Growth Hormone-induced Diacylglycerol and Ceramide Formation via Gαi3 and Gβγ in GH4 Pituitary Cells

Growth hormone (GH) secretion is regulated by indirect negative feedback mechanisms. To address whether GH has direct actions on pituitary cells, lipid signaling in GH4ZR7 somatomammotroph cells was examined. GH (EC50 = 5 nm) stimulated diacylglycerol (DAG) and ceramide formation in parallel by over 10-fold within 15 min and persisting for >3 h. GH-induced DAG/ceramide formation was blocked by pertussis toxin (PTX) implicating Gi/Goproteins and was potentiated 1.5-fold by activation of Gi/Go-coupled dopamine-D2S receptors, which had no effect alone. Following PTX pretreatment, only PTX-resistant Gαi3, not Gαo or Gαi2, rescued GH-induced DAG/ceramide signaling. GH-induced DAG/ceramide formation was also blocked in cells expressing Gβγ blocker GRK-ct. In GH4ZR7 cells, GH induced phosphorylation of JAK2 and STAT5, which was blocked by PTX and mimicked by ceramide analogue C2-ceramide or sphingomyelinase treatment to increase endogenous ceramide. We conclude that in GH4 pituitary cells, GH induces formation of DAG/ceramide via a novel Gαi3/Gβγ-dependent pathway. This novel pathway suggests a mechanism for autocrine feedback regulation by GH of pituitary function.

Pituitary somatotrophs synthesize and secrete GH, 1 which acts at the liver and other tissues to stimulate IGF formation, promoting somatic growth throughout the body (1,2). Secretion of GH is stimulated by hypothalamic GH-releasing hormone and inhibited by the hypothalamic tetradecapeptide somatostatin and by IGF. In addition, somatostatin agonists (e.g. octreotide) and dopamine-D2 agonists (e.g. bromocryptine) are used clinically to treat acromegaly (a syndrome produced by hypersecretion of GH) and to inhibit somatomammotroph growth and GH production (3). Negative feedback via autocrine actions of GH at the pituitary has been postulated (4) but is yet to be clearly demonstrated.
The GH receptor is a member of the type I cytokine receptor superfamily, related to PRL and erythropoietin receptors that homodimerize to initiate signaling (5). The GH receptor signals through the JAK2 tyrosine kinase-signal transducer and activator of transcription 5 (STAT5) transcription factor pathway to induce gene expression (6 -9). Phosphorylation on residue Tyr-694 by JAK2 is obligatory for STAT5 activation (10). The two STAT5 variants, STAT5a and STAT5b, have 90% identical protein sequences and are independently regulated and activated in various cell types (11). Studies using STAT5a or STAT5b knockout mice have demonstrated that STAT5b, but not STAT5a, is required for GH-induced regulation of IGF1 and sex-specific steroidogenic enzymes in liver (11)(12)(13). While STAT5 activation is implicated in many GH actions, other signaling pathways not involving STAT5 appear to be recruited for GH-induced stimulation of other pathways including MAPK phosphorylation and phosphatidyl inositol 3Ј-kinase or protein kinase C activation (14,15) in a cell type-dependent manner (16). Ceramide is a novel second messenger implicated in regulation of cell differentiation, proliferation, inflammation, and apoptosis (17)(18)(19). Ceramide plays an important role in signaling of a subgroup of cytokine receptors that includes tumor necrosis factor and interleukin-1 receptors (5,15,20,21). However, the coupling of the GH/PRL-related family of receptors to ceramide has not been reported. We therefore examined whether GH might influence ceramide formation in pituitary cells as part of an autocrine feedback pathway and whether dopamine-D2 agonists would influence GH action.
Rat pituitary tumor GH4C1 cells synthesize and secrete PRL and GH and provide an excellent model of pituitary somatomammotrophs used for over 30 years (22). In this report we have identified a novel induction of DAG and ceramide formation by GH that is blocked by PTX, implicating the involvement of G i /G o proteins (23). The contribution of specific G␣ subunits to GH autocrine signaling pathways was addressed using PTXinsensitive mutants of G␣ i 2, G␣ i 3, and G␣ o individually transfected into GH 4 ZR 7 pituitary cells (GH 4 C 1 cells transfected with the dopamine-D2S receptor (24,25)). In PTX-insensitive G protein mutants, the carboxyl-terminal ribosyl-acceptor cysteine was changed to a nonacceptor serine. The Cys-to-Ser mutation is a structurally conservative change, and the mutant G proteins remain functional following PTX pretreatment (26 -28). The role of G␤␥ subunits was evaluated by using the carboxyl-terminal domain of G protein-coupled receptor kinase (GRK-ct), a selective G␤␥ scavenger (29). In GH 4 ZR 7 cells, dopamine-D2S receptor activation potentiated GH-induced DAG and ceramide formation. We have identified G␣ i 3 and G␤␥ as crucial for both GH-induced ceramide formation and dopamine-D2-induced potentiation of the GH response. * This work was supported by The National Cancer Institute of Canada. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Plasmid Construction-As previously described (27), PTX-insensitive G␣ i/o mutants were generated by point mutation of rat cDNAs (30) encoding G␣ i 2 and G␣ i 3 subunits at cysteine 351 (352 for G␣ i 2). The TGT (cysteine codon) was mutated to TCT (serine) and confirmed by Sanger dideoxynucleotide sequencing. The mutant G␣ i 2 and G␣ i 3 cDNAs were FLAG-tagged at the initiator ATG codon, and the cDNAs were subcloned in KpnI-EcoRI-cut pcDNA3 (Invitrogen) to generate G␣ i 2-PTX, G␣ i 3-PTX, and G␣ o -PTX. The carboxyl-terminal domain of OK-GRK2 cDNA (31), beginning from Thr-493, was tagged at the amino-terminal with RGS-His 6 , and the His-GRK-ct fragment was cloned into pcDNA3 to produce the GRK-ct construct.
Cell Culture and Transfection-GH 4 ZR 7 cells and derivative clones were maintained in Ham's F10 medium with 8% fetal bovine serum (FBS) at 37°C, 5% CO 2 . G␣ i 2-PTX, G␣ i 3-PTX, and G␣ o -PTX (20 g) were cotransfected individually with pGK-puro (2 g) into GH4ZR7 cells using calcium phosphate co-precipitation. The transfected cells were cultured in F10 ϩ 8% FBS containing puromycin (20 g/ml) for 3-4 weeks. Antibiotic-resistant clones were picked (24 clones/transfection) and tested for expression of the corresponding G␣ i/o proteins by Western blot analysis. The following G␣ i 2-PTX, G␣ i 3-PTX, G␣ o -PTX and GRK-ct clones were selected for analysis, respectively: G i 2Z 24, G i 3Z 15, G␣ o 15, and GRKZ 17. These cells express about 2-to 3-fold times the endogenous level of total G␣ protein in GH 4 ZR 7 cells, suggesting that the ratio of PTX-insensitive/endogenous G␣ proteins in the clones was 1-2-fold (25).
Lipid Extraction-Equivalent numbers of cells were cultured in ten 10-cm plates with Ham's F10 medium plus 8% FBS in a humidified atmosphere of 5% CO 2 , 95% air at 37°C, grown to 80 -90% confluence, and placed in serum-free F10 medium for 16 h. For PTX treatment, the cells were treated with 10 ng/ml PTX for 16 h prior to experimentation. Cells were rinsed with serum-free F10 medium and treated with experimental compounds at 37°C as indicated. Following incubations, cells were twice rinsed with ice-cold PBS and lipids extracted (32). After centrifugation at 500 ϫ g for 1 min at 4°C, the supernatants were aspirated and the cells were lyzed with 0.5 ml of chloroform/methanol/ HCl (20:40:1, v/v/v), and sonicated in 5-s intervals ϫ 6 on ice. Cells were rinsed with 1 ml of chloroform and 0.3 ml of 1 M NaCl and spun at 14,000 ϫ g for 15 min at 4°C. The upper aqueous layer was discarded, and the lower lipid-containing layer was transferred to a 1-ml glass Chrompack vial, dried under a stream of O 2 -free N 2 gas, and redissolved in 200 l of chloroform. The samples were stored at Ϫ80°C until analysis. The particulate protein interface was air-dried, dissolved in 0.5 ml of 2 M NaOH, and assayed for protein according to Lowry's method.
Quantification of DAG and Ceramide-DAG and ceramide were quantified using the DAG kinase method (33,34). A blank tube and a standard ceramide/DAG tube were included as controls. For each sample, 10 l of DAG kinase (20 milliunits), 50 l of reaction buffer (100 mM imidazole, pH 6.6, 25 mM MgCl 2 and 2 mM EGTA), 10 l of 20 mM dithiothreitol, and 10 l of [␥-32 P]ATP (2.5 ϫ 10 5 dpm/nmol) were added and incubated at 25°C for 30 min, and the reactions terminated by addition of 0.5-ml ice-cold chloroform/methanol (1:2 v/v). The lipids were separated and extracted by addition of 0.5 ml of chloroform and 0.5 ml of 1 M NaCl spun at 14,000 ϫ g for 3 min, and the upper aqueous phase was discarded. The organic phase was sequentially washed with 0.5 ml of 1% perchloric acid, 0.3 ml of chloroform/methanol (1:2 v/v), 0.2 ml of chloroform, and 0.2 ml of water, dried under N 2 , and reconstituted in 25 l of chloroform/methanol (95:5, v/v). The samples were spotted onto a Silica Gel 60 TLC plate, heat-activated, and developed in a solvent mixture of chloroform/acetone/methanol/acetic acid/water (10: 4:3:2:1, v/v/v/v/v). Since DAG kinase can use ceramide or DAG as substrate, [ 32 P]ceramide-phosphate represented ceramide production and [ 32 P]phosphatidic acid represented DAG production. The TLC plates were exposed to phosphor screens for 18 h, and [ 32 P]ceramidephosphate and [ 32 P]phosphatidic acid were quantified using the Molecular Dynamics System ImageQuaNT computer software. Results are expressed as percentage of control.
GH Binding Assay-The binding assay was performed using 50 g of protein and 15,000 cpm/sample of 125 I-hGH (2150 Ci/mmol, PerkinElmer Life Sciences) in a final volume of 300 l (0.021 nM final concentration) of TME buffer (75 mM Tris, pH 7.4, 12.5 mM MgCl 2 , 1 mM EDTA) containing 0.1% BSA (35). To assess nonspecific binding 1 nM unlabeled hGH was added to the reaction. Incubation at room temperature was stopped after 30 min by the addition of 500 l of cold 100 mM Tris, pH 7.4. The reaction was then filtered through GF/C filters and washed three times with 5 ml of cold 100 mM Tris, pH 7.4. Triplicate measurements were performed for all samples.
Western Blot Analysis-Cells were treated as described above. Cell pellets were frozen on dry ice/ethanol and stored at -80°C. Samples were sonicated 10 -15 s, heated at 95°C for 5 min, and centrifuged, and 40 l/sample loaded onto SDS-PAGE gel and electrotransferred to polyvinylidene difluoride membrane. The membrane was blocked (1 h, room temperature), probed with primary antibody (1:1000, overnight, 4°C), washed in TBST (10 mM Tris-HCl, pH 8, 150 mM NaCl, and 0.05% Tween 20) and incubated with horseradish peroxidase-conjugated secondary antibody (1:2000) and horseradish peroxidase-conjugated antibiotin antibody (1:1000) to detect biotinylated protein markers (2 h at room temperature). The blot was then washed, incubated with Lumi-GLO (1 min), and exposed to x-ray film. Exposures in the linear range (gray scale) were scanned and quantified using the UnScanIt program (Silk Scientific Inc., Orem, Utah).
Statistical Analysis-The data were analyzed by repeated measure using analysis of variance for each set of experiments. Differences of p Ͻ 0.05 were considered statistically significant.

Concentration-and Time-dependent Increase in DAG/Ceramide Formation Induced by GH-
The acute action of GH on endogenous DAG and ceramide levels in GH 4 ZR 7 pituitary cells was assessed by the DAG kinase assay. The cells were washed to remove extracellular (secreted) GH and assayed in serumfree medium. GH induced a 10-fold increase in both DAG and ceramide production in a concentration-dependent manner from 10 Ϫ10 to 10 Ϫ6 M at 20 min with an EC 50 of ϳ5 nM (Fig. 1). Addition of exogenous SMase (0.1 units/ml) was included as a positive control to demonstrate the hydrolysis of endogenous SM to form ceramide. The phosphorylated DAG and ceramide species co-migrated with the respective standards, confirming the identity of the products. GH (10 Ϫ7 M) robustly increased both DAG and ceramide production in parallel, which was maximal within 15 min and declined but remained significantly elevated at 3 h (Fig. 2). Low levels of GH are secreted by GH 4 C 1 cells at a rate of 0.2 ng/ml/min or 10 Ϫ11 mol/liter/min (36), sufficient to reach a threshold concentration (10 Ϫ9 M) for DAG/ceramide formation in 1.5 h following initiation of treatments (see "Materials and Methods"). However, GH is also metabolized, hence the actual GH concentration under culture conditions may be lower and did not appear to interfere with actions of exogenous GH.
PTX Blocks GH-induced Ceramide Production in GH 4 ZR 7 Cells-We recently showed that in Balb/c-3T3 fibroblasts, activation of the D2S receptor induces DAG and ceramide formation that is blocked by PTX, which inactivates G i /G o proteins. 2 Cells were pretreated with or without 10 ng/ml PTX for 16 h, a concentration that blocks G i /G o -mediated signaling in these cells (22). PTX treatment blocked GH-induced DAG and ceramide formation, thus implicating G i /G o proteins (Fig. 3). By contrast, PTX or dopamine-D2 agonist apomorphine (10 Ϫ6 M) alone did not alter DAG or ceramide formation. Importantly, PTX treatment did not change the level of specific 125 I-GH binding sites measured in crude membranes from GH4ZR7 cells. Specific 125 I-GH binding was 118 Ϯ 45 fmol/mg in GH 4 ZR 7 cells (mean Ϯ S.E., n ϭ 3), and binding in PTX-treated cells was 104 Ϯ 8% of control binding, indicating that blockade of GH-induced ceramide by PTX was not due to loss of receptor sites.
Apomorphine Potentiates GH-induced DAG and Ceramide Production in GH 4 ZR 7 Cells-To examine further whether activation of the D2S receptor modulates DAG or ceramide formation, GH 4 ZR 7 cells were incubated with GH, apomorphine (a D2 receptor agonist) or both GH and apomorphine (Fig. 4). Although dopamine-D2S receptor activation alone did not in-fluence DAG or ceramide formation, apomorphine potentiated by 1.5-to 2-fold times the GH-induced formation of DAG and ceramide. In parental GH 4 C 1 cells, which lack dopamine receptors, GH induced both DAG and ceramide formation but this effect was not enhanced by apomorphine (data not shown), indicating that apomorphine-induced potentiation is mediated via activation of dopamine-D2S receptors present on GH 4 ZR 7 cells. Pretreatment with PTX blocked GH-induced ceramide production in GH 4 ZR 7 and GH 4 C 1 cells and also completely blocked DAG/ceramide production by apomorphine/GH (Fig. 5), indicating that D2S-induced potentiation of GH action involves G i /G o proteins.
Rescue of GH-induced DAG and Ceramide Production by PTX-insensitive G␣ i 3, but Not G␣ i 2 or G␣ o and Involvement of G␤␥ Subunits-We examined which subunit(s) of G proteins mediate DAG or ceramide signaling induced by GH or apomorphine/GH in combination using GH 4 ZR 7 cells stably transfected with PTX-insensitive G␣ mutants (G␣ i 2-PTX and G␣ i 3-PTX cells) (25). As observed in wild-type GH4ZR7 cells and in G␣ i 2-PTX and G␣ i 3-PTX clones, the level of ceramide production induced by combination of apomorphine and GH was greater than for GH alone ( Fig. 6 and data not shown). To examine the importance of G␣ i 2-PTX and G␣ i 3-PTX, cells were pretreated with PTX to block endogenous G i/o proteins and challenged with GH or apomorphine/GH in combination. PTX blocked completely DAG and ceramide production stimulated by GH or apomorphine/GH in G␣ i 2-PTX cells (Fig. 6). However in G␣ i 3-PTX cells, both DAG and ceramide production were at least 50% resistant to PTX pretreatment (Fig. 7). Since ϳ50% of the total G␣ i 3 was PTX-sensitive endogenous protein (25), a recovery of 50% of the response would be expected from the remaining fraction of PTX-insensitive G i 3 proteins. Thus G␣ i 3, but not G␣ i 2, plays a crucial role in both GH-and apomorphine/ GH-induced DAG and ceramide formation. To examine the role of G␣ o subunits in GH-induced lipid signaling, GH4ZR7 cells were stably transfected with G␣ o -PTX. In these cells, PTX completely blocked DAG and ceramide formation induced by the combination of apomorphine and GH (Fig. 8), indicating that like G␣ i 2-PTX, G␣ o -PTX does not rescue GH-induced lipid signaling. As a selective G␤␥ scavenger (29), the carboxyl-terminal domain of G protein-coupled receptor kinase (GRK-ct) was used to examine the role of G␤␥ subunits in signaling to ceramide formation. We have transfected GRK-ct into GH 4 ZR 7 cells and identified expression of GRK-ct by Western blot (25). Neither apomorphine nor apomorphine/GH induced DAG or ceramide formation in GRK-ct cells (Fig. 9). This suggests that G␤␥ subunits are necessary for ceramide formation induced by the combination of apomorphine and GH.
GH and Ceramide Enhance JAK2/STAT5 Phosphorylation in GH 4 ZR 7 Cells-Based on the results above, we examined the influence of GH, apomorphine, PTX, and ceramide on well known and potential downstream pathways of the GH receptor including phosphorylation of JAK2, STAT5 (Fig. 10), STAT3 or MAPK. In GH 4 ZR 7 cells, GH alone increased phosphorylation of JAK2 (100% increase over basal) and STAT5 (40% increase), which was more strongly enhanced with both apomorphine and GH (160% increase over basal for phospho-JAK2, 90% increase for phospho-STAT5). Treatment with a ceramide analogue (C2ceramide) or SMase (to increase endogenous ceramide) also increased JAK2 phosphorylation by 90 and 150%, and STAT5 phosphorylation by 60 and 90%, respectively. Interestingly, PTX-blocked apomorphine/GH-induced STAT5 phosphorylation by 50%, further supporting a role for the PTX-sensitive ceramide pathway in GH-induced STAT5 phosphorylation in these cells. By contrast, these compounds elicited no changes in STAT3 or MAPK phosphorylation (data not shown).

GH-induced Coupling to G Proteins and Lipid Signaling-
Our results indicate that GH induces a G protein-dependent increase in lipid metabolism to generate DAG and ceramide in GH 4 cells. Previous studies in pre-adipocyte Ob1771 cells (37,38) and in pancreatic ␤-cells (39) have shown that GH induces DAG formation via activation of PC-PLC. By analogy, GH may activate PLC in GH 4 cells to induce DAG formation. Both DAG and ceramide formation were induced in parallel, suggesting interconversion between these lipids possibly via SM synthase, which can convert DAG into ceramide, leading to depletion of SM (40,41). Alternately, DAG can activate acidic SMase to generate ceramide (42,43). Interconversion of DAG to ceram-ide would account for the identical G␣ i 3 and G␤␥ dependencies of GH-mediated lipid formation.
The actions of GH in GH 4 cells were sensitive to PTX pretreatment, indicating a role for G i /G o proteins. Upon activation, GH receptors dimerize, associate with JAK2, and recruit a family of negative regulators, the SOCS (suppressors of cytokine signaling) proteins (44,45). Coupling of the GH receptor to PTX-sensitive G proteins is relatively unexplored, and potential interactions of GH receptors or associated proteins such as SOCS proteins with G proteins remain elusive. There is some evidence that GH-like receptors interact with G proteins. In Nb2 cells, G␣ i proteins labeled by PTX-mediated ADP ribosylation were cross-linked to the PRL receptor using a 16-Å cross-linking agent, but not cross-linkers with shorter molecular lengths, consistent with a direct physical interaction (46). In addition, some PTX-sensitive GH-induced responses have been reported. For example, GH-induced PC-PLC activation in Ob1771 preadipocytes (37,38) and GH-mediated DAG formation and mitogenesis in pancreatic ␤-cells (39) are PTX-sensitive actions. Similarly, activation of the homologous PRL receptor in Nb2 lymphoma cells enhances PTX labeling of G i proteins (suggesting activation) and induces PTX-sensitive mitogenesis (47)(48)(49). Taken together, these results are consistent with coupling of the GH receptor to PTX-sensitive G i proteins to activate PLC thereby generating DAG, which can be converted to ceramide.
Although coupled to G i /G o proteins, GH signaled differently from the G i /G o -coupled dopamine D2 receptor to induce PTXsensitive DAG and ceramide formation since apomorphine alone had no effect. Nevertheless there was an interaction between GH and D2 signaling since apomorphine potentiated GH-induced lipid signaling and JAK2/STAT5 activation. Furthermore, GH-and apomorphine/GH-induced DAG and cera- Pretreatment with 10 ng/ml of PTX (P) was for 16 h. After harvesting cells, Western immunoblotting using antibodies to the indicated phospho-protein or ␤-actin (loading control) was carried out. A, representative blots probed for phospho-JAK2 (p-JAK2), phospho-STAT5 (p-STAT5) and ␤-actin as indicated. B and C, data from three independent experiments probed for phospho-JAK2 (B) or phospho-STAT5 (C) were quantified using the Un-Scan-It program and are presented as percent of control (mean Ϯ S.E.); *, p Ͻ 0.03 and **, p Ͻ 0.01 compared with control. mide formation were both rescued by G␣ i 3-PTX and blocked by GRK-ct, suggesting a crucial role for G␣ i 3/G␤␥ for both receptors. The dopamine-D2 receptor utilizes G i 3 to mediate activation of potassium channels in pituitary cells (50) via binding of G␤␥ to the GIRK potassium channel (51), and is likely to couple to G␣ i 3/G␤␥ in GH4 pituitary cells. The mechanism by which GH receptors couple to G i 3 remains to be elucidated, but GH receptors appear to interact with G i proteins differently from G i -coupled heptahelical receptors (such as adenosine or D2S receptors). In adipocytes, GH prevented coupling of adenosine receptor-mediated inhibition of cAMP and activation of phosphatidylinositol-specific-PLC and blocked PTX-induced ADPribosylation (52,53). GH may induce relocalization of G␣ i subunits, prevent their coupling to adenylyl cyclase (53,54), and allow efficient coupling to DAG/ceramide formation. Since sites of ceramide synthesis display discrete subcellular localization (55), differences in the localization of D2S-and GH-receptor coupling might account for their differing effectiveness to induce ceramide formation in GH4 pituitary cells.
A Novel Pathway for GH-induced JAK2/STAT5 Activation-Our data show that C2-ceramide and sphingomyelinase induce JAK2/STAT5 activation in GH4 cells, suggesting a link between GH-induced changes in DAG/ceramide and the classical GH-receptor-mediated JAK/STAT pathway. Consistent with our results, sphingomyelinase was shown to increase ceramide levels and was found to activate JAK2 and STAT1/3 in cultured human fibroblasts (56). Importantly, as observed for GH-mediated ceramide formation, GH-induced JAK2/STAT5 activation was enhanced by apomorphine and was partially blocked by PTX, suggesting that both G protein-dependent and -independent pathways lead to JAK2/STAT5 activation in these cells. Thus G i -mediated ceramide signaling regulates GH-induced JAK2/STAT5 activation.
In addition to regulating JAK2/STAT5, GH-induced ceramide formation may activate other signaling cascades (19). Both GH (16,57) and ceramide (19,20) have been shown to activate the MAPK cascade in other cell types, but we observed no induction of p42/44-MAPK by either GH or ceramide in GH4 cells. Ceramide regulates other pathways including the SAPK/ JNK cascade, and several proapoptotic pathways, but the roles of these pathways in GH4 cells is not known.
GH-mediated Autocrine Regulation of Pituitary Cells-Multiple negative feedback pathways regulate GH secretion at the level of the hypothalamus and pituitary. At the level of the hypothalamus, GH inhibits GH-releasing hormone synthesis and enhances somatostatin release, resulting in decreased GH secretion at the pituitary (58 -60). GH-induced IGF formation is believed to be the primary negative feedback pathway to inhibit GH synthesis in somatotrophs (1,2). In addition, G i /G ocoupled dopamine-D2 and somatostatin receptors also inhibit GH secretion and somatomammotroph growth (3). It is tempting to speculate that GH may negatively regulate its own secretion; however, evidence for a non-IGF-mediated autocrine pituitary feedback by GH is indirect (4,61,62). The GH receptor is expressed in rat and human anterior pituitary and binds and internalizes radiolabeled GH, suggesting a role for GH to regulate its secretion from the pituitary (63)(64)(65)(66). However, the signaling of the GH receptor in pituitary cells has not been investigated. Our finding of a novel G protein-mediated action of GH to induce DAG/ceramide as well as JAK2/STAT5 activation in GH4 cells suggests a role for GH in regulation of pituitary function. GH4 cells are a pituitary cell strain that has provided an important model of somatotrophs that synthesize and secrete levels of GH that are sufficient to mediate autocrine GH-induced actions (22). Interestingly, C2-ceramide has been shown to inhibit GH secretion from rat anterior pituitary cells (67), suggesting that GH-induced ceramide formation could mediate negative feedback inhibition of GH secretion. Previously unexplored direct actions of GH on DAG and ceramide may provide a more sensitive method to address direct actions of GH in regulation of somatotroph function in vivo.