- Email: [email protected]
Insulin and heregulin-β1 upregulate guanylyl cyclase C expression in rat hepatocytes
Cellular Signalling 13 (2001) 665 – 672
Insulin and heregulin-b1 upregulate guanylyl cyclase C expression in rat hepatocytes Reversal by phosphodiesterase-3 inhibition Lawrence A. Schevinga,*, William E. Russellb,c a
Division of Pediatric Gastroenterology and Nutrition, Vanderbilt University School of Medicine, Nashville, TN 37232, USA b Division of Pediatric Endocrinology, Vanderbilt University School of Medicine, Nashville, TN 37232, USA c Department of Cell Biology, Vanderbilt University School of Medicine, Nashville, TN 37232, USA Received 6 April 2000; accepted 28 March 2001
Abstract Guanylyl cyclase C (GC-C) is the receptor for the hormones guanylin and uroguanylin. Although primarily expressed in the rat intestine, GC-C is also expressed in the liver during neonatal or regenerative growth or during the acute phase response. Little is known about the hepatic regulation of GC-C expression. The influence of various hepatic growth or acute phase regulators on GC-C expression was evaluated by immunoblot analysis of protein from primary rat hepatocytes grown in a serum-free medium. Insulin and heregulin-b1 strongly stimulated GC-C expression by 24 h of cell culture. Several different hormones and agents suppressed this action, including transforming growth factor b (TGF-b), as well as inhibitors of phosphatidylinositol 3-kinase (PI-3-kinase) and phosphodiesterase 3 (PDE-3, an insulin- and PI-3-kinasedependent enzyme). The compartmental downregulation of cAMP levels by PDE-3 may be a critical step in the hormonal action that culminates in GC-C synthesis. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Guanylyl cyclase C; Guanylin; Uroguanylin; Phosphodiesterase; Insulin; Liver
1. Introduction Guanylyl cyclases (GCs) consist of a family of enzymes that form cyclic GMP (cGMP) from GTP. Very little is known about the regulation or role of GCs in the liver. Earp [1] originally reported 20 years ago that primary cultures of adult rat hepatocytes displayed increased GC activity when cultured for 22 h in the presence of insulin, glucagon, and cyclic AMP (cAMP). These hormones synergistically stimulated cyclase activity, but the identity of the respon-
Abbreviations: Guanylyl cyclase C (GC-C); Phosphatidylinositol 3kinase (PI-3-kinase); Phosphodiesterase 3 (PDE-3); Phosphodiesterase 4 (PDE-4); Interleukin-6 (IL-6); Epidermal growth factor (EGF); Transforming growth factor b (TGF-b); Heregulin-b1 (HRG-b1); Inducible NO synthase (NOSi) * Corresponding author. Division of Pediatric Endocrinology, T-0107 Medical Center North, Vanderbilt University School of Medicine, Nashville, TN 37232-2579, USA. Tel.: +1-615-343-0698; fax: +1-615-343-5845. E-mail address: [email protected] (L.A. Scheving).
sible enzyme could not then be determined. It is now known that the adult rat liver makes several different GCs [2 –4], including soluble GC, which is activated by nitric oxide, and Gc-A and Gc-B, which are the membrane-bound receptors for the atrial and brain natriuretic peptides. A third membrane-bound GC — guanylyl cyclase C (GC-C) — is developmentally regulated in the liver. It is normally synthesized by the late foetal and early neonatal rat liver and by small and large intestine epithelial cells irrespective of age. The adult rat liver does not express GC-C except under special conditions. These include liver regeneration following partial hepatectomy or during the acute phase response to tissue injury and infection [3]. The regulation of hepatic GC-C expression is important because the intestine releases the GC-C ligands — guanylin and uroguanylin — into the portal blood, creating a potential regulatory link between the gut and liver [3– 6]. To study the regulation and function of GC-C, it would be advantageous to identify a primary cell type capable of expressing GC-C in long-term culture. This cannot be
0898-6568/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 8 9 8 - 6 5 6 8 ( 0 1 ) 0 0 1 7 9 - 6
666
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
achieved with primary intestinal epithelial cells because they undergo rapid programmed cell death within several hours of culture [6]. Recently, others have shown that GC-C expression can be induced in primary hepatocyte cultures derived from adult rat liver [7]. Balint et al. showed that interleukin-6 (IL-6) induced GC-C expression in hepatocytes cultured in the presence of dexamethasone (10 4 M) and foetal calf serum (5%). In this paper, we examine the regulation of GC-C expression in primary hepatocytes in a serum-free medium containing a lower concentration of dexamethasone (10 8 M), focusing on hormones and agents that modulate hepatic GC activity, cell proliferation [3,4], or the acute phase response [3]. We find that insulin and heregulin-b1 each strongly stimulate GC-C expression after 24 h in culture. Glucagon and cAMP potentiate the insulin action, suggesting a molecular basis for their previously reported stimulation of hepatocellular GC activity. Interestingly, nitric oxide depletion (which lowers cGMP by decreasing the NO ligand for soluble GC) enhances the GC-C induction, whereas inactivation of a specific membrane-associated phosphodiesterase — phosphodiesterase 3 (PDE-3) (which is a cGMP inhibited, cAMP-PDE) — blocks it. Collectively, our findings implicate hormonally mediated cyclic nucleotide feedback loops in the mechanism by which hormones, such as insulin, and growth factors, such as heregulin-b1, upregulate hepatic GC-C expression. Our results also provide additional evidence for the importance of subcellular compartmentalization in mechanism of cyclic nucleotide action [8].
2. Materials and methods 2.1. Chemicals and reagents Bovine serum albumin (BSA), STa, tris, leupeptin, aprotinin, chymostatin, PMSF, dexamethasone, pyruvate, and percoll were from Sigma (St. Louis, MO). Ammonium persulfate, N 0, N 0 methylenebisacrylamide, polyacrylamide, N 0, N 0, N 0, N 0 tetra-ethylethylenediamine, protein standards, and Tween-20 were from Bio-Rad (Richmond, CA). Nitrocellulose (nitropure) was from Micron Separation (Westboro, MA). X-ray film (Hyperfilm) and horseradish peroxidase (HRP-linked secondary antibody (donkey antirabbit) were from Amersham (Arlington Heights, IL). Chemiluminescence solutions were from New England Nuclear Research Products (Boston, MA). Human recombinant HRG-b1 (amino acids 177 – 244) was from R&D Systems (Minneapolis, MN). Insulin was from Eli Lilly and Company (Indianapolis, IN). Recombinant rat IL-6 was from Research Diagnostics (Fanres, NJ). LY 294002, wortmannin, and PD 98059 were from Calbiochem (San Diego, CA). Cilostamide and trequinsin were from BIOMOL (Ply-
mouth Meeting, PA). Rolipram, SB203580, and rapamycin were from Sigma. 2.2. Culture media and supplies Serum-free Williams’ Medium E, supplemented with 20 mM pyruvate, 10 nM dexamethasone, and 50 mg/ml gentamicin, was the standard medium used for all culture studies. Medium was purchased from Life Technologies. Type I collagenase was from Waco Pure Chemical Industries (Richmond, VA). 2.3. Primary culture of hepatocytes Hepatocytes were isolated from the livers of male Sprague – Dawley rats (175 –250 g; Harlan Sprague – Dawley, Indianapolis, IN) with modifications of our previously described methods [9,10]. The livers of ether-anesthetized rats were perfused through the portal vein with a calciumfree solution consisting of 150 mM NaCl, 2.8 mM KCl, 5.5 mM glucose, and 25 mM HEPES (pH 7.6) for 10 min followed by the same solution containing 3.8 mM CaCl2, 10 mg/ml soybean trypsin inhibitor, and 0.5 mg/ml type I collagenase for 8 min. The cells were dispersed in medium supplemented with 10% calf serum and filtered through 61-mm nylon mesh. The hepatocytes were then purified by a 5-min sedimentation at 1 g in serumcontaining medium followed by centrifugation at 50 g in isotonic percoll (specific gravity = 1.06) to reduce contamination by nonparenchymal cells. Percoll was removed by two washes in serum-containing medium, and the hepatocytes were assessed for viability by trypan blue exclusion ( > 95% viable). Cells (1 106 cells/well) were plated in type-1 collagen-coated 60-mm wells. After a 60-min attachment period, the serum-containing medium was replaced with serum-free medium containing hormones, cytokines, amino acids, and signaling pathway inhibitors as indicated. 2.4. Hepatic and intestinal cell membrane preparation The washed cell pellet (or in some experiments, cells that had been snap-frozen in liquid N2 and stored at 70C) was immediately resuspended in 2 ml cell homogenization buffer (10 mM Tris, 1 mM EGTA, 40 mg/ml PMSF, 10 mg/ ml aprotinin, 10 mg/ml leupeptin, 10% glycerol) homogenized with a Tissue-Tearor (Biospec Products) for 30 s at setting 2. The homogenate had final protein concentrations between 3 and 6 mg/ml as determined by the Coomassie protein assay [4]. 2.5. Immunoblots GC-C antiserum was generated as previously described against the peptide CNNSDHDSTYF, coupled to keyhole limpet antigen at the C residue [4]. Samples were subjected
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
to SDS – PAGE, transblotted, and immunoblotted as previously described [11].
3. Results and discussion In this paper, we examined the ability of different hormones to regulate GC-C expression in primary hepatocytes derived from the adult rat liver. We evaluated hormones that regulate physiological states associated with increased GC-C expression, namely, hepatocyte proliferation (insulin, epidermal growth factor (EGF), heregulin-b1, transforming growth factor b (TGF-b), and interleukin-6) and the acute phase response (dexamethasone and interleukin-6). We also evaluated the independent and collective influence of insulin, glucagon, and cAMP on GC-C synthesis since they have been previously shown to increase GC activity in cultured hepatocytes. The hormonal results prompted us to test specific pharmacologic inhibitors (inducible NO synthase (NOSi), phosphatidylinositol 3-kinase (PI-3-kinase), MEK, p38 MAP kinase, and PDE-3 inhibitors) on GC-C expression as well. 3.1. Insulin induces GC-C in normal liver hepatocytes Because GC-C expression increases during periods of increased liver cell proliferation, such as regeneration and neonatal growth [3,4], we hypothesized that growth factor stimulation of hepatocellular proliferation in culture would increase GC-C expression. We cultured primary hepatocytes for 18 h in control medium containing different regulators of hepatic growth, including TGF-b (4 ng/ml), EGF (10 ng/ml), heregulin-b1 (30 ng/ml), and insulin (150 nM), and then analyzed GC-C expression by immunoblot (Fig. 1A). In this model, insulin and heregulin-b1 are weak mitogens, EGF, a strong one, and TGF-b, a growth inhibitor. For intestinal membrane proteins, an antibody against the C-terminal amino acid sequence of GC-C resolved on immunoblot three bands of varying electrophoretic mobility (Fig. 1A, rightmost lane). These bands represent the immature (130 kDa) and mature glycoforms (140 kDa), as well as a brush-border-specific, proteolytically processed, 85-kDa mature form [12]. In cultured hepatocytes, insulin strongly stimulated the 130- and 140-kDa isoforms, but there was no evidence for the 85-kDa form. Heregulin-b1, an EGF-like ligand for the receptors ErbB3 and ErbB4, weakly stimulated GC-C expression in this experiment; EGF had little or no effect; and TGF-b1 showed no immunoreactivity. Immunohistochemical experiments revealed that hepatocellular GC-C had a diffuse pattern of staining (data not shown), much of which localized to intracellular membranes, as reported previously during liver regeneration [4]. Time course experiments confirmed the effect of insulin and showed that heregulin-b1 stimulated GC-C to nearly the
667
same extent as insulin (Fig. 1B). These results also suggested that some GC-C appeared during culture even in the absence of insulin or heregulin-b1; however, the basal GC-C level varied from one hepatocyte preparation to the next and may have been triggered by exposure to 5% serum during the initial 60-min plating period. Since weak mitogens (insulin and heregulin-b1) strongly stimulated GC-C expression, whereas a strong one (EGF) had no effect, GC-C expression in cultured hepatocytes is not a direct consequence of cell cycle progression. Kinetically, insulin induced the immature 130-kDa and mature 140-kDa GC-C glycoforms with a sequence and time course that resembled that of the regenerating liver (Fig. 1C) [4]. Insulin (+; 100 nM) first induced the immature glycoform by 12 h, followed by the mature one between 12 and 24 h. Scanning laser densitometry revealed the GC-C signal in the insulin-treated samples to be 0, 1, 10, and 5.2 absorbance units at 0, 12, 24, and 60 h, respectively. Maximal GC-C expression was observed at 24 h of culture, when GC-C resolved into nearly equal amounts of the immature and mature glycoforms (we set 24 h of culture as the focal time point in subsequent experiments). By 60 h, the mature GC-C isoform predominated, the ratio of the mature to immature form being 50% higher at 60 h (1.8) compared to 24 h (1.2). Insulin stimulated GC-C expression over a broad concentration range (Fig. 1D). Near-maximal stimulation of expression was observed at a concentration as low as 15 nM. Increasing the concentration to 50 nM slightly increased total GC-C. Expression remained elevated at 150 nM. The concentration dependency curve indicates that the insulin action was physiologic since the actual concentration of insulin in these cultures is at least two orders of magnitude lower due to the presence of an insulin degrading enzyme and nonspecific binding of insulin to the culture plate (W.E. Russell, unpublished results). This is well within the concentration range reported for insulin in the portal blood. Since dexamethasone is a component of our serum-free culture medium and since there is a glucocorticoid response sequence in the 50-flanking region of the human GC-C gene [13], we examined the effect of dexamethasone on GC-C induction. Glucocorticoids modulated GC-C induction by insulin (Fig. 1E). Although insulin (100 nM) alone increased GC-C expression at 24 h, this effect was markedly enhanced by addition of dexamethasone (10 6 and 10 8 M; Fig. 1E) in a dose-dependent manner. Dexamethasone by itself was without effect. 3.2. Insulin, cAMP, glucagon, and dexamethasone synergistically induce GC-C Because glucagon and cAMP increased the GC activity of primary hepatocytes [1], particularly in the presence of insulin, we examined their ability to modulate GC-C expression. Both agents alone slightly increased GC-C
668
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
Fig. 1. Insulin induces GC-C protein expression in primary cultures of hepatocytes. (A) Hepatocytes were cultured in a serum-free medium (WE) containing dexamethasone (10 6 M) in the presence of TGF-b (4 ng/ml), EGF (10 ng/ml), insulin (150 nM), and heregulin-b1 (30 ng/ml). The cells were harvested 18 h later. Homogenates were prepared and immunoblotted using a monospecific GC-C polyclonal antibody. (B) Densitometric analysis of time course experiments showing GC-C levels in hepatocytes cultured in the presence of insulin (150 nm; open diamonds), heregulin-b1 (30 ng/ml; open circles), EGF (10 ng/ml;closed triangles), or WE (closed squares) at 12, 24, and 60 h after culture. Values are means of two experiments. (C) Time course experiments showing the two main glycoforms induced by insulin in primary hepatocytes at 12, 24, and 60 h of culture. (D) Concentration dependency of insulin at 24 h; note that insulin induces GC-C expression at concentrations as low as 15 nM. (E) Concentration dependency of dexamethasone on insulin action (100 nM).
but not to the levels attained by insulin treatment (Fig. 2A). For example, dibutryl cAMP at 100 and 200 mM stimulated GC-C expression to levels 20 – 25% that observed with insulin. Lower concentrations had little or no effect by themselves; however, the combination of insulin (100 nM), glucagon (100 nM), and dibutryl cAMP (5 mM) synergistically stimulated GC-C expression, increasing it in one experiment 16-fold over that caused by insulin alone. This was twice that achieved by insulin in the presence of the highest concentration of dexamethasone tested (10 6 M; Fig. 2B). Our results suggest that the previously reported increase in GC activity in response to insulin, glucagon, and dibutryl cAMP at the same concentrations requires de novo synthesis of GC-C (probably activated nonspecifically in the original report by the
Mn2 + in the enzyme assay buffer). This would be consistent with Earp’s observation that cycloheximide, a protein synthesis inhibitor, blocked the cyclase increase [1]. Further investigation is required to determine whether the potentiating effect of glucagon on insulin stimulated GC-C expression results from further elevations of cAMP or the additional effect of glucagon to raise intracellular Ca2 + concentrations [14]. 3.3. Insulin action is blocked by TGF-b and potentiated by and IL-6
L-NAME
Various cytokines and pharmacological agents were evaluated for their ability to modulate the induction of GC-C by insulin (Fig. 3). TGF-b1 was tested because it is
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
669
Fig. 2. Independent and synergistic induction of GC-C by glucagon and dibutryl cAMP. (A) Cells were treated for 24 h with a serum free media (WE) with or without glucagon or dibutryl cAMP at the indicated concentrations. Cell homogenates were prepared and immunoblotted as described in Materials and methods. (B) Cells were treated with dibutryl cAMP (5 mM), glucagon (100 nM), or both in the presence or absence of insulin (100 nM). All cells were treated with 10 8 M dexamethasone, except for the lane designated 10 6 M. Note increased GC-C expression in the presence of 10 6 M dexamethasone and a synergistic induction of GC-C by insulin, glucagon, and dibutryl cAMP in the presence of 10 8 M dexamethasone. (C) Quantitative densitometric analysis of experiments showing the individual and combined effects of cAMP (5 mM), glucagon (100 nM), and insulin (100 nM). Values are the means ± S.D. of GC-C protein levels. The combination of all three hormones on GC-C levels was different from the insulin treated hepatocytes ( P < .05, Student’s t test).
a potent inhibitor of hepatocyte growth and because it abolished the basal immunoreactivity seen in the original studies (Fig. 1). TGF-b1, at concentrations as low as 0.15 ng/ml, suppressed GC-C induction by insulin (100 nM). Maximal inhibition was seen at 5 ng/ml (Fig. 3A). Several agents capable of potentiating insulin’s upregulation of GC-C (i.e. dexamethasone, glucagon, and cAMP) also reduce NO production indirectly by inhibiting the synthesis of NOSi [15,16]. This enzyme makes nitric oxide, which is the only known activating ligand for soluble GC, a major source of cytoplasmic cGMP in cultured hepatocytes. Downregulation of NOSi results in a decreased production not only of NO but also of soluble GC-generated cGMP, the catalytic end-product of GC-C. We speculated that decreased NO production enhances GC-C expression as part of a feedback mechanism to normalize cellular cGMP levels. To test this idea, we cultured hepatocytes in the presence of L-NAME (1 mM), a slowly reversible inhibitor of the NOSi enzyme. As predicted, L-NAME treated cells increased their expression of GC-C, particularly in the presence of insulin (Fig. 3B). Notably, TGF-b (10 ng/ml) blocked the strong GC-C induction by L-NAME, insulin, and 10 6 M dexamethasone (rightmost lane of Fig. 3B).
IL-6 is an activator of the acute phase response [17] and may be required for proper liver regeneration [18]. Balint et al. [7] previously reported that IL-6 promotes GC-C expression in primary hepatocytes. Like L-NAME, we found that IL-6 (25 ng/ml) had little effect on GC-C expression by itself (Fig. 3B); however, when combined with insulin (100 nM) and L-NAME (1 mM), it potentiated the GC-C induction, particularly at the high dexamethasone concentration. These findings confirm the previous work but also suggest that the prior culture conditions, including a high concentration of dexamethasone (10 4 M), as well as 5% calf serum (which contains insulin), played a permissive role in the IL-6 action. 3.4. The action of insulin on GC-C requires both PI-3kinase and PDE-3 activation Some of the late actions of insulin in hepatocytes are mediated by PI-3-kinase activation [10]. To determine whether PI-3-kinase activation was required for GC-C induction, we incubated cells with two different selective inhibitors of this enzyme, wortmannin (100 nM) and LY294002 (50 mM). Both agents at the tested concentrations
670
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
Fig. 3. TGF-b inhibits the induction of GC-C, whereas IL-6 and L-NAME stimulate it. (A) TGF-b was added at the specified concentrations to cells along with insulin (100 nM) in a serum-free medium. Twenty-four hours later, homogenates were prepared and immunoblotted as previously described. Note that TGF-b inhibited the induction of GC-C by insulin (100 nM) in a dose-dependent manner. (B) IL-6 (25 ng/ml) and L-NAME (1 mM) were added along with insulin (100 nM) to hepatocyte cultures. Both agents synergistically increased the induction of GC-C by insulin at the low dexamethasone (10 8 M) concentration. TGF-b (10 ng/ml) decreased the induction of GC-C by L-NAME and insulin at the high dexamethasone concentration.
inhibited GC-C induction by insulin (150 nM), suggesting a requirement for PI-3-kinase activation (Fig. 4A). One of the PI-3-kinase-dependent enzymes that insulin rapidly activates in cultured hepatocytes is the cGMP inhibited, cAMP-phosphodiesterase (PDE-3) [19]. This enzyme, which causes cAMP hydrolysis, is also activated by glucagon and cAMP through other non-PI-3-kinasedependent pathways. For example, removal of dibutryl cAMP or insulin from cell culture medium causes a 30 – 40% decrease of PDE-3 activity within only 5 min [20]. Since PI-3-kinase was required for maximal GC-C induction, and since insulin, glucagon, and cAMP each activate PDE-3 to varying degrees, we hypothesized that PDE-3 inhibition would block GC-C induction. To test this, we treated cells with two different selective PDE-3 inhibitors, cilostamide and trequinsin (Fig. 4B). Several different concentrations were examined for each inhibitor (0.5, 5, and 25 mM). PDE-3 inhibition reduced the insulin-mediated GC-C induction in a dosedependent manner. The highest concentration tested (25 mM) completely blocked maximal GC-C induction by insulin (150 nM), glucagon (100 nM), and dibutryl cAMP (5 mM). This result suggests that the activation of PDE-3 plays a pivotal role in the hormonal induction of GC-C in rat hepatocytes. The results of this experiment prompted us to examine the effects of other pharmacological inhibitors. Since PDE-3 contributes to only about 30% of the PDE activity in hepatocytes, most of which is localized to cell membranes [21], we examined whether the inhibition of the largely
cytosolic phosphodiesterase 4 (PDE-4) similarly affected GC-C synthesis. PDE-4 is the other major PDE in primary hepatocytes [21]. To test this, we compared the effects of a nonselective PDE inhibitor (IBMX), two PDE-3 inhibitors (trequinsin and cilostamide), and a PDE-4 inhibitor rolipram). We also evaluated several agents known to block several other intermediary enzymes involved in insulin’s actions, including ERK (PD98059), p38 MAPK (SB203580), and p70 56K (rapamycin) (Fig. 4C). Both IBMX and the specific PDE-3 inhibitors had a more pronounced inhibitory effect on the GC-C induction than rolipram, suggesting that PDE-3 but not PDE-4 played the major role in the action of insulin (Fig. 4C). These results are consistent with the concept that the cellular actions of cyclic nucleotides derive from local, subcellular changes in their levels (a consequence of the relative balance of cyclases and PDEs at a particular subcellular site) and accessibility to binding proteins [8]. Moreover, although all of the enzyme inhibitors caused a variable decrease in GC-C expression, p38 MAPK inhibition with SB 203580 was notably effective in blocking the action of insulin (Fig. 4C). Still, the inhibition achieved by this agent at the tested concentration was less than that seen with PI-3-kinase or PDE-3 inhibition. Finally, since heregulin-b1 induced GC-C to the same extent as insulin, we examined whether the PDE and p38 MAPK inhibitors and TGF-b had the same inhibitory actions on heregulin-b1 as insulin. As shown in Fig. 4D, these inhibitors were as effective or more effective in inhibiting this heregulin-b1 action. Wortmannin, the PI-3-
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
671
Fig. 4. Interference with PI-3-kinase or PDE-3 blocks GC-C induction by insulin. (A) Cells were treated with insulin ± specific inhibitors of PI-3 kinase, including LY 294002 (50 mM) and wortmannin (100 nm). (B) Cells were treated with the specific inhibitors of PDE-3 — cilostamide or trequinsin — at several different concentrations. Note that both inhibitors blocked in a concentration-dependent manner the ability of insulin (150 nM) or insulin, glucagon (100 nM), and dibutryl cAMP (5 mM) to induce GC-C. (C) Cells were treated IBMX (2 mM), trequinsin (Treq.; 25 mM), cilostamide (Cilos.; 25 mM), rolipram (Roli.; 25 mM), PD98059 (30 mM), SB203580 (10 mM), or rapamycin (50 nM) for 24 h in the presence of insulin (150 nM). Homogenates were immunoblotted for GC-C. Two exposure times (Exp.times) differentiate the amount of inhibition by the PDE and p38 MAP kinase inhibitors. (D) Cells were treated with insulin (150 nm) or heregulin-b1 (30 nM) for 12 or 24 h in the presence or absence of IBMX (2 mM), TGF-b (5 ng/ml), or SB203580 (10 mM). Homogenates were immunoblotted for GC-C.
kinase inhibitor, was also effective in blocking the heregulin-b1-mediated GC-C induction (data not shown). Thus, both insulin and heregulin-b1, which interact through separate receptors, are capable of inducing GCC through their actions on PDE-3, presumably through the phosphorylation of PDE-3 by PI-3-kinase. Although we have shown that GC-C is a heregulin-b1-regulated protein, the actual erbB-signalling complex remains to be defined. Heregulins are ligands for the receptors ErbB3 and ErbB4,
but dimeric associations of the various ErbB proteins result in multiple potential signaling complexes. Since cultured adult hepatocytes express ErbB3 and EGFR (ErbB1), but not ErbB2 or ErbB4 [9,10], we hypothesize that the active ErbB signalling complex is mediated by dimer composed of ErbB3 (inactive kinase) and EGFR (active kinase). These studies indicate that ErbB receptor signalling in hepatocytes involves PDE-3 activation, as well as GC-C expression.
672
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
The functional role of GC-C in hepatocytes during liver regeneration, the acute phase response, or cell culture is not known. Since GC-C generates cGMP, an understanding of GC-C function is linked to that of cGMP, which is generated and broken down by several different cyclases and PDEs. Through its direct actions on protein kinase G, ligand-gated channels, and phosphodiesterases [22], cGMP may regulate assorted processes including cell volume regulation [23], gluconeogenesis [2], cell proliferation [24], and apoptosis [25,26]. Since agents capable of altering the turnover of cGMP (NOSi) or cAMP (PDE-3), also regulate GC-C expression, hormonally sensitive cyclic nucleotide feedback loops may regulate hepatocellular GC-C expression. The enzyme critical to the GC-C induction, PDE-3, is itself inhibited by cGMP, providing a potential mechanism for hepatocytes to resist the antigluconeogenic PDE-3-mediated actions of insulin at a postreceptor level. Similar cyclic nucleotide-dependent feedback loops mediate some actions of insulin in platelets and vascular smooth muscle cells [27]. Because compartmentalization of cyclic nucleotides can modulate their cellular actions, it will be important to define the precise location of GC-C in hepatocytes and its potential for dynamic movement in response to different stressors or hormonal stimuli. In summary, cultured hepatocytes in a defined, serumfree medium synthesize GC-C in response to insulin and heregulin-b1. The reported findings provide a foundation to analyze the hormonal and biochemical regulation of hepatic GC-C and to resolve the function (s) of this enzyme in these cells.
Acknowledgments The authors wish to acknowledge the expert technical assistance of Mary Stevenson and Kang-mei Chong. This work was supported by grants from the National Institutes of Health: DK45925 (L.A.S.) and DK53804 (W.E.R.).
References [1] Earp HS. J Biol Chem 1980;255:8979 – 82. [2] Patel TB, Nair BG, Padmini E, Rashed HM, Sun H. Hepatology 1995;21:1682 – 9. [3] Laney DWJ, Bezerra JA, Kosiba JL, Degen SJF, Cohen MB. Am J Physiol 1994;266:G899 – 906. [4] Scheving LA, Russell WE. Cancer Res 1996;56:5186 – 91. [5] Vaandrager AB, van der Wiel E, Hom ML, Luthjens LH, de Jonge HR. J Biol Chem 1994;269:16409 – 15. [6] Scheving LA, Jin WH, Chong KM, Gardner W, Cope FO. Am J Physiol 1998;43:C1363 – 72. [7] Balint JP, Kosiba JL, Cohen MB. J Recept Signal Transduction Res 1997;17:609 – 30. [8] Houslay MD. Cell Dev Biol 1998;9:161 – 7. [9] Carver RS, Sliwkowski MX, Sitaric S, Russell WE. J Biol Chem 1996;271:13491 – 6. [10] Carver RS, Mathew PM, Russell WE. Endocrinology 1997;138: 5195 – 201. [11] Scheving LA, Russell WE, Chong KM. Am J Physiol 1996;271: G959 – 68. [12] Scheving LA, Chong KM. Am J Physiol 1997;272:C1995 – 2004. [13] Mann EA, Jump ML, Giannella RA. Biochim Biophys Acta 1996; 1305:7 – 10. [14] Irvine F, Wallace AV, Sarawak SR, Houslay MD. Biochem J 1993; 293:249 – 253. [15] De Vera ME, Taylor BS, Wang Q, Shapiro RA, Billiar TR, Geller DA. Am J Physiol 1997;273:G1290 – 6. [16] Harbrecht BG, Billiar TR, Stadler J, Demetris AJ, Ochoa J, Curran RD, Simmons RL. J Leukocyte Biol 1992;52:390 – 4. [17] Gauldie J, Northemann W, Fey GH. J Immunol 1990;144:3804 – 8. [18] Cressman DE, Greenbaum RA, DeAngelis RA, Ciliberto G, Furth EE, Poli V, Taub R. Science 1996;274:1379 – 83. [19] Manganiello VC, Degerman E, Taira M, Kono T, Belfrage P. Curr Top Cell Regul 1996;34:63 – 100. [20] Hermsdorf T, Dettmer D. Cell Signal 1998;10:629 – 35. [21] Hermsdorf T, Richter W, Dettmer D. Cell Signal 1999;11:685 – 90. [22] Lincoln TM, Cornwell TL. FASEB J 1993;7:328 – 38. [23] Clemo HF, Baumgarten CM. Circ Res 1995;77:741 – 9. [24] Miyazaki M, Wahid S, Bai L, Namba M. Exp Cell Res 1992;200: 404 – 9. [25] Kim YM, de Vera ME, Watkins SC, Billiar TR. J Biol Chem 1997; 272:1402 – 11. [26] Wu CF, Bishopric NH, Pratt RE. J Biol Chem 1997;272:14860 – 6. [27] Trovati M, Anfossi G. Diabetologia 1998;41:609 – 22.
Insulin and heregulin-b1 upregulate guanylyl cyclase C expression in rat hepatocytes Reversal by phosphodiesterase-3 inhibition Lawrence A. Schevinga,*, William E. Russellb,c a
Division of Pediatric Gastroenterology and Nutrition, Vanderbilt University School of Medicine, Nashville, TN 37232, USA b Division of Pediatric Endocrinology, Vanderbilt University School of Medicine, Nashville, TN 37232, USA c Department of Cell Biology, Vanderbilt University School of Medicine, Nashville, TN 37232, USA Received 6 April 2000; accepted 28 March 2001
Abstract Guanylyl cyclase C (GC-C) is the receptor for the hormones guanylin and uroguanylin. Although primarily expressed in the rat intestine, GC-C is also expressed in the liver during neonatal or regenerative growth or during the acute phase response. Little is known about the hepatic regulation of GC-C expression. The influence of various hepatic growth or acute phase regulators on GC-C expression was evaluated by immunoblot analysis of protein from primary rat hepatocytes grown in a serum-free medium. Insulin and heregulin-b1 strongly stimulated GC-C expression by 24 h of cell culture. Several different hormones and agents suppressed this action, including transforming growth factor b (TGF-b), as well as inhibitors of phosphatidylinositol 3-kinase (PI-3-kinase) and phosphodiesterase 3 (PDE-3, an insulin- and PI-3-kinasedependent enzyme). The compartmental downregulation of cAMP levels by PDE-3 may be a critical step in the hormonal action that culminates in GC-C synthesis. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Guanylyl cyclase C; Guanylin; Uroguanylin; Phosphodiesterase; Insulin; Liver
1. Introduction Guanylyl cyclases (GCs) consist of a family of enzymes that form cyclic GMP (cGMP) from GTP. Very little is known about the regulation or role of GCs in the liver. Earp [1] originally reported 20 years ago that primary cultures of adult rat hepatocytes displayed increased GC activity when cultured for 22 h in the presence of insulin, glucagon, and cyclic AMP (cAMP). These hormones synergistically stimulated cyclase activity, but the identity of the respon-
Abbreviations: Guanylyl cyclase C (GC-C); Phosphatidylinositol 3kinase (PI-3-kinase); Phosphodiesterase 3 (PDE-3); Phosphodiesterase 4 (PDE-4); Interleukin-6 (IL-6); Epidermal growth factor (EGF); Transforming growth factor b (TGF-b); Heregulin-b1 (HRG-b1); Inducible NO synthase (NOSi) * Corresponding author. Division of Pediatric Endocrinology, T-0107 Medical Center North, Vanderbilt University School of Medicine, Nashville, TN 37232-2579, USA. Tel.: +1-615-343-0698; fax: +1-615-343-5845. E-mail address: [email protected] (L.A. Scheving).
sible enzyme could not then be determined. It is now known that the adult rat liver makes several different GCs [2 –4], including soluble GC, which is activated by nitric oxide, and Gc-A and Gc-B, which are the membrane-bound receptors for the atrial and brain natriuretic peptides. A third membrane-bound GC — guanylyl cyclase C (GC-C) — is developmentally regulated in the liver. It is normally synthesized by the late foetal and early neonatal rat liver and by small and large intestine epithelial cells irrespective of age. The adult rat liver does not express GC-C except under special conditions. These include liver regeneration following partial hepatectomy or during the acute phase response to tissue injury and infection [3]. The regulation of hepatic GC-C expression is important because the intestine releases the GC-C ligands — guanylin and uroguanylin — into the portal blood, creating a potential regulatory link between the gut and liver [3– 6]. To study the regulation and function of GC-C, it would be advantageous to identify a primary cell type capable of expressing GC-C in long-term culture. This cannot be
0898-6568/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 8 9 8 - 6 5 6 8 ( 0 1 ) 0 0 1 7 9 - 6
666
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
achieved with primary intestinal epithelial cells because they undergo rapid programmed cell death within several hours of culture [6]. Recently, others have shown that GC-C expression can be induced in primary hepatocyte cultures derived from adult rat liver [7]. Balint et al. showed that interleukin-6 (IL-6) induced GC-C expression in hepatocytes cultured in the presence of dexamethasone (10 4 M) and foetal calf serum (5%). In this paper, we examine the regulation of GC-C expression in primary hepatocytes in a serum-free medium containing a lower concentration of dexamethasone (10 8 M), focusing on hormones and agents that modulate hepatic GC activity, cell proliferation [3,4], or the acute phase response [3]. We find that insulin and heregulin-b1 each strongly stimulate GC-C expression after 24 h in culture. Glucagon and cAMP potentiate the insulin action, suggesting a molecular basis for their previously reported stimulation of hepatocellular GC activity. Interestingly, nitric oxide depletion (which lowers cGMP by decreasing the NO ligand for soluble GC) enhances the GC-C induction, whereas inactivation of a specific membrane-associated phosphodiesterase — phosphodiesterase 3 (PDE-3) (which is a cGMP inhibited, cAMP-PDE) — blocks it. Collectively, our findings implicate hormonally mediated cyclic nucleotide feedback loops in the mechanism by which hormones, such as insulin, and growth factors, such as heregulin-b1, upregulate hepatic GC-C expression. Our results also provide additional evidence for the importance of subcellular compartmentalization in mechanism of cyclic nucleotide action [8].
2. Materials and methods 2.1. Chemicals and reagents Bovine serum albumin (BSA), STa, tris, leupeptin, aprotinin, chymostatin, PMSF, dexamethasone, pyruvate, and percoll were from Sigma (St. Louis, MO). Ammonium persulfate, N 0, N 0 methylenebisacrylamide, polyacrylamide, N 0, N 0, N 0, N 0 tetra-ethylethylenediamine, protein standards, and Tween-20 were from Bio-Rad (Richmond, CA). Nitrocellulose (nitropure) was from Micron Separation (Westboro, MA). X-ray film (Hyperfilm) and horseradish peroxidase (HRP-linked secondary antibody (donkey antirabbit) were from Amersham (Arlington Heights, IL). Chemiluminescence solutions were from New England Nuclear Research Products (Boston, MA). Human recombinant HRG-b1 (amino acids 177 – 244) was from R&D Systems (Minneapolis, MN). Insulin was from Eli Lilly and Company (Indianapolis, IN). Recombinant rat IL-6 was from Research Diagnostics (Fanres, NJ). LY 294002, wortmannin, and PD 98059 were from Calbiochem (San Diego, CA). Cilostamide and trequinsin were from BIOMOL (Ply-
mouth Meeting, PA). Rolipram, SB203580, and rapamycin were from Sigma. 2.2. Culture media and supplies Serum-free Williams’ Medium E, supplemented with 20 mM pyruvate, 10 nM dexamethasone, and 50 mg/ml gentamicin, was the standard medium used for all culture studies. Medium was purchased from Life Technologies. Type I collagenase was from Waco Pure Chemical Industries (Richmond, VA). 2.3. Primary culture of hepatocytes Hepatocytes were isolated from the livers of male Sprague – Dawley rats (175 –250 g; Harlan Sprague – Dawley, Indianapolis, IN) with modifications of our previously described methods [9,10]. The livers of ether-anesthetized rats were perfused through the portal vein with a calciumfree solution consisting of 150 mM NaCl, 2.8 mM KCl, 5.5 mM glucose, and 25 mM HEPES (pH 7.6) for 10 min followed by the same solution containing 3.8 mM CaCl2, 10 mg/ml soybean trypsin inhibitor, and 0.5 mg/ml type I collagenase for 8 min. The cells were dispersed in medium supplemented with 10% calf serum and filtered through 61-mm nylon mesh. The hepatocytes were then purified by a 5-min sedimentation at 1 g in serumcontaining medium followed by centrifugation at 50 g in isotonic percoll (specific gravity = 1.06) to reduce contamination by nonparenchymal cells. Percoll was removed by two washes in serum-containing medium, and the hepatocytes were assessed for viability by trypan blue exclusion ( > 95% viable). Cells (1 106 cells/well) were plated in type-1 collagen-coated 60-mm wells. After a 60-min attachment period, the serum-containing medium was replaced with serum-free medium containing hormones, cytokines, amino acids, and signaling pathway inhibitors as indicated. 2.4. Hepatic and intestinal cell membrane preparation The washed cell pellet (or in some experiments, cells that had been snap-frozen in liquid N2 and stored at 70C) was immediately resuspended in 2 ml cell homogenization buffer (10 mM Tris, 1 mM EGTA, 40 mg/ml PMSF, 10 mg/ ml aprotinin, 10 mg/ml leupeptin, 10% glycerol) homogenized with a Tissue-Tearor (Biospec Products) for 30 s at setting 2. The homogenate had final protein concentrations between 3 and 6 mg/ml as determined by the Coomassie protein assay [4]. 2.5. Immunoblots GC-C antiserum was generated as previously described against the peptide CNNSDHDSTYF, coupled to keyhole limpet antigen at the C residue [4]. Samples were subjected
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
to SDS – PAGE, transblotted, and immunoblotted as previously described [11].
3. Results and discussion In this paper, we examined the ability of different hormones to regulate GC-C expression in primary hepatocytes derived from the adult rat liver. We evaluated hormones that regulate physiological states associated with increased GC-C expression, namely, hepatocyte proliferation (insulin, epidermal growth factor (EGF), heregulin-b1, transforming growth factor b (TGF-b), and interleukin-6) and the acute phase response (dexamethasone and interleukin-6). We also evaluated the independent and collective influence of insulin, glucagon, and cAMP on GC-C synthesis since they have been previously shown to increase GC activity in cultured hepatocytes. The hormonal results prompted us to test specific pharmacologic inhibitors (inducible NO synthase (NOSi), phosphatidylinositol 3-kinase (PI-3-kinase), MEK, p38 MAP kinase, and PDE-3 inhibitors) on GC-C expression as well. 3.1. Insulin induces GC-C in normal liver hepatocytes Because GC-C expression increases during periods of increased liver cell proliferation, such as regeneration and neonatal growth [3,4], we hypothesized that growth factor stimulation of hepatocellular proliferation in culture would increase GC-C expression. We cultured primary hepatocytes for 18 h in control medium containing different regulators of hepatic growth, including TGF-b (4 ng/ml), EGF (10 ng/ml), heregulin-b1 (30 ng/ml), and insulin (150 nM), and then analyzed GC-C expression by immunoblot (Fig. 1A). In this model, insulin and heregulin-b1 are weak mitogens, EGF, a strong one, and TGF-b, a growth inhibitor. For intestinal membrane proteins, an antibody against the C-terminal amino acid sequence of GC-C resolved on immunoblot three bands of varying electrophoretic mobility (Fig. 1A, rightmost lane). These bands represent the immature (130 kDa) and mature glycoforms (140 kDa), as well as a brush-border-specific, proteolytically processed, 85-kDa mature form [12]. In cultured hepatocytes, insulin strongly stimulated the 130- and 140-kDa isoforms, but there was no evidence for the 85-kDa form. Heregulin-b1, an EGF-like ligand for the receptors ErbB3 and ErbB4, weakly stimulated GC-C expression in this experiment; EGF had little or no effect; and TGF-b1 showed no immunoreactivity. Immunohistochemical experiments revealed that hepatocellular GC-C had a diffuse pattern of staining (data not shown), much of which localized to intracellular membranes, as reported previously during liver regeneration [4]. Time course experiments confirmed the effect of insulin and showed that heregulin-b1 stimulated GC-C to nearly the
667
same extent as insulin (Fig. 1B). These results also suggested that some GC-C appeared during culture even in the absence of insulin or heregulin-b1; however, the basal GC-C level varied from one hepatocyte preparation to the next and may have been triggered by exposure to 5% serum during the initial 60-min plating period. Since weak mitogens (insulin and heregulin-b1) strongly stimulated GC-C expression, whereas a strong one (EGF) had no effect, GC-C expression in cultured hepatocytes is not a direct consequence of cell cycle progression. Kinetically, insulin induced the immature 130-kDa and mature 140-kDa GC-C glycoforms with a sequence and time course that resembled that of the regenerating liver (Fig. 1C) [4]. Insulin (+; 100 nM) first induced the immature glycoform by 12 h, followed by the mature one between 12 and 24 h. Scanning laser densitometry revealed the GC-C signal in the insulin-treated samples to be 0, 1, 10, and 5.2 absorbance units at 0, 12, 24, and 60 h, respectively. Maximal GC-C expression was observed at 24 h of culture, when GC-C resolved into nearly equal amounts of the immature and mature glycoforms (we set 24 h of culture as the focal time point in subsequent experiments). By 60 h, the mature GC-C isoform predominated, the ratio of the mature to immature form being 50% higher at 60 h (1.8) compared to 24 h (1.2). Insulin stimulated GC-C expression over a broad concentration range (Fig. 1D). Near-maximal stimulation of expression was observed at a concentration as low as 15 nM. Increasing the concentration to 50 nM slightly increased total GC-C. Expression remained elevated at 150 nM. The concentration dependency curve indicates that the insulin action was physiologic since the actual concentration of insulin in these cultures is at least two orders of magnitude lower due to the presence of an insulin degrading enzyme and nonspecific binding of insulin to the culture plate (W.E. Russell, unpublished results). This is well within the concentration range reported for insulin in the portal blood. Since dexamethasone is a component of our serum-free culture medium and since there is a glucocorticoid response sequence in the 50-flanking region of the human GC-C gene [13], we examined the effect of dexamethasone on GC-C induction. Glucocorticoids modulated GC-C induction by insulin (Fig. 1E). Although insulin (100 nM) alone increased GC-C expression at 24 h, this effect was markedly enhanced by addition of dexamethasone (10 6 and 10 8 M; Fig. 1E) in a dose-dependent manner. Dexamethasone by itself was without effect. 3.2. Insulin, cAMP, glucagon, and dexamethasone synergistically induce GC-C Because glucagon and cAMP increased the GC activity of primary hepatocytes [1], particularly in the presence of insulin, we examined their ability to modulate GC-C expression. Both agents alone slightly increased GC-C
668
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
Fig. 1. Insulin induces GC-C protein expression in primary cultures of hepatocytes. (A) Hepatocytes were cultured in a serum-free medium (WE) containing dexamethasone (10 6 M) in the presence of TGF-b (4 ng/ml), EGF (10 ng/ml), insulin (150 nM), and heregulin-b1 (30 ng/ml). The cells were harvested 18 h later. Homogenates were prepared and immunoblotted using a monospecific GC-C polyclonal antibody. (B) Densitometric analysis of time course experiments showing GC-C levels in hepatocytes cultured in the presence of insulin (150 nm; open diamonds), heregulin-b1 (30 ng/ml; open circles), EGF (10 ng/ml;closed triangles), or WE (closed squares) at 12, 24, and 60 h after culture. Values are means of two experiments. (C) Time course experiments showing the two main glycoforms induced by insulin in primary hepatocytes at 12, 24, and 60 h of culture. (D) Concentration dependency of insulin at 24 h; note that insulin induces GC-C expression at concentrations as low as 15 nM. (E) Concentration dependency of dexamethasone on insulin action (100 nM).
but not to the levels attained by insulin treatment (Fig. 2A). For example, dibutryl cAMP at 100 and 200 mM stimulated GC-C expression to levels 20 – 25% that observed with insulin. Lower concentrations had little or no effect by themselves; however, the combination of insulin (100 nM), glucagon (100 nM), and dibutryl cAMP (5 mM) synergistically stimulated GC-C expression, increasing it in one experiment 16-fold over that caused by insulin alone. This was twice that achieved by insulin in the presence of the highest concentration of dexamethasone tested (10 6 M; Fig. 2B). Our results suggest that the previously reported increase in GC activity in response to insulin, glucagon, and dibutryl cAMP at the same concentrations requires de novo synthesis of GC-C (probably activated nonspecifically in the original report by the
Mn2 + in the enzyme assay buffer). This would be consistent with Earp’s observation that cycloheximide, a protein synthesis inhibitor, blocked the cyclase increase [1]. Further investigation is required to determine whether the potentiating effect of glucagon on insulin stimulated GC-C expression results from further elevations of cAMP or the additional effect of glucagon to raise intracellular Ca2 + concentrations [14]. 3.3. Insulin action is blocked by TGF-b and potentiated by and IL-6
L-NAME
Various cytokines and pharmacological agents were evaluated for their ability to modulate the induction of GC-C by insulin (Fig. 3). TGF-b1 was tested because it is
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
669
Fig. 2. Independent and synergistic induction of GC-C by glucagon and dibutryl cAMP. (A) Cells were treated for 24 h with a serum free media (WE) with or without glucagon or dibutryl cAMP at the indicated concentrations. Cell homogenates were prepared and immunoblotted as described in Materials and methods. (B) Cells were treated with dibutryl cAMP (5 mM), glucagon (100 nM), or both in the presence or absence of insulin (100 nM). All cells were treated with 10 8 M dexamethasone, except for the lane designated 10 6 M. Note increased GC-C expression in the presence of 10 6 M dexamethasone and a synergistic induction of GC-C by insulin, glucagon, and dibutryl cAMP in the presence of 10 8 M dexamethasone. (C) Quantitative densitometric analysis of experiments showing the individual and combined effects of cAMP (5 mM), glucagon (100 nM), and insulin (100 nM). Values are the means ± S.D. of GC-C protein levels. The combination of all three hormones on GC-C levels was different from the insulin treated hepatocytes ( P < .05, Student’s t test).
a potent inhibitor of hepatocyte growth and because it abolished the basal immunoreactivity seen in the original studies (Fig. 1). TGF-b1, at concentrations as low as 0.15 ng/ml, suppressed GC-C induction by insulin (100 nM). Maximal inhibition was seen at 5 ng/ml (Fig. 3A). Several agents capable of potentiating insulin’s upregulation of GC-C (i.e. dexamethasone, glucagon, and cAMP) also reduce NO production indirectly by inhibiting the synthesis of NOSi [15,16]. This enzyme makes nitric oxide, which is the only known activating ligand for soluble GC, a major source of cytoplasmic cGMP in cultured hepatocytes. Downregulation of NOSi results in a decreased production not only of NO but also of soluble GC-generated cGMP, the catalytic end-product of GC-C. We speculated that decreased NO production enhances GC-C expression as part of a feedback mechanism to normalize cellular cGMP levels. To test this idea, we cultured hepatocytes in the presence of L-NAME (1 mM), a slowly reversible inhibitor of the NOSi enzyme. As predicted, L-NAME treated cells increased their expression of GC-C, particularly in the presence of insulin (Fig. 3B). Notably, TGF-b (10 ng/ml) blocked the strong GC-C induction by L-NAME, insulin, and 10 6 M dexamethasone (rightmost lane of Fig. 3B).
IL-6 is an activator of the acute phase response [17] and may be required for proper liver regeneration [18]. Balint et al. [7] previously reported that IL-6 promotes GC-C expression in primary hepatocytes. Like L-NAME, we found that IL-6 (25 ng/ml) had little effect on GC-C expression by itself (Fig. 3B); however, when combined with insulin (100 nM) and L-NAME (1 mM), it potentiated the GC-C induction, particularly at the high dexamethasone concentration. These findings confirm the previous work but also suggest that the prior culture conditions, including a high concentration of dexamethasone (10 4 M), as well as 5% calf serum (which contains insulin), played a permissive role in the IL-6 action. 3.4. The action of insulin on GC-C requires both PI-3kinase and PDE-3 activation Some of the late actions of insulin in hepatocytes are mediated by PI-3-kinase activation [10]. To determine whether PI-3-kinase activation was required for GC-C induction, we incubated cells with two different selective inhibitors of this enzyme, wortmannin (100 nM) and LY294002 (50 mM). Both agents at the tested concentrations
670
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
Fig. 3. TGF-b inhibits the induction of GC-C, whereas IL-6 and L-NAME stimulate it. (A) TGF-b was added at the specified concentrations to cells along with insulin (100 nM) in a serum-free medium. Twenty-four hours later, homogenates were prepared and immunoblotted as previously described. Note that TGF-b inhibited the induction of GC-C by insulin (100 nM) in a dose-dependent manner. (B) IL-6 (25 ng/ml) and L-NAME (1 mM) were added along with insulin (100 nM) to hepatocyte cultures. Both agents synergistically increased the induction of GC-C by insulin at the low dexamethasone (10 8 M) concentration. TGF-b (10 ng/ml) decreased the induction of GC-C by L-NAME and insulin at the high dexamethasone concentration.
inhibited GC-C induction by insulin (150 nM), suggesting a requirement for PI-3-kinase activation (Fig. 4A). One of the PI-3-kinase-dependent enzymes that insulin rapidly activates in cultured hepatocytes is the cGMP inhibited, cAMP-phosphodiesterase (PDE-3) [19]. This enzyme, which causes cAMP hydrolysis, is also activated by glucagon and cAMP through other non-PI-3-kinasedependent pathways. For example, removal of dibutryl cAMP or insulin from cell culture medium causes a 30 – 40% decrease of PDE-3 activity within only 5 min [20]. Since PI-3-kinase was required for maximal GC-C induction, and since insulin, glucagon, and cAMP each activate PDE-3 to varying degrees, we hypothesized that PDE-3 inhibition would block GC-C induction. To test this, we treated cells with two different selective PDE-3 inhibitors, cilostamide and trequinsin (Fig. 4B). Several different concentrations were examined for each inhibitor (0.5, 5, and 25 mM). PDE-3 inhibition reduced the insulin-mediated GC-C induction in a dosedependent manner. The highest concentration tested (25 mM) completely blocked maximal GC-C induction by insulin (150 nM), glucagon (100 nM), and dibutryl cAMP (5 mM). This result suggests that the activation of PDE-3 plays a pivotal role in the hormonal induction of GC-C in rat hepatocytes. The results of this experiment prompted us to examine the effects of other pharmacological inhibitors. Since PDE-3 contributes to only about 30% of the PDE activity in hepatocytes, most of which is localized to cell membranes [21], we examined whether the inhibition of the largely
cytosolic phosphodiesterase 4 (PDE-4) similarly affected GC-C synthesis. PDE-4 is the other major PDE in primary hepatocytes [21]. To test this, we compared the effects of a nonselective PDE inhibitor (IBMX), two PDE-3 inhibitors (trequinsin and cilostamide), and a PDE-4 inhibitor rolipram). We also evaluated several agents known to block several other intermediary enzymes involved in insulin’s actions, including ERK (PD98059), p38 MAPK (SB203580), and p70 56K (rapamycin) (Fig. 4C). Both IBMX and the specific PDE-3 inhibitors had a more pronounced inhibitory effect on the GC-C induction than rolipram, suggesting that PDE-3 but not PDE-4 played the major role in the action of insulin (Fig. 4C). These results are consistent with the concept that the cellular actions of cyclic nucleotides derive from local, subcellular changes in their levels (a consequence of the relative balance of cyclases and PDEs at a particular subcellular site) and accessibility to binding proteins [8]. Moreover, although all of the enzyme inhibitors caused a variable decrease in GC-C expression, p38 MAPK inhibition with SB 203580 was notably effective in blocking the action of insulin (Fig. 4C). Still, the inhibition achieved by this agent at the tested concentration was less than that seen with PI-3-kinase or PDE-3 inhibition. Finally, since heregulin-b1 induced GC-C to the same extent as insulin, we examined whether the PDE and p38 MAPK inhibitors and TGF-b had the same inhibitory actions on heregulin-b1 as insulin. As shown in Fig. 4D, these inhibitors were as effective or more effective in inhibiting this heregulin-b1 action. Wortmannin, the PI-3-
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
671
Fig. 4. Interference with PI-3-kinase or PDE-3 blocks GC-C induction by insulin. (A) Cells were treated with insulin ± specific inhibitors of PI-3 kinase, including LY 294002 (50 mM) and wortmannin (100 nm). (B) Cells were treated with the specific inhibitors of PDE-3 — cilostamide or trequinsin — at several different concentrations. Note that both inhibitors blocked in a concentration-dependent manner the ability of insulin (150 nM) or insulin, glucagon (100 nM), and dibutryl cAMP (5 mM) to induce GC-C. (C) Cells were treated IBMX (2 mM), trequinsin (Treq.; 25 mM), cilostamide (Cilos.; 25 mM), rolipram (Roli.; 25 mM), PD98059 (30 mM), SB203580 (10 mM), or rapamycin (50 nM) for 24 h in the presence of insulin (150 nM). Homogenates were immunoblotted for GC-C. Two exposure times (Exp.times) differentiate the amount of inhibition by the PDE and p38 MAP kinase inhibitors. (D) Cells were treated with insulin (150 nm) or heregulin-b1 (30 nM) for 12 or 24 h in the presence or absence of IBMX (2 mM), TGF-b (5 ng/ml), or SB203580 (10 mM). Homogenates were immunoblotted for GC-C.
kinase inhibitor, was also effective in blocking the heregulin-b1-mediated GC-C induction (data not shown). Thus, both insulin and heregulin-b1, which interact through separate receptors, are capable of inducing GCC through their actions on PDE-3, presumably through the phosphorylation of PDE-3 by PI-3-kinase. Although we have shown that GC-C is a heregulin-b1-regulated protein, the actual erbB-signalling complex remains to be defined. Heregulins are ligands for the receptors ErbB3 and ErbB4,
but dimeric associations of the various ErbB proteins result in multiple potential signaling complexes. Since cultured adult hepatocytes express ErbB3 and EGFR (ErbB1), but not ErbB2 or ErbB4 [9,10], we hypothesize that the active ErbB signalling complex is mediated by dimer composed of ErbB3 (inactive kinase) and EGFR (active kinase). These studies indicate that ErbB receptor signalling in hepatocytes involves PDE-3 activation, as well as GC-C expression.
672
L.A. Scheving, W.E. Russell / Cellular Signalling 13 (2001) 665–672
The functional role of GC-C in hepatocytes during liver regeneration, the acute phase response, or cell culture is not known. Since GC-C generates cGMP, an understanding of GC-C function is linked to that of cGMP, which is generated and broken down by several different cyclases and PDEs. Through its direct actions on protein kinase G, ligand-gated channels, and phosphodiesterases [22], cGMP may regulate assorted processes including cell volume regulation [23], gluconeogenesis [2], cell proliferation [24], and apoptosis [25,26]. Since agents capable of altering the turnover of cGMP (NOSi) or cAMP (PDE-3), also regulate GC-C expression, hormonally sensitive cyclic nucleotide feedback loops may regulate hepatocellular GC-C expression. The enzyme critical to the GC-C induction, PDE-3, is itself inhibited by cGMP, providing a potential mechanism for hepatocytes to resist the antigluconeogenic PDE-3-mediated actions of insulin at a postreceptor level. Similar cyclic nucleotide-dependent feedback loops mediate some actions of insulin in platelets and vascular smooth muscle cells [27]. Because compartmentalization of cyclic nucleotides can modulate their cellular actions, it will be important to define the precise location of GC-C in hepatocytes and its potential for dynamic movement in response to different stressors or hormonal stimuli. In summary, cultured hepatocytes in a defined, serumfree medium synthesize GC-C in response to insulin and heregulin-b1. The reported findings provide a foundation to analyze the hormonal and biochemical regulation of hepatic GC-C and to resolve the function (s) of this enzyme in these cells.
Acknowledgments The authors wish to acknowledge the expert technical assistance of Mary Stevenson and Kang-mei Chong. This work was supported by grants from the National Institutes of Health: DK45925 (L.A.S.) and DK53804 (W.E.R.).
References [1] Earp HS. J Biol Chem 1980;255:8979 – 82. [2] Patel TB, Nair BG, Padmini E, Rashed HM, Sun H. Hepatology 1995;21:1682 – 9. [3] Laney DWJ, Bezerra JA, Kosiba JL, Degen SJF, Cohen MB. Am J Physiol 1994;266:G899 – 906. [4] Scheving LA, Russell WE. Cancer Res 1996;56:5186 – 91. [5] Vaandrager AB, van der Wiel E, Hom ML, Luthjens LH, de Jonge HR. J Biol Chem 1994;269:16409 – 15. [6] Scheving LA, Jin WH, Chong KM, Gardner W, Cope FO. Am J Physiol 1998;43:C1363 – 72. [7] Balint JP, Kosiba JL, Cohen MB. J Recept Signal Transduction Res 1997;17:609 – 30. [8] Houslay MD. Cell Dev Biol 1998;9:161 – 7. [9] Carver RS, Sliwkowski MX, Sitaric S, Russell WE. J Biol Chem 1996;271:13491 – 6. [10] Carver RS, Mathew PM, Russell WE. Endocrinology 1997;138: 5195 – 201. [11] Scheving LA, Russell WE, Chong KM. Am J Physiol 1996;271: G959 – 68. [12] Scheving LA, Chong KM. Am J Physiol 1997;272:C1995 – 2004. [13] Mann EA, Jump ML, Giannella RA. Biochim Biophys Acta 1996; 1305:7 – 10. [14] Irvine F, Wallace AV, Sarawak SR, Houslay MD. Biochem J 1993; 293:249 – 253. [15] De Vera ME, Taylor BS, Wang Q, Shapiro RA, Billiar TR, Geller DA. Am J Physiol 1997;273:G1290 – 6. [16] Harbrecht BG, Billiar TR, Stadler J, Demetris AJ, Ochoa J, Curran RD, Simmons RL. J Leukocyte Biol 1992;52:390 – 4. [17] Gauldie J, Northemann W, Fey GH. J Immunol 1990;144:3804 – 8. [18] Cressman DE, Greenbaum RA, DeAngelis RA, Ciliberto G, Furth EE, Poli V, Taub R. Science 1996;274:1379 – 83. [19] Manganiello VC, Degerman E, Taira M, Kono T, Belfrage P. Curr Top Cell Regul 1996;34:63 – 100. [20] Hermsdorf T, Dettmer D. Cell Signal 1998;10:629 – 35. [21] Hermsdorf T, Richter W, Dettmer D. Cell Signal 1999;11:685 – 90. [22] Lincoln TM, Cornwell TL. FASEB J 1993;7:328 – 38. [23] Clemo HF, Baumgarten CM. Circ Res 1995;77:741 – 9. [24] Miyazaki M, Wahid S, Bai L, Namba M. Exp Cell Res 1992;200: 404 – 9. [25] Kim YM, de Vera ME, Watkins SC, Billiar TR. J Biol Chem 1997; 272:1402 – 11. [26] Wu CF, Bishopric NH, Pratt RE. J Biol Chem 1997;272:14860 – 6. [27] Trovati M, Anfossi G. Diabetologia 1998;41:609 – 22.