Prostaglandin F_2α Facilitates Hepatic Glucose Production Through CaMKIIγ/p38/FOXO1 Signaling Pathway in Fasting and Obesity

Type 2 diabetes constitutes a major worldwide public health burden and is expected to affect >642 million adults by 2040 (1). Type 2 diabetes is characterized by hyperglycemia, insulin resistance, and β-cell dysfunction, and often is associated with low-grade chronic inflammation (2). Blood glucose homeostasis is maintained by the balance between hepatic glucose production (HGP) (glycogenolysis and gluconeogenesis) and glucose utilization by peripheral tissues. In patients with type 2 diabetes, hepatic gluconeogenesis is considerably elevated and contributes to both fasting and postprandial hyperglycemia, and suppressing hepatic gluconeogenesis improves insulin sensitivity and glucose homeostasis, making it an attractive target for the treatment of diabetes (3). Hepatic gluconeogenesis is tightly controlled by rate-limiting gluconeogenic enzymes, including PEPCK (PCK1), glucose-6-phosphatase (G6Pase), and fructose-1,6-bisphosphatase. The quantity and activity of these essential enzymes are mainly regulated by the pancreatic hormones insulin and glucagon (4). In addition to hormonal control, hepatic gluconeogenesis is also directly regulated by inflammatory mediators, such as cytokines (5) and phospholipid derivatives (6,7).

Prostaglandin (PG) F_2α is a bioactive lipid metabolite of arachidonic acid produced through the sequential reaction of cyclooxygenases and PGF synthase. PGF_2α exerts a wild range of pathological and physiological functions, including reproduction (8), blood pressure (9), and bone remodeling (10), by binding to its receptor (F-prostanoid [FP]). Experimental and clinical studies showed that PGF_2α is involved in different inflammatory conditions, including rheumatic diseases, asthma, atherosclerosis, ischemia-reperfusion, septic shock, and obesity (11). Interestingly, increased PGF_2α metabolite production is found in urine from patients with both type 1 diabetes (12) and type 2 diabetes (13), and plasma PGF_2α levels can be used as a marker to predict glycemic control and oxidation status (14). Moreover, hepatic PGF_2α production is also significantly increased in fasted, high-fat diet (HFD)–fed, or diabetic animals (7). FP receptor is expressed in hepatocytes (15); however, whether an activated PGF_2α/FP axis is implicated in aberrant glucose metabolism in diabetes remains largely unclear.

In this study, we demonstrated that the activation of FP receptor increased fasting-induced hepatic gluconeogenesis in mice by upregulating gluconeogenic genes (PCK1 and G6Pase) in a FOXO1-dependent manner. FP activation promoted FOXO1 nuclear translocation by stimulating G_q-mediated Ca²⁺/calmodulin-activated protein kinase IIγ (CaMKIIγ) signaling, with p38 required for FP/CaMKIIγ-mediated FOXO1 nuclear translocation. Moreover, the silencing of hepatic FP receptor–ameliorated glucose metabolism in obese mice. Therefore, FP activation facilitated hepatic gluconeogenesis through the CaMKIIγ/p38/FOXO1 signaling pathway under both fasting and diabetic conditions.

Human FP (hFP) hepatocyte transgenic (HCTG) and nontransgenic (NTG) littermates were obtained from the mating of Tg-hFP-STOP^Flox (16) and Alb^Cre mice. FP exon2-floxed mice (FP^F/F) were generated through CRISPR/Cas9 technology by Shanghai Biomodel Organism Science & Technology Development Co., Ltd. (Shanghai, People’s Republic of China) and crossed with Alb^Cre mice to obtain hepatic FP-deficient mice (FP^F/FAlb^cre). All mice used were from a C57BL/6 background. For inhibitor treatment, 8-week-old HCTG and NTG mice were intraperitoneally injected with KN-93 (3 mg/kg) (MedChem Express, Monmouth Junction, NJ) three times weekly for 1 week (17) or with SB202190 (2.5 μmol/L/kg) (Sigma-Aldrich, St. Louis, MO) daily for 1 week (18). All animals were maintained and used in accordance with the guidelines of the Institutional Animal Care and Use Committee of the Institution for Nutritional Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, People’s Republic of China.

DNA was extracted from mice tails using the proteinase-K-chloroform method, and PCR was performed on an S1000 thermal cycler (Bio-Rad, Hercules, CA). RNA was isolated from tissues or cultured cells using TRIzol (Invitrogen, Carlsbad, CA) and reverse transcribed to cDNA using an RT Reagent Kit (TaKaRa, Dalian, People’s Republic of China). The resulting cDNA was amplified for 40 cycles on a C1000 Thermal Cycler (Bio-Rad) using SYBR Green PCR mix (TaKaRa). Values were normalized to actin mRNA or 18S rRNA. See Supplementary Tables 1 and 2 for primer sequences.

The frozen sections from liver or chamber slides with hepatocytes were washed with PBS and fixed in cold acetone. To block nonspecific binding of antibodies, samples were incubated with PBS containing 3% BSA and 0.1% Triton X-100 (for permeabilization) for 30 min. The samples were then incubated with mCherry fluorescent protein (1:200; Abbkine, Redlands, CA) or FOXO1 (1:500; Cell Signaling Technology, Danvers, MA) overnight at 4°C. Slides were then washed with PBS three times and incubated with secondary antibodies conjugated with Alexa Fluor 594 or Alexa Fluor 633 (Invitrogen) for 1 h at room temperature. ProLong Gold Antifade Reagent with DAPI (Invitrogen) was applied to mount and counterstain the slides. Images were obtained using confocal microscopy (IX51; Olympus, Center Valley, PA).

Total protein and cytosolic and nuclear protein fractions were isolated from tissues or cultured cells using a protein extraction kit (Beyotime, Shanghai, People’s Republic of China), followed by separation by SDS-PAGE, transfer to nitrocellulose membranes, and probing with different primary antibodies against PCK1, p38, and lamin B (Proteintech, Wuhan, People’s Republic of China); G6Pase and CaMKIIγ (Santa Cruz Biotechnology, Dallas, TX); α-tubulin, phospho-Thr287 CaMKIIγ, phospho-p38, FOXO1, and hemagglutinin (HA)-Tag (Cell Signaling Technology); hFP (Epitomics, Burlingame, CA); and β-actin (Sigma-Aldrich). The membranes were then conjugated with a horseradish peroxidase–labeled secondary antibody in blocking buffer for 2 h at room temperature. Blots were developed using enhanced chemiluminescence reagents (Pierce, Rockford, IL), followed by densitometric quantification using ImageJ (National Institutes of Health, Bethesda, MD). See Supplementary Table 3 for antibodies.

Male mice (8 weeks old) were injected with G6P-Luc adenoviruses (1 × 10⁸ plaque-forming units) by tail vein. On day 5 after adenovirus delivery, mice were subjected to fasting for 8 h before imaging, followed by intraperitoneal injection with 100 mg/kg sterile firefly d-luciferin (Biosynth, Itasca, IL). After 10 min, mice were imaged on the IVIS 100 Imaging System, and images were analyzed with Living Image software (Xenogen, Alameda, CA).

Adenoviruses were constructed using the AdEasy Adenoviral System (Qibogene, Irvine, CA) as previously described (19). For short hairpin RNA (shRNA) adenovirus, synthetic hFP, p38, and CaMKIIγ shRNA was inserted into the pGE1 plasmid under the control of U6 promoter, then a fragment containing shRNA and U6 promoter was digested and cloned into the pAd-Track-cytomegalovirus (CMV) construct [pAd-Track-CMV] containing CMV-green fluorescent protein cassette. The empty vector [pAd-Track-CMV] was used as a control. For dominant-negative (DN)-FOXO1 adenovirus, DN-FOXO1 cDNA was inserted into the pAd-Track-CMV(+) construct under the control of the CMV promoter. The empty vector [pAd-Track-CMV(+)] was used as a DN-FOXO1 cDNA control. For WT-Flag-FOXO1and 9A-Flag-FOXO1 expressing adenovirus (20), cDNAs were first inserted into a pAd-Track-CMV(+) plasmid using CMV promoter, and then the constructs were subcloned into the pAd-Easy-1 adenoviral-backbone vector through homologous recombination in BJ5183. Adenoviral DNA was linearized by PacI restriction digestion and transfected into HEK293 cells using Lipofectamine 2000 (Invitrogen) for adenovirus packaging. After several rounds of propagation, recombinant adenovirus was purified by ultracentrifugation (Beckman, Urbana, IL) in a cesium chloride gradient. See Supplementary Table 4 for shRNA sequences.

Insulin tolerance tests were performed by intraperitoneal injection of 0.8 units/kg insulin after 6 h of fasting, and pyruvate tolerance tests (PTTs) were performed by intraperitoneal injection of 2 g/kg sodium pyruvate after overnight fasting. Fasting blood glucose was measured by obtaining a tail-blood sample after 12 h of fasting. Serum triglyceride (TG) and total cholesterol, as well as liver glycogen levels, were measured using assay kits (BJKT, Beijing, People’s Republic of China).

Primary mouse hepatocytes were isolated from 8- to 12-week-old mice using a two-step collagenase-perfusion technique as described previously (21). Hepatocytes were cultured in DMEM supplemented with 25 mmol/L glucose, 10% FBS, and 50 μg/mL penicillin and streptomycin at 37°C and 5% CO₂/95% air. Glucagon (100 nmol/L) and PGF_2α (100 nmol/L) from Cayman Chemical Company (Ann Arbor, MI), and pertussis toxin (PTX), Wortmannin, U73122, KN-93, and SB202190 from Sigma-Aldrich were used to treat hepatocytes, as indicated.

Primary hepatocyte glucose production was measured as previously described (7,22). Briefly, isolated hepatocytes were cultured in monolayer in glucose- and phenol-free DMEM with 20 mmol/L sodium lactate and 1 mmol/L sodium pyruvate for 6 h after overnight FBS starvation, and glucose levels were measured using a glucose assay kit (Sigma-Aldrich). Total protein in the cell lysate was used for normalization. For adenovirus experiments, a hepatocyte glucose production test was performed 48 h after adenovirus infection.

Ca²⁺ transients were recorded using a laser-scanning confocal microscope (Carl Zeiss, Oberkochen, Germany) as previously described (23). Briefly, hepatocytes seeded on chamber slides (Thermo Fisher Scientific, Waltham, MA) were loaded with 5 μmol/L Fluo 3-AM (Dojindo Laboratories, Kumamoto, Japan) for 30 min, washed, and imaged under an IX71 Fluorescence Microscope (Olympus). Fluo-3 was excited at 488 nm.

Results are represented as the mean ± SEM. All statistical analyses were subjected to Student t test or two-way ANOVA, followed by the Bonferroni post hoc test using GraphPad Prism 5 software (GraphPad, La Jolla, CA). A P value of <0.05 was considered statistically significant.

PGF_2α production in the liver is markedly elevated in mice in response to fasting and HFD treatment or those with obesity (7), conditions associated with augmented hepatic gluconeogenesis. Accordingly, we observed significantly upregulated FP expression in livers from fasted, ob/ob, and HFD-treated mice (Fig. 1A), indicating that the PGF_2α/FP axis was activated in the liver upon fasting and dia-betic stress. Interestingly, hepatic knockdown of FP receptor by intravenous delivery of shRNA adenovirus (Ad-FP-shRNA) (Fig. 1B and C) markedly reduced circulating glucose concentrations after an 8- to 12-h fast (Fig. 1D) and attenuated gluconeogenic capacity in mice after pyruvate infusion (Fig. 1E), but did not affect hepatic glycogen content under fasting conditions (Supplementary Fig. 1). Consistently, FP knockdown led to significant decreases in the expression of hepatic gluconeogenic genes, such as PCK1 and G6Pase, without altering the expression of the hepatic glycogenolysis gene glycogen phosphorylase (Pygl) under fasting conditions in mice (Fig. 1F–H). Similarly, in isolated primary hepatocytes, FP knockdown significantly reduced glucose output after pyruvate challenge (Fig. 1I) and suppressed gluconeogenic gene expression (G6pase and Pck1) without affecting glycogenolytic gene expression (Pygl) (Fig. 1J–L). Furthermore, FP knockdown improved insulin sensitivity in mice (Fig. 1M). These results suggested that FP is directly involved in hepatic gluconeogenesis by promoting hepatic gluconeogenic gene expression under both physiological and pathological conditions.

To examine whether FP overexpression in hepatocytes influences gluconeogenesis in vivo, we generated hFP transgenic mice using a Cre-loxp strategy. Two strains exhibiting hepatic hFP overexpression (HCTG-A and HCTG-E) were obtained, with NTG littermates serving as controls (Supplementary Fig. 2A and B). In HCTG mice, the hFP gene was expressed under the control of the albumin promoter in hepatocytes by Cre-mediated mCherry sequence excision (Supplementary Fig. 2C), leading to highly effective and selective expression in livers (Supplementary Fig. 2D and E). There were no differences in body weight, body composition, and plasma TG and cholesterol levels between wild-type and HCTG mice (Supplementary Fig. 3A–C). Notably, HCTG mice showed significantly higher blood glucose concentrations after 8–12 h of fasting compared with NTG mice (Fig. 2A) and exhibited enhanced HGP according to PTT results (Fig. 2B) along with upregulated hepatic gluconeogenic gene expression (G6pase and Pck1) (Fig. 2C–G). Importantly, hFP transgenic mice also showed less sensitivity to insulin challenge (Fig. 2H). We also examined hepatic transcription activity of gluconeogenic genes in HCTG mice by infusion of the adenoviral G6Pase-luciferase (Ad-G6Pase-Luc) reporter containing both FOXO1-binding and cAMP-responsive element–binding sites that mediate upregulation of gluconeogenic genes during fasting (24). During fasting, Ad-G6Pase-Luc activity was markedly elevated in HCTG mice compared with control mice (Fig. 2I and J). Moreover, forced expression of the hFP gene increased basal HGP in culture, and PGF_2α treatment further augmented HGP in HCTG hepatocytes (Fig. 3A). Furthermore, PGF_2α induced gluconeogenic gene expression (G6pase and Pck1) at both the mRNA and protein levels in NTG and HCTG hepatocytes without altering glycogenolytic gene (Pygl) expression (Fig. 3B–E). Consistently, Ad-G6Pase-Luc activity was markedly increased in response to PGF_2α in cultured HCTG hepatocytes compared with levels observed in NTG cells (Fig. 3F). Therefore, these results indicated that PGF_2α/FP promoted hepatic gluconeogenesis by upregulating gluconeogenic genes.

The FP receptor might couple to different G-proteins to trigger downstream-signaling molecules, such as G_q to mobilize cytosolic Ca²⁺ or G_i to reduce intracellular cAMP levels (25). To determine which G-protein signal is involved in FP-mediated gluconeogenesis, primary hepatocytes were treated with the phospholipase C inhibitor U73122 (inhibits Ca²⁺ release), the G_i inhibitor pertussis toxin, or the phosphoinositide 3-kinase inhibitor Wortmannin in the presence of PGF_2α, followed by the monitoring of glucose production after pyruvate challenge. U73122 significantly suppressed PGF_2α-induced glucose output and gluconeogenic gene expression (G6pase and Pck1) in HCTG hepatocytes (Fig. 4A and B). Additionally, hFP overexpression resulted in significantly elevated intercellular Ca²⁺ concentrations in hepatocytes, which was blocked by U73122 treatment (Fig. 4C). Because the Ca²⁺-sensing kinase CaMKII is essential for gluconeogenesis in fasting conditions and obesity (26), we then examined the expression levels of CaMKII isoforms (α, β, γ, and δ) in mouse hepatocytes, finding only abundant expression of CaMKIIγ (Fig. 4D and E). FP knockdown significantly suppressed CaMKIIγ phosphorylation in cultured hepatocytes and livers in mice (Fig. 4F–I), whereas hFP overexpression dramatically increased CaMKIIγ phosphorylation in cultured hepatocytes and livers from HCTG mice (Fig. 4J–L). Blockage of Ca²⁺ influx by U-73122 effectively attenuated increased CaMKIIγ phosphorylation levels in hFP-overexpressing hepatocytes from HCTG mice (Fig. 4M). Furthermore, treatment with the specific CaMKII inhibitor KN-93 suppressed CaMKIIγ phosphorylation, diminished the increased HGP in hFP-overexpressing hepatocytes in response to pyruvate challenge (Fig. 4N), and blunted the upregulated gluconeogenic protein expression in hepatocytes from HCTG mice (Fig. 4O). Consistent with in vitro observations, KN-93 treatment also abrogated elevated blood glucose concentrations in HCTG mice observed after 8–12 h of fasting (Fig. 4P). Collectively, these results showed that CaMKIIγ mediated PGF_2α-induced hepatic gluconeogenesis in mice upon fasting.

CaMKII modulates glucose metabolism and hepatic insulin signaling through its downstream mediator p38 mitogen-activated protein kinase (26,27), and p38 is involved in fasting-induced gluconeogenesis (27,28). Strikingly, FP knockdown significantly inhibited p38 phosphorylation, which represents the activated form of p38, in primary hepatocytes (Fig. 5A). By contrast, hFP overexpression enhanced p38 phosphorylation in primary hepatocytes in either the presence or absence of exogenous PGF_2α stimulation (Fig. 5B). Additionally, the inhibition of intracellular Ca²⁺ influx by U-73122 or CaMKII activity by KN-93 abrogated induction of p38 phosphorylation in HCTG hepatocytes (Fig. 5C and D). In agreement with the in vitro observations, phosphorylated p38 levels in mouse livers were reduced by Ad-FP-shRNA infection (Fig. 5E and F), whereas they were elevated by hepatic overexpression of hFP receptor (Fig. 5G). Importantly, the administration of SB202190, which suppresses p38 activity according to reductions in phosphorylated mitogen-activated protein kinase–activated protein kinase 2 (p-MK2; a p38 kinase target), attenuated upregulated gluconeogenic protein expression (G6pase and Pck1) in HCTG hepatocytes (Fig. 5H), thereby blunting increased HGP in HCTG hepatocytes (Fig. 5I). Consistently, SB202190 treatment also eliminated elevations in plasma glucose concentrations in HCTG mice after fasting for 8–12 h (Fig. 5J). Moreover, silencing of CaMKIIγ, or p38α, a dominant expressed isoform in liver (28), significantly attenuated the glucose production and suppressed the upregulation of gluconeogenic gene expression (G6pase and Pck1) in HCTG hepatocytes, respectively (Supplementary Fig. 4A–F). Therefore, PGF_2α promoted hepatic gluconeogenesis by activating CaMKIIγ/p38 signaling.

FOXO1 is a major transcription factor involved in glucose metabolism and is regulated mainly through changes in its cellular localization between the cytoplasm and nucleus (29). p38 can directly phosphorylate FOXO1 at five different sites (20), thereby promoting FOXO1 nuclear translocation and facilitating HGP in hepatocytes (26). We hypothesized that FP might facilitate HGP through P38-mediated FOXO1 nuclear localization. Interestingly, FP knockdown led to relative cytoplasmic accumulation of FOXO1 (Fig. 6A and B), whereas hFP overexpression markedly increased nuclear translocation of FOXO1 in both cultured hepatocytes and livers in mice (Fig. 6C–E). Of note, either knockdown or overexpression of FP did not affect the total FOXO1 protein level in mouse hepatocytes and liver tissues (Fig. 6A–D). Inhibition of FOXO1 by adenoviral delivery of HA-tagged DN-FOXO1, a truncated mutant of FOXO1 containing the entire DNA-binding domain but lacking the transactivation domain (30), abrogated the elevated expression of G6pase and Pck1 in HCTG hepatocytes and suppressed the induction of PGF_2α-induced HGP in HCTG hepatocytes after pyruvate challenge (Fig. 6F and G). In vivo transfection of the DN-FOXO1 adenovirus also completely eliminated elevated plasma glucose concentrations in HCTG mice after an 8- to 12-h fast (Fig. 6H). Furthermore, the blockade of intracellular Ca²⁺ flux (via CaMKII) or p38 activity impeded nuclear accumulation of FOXO1 in HCTG hepatocytes in response to PGF_2α (Fig. 6I and K). Finally, mutation of p38 phosphorylation sites in FOXO1 protein (20) led to a nuclear translocation defect in response to PGF_2α in HCTG hepatocytes (Supplementary Fig. 5). Collectively, these results indicated that FP promoted hepatic gluconeogenesis by enhancing gluconeogenic gene expression through CaMKIIγ/p38 signaling-mediated FOXO1 nuclear translocation.

To exclude the effect of FP in other tissues on HGP, we created hepatocyte-specifc FP-deficient mice using FP^F/F mice crossed with Alb^cre mice, which express Cre under the control of the albumin promoter (Supplementary Fig. 6A and B). This resulted in FP^F/FAlb^cre mice in which the FP gene was specifically deleted in the liver (Supplementary Fig. 6C). FP^F/FAlb^cre mice exhibited significantly lower blood glucose levels after 8–12 h of fasting (Fig. 7A), less HGP after pyruvate infusion (Fig. 7B), and increased sensitivity to insulin challenge compared with control mice (Fig. 7C). Moreover, hepatic deletion of FP significantly attenuated the phosphorylation of CaMKIIγ and p38, reduced nuclear translocation of FOXO1 without changing its total expression, and suppressed gluconeogenic gene expression (G6pase and Pck1) in the livers of FP^F/FAlb^cre mice (Fig. 7D–F). These results showed that FP mediated hepatic gluconeogenesis via the CaMKIIγ/p38/FOXO1 signaling pathway (Fig. 7H).

Hepatic glucose output, such as that resulting from gluconeogenesis, is higher in subjects who are obese and have diabetes compared with that observed in subjects who are lean and do not have diabetes (31,32). To test whether silencing FP could improve glucose metabolism during obesity, ob/ob mice were infected with Ad-FP shRNA through tail-vein injection. As shown in Fig. 8A–C, hepatic knockdown of the FP receptor (Fig. 8A) significantly reduced blood glucose levels after 4-, 8-, and 12-h fasts (Fig. 8B), suppressed glucose production after PTT challenge (Fig. 8C), and improved insulin sensitivity in ob/ob mice (Fig. 8D). Accordingly, FP silencing also substantially inhibited the phosphorylation of CaMKIIγ and p38, impeded nuclear translocation of FOXO1, and suppressed gluconeogenic protein expression (G6pase and Pck1) in the livers of ob/ob mice (Fig. 8E–G).

In obesity and type 2 diabetes, circulating levels of PGF_2α are markedly elevated (11), but the role of PGF_2α in glucose metabolism remains unknown. Here, we observed that the FP receptor was upregulated in the livers of fasting and obese mice and that hepatic FP deletion in hepatocytes suppressed HGP by downregulating the expression of gluconeogenic genes, whereas hFP overexpression in hepatocytes had an opposite effect. Mechanistically, FP activation enhanced the nuclear localization of FOXO1 and gluconeogenic gene translation through the Ca²⁺/CaMKIIγ/p38 signaling pathway. These observations indicated that the FP-mediated gluconeogenic pathway might represent a potential therapeutic target for type 2 diabetes.

PGF_2α is a potent luteolytic agent that is also involved in modulating intraocular pressure and smooth muscle contraction in the uterus (11), with PGF_2α analogs, such as latanoprost, used in medicine to treat glaucoma (33). However, exaggerated PGF_2α production has been observed in patients experiencing rheumatic diseases (34), essential hypertension (35), obesity (36), and diabetes (37). Furthermore, FP deficiency reduces blood pressure and retards attendant atherogenesis in hyperlipidemic mice (9). Interestingly, in this study, we observed that hepatic FP deletion markedly suppressed gluconeogenesis, whereas hepatic overexpression of hFP promoted gluconeogenesis in mice. In vitro, PGF_2α infusion directly stimulates hepatic glucose output in isolated rat livers (38). By contrast, endotoxin treatment attenuates hepatic gluconeogenesis in rats by downregulating expression of the FP receptor (39). Here, we found that FP silencing significantly improved insulin sensitivity in ob/ob mice, probably because of suppression of hepatic gluconeogenesis. Indeed, aspirin administration, which reduces PG production by inhibiting cyclooxygenase activity, dramatically reduces fasting plasma glucose in patients with type 2 diabetes and diabetic rats (40,41). Taken together, increased PGF_2α biosynthesis might lead to increased fasting or postprandial glucose levels in patients with type 2 diabetes.

In mammals, gluconeogenesis occurs mainly in the liver (42). FP receptor promotes G_q/Ca²⁺ signaling in hepatocytes and plays an important role in metabolism (43). In preadipocytes, PGF_2α binds G_q to modulate adipocyte differentiation by modulating intracellular Ca²⁺ signaling (44). Luteal regression in mammals mediated by PGF_2α also relies upon Ca²⁺-dependent mechanism (45), suggesting the pivotal role of Ca²⁺ signaling in FP-mediated biological functions. Here, we found that FP receptor activated Ca²⁺-dependent CaMKIIγ isoforms in hepatocytes and that the inhibition of CaMKIIγ abrogated PGF_2α-induced HGP. Notably, FP promoted gluconeogenic gene expression particularly through enhancing nuclear translocation of FOXO1 without influencing glycogenolytic gene expression. Indeed, Ca²⁺/CaMKIIγ signaling promotes glucose production by promoting FOXO1 nuclear translocation in hepatocytes (26). Interestingly, we and others observed that genetic deficiency or inhibition of CaMKIIγ significantly decreases fasting-induced gluconeogenesis (26), suggesting that CaMKIIγ maybe serve as a potential therapeutic target for hyperglycemia and type 2 diabetes. Additionally, p38 is involved in the mediation of CaMKIIγ-driven FOXO1 nuclear translocation (26,27) and regulation of hepatic gluconeogenesis (28,46). In this study, we demonstrated that FP activation promoted p38 activity in livers and that p38 inhibition suppressed FP-mediated hepatic gluconeogenesis by reducing FOXO1 nuclear translocation. In agreement with our findings, p38 is a key signaling molecule involved in various FP-mediated physiological functions, such as myometrial contractility during gestation (47), corpus luteum regression (48), cardiomyocyte hypertrophy (49), and artery contraction (50). Therefore, our findings suggest that FP promotes hepatic gluconeogenesis by activating the CaMKIIγ/p38/FOXO1 signaling pathway.

In summary, we demonstrated that the PGF_2α/FP axis facilitates hepatic gluconeogenesis during fasting and obesity through CaMKIIγ/p38 signaling, suggesting that targeting the PGF_2α/FP signaling pathway might represent a promising therapeutic strategy for treating type 2 diabetes.

Acknowledgments. The authors thank Dr. Akiyoshi Fukamizu (University of Tsukuba, Tsukuba, Japan) for providing WT-FLAG-FOXO1 and 9A-FlAG-FOXO1 plasmids.

Funding. This work was supported by the National Natural Science Foundation of China (81525004, 91439204, 91639302, 31771269, and 31200860) and the National Key R&D Program from the Ministry of Science and Technology of China (2017YFC1307404 and 2017YFC1307402). Yi.Y. is a Fellow at the Jiangsu Collaborative Innovation Center for Cardiovascular Disease Translational Medicine.

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

Author Contributions. Y.W. acquired the data, performed statistical analysis, and drafted the manuscript. S.Y., B.X., S.Z., Q.Z., G.C., Q.L., Y.L., Yu Y., and D.C. acquired the data. Yi.Y. handled funding and supervision. Yi.Y. and Y.S. conceived and designed the research. Yi.Y. and Y.S. made critical revision of the manuscript for key intellectual content. Yi.Y. 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.