-fed rats. The protein expression levels of Dvl1 and the unphosphorylated form of β-catenin were also reduced in the SHRs and fructose-fed rats; however, the phosphorylated levels of β-catenin^Ser33 were induced in the SHRs and fructose-fed rats (Fig. 1A–C). Figure 1D–F shows that there were fewer P-LRP6^S1490–positive cells in the NTS of SHRs and fructose-fed rats. Then, we examined whether central administration of Wnt3a and/or DKK1, a negative regulator of Wnt signaling, affects hypertension and found that SHRs that were intracerebroventricularly infused with Wnt3a exhibited a depressor response. Coadministration of Wnt3a and DKK1 blocked the Wnt3a-induced depressor effect and NO release in the NTS of SHRs (Fig. 1G and J). In contrast, Fig. 1H and K shows that intracerebroventricular infusion of DKK1 increased BP and decreased NO production in the NTS of WKY rats. In the fructose-fed–induced hypertension model, the increase in BP was attenuated by intracerebroventricular infusion of Wnt3a for 2 weeks (Fig. 1I). In addition, NO release in the NTS was found to be restored following treatment with Wnt3a for 2 weeks (Fig. 1L). As shown in Fig. 1, elevating the levels of Wnt3a in the brain resulted in an improvement of hypertension. The reported results of the assay and immunofluorescence staining (Fig. 2A–C) demonstrate that the level of nuclear β-catenin and its mediated transcriptional activity significantly increased (active) in the NTS of SHRs. In Fig. 2A and B, results from the confocal microscopy experiments showed that active β-catenin–positive cells (labeled with green fluorescence) were scarcely detected in the NTSs of SHRs. (Fig. 2A and B, lane 1). Strikingly, the administration of Wnt3a resulted in an increase in active β-catenin immunofluorescence (Fig. 2A and B, lane 2). Moreover, coadministration of DKK1 and Wnt3a did not restore the active β-catenin levels in nuclei (Fig. 2A and B, lane 3). The data from the TOPflash/FOPflash Wnt reporter assay showed that treatment with Wnt3a can significantly increase the β-catenin–mediated transcriptional activity in the NTSs of SHRs compared with the control group (SHRs) (Fig. 2C, lane 2). Intracerebroventricular infusion of both DKK1 and Wnt3a can reduce the Wnt3a-induced increase in the β-catenin–mediated transcriptional activity (Fig. 2C, lane 3). Next, we investigated the expression pattern of Wnt signaling factors in the NTSs of vehicle-, Wnt3a-, and DKK1-infused rats. Treatment with Wnt3a significantly increased the phosphorylation levels of LRP6 and GSK-3β, the protein expression levels of Dvl1, and the unphosphorylated levels of β-catenin; however, the phosphorylation levels of β-catenin did not change in the NTSs of SHRs. Intracerebroventricular infusion of both DKK1 and Wnt3a reduced the Wnt3a-induced increase in the phosphorylation and expression levels of LRP6, GSK-3β, and Dvl1 and the unphosphorylated levels of β-catenin, but the phosphorylation levels of β-catenin^Ser33 were reversed (Fig. 2D and Supplementary Fig. 2A). As shown in Fig. 2E, the simultaneous injection of DKK1 and Wnt3a reduced canonical Wnt signaling activity in the NTS of normotensive WKY rats. Interestingly, fructose administration resulted in a significant decrease in the serine phosphorylation levels of LRP6 and GSK-3β and in the expression levels of Dvl1 and unphosphorylated levels of β-catenin; however, there was a significant increase in the serine phosphorylation levels of β-catenin. Moreover, the inhibitory effect of fructose on Wnt pathway activity was enhanced following Wnt3a infusion in the brain (Fig. 2F). These results suggest that the activation of Wnt signaling may be required for NO release and BP regulation in the NTS.

We found that rats exhibit a depressor response following microinjection of insulin into the NTS (1). In addition, the activation of Wnt signaling leads to an upregulation of the insulin pathway (23). Therefore, to determine whether Wnt signaling interferes with the insulin pathway in the NTS, we injected Wnt3a into the brain. Wnt3a administration and coadministration of DKK1 and Wnt3a did not affect the phosphorylation of IRS2 at Serine (Ser) 731 in the NTS of SHRs, WKY rats, and fructose-fed rats (Supplementary Fig. 1A–C). These results indicate that IRS2 may not be involved in Wnt induced insulin signaling. Wnt3a administration significantly decreased the phosphorylation of IRS1 at Ser332 and increased downstream insulin signaling in the NTSs of SHRs and fructose-fed rats (Fig. 3A and C). Similarly, a reduction in the insulin levels in the NTSs of SHRs and fructose-fed rats was detected following Wnt3a injection (Fig. 3D and F). Furthermore, coadministration of DKK1 and Wnt3a enhanced the IRS1^S332 phosphorylation and insulin levels and significantly decreased insulin signaling in the NTSs of WKY rats (Fig. 3B and E and Supplementary Fig. 2B). These results indicate that Wnt promotes the activation of IR signaling, which, in turn, downregulates insulin production in the NTSs of hypertensive rats.

As shown above, our experiments indicated that treatment with Wnt3a leads to a decrease in IRS1^S332 phosphorylation in the NTSs of hypertensive rats. Next, we examined the endogenous IRS1^S332 phosphorylation levels in the NTSs of hypertensive and normotensive rats. As shown in Fig. 4A, there was a significant increase in the P-IRS1^S332 levels in the NTSs of SHRs and fructose-fed rats compared with normotensive WKY rats. In addition, GSK-3β phosphorylates IRS1 at Ser332 to attenuate downstream insulin signaling in cultured cells (25). To determine whether Wnt regulates the interaction between GSK-3β and P-IRS1^S332 in the NTS, we double immunolabeled for P-IRS1^S332 and GSK-3β. Using confocal microscopy, P-IRS1^S332 (red) immunoreactivity was abundantly detected in the NTSs of SHRs (Fig. 4B, lane 1). GSK-3β (green) immunofluorescence colocalized to P-IRS1^S332 immunoreactivity (Fig. 4B, lane 1). Strikingly, the administration of Wnt3a resulted in a decrease in P-IRS1^S332 immunofluorescence and in the colocalization of P-IRS1^S332 with GSK-3β (Fig. 4B, lane 2). Moreover, coadministration of DKK1 and Wnt3a restored the P-IRS1^S332 levels and the colocalization of P-IRS1^S332 with GSK-3β (Fig. 4B, lane 3). Then, we performed coimmunoprecipitation assays to analyze whether GSK-3β binds to P-IRS1^S332. The NTS lysate was immunoprecipitated using either an anti–GSK-3β or anti–P-IRS1^S332 antibody and was subsequently probed using anti–GSK-3β and anti–P-IRS1^S332 antibodies. The results revealed that GSK-3β coimmunoprecipitated with P-IRS1^S332 (Fig. 4C, lanes 2 and 3). Next, an in vitro kinase assay was performed to determine whether GSK-3β directly phosphorylates IRS1^S332 in vitro. We incubated purified IRS1 protein in GSK-3β immunoprecipitated from the NTS lysate in the presence of ATP for the indicated reaction period. Figure 4D shows the time-dependent increase in phosphorylation of IRS1 at Ser332 by GSK-3β. These results suggest that GSK-3β phosphorylates IRS1^S332 in the NTS. In addition, we determined whether GSK-3β is involved in the Wnt-induced depressor response and NO production. Intracerebroventricular infusion of the GSK-3β–specific inhibitor TWS119 decreased the BP in SHRs and partially blocked the fructose-induced depressor response in fructose-fed rats (Fig. 4E and F). As shown in Fig. 4G, after intracerebroventricular treatment with TWS119, the phosphorylation levels of IRS1 in the NTSs of hypertensive rats were significantly decreased compared with the control rats; however, neuronal NOS (nNOS), Akt, ERK, and RSK in the NTSs of hypertensive rats were significantly increased compared with the control rats in SHRs and fructose-fed rats.

Essential hypertension is the most common cardiovascular disease worldwide, being estimated to cause 13% of all deaths (26). The SHR is a widely used animal model of human essential hypertension and insulin resistance (27). It has been reported (28–30) that various animal species that are fed a high-fructose diet or fructose-containing water for several weeks develop the metabolic syndrome, which includes insulin resistance, impaired glucose homeostasis, and hypertension. Therefore, we used SHRs and fructose-fed rats as animal models in our current study to investigate the signaling mechanisms involved in Wnt-mediated central BP regulation in the NTS. In addition, our present results reveal endogenous aberrant canonical Wnt signaling in the NTSs of hypertensive rats. Phosphorylation of LRP6 promotes canonical Wnt signaling activation. As shown in Fig. 1A and D, on the basis of immunoblotting and immunohistochemistry analyses, the LRP serine phosphorylation levels in both the SHRs and the fructose-fed rats were lower than in the normotensive rats. Interestingly, our results show that central administration of Wnt3a induced NO release from the NTS and a depressor effect in the SHRs (Fig. 1G). The fructose-mediated pressor effect was also abolished by Wnt3a administration (Fig. 1I). The inhibition of LRP6 activity using DKK1 in the presence of Wnt3a decreased LRP6 phosphorylation at Ser1490 in the NTS after the induction of IRS1 serine phosphorylation and the development of hypertension (Fig. 1H). Our findings in murine models support those obtained from a clinical study, in which the subjects who carried a loss-of-function mutation in LRP6 displayed hypertension, insulin resistance, and diabetes (19). Krüger et al. (31) revealed that stimulation of insulin increases LRP6 activation. Additionally, a mutation in LRP6 impairs insulin signaling in skeletal muscle and leads to the development of insulin resistance (32). In contrast to our findings, in homozygotic LRP6 knockout mice, insulin signaling and IRS1 activity are enhanced in brown adipose tissue (33). Cruciat et al. (34) revealed that the (pro)renin receptor binds to the Fzd receptor, phosphorylating LRP6 and activating LRP6 phosphorylation. The mRNA levels of the (pro)renin receptor were 45% higher in the NTSs of SHRs than in those of WKY rats, suggesting that the (pro)renin receptor may contribute to the regulation of BP (35). The possible mechanism connecting Wnt and the (pro)renin receptor in the NTS for BP regulation remains to be elucidated. Our findings extend the findings of previous reports that Wnt signaling in the NTS might participate in BP regulation.

IRS1 is a docking protein for the IR and insulin growth factor-1 receptor. IRS1 contains several serine/threonine and tyrosine phosphorylation sites. The phosphorylation of IRS1 at tyrosine residues occurs in response to insulin-stimulated activation. Moreover, most phosphorylation events at serine/threonine residues are recognized as negative feedback regulators of insulin signaling. Many serine/threonine kinases, including GSK-3β, S6 kinase, ERK1/2, c-Jun NH₂-terminal kinase, mammalian target of rapamycin, AMPK, and protein kinase C (36,37), have been shown to phosphorylate IRS1. GSK-3β has been identified to phosphorylate IRS1 at Ser332, and high-glucose stimulation activates GSK-3β to promote IRS1^S332 phosphorylation (25,38). The current study also showed that GSK-3β directly increases P-IRS1^S332 in a time-dependent manner and identified GSK-3β as a key factor in the downregulation of IRS1 phosphorylation. Treatment with Wnt3a reduced the interaction between GSK-3β and P-IRS1^S332 in the NTSs of SHRs. Simultaneous intracerebroventricular infusion of DKK1 and Wnt3a restored this interaction between GSK-3β and P-IRS1^S332, suggesting that the administration of Wnt3a downregulated the GSK-3β-P-IRS1^S332 signaling cascade in the NTSs of hypertensive rats. TWS119 is a potent and cell-permeable GSK-3β inhibitor; it has been reported (39) to promote Wnt signaling in CD8⁺ T cells. Treatment with TWS119 decreased the SBP of SHRs and ameliorated the fructose-induced pressor response (Fig. 4E and F). Our data revealed that inhibiting GSK-3β activity using TWS119 prevents IRS1^S332 phosphorylation in the NTS. However, it has been reported (40) that membrane-associated GSK-3β activates LRP6 via phosphorylation at Ser1490. GSK-3β is also negatively regulated by insulin. Furthermore, insulin activates the PI3K-Akt cascade, which leads to the phosphorylation of GSK-3β at Ser9, resulting in its inactivation (41). Together, these findings indicate that GSK-3β is a critical factor in the cross talk between Wnt and insulin signaling. Our results indicate that Wnt3a-mediated BP regulation may occur via the downregulation of the GSK-3β-P-IRS1^S332 interaction, subsequently activating insulin signaling in the NTS. Further studies are required to determine the mechanisms by which the stimulation of insulin signaling is associated with GSK-3β activity and the Wnt signaling cascade in the NTS to regulate BP.

Epidemiologic studies (42) have established that diabetes is estimated to affect nearly 24 million people in the U.S. This significant disease burden translates into a major economic impact. Clinical intervention trials have established the key roles of the transcription factor 7-like 2 (TCF7L2) gene with type 2 diabetes in African (43) and African-derived populations (i.e., African Americans) (44). The TCF7L2 gene product is a high-mobility transcription factor that is associated with blood glucose homeostasis. It is via the Wnt signaling pathway that TCF7L2 regulates the proglucagon gene expression in enteroendocrine cells (20). Recent reports (22) showed that Wnt stimulation enhances IRS1 expression and triggers Akt activation. Besides, Cross et al. (41) demonstrated that insulin is thought to stimulate the dephosphorylation of glycogen synthase at these residues by inducing the inactivation of GSK-3α and GSK-3β via phosphorylation of an N-terminal Ser residue (Ser21 in GSK-3α and Ser9 in GSK-3β), which is catalyzed by Akt in muscle. McManus et al. (45) also found that Wnt3a induces normal inactivation of GSK-3 at Ser21/Ser9, as judged by the stabilization of β-catenin and the stimulation of Wnt-dependent transcription. These results establish the function of Ser21/Ser9 phosphorylation in several processes in which GSK-3 inactivation has previously been implicated (45). In Fig. 2D and E, our data found that intracerebroventricular infusion of Wnt3a increases the phosphorylation and expression levels of GSK-3β^S9, which may be associated with Akt or Wnt3a stimulation.

Inhibition of DKK improves insulin sensitivity and glucose tolerance in diabetic mice but not in nondiabetic mice (46). Epidemiological studies showed that mutations of LRP5 may cause type 1 diabetes and obesity (21,47). Furthermore, Wnt3a-stimulated insulin secretion is mediated by LRP5 (48). LRP5 has also been shown to regulate LRP6 serine phosphorylation. Additionally, LRP5 may serve as a coreceptor in the Wnt and insulin signaling pathways to upregulate the insulin signaling cascade in preadipocytes (23). In the current study, we found decreased insulin levels in the NTS following intracerebroventricular infusion of Wnt3a, which was restored by cotreatment with DKK1. Central administration of Wnt3a into the brain activated critical mediators of the insulin signaling network, including phosphorylated Akt^S473, ERK1/2^T202/Y204, and RSK^T359/S363 in the NTS of hypertensive rats, and these phosphorylation events were blocked by the administration of DKK1 (Fig. 3A and B). Moreover, antagonizing LRP6 phosphorylation prevented PI3K-Akt and ERK1/2-RSK signaling activation and led to insulin production in the NTS of WKY rats. A mutation in LRP6 underlies insulin resistance and diminishes Akt^S473 phosphorylation (32). Taken together, these data suggest that Wnt administration reverses insulin signaling and improves insulin release via the Wnt coreceptor LRP. However, Mani et al. (19) identified a missense mutation in LRP6, which substitutes cysteine for arginine at a highly conserved residue of an epidermal growth factor–like domain, that impairs Wnt signaling in vitro. These results link a single gene defect in Wnt signaling to coronary artery disease and multiple cardiovascular risk factors (19). Similarly, rare mutations in other Wnt-related transcription factors cause maturity-onset diabetes of youth (49). In a further perspective on patients with early coronary artery disease, metabolic syndrome, and its components, we suggest that Wnt signaling may provide new insight into disease pathophysiology or into approaches to prevent these disorders and may be a potential therapeutic target for treating type 2 diabetes.

In conclusion, we suggest that Wnt modulates central BP in the NTS and that this regulation is accomplished by activating IRS1-PI3K-Akt and IRS1-ERK1/2-RSK signaling–mediated NO release. We have shown for the first time that Wnt3a induces a depressor effect and reduces the interaction between GSK-3β and P-IRS1^S332 in the NTSs of hypertensive rats (Fig. 5). The present data demonstrate that Wnt3a modulates central BP via inhibition of the GSK-3β-IRS1 pathway and the consequent activation of insulin signaling and downstream NO release.

Funding. This work was supported by grants from the National Science Council (NSC100-2321-B-075B-002, NSC 101-2320-B-075B-002, and NSC102-2320-B-075B-002) and the Kaohsiung Veterans General Hospital (VGHKS100-101, VGHKS101-024, VGHKS102-001, and VGHKS 103-104) (to C.-J.T.).

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

Author Contributions. P.-W.C. conducted most of the studies and wrote the manuscript. Y.-Y.C., W.-H.C., and B.-R.C. assisted in the conduct of most of the studies. P.-J.L., M.H., and C.-J.T. conceived and designed the study. H.-H.C., T.-C.Y., and G.-C.S. contributed to the writing of the manuscript. C.-J.T. 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.