Abstract
Presynaptic, cocaine- and antidepressant-sensitive norepinephrine (NE) transporters (NETs) dictate levels of extracellular NE after vesicular release. Recent studies suggest that G protein-coupled receptors linked to protein kinase C (PKC) down-regulate cell surface NET protein levels and diminish NE uptake capacity. We identified distinct phosphatidylinositol 3-OH kinase (PI3K)-linked pathways supporting basal and insulin-triggered NE transport in the human noradrenergic neuroblastoma, SK-N-SH. Acute (0–60 min) insulin treatments produced a time- and concentration-dependent stimulation of NE transport, resolved in kinetic studies as an enhancement of NE transport capacity (Vmax) without an alteration in NEKm. Basal and insulin-modulated NET activities were reduced by the tyrosine kinase inhibitor genistein and the PI3K inhibitors wortmannin and LY-294002, but not by the PKC inhibitor staurosporine. PI3K activation was found to support phosphorylation of p38 mitogen-activated protein kinase (p38 MAPK). However, basal and insulin-stimulated NET activities were differentiated by their reliance on p38 MAPK activation. Thus, the p38 MAPK inhibitor SB203580 and SB202190 abolished insulin activation of NE transport yet failed to impact basal NET activity. Moreover, p38 MAPK activation and insulin activation of NETs were found to be sensitive to external Ca2+ depletion, blockade of voltage-sensitive Ca2+ channels, and inhibition of protein phosphatase 2A. Effects of tyrosine kinase and PI3K inhibitors on basal NET uptake appear to arise from a loss of cell surface NET protein, whereas the p38 MAPK-dependent enhancement of NE transport occurs without a detectable enhancement of surface NET. Our findings establish two distinct pathways for regulation of NE uptake involving PI3K, one linked to transporter trafficking and a second linked to Ca2+-dependent, p38 MAPK phosphorylation that promotes activation of cell surface NETs.
At peripheral and central noradrenergic synapses, the effects of norepinephrine (NE) are terminated primarily by active reuptake of the catecholamine via a neuronal membrane transporter (Barker and Blakely, 1995). The NE transporter (NET) is a member of the Na+- and Cl−-coupled cotransporter gene family, which includes the transporters for biogenic monoamines (dopamine, NE, and serotonin) and transporters for amino acids GABA and glycine (Barker and Blakely, 1995). Tricyclic antidepressants and psychoactive agents, including desipramine, amphetamine, and cocaine act on NET and modulate noradrenergic neurotransmission in the CNS and peripheral nervous system (Barker and Blakely, 1995). The importance of biogenic monoamine transporters in the control of neurotransmitter clearance is well illustrated by the diminished rate of dopamine clearance in dopamine transporter knockout mice and comparable findings in NET knockout mice (Giros et al., 1996; Xu et al., 2000). Impaired NET function or expression has been reported in cardiac failure, diabetic cardiomyopathy (Ganguly et al., 1986), and hypertension (Barker and Blakely, 1995). In the CNS, disorders such as depression (Klimek et al., 1999) may involve changes in NET activity. Recently, we identified a loss of function mutation in NET associated with symptoms of orthostatic intolerance (Shannon et al., 2000). These findings underscore the need to understand the pathways supporting NET regulation.
Similar to other members of the Na+-/Cl−-cotransporter gene family, NETs can be regulated by neuronal activity, neurotransmitters, peptide hormones, and second messengers (Vatta et al., 1991; Apparsundaram et al., 1998a,b). Recently, we reported evidence for both PKC-dependent and -independent regulation of NE transport. Specifically, we have shown that activation of M3 muscarinic acetylcholine receptors in SK-N-SH cells reduces NE transport by reducing the surface density of NETs, providing the first evidence for altered trafficking of NETs by G protein-coupled receptors (Apparsundaram et al., 1998a,b). Interestingly, although PKC inhibitors block the down-regulation of NETs by phorbol esters, muscarinic receptor effects on NE transport are only partially attenuated. These findings suggest the existence of PKC-independent regulatory pathways supporting NET surface expression and/or transporter activation.
Hormones and trophic factors acting through tyrosine kinase-linked receptors have also been implicated in NET regulation. Autonomic dysfunction seen in conditions of insulin deficiency and hyperinsulinemia is associated with impaired noradrenergic function (Ferrari and Weidmann, 1990; Moreau et al., 1995). Insulin has been shown to alter NE release and clearance (Bhagat et al., 1981;Christensen, 1983; Shimosawa et al., 1992; Townsend et al., 1992), and activation of insulin receptors (IRs) alters NE uptake in PC12 cells and hypothalamic slices (Boyd et al., 1986; Figlewicz et al., 1993). In the present study, we evaluate the action of insulin on NE transport in SK-N-SH cells. We provide evidence that NET surface expression is dependent on a phosphatidylinositol 3-OH kinase (PI3K)-linked pathway that is active under basal conditions and that an additional pathway, also PI3K-dependent, leads to alterations in transporter catalytic activity. We discuss our findings in the context of a model whereby the resting tone of NE clearance is established through the modulation of NET surface trafficking and the intrinsic activation of surface-resident NET proteins.
Experimental Procedures
Materials.
Reagents used to modify receptors and second messengers were obtained from the following sources: insulin and IGF-1 (Collaborative Research, Bedford, MA); actinomycin D, agatoxin, ω-conotoxin, cycloheximide, genistein, genistin, pargyline, staurosporine, and verapamil (Sigma Chemical, St. Louis, MO); cyclosporin A, calyculin A, okadaic acid, norokadone, LY-294002, microcystin LR, SB202190, tautomycin, and wortmannin (Alexis Biochemicals, San Diego, CA); β-PMA, BAPTA-AM, PD 908059, SB203580, and SB202474 (Calbiochem, San Diego, CA); and U-0521 (Upjohn, Kalamazoo, MI). Sulfo-NHS-biotin and immobilized monomeric avidin beads were obtained from Pierce Chemical (Rockford, IL). Fura-2 acetoxymethyl ester (fura-2 AM) and pluronic acid were purchased from Molecular Probes (Eugene, OR). l-[7,8-[3H]NE (37 Ci/mmol), l-[3H]glycine, and [N-methyl-3H]nisoxetine (86 Ci/mmol) were obtained from Amersham Pharmacia Biotech (Arlington Heights, IL). Monoclonal NET-17 antibody was obtained from MAb Technologies (Atlanta, GA), rabbit polyclonal phospho-specific protein kinase B (PKB/Akt, Ser 473) and p38 MAPK antibody were obtained from New England Biolabs (Beverly, MA), whereas the polyclonal goat antibody for detecting total PKB/Akt (Akt1/2) was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Other reagents were of analytical purity and were obtained from standard sources.
Cell Culture and NE Uptake Assays.
SK-N-SH cells (American Type Culture Collection, Manassas, VA) were maintained in culture medium containing RPMI 1640 supplemented with 10% fetal bovine serum, 2 mM glutamine, 100 IU/ml penicillin, and 100 g/ml streptomycin. Cells were plated at 300,000 cells/well in 24-well plates and incubated with culture medium for 24 h. Then culture medium was removed and cells were incubated with serum-free medium (RPMI 1640 supplemented with 2 mM glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin) for 24 h. For uptake assays, serum starvation cells were washed with 2 ml of Krebs-Ringer-HEPES (KRH) buffer containing 130 mM NaCl, 1.3 mM KCl, 2.2 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, and 10 mM HEPES, pH 7.4). Cells were then equilibrated in assay buffer (KRH, 10 mMd-glucose, 100 μM pargyline, 10 μM U-0521, and 100 μM ascorbic acid) at 37°C for 10 min. After the equilibration period, cells were incubated in assay buffer containing either modulating agents or appropriate vehicle as described in figure legends. NE transport assays were initiated by the addition of [3H]NE for 10 min at 37°C and terminated by three rapid washes with ice-cold KRH buffer. In some experiments, cells were treated with either 10 μM actinomycin D (20 min) or 10 μM cycloheximide (2 h) prior to the addition of modulating agents. After incubation with labeled substrate, cells were lysed in 1 ml of Optiphase Supermix scintillation cocktail (Wallac, Gaithersburg, MD) or 0.5 ml of 1% SDS and accumulated radioactivity quantified using Microbeta counter (Wallac) or scintillation spectrophotometer (Beckman Coulter, Inc., Fullerton, CA). An aliquot of the SDS extract was used to determine protein level (Bradford assay; Bio-Rad, Hercules, CA). Nonspecific [3H]NE uptake was determined using 1 μM desipramine and was subtracted from total uptake to define hNET specific accumulation. Nonlinear curve fits of saturation data (KaleidaGraph; Synergy Software, Reading, PA) used the Michaelis-Menten model V =Vmax[S]n/[S]n+ [K]n. All uptake assays were performed in triplicate and each experiment repeated at least three times. Statistical analyses were performed comparing mean transport or mean percentage values derived from three experiments. Student'st test was used for statistical analysis of data involving comparisons between two groups, whereas one-way analysis of variance followed by Tukey's test was used for deducing statistical differences in experiments involving multiple comparisons (GraphPad Software, San Diego, CA). A level of significance of 0.05 was used in all statistical analysis.
Evaluation of hNET Surface Density by Using Radioligand Binding Assays.
To assess changes in hNET surface density in SK-N-SH cells, and to evaluate direct interaction of modulating agents on NETs, radioligand binding assays were carried out in both intact SK-N-SH cells as well as isolated membrane fractions by using [3H]nisoxetine as described previously (Apparsundaram et al., 1998a). Briefly, cells were incubated with modulating agents as described above and washed with ice-cold binding buffer (100 mM NaCl, 50 mM Tris, 100 μM ascorbic acid, pH 8). Cells were incubated with [3H]nisoxetine (0.01–10 nM) in ice-cold binding buffer at 4°C for 2 h. Assays were terminated using ice-cold binding buffer. Cell extracts were prepared with 0.5 ml of 1% SDS and bound radioactivity quantified by scintillation counting (Beckman Coulter, Inc.). A portion of the cell extract was used to quantify protein content (Bradford assay; Bio-Rad). Nonspecific binding was determined as the whole cell binding evident in the presence of 100 μM dopamine and subtracted from total counts to identify surface and NET specific labeling. For radioligand binding in membranes, cells were washed with ice-cold phosphate-buffered saline and then scraped off the dishes in ice-cold PBS. Cells were pelleted at 1600g, and the pellet homogenized with 3 ml of ice-cold binding buffer with a Polytron (Brinkman, Westbury, NY) at 25,000 rpm for 5 s. Homogenate was centrifuged (20,000g; 20 min; 4°C) and the pellet resuspended in ice-cold binding buffer. An aliquot of sample was used for protein determination by the Bradford method (Bio-Rad). Eighty micrograms per tube of membrane protein fraction was incubated with [3H]nisoxetine (0.01–10 nM) for 4 h at 4°C. For the determination of direct interaction of modulating agents on hNETs, membranes were incubated with modulating agents at 37°C prior to incubation with 5 nM [3H]nisoxetine at 37°C for 1 h in KRH buffer. In all experiments, membrane-bound radioactivity was recovered on GF/B glass-fiber filters (Whatman, Clifton, NJ) and bound radioactivity determined by liquid scintillation counting. Nonspecific binding was determined using 10 μM desipramine. All binding assays were performed in triplicate and statistical analyses were performed comparing mean binding values of at least three experiments by using GraphPad Prism software.
Immunoblotting of hNETs and Signal Transduction Complexes.
We also assessed changes in cell surface distribution of hNETs in SK-N-SH cells by using the cell surface biotinylation method (Apparsundaram et al., 1998b). Briefly, cells were cultured in six-well plates (1 million cells/well) for 48 h, serum-starved, and treated with insulin as described above. Under these conditions, insulin stimulates NE transport similar to that seen in 24-well plate assays (data not shown). After drug treatment, cells were washed quickly with KRH and then treated with sulfosuccinimidobiotin (1.5 mg/ml) at 4°C for 1 h in PBS/Ca2+-Mg2+. The biotinylating reagent was quenched by incubation with 100 mM glycine in PBS/Ca2+-Mg2+ for 30 min and then cells were washed with PBS/Ca2+-Mg2+ before lysis with 350 μl/well radioimmunoprecipitation assay (RIPA) buffer supplemented with protease inhibitors (1 μM pepstatin A, 250 μM phenylmethylsulfonyl fluoride, 1 μg/ml of leupeptin, and 1 μg/ml aprotinin) for 1 h at 4°C with constant shaking. Lysates were centrifuged at 20,000g for 30 min at 4°C and supernatant incubated with monomeric avidin beads for 1 h at room temperature. Beads were washed three times with RIPA buffer and absorbed proteins eluted with 50 μl of Laemmli buffer (62.5 mM Tris, pH 6.8, 20% glycerol, 2% SDS, 5% β-mercaptoethanol, and 0.01% bromophenol blue) for 30 min at room temperature. Eluted proteins were separated by SDS-gel electrophoresis (10%) and transferred to Immobilon-P membrane (0.45-μm pore size; Millipore Corporation, Bedford, MA) and immunoblotted with NET-17 monoclonal antibody (1:1000) and sheep-antimouse peroxidase-conjugated secondary antibody (1:3000). Immunoreactive bands were visualized by enhanced chemiluminescence on Hypersensitive film according to the manufacturer's protocol (Amersham Pharmacia Biotech).
For immunoblotting of total and phosphorylated PKB/Akt, insulin-treated cells were washed with PBS and extracted with buffer containing 10 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100 supplemented with protease inhibitors, and phosphatase inhibitors (10 mM sodium fluoride, 50 mM sodium pyrophosphate, and 1 μM okadaic acid) for 1 h at 4°C with agitation. Extracts were centrifuged at 20,000g for 30 min at 4°C, protein content of supernatant was determined using the bicinchoninic acid reagent (Pierce Chemical), and an aliquot of sample containing 200 μg/well was mixed with Laemmli sample buffer, incubated for 20 min at 22°C, and then resolved by SDS-PAGE, proteins transferred and immunoblotted with primary antibodies (1:1000) recognizing total and phosphorylated PKB/Akt.
Activation of p38 MAPK was determined as previously described (Sweeney et al., 1999). Briefly, cells were treated with insulin in the presence and absence of SB203580. After treatment, cells were lysed in RIPA buffer supplemented with protease inhibitors. Cell lysates were centrifuged at 20,000g as indicated above. Anti-phosphotyrosine antibody was then added to supernatants and incubated overnight at 4°C, followed by addition of 3 mg of protein A-Sepharose for 1 h at room temperature. The beads were washed three times with lysis buffer and solubilized in 50 μl of Laemmli sample buffer and eluted proteins separated by SDS-PAGE, transferred and immunoblotted with p38 MAPK polyclonal antibody (1:1000), and detected by enhanced chemiluminescence. A portion of total cell extracts was processed in parallel to reveal total immunoreactive p38 MAPK.
Measurement of Ca2+ Flux in Response to Insulin and Muscarinic Receptor Activation.
Cells were plated on 35-mm tissue culture plates, incubated in serum-free Dulbecco's modified Eagle's medium overnight, and then loaded with fura-2 by incubation in 0.5 mM fura-2-AM in Hanks' balanced salt solution (HBSS) containing 10 mM HEPES, 140 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 5 mM d-glucose for 60 min at room temperature, followed by two washes with HBSS. Cells were superfused with HBSS, or insulin, or muscarine. [Ca2+]i was measured in individual cells by dual-wavelength spectrofluorometry with a Nikon inverted microscope attached to a Compix Calcium Imaging System consisting of a charge-coupled device camera (Dage-MTI CCD-72; Michigan, IN) attached to an IBM compatible computer executing SIMCA C-Imaging software (Compix, Cranberry Township, PA). Cells were exposed to excitation wavelengths of 340 and 380 nm every 2 s, and the emitted fluorescence was measured in real time at 510 nm. The ratio of emission at 340- and 380-nm excitation was used as an index of [Ca2+]i.
Results
Insulin Enhances NE Transport in SK-N-SH Cells.
Treatment of SK-N-SH cells with insulin produced a time- and concentration-dependent enhancement of NE uptake (Fig. 1, A and B). A significant increase in NE uptake was observed by 30 min after the initiation of insulin (3 nM) treatment, reaching a maximal increase of 51 ± 6% at 60-min exposure to 100 nM insulin. SK-N-SH cells also possess high-affinity IGF-1 receptors to which IGF-1 binds with aKd of ∼0.5 nM. Because insulin and IGF-1 receptors are structurally similar, and both IGF-1 and insulin can activate each other's receptor (Ota et al., 1988b), we tested whether the stimulation of NE transport was specific to insulin. We found that IGF-1 does not alter NE uptake in SK-N-SH cells across a similar time and concentration range (Fig. 1, A and B). Kinetic analyses indicate that insulin significantly enhancedVmax [control = 3.61 ± 0.46 pmol/106 cells/min; insulin (10 nM; 60 min) = 4.48 ± 0.50 pmol/106 cells/min;p < 0.05, Student's paired t test] of NE transport without altering Km(control = 486 ± 40 nM; insulin = 451 ± 55 nM) for NE (Fig. 1C). Whereas the effects of insulin on [3H]NE transport were very reproducible, the peptide produced no change in Na+- and Cl−-dependent glycine transport assayed in parallel (data not shown). Thus, the effect of insulin on NE transport is also not likely due to global changes in membrane potential or ion gradients. Although insulin can increase glucose uptake in SK-N-SH, the effect of insulin on NE transport appears to not be secondary to changes in glucose transport because insulin increases NE transport to a similar magnitude regardless of the concentration of glucose in the assay buffer (data not shown). Transcription and protein synthesis inhibitors, cycloheximide and actinomycin D, failed to alter insulin stimulated NET activity (NE uptake as % of control: control: vehicle = 100%, actinomycin D (10 μM; 20 min) = 102 ± 3%; cycloheximide (10 μM; 2 h) = 101 ± 5%; insulin (10 nM; 60 min): vehicle = 127 ± 5%; insulin + actinomycin D = 123 ± 3%∗; insulin + cycloheximide = 126 ± 4%∗; p < 0.05; ANOVA, Tukey's test).
Involvement of Tyrosine Kinase and PI3K in Control of Basal NET Activity.
To explore further the signaling pathways involved in insulin-mediated activation of NE uptake, we first evaluated NET activity in the presence and absence of tyrosine kinase and PI3K-modulating agents. First, we established the activity of constitutive tyrosine kinase and PI3K-linked pathways supporting basal NET activity. The tyrosine kinase inhibitor genistein produced a concentration- and time-dependent decrease in NE uptake (Fig.2, A and B). A significant decrease in NE uptake is apparent after incubation of cells with 30 μM genistein for 20 min, which reached 40% inhibition by 1 h. Kinetic analysis of NE transport indicated that genistein significantly reducedVmax [control = 3.20 ± 0.26 pmol/106 cells/min; genistein (30 μM; 30 min) = 1.68 ± 0.36 pmol/106 cells/min;p < 0.05, Student's paired t test] with no significant effect on NE Km(control = 355 ± 30 nM; genistein = 382 ± 33 nM). On the other hand, the inactive analog of genistein, genistin, failed to alter to NE transport in parallel experiments (data not shown).
PI3K is known to participate in the insulin-dependent modulation of glucose transport (Shepherd et al., 1998). In SK-N-SH cells, the PI3K inhibitor wortmannin produced a time- and concentration-dependent reduction in NE transport (Fig. 3, A and B). Wortmannin (100 nM) after a 50-min incubation produced a 38 ± 4% decrease in NE uptake (Fig. 3, A and B). A structurally distinct PI3K inhibitor, LY-294002, had similar effects on NE transport as wortmannin (Fig. 3, C and D). Thus, 30 μM LY-294002 produced a 37 ± 6% reduction in NE transport after 50-min incubation. The effects of genistein, wortmannin, and LY-294002 on NE transport could be mediated by a nonspecific disruption of ion gradients required for transport or by binding of these agents to NETs directly, compromising NE recognition. However, these agents alter neither the Na+-/Cl−-dependent uptake of glycine measured in intact SK-N-SH cells (data not shown) nor [3H]nisoxetine binding measured in membrane fractions of SK-N-SH cells (Table 1). Together, these results indicate that basal NE transport in SK-N-SH cells is supported by constitutive tyrosine kinase and PI3K activity.
PI3K Is Required for Insulin-Stimulated NET Activity.
To validate increased activation of PI3K by insulin in SK-N-SH cells, we examined the phosphorylation of PKB/Akt, a downstream target of PI3K by using a PKB/Akt phosphospecific antibody. We find that insulin (10 nM) produced a time-dependent enhancement of PKB/Akt phosphorylation (Fig.4A) without altering the total PKB/Akt levels. Wortmannin abolished insulin-evoked PKB/Akt phosphorylation (Fig. 4B), confirming that PI3K activation is required for PKB/Akt activation. To examine whether the action of insulin on NE transport is similarly supported by tyrosine kinase and PI3K-linked pathways, we treated cells with insulin before and after exposure to genistein, wortmannin, or LY-294002. As previously noted, pretreatment of cells with genistein, wortmannin, or LY-294002 for 60 min produced significant reduction in NE transport (Fig.5). More importantly, the presence of genistein, wortmannin, and LY-294002, abolished insulin's stimulation of NE uptake. In contrast, the protein kinase C inhibitor staurosporine, at concentrations that effectively block β-PMA-mediated activation of PKC (Apparsundaram et al., 1998a), failed to alter either basal or insulin-stimulated NE transport. Additionally, chronic β-PMA treatment, under conditions that abolished acute β-PMA-mediated inhibition of NE transport presumably by down-regulation of PKC (Apparsundaram et al., 1998a), failed to alter insulin-stimulated NET activity (insulin-stimulated NE transport as percentage of control: vehicle-treated = 124 ± 3%; β-PMA-treated = 126 ± 2%). Together, these results indicate that insulin's augmentation of NE transport involves a PI3K-linked pathway, a pathway distinct from that supporting PKC-dependent NET modulation.
p38 MAPK Is Required for Insulin-Enhanced NET Activity.
The lack of involvement of a PKC-dependent trafficking pathway in NET regulation suggested to us that catalytic function rather than trafficking might be involved in insulin regulation. Notably, insulin increases the intrinsic activity of GLUT4 via p38 MAPK (Sweeney et al., 1999) whose activation is supported by PI3K (Chun et al., 2000;Mockridge et al., 2000). To gain evidence for a role of p38 MAPK in NET expression and regulation, we tested the specific p38 MAPK inhibitor SB203580 and its inactive analog SB202474 for their ability to influence basal and insulin-stimulated NE transport. Unlike PI3K inhibitors, SB203580 (10 μM; 60 min) does not affect basal NE transport. However, the compound abolished insulin's ability to increase NET activity (Fig. 6A). Moreover, the inactive analog SB202474 failed to alter either basal or insulin-stimulated NE transport, consistent with a specific role of p38 MAPK in insulin actions. Another potent p38 MAPK inhibitor SB202190 also produced effects comparable with SB203580 (control 100%; SB202190 = 97 ± 4%; insulin = 128 ± 3%∗; insulin + SB202190 = 102 ± 7%, *p < 0.05; ANOVA, Tukey's test). On the other hand, the ERK1, ERK2 (p42,44) inhibitor PD 908059 failed to alter basal or insulin modulated NET activity (data not shown), consistent with a specific role for the p38 MAPK pathway in NET regulation. To validate that insulin activates p38 MAPK, we used phosphospecific antibodies that recognize the activated form of p38 MAPK. Immunoblot analyses reveal that insulin enhances the basal phosphorylation of p38 MAPK (Fig. 6B), and this activation is inhibited by SB203580. The PI3K inhibitors wortmannin and LY-294002 also inhibit p38 MAPK phosphorylation (data not shown). SB203580 also abolished insulin-evoked phosphorylation of PKB/Akt without altering total PKB/Akt (Fig. 6C). These results suggest that PI3K leads to phosphorylation and activation of p38 MAPK, which then participates in both PKB/Akt activation and NET regulation.
External Ca2+ Is Required for Insulin-Mediated Effects on NET Activity.
Previously (Apparsundaram et al., 1998a), we reported that acute pretreatment of SK-N-SH cells with the membrane-permeant Ca2+ chelator BAPTA-AM diminished basal NE uptake, suggesting a role for Ca2+ in supporting the activity and/or surface expression of NETs. Interestingly, when cells were pretreated with BAPTA-AM (in the presence of external Ca2+, the insulin-mediated enhancement in NE uptake was abolished (Fig.7A). To explore this issue further, we examined the effects of insulin in the presence or absence of external Ca2+ (2.2 mM) in the assay buffer. Removal of external Ca2+ in this manner produced a significant reduction in insulin-mediated stimulation of NE uptake (Fig. 7B). Next, we used the sarcolemmal Ca2+pump inhibitor thapsigargin in the absence of external Ca2+ to deplete internal Ca2+ and then tested the effects of insulin. Thapsigargin treatment in the absence of external Ca2+, like depletion of external Ca2+ alone, blocked insulin's action (data not shown). However, restoring external Ca2+ to thapsigargin-treated cells restored the ability of insulin to stimulate NE transport activity (Fig. 7C). The latter findings suggested that insulin's actions on NE transport are supported by basal or insulin-triggered influx of Ca2+ across the plasma membrane. To test this hypothesis, we first examined the effect of Ca2+ channel blockers on insulin-stimulated NET activity. Blockade of voltage-gated Ca2+channels by verapamil (L-type), ω-conotoxin (N-type), and agatoxin (P- and Q-type) produced a nonsignificant decrease in NE transport (Fig. 7D). More importantly, in the presence of these agents, insulin lost efficacy to enhance NE transport (Fig. 7D). Further studies reveal that coapplication of verapamil and ω-conotoxin are sufficient to abolish insulin-stimulated NE transport (data not shown), suggesting that Ca2+ influx via L- and N-type Ca2+ channels supports insulin-mediated stimulation of NET activity.
The Ca2+ dependence of insulin's action could involve either a basal requirement for cytosolic Ca2+ or arise from insulin activation of Ca2+ influx. We asked whether insulin could trigger an intracellular Ca2+ elevation in SK-N-SH cells, measuring changes in intracellular Ca2+ by using ratiometric imaging of the Ca2+-sensitive probe fura-2 AM. As a positive control, we used methacholine, a muscarinic agonist that elicits a calcium elevation through M3 receptors in these cells. Under conditions in which we readily detect an elevation of triggered [Ca2+]i by the M3 receptor agonist methacholine, insulin (10 nM) failed to induce changes in [Ca2+]i (Fig.8). These findings indicate that insulin's activation of NE transport is most likely supported by constitutive Ca2+ entry via voltage-dependent Ca2+ channels rather than arising from direct Ca2+ channel activation.
The involvement of Ca2+ in insulin's regulation of NET activity could occur at multiple levels. To test whether Ca2+ is required upstream or downstream of p38 MAPK, we reevaluated the ability of insulin to trigger p38 MAPK phosphorylation in the presence of Ca2+ channel blockers. Pretreatment of cells with verapamil, ω-conotoxin, and agatoxin abolished insulin-mediated activation p38 MAPK (Fig.9) consistent with a role of ion channel-supported cytosolic Ca2+ in p38 MAPK activation. However, M3 receptor stimulation that triggers an increase in cytosolic Ca2+ from intracellular stores fails to alter p38 MAK phosphorylation (data not shown). These findings are consistent with the presence of distinct pathways supporting NET regulation by M3receptors and insulin, differentially influenced by tonic, ion channel-supported Ca2+ influx and G protein-coupled receptor-elevated [Ca2+]i arising from intracellular stores.
PP2A Is Required for Insulin-Triggered Stimulation of NET.
Recently, we have shown that endogenous NET proteins in rat vas deferens exist as physical complexes with the catalytic subunit of protein phosphatase 2A (PP2Ac) in an okadaic acid-sensitive manner (Bauman et al., 2000). Studies by Westermarck et al. (2001) reveal that activated p38 MAPK triggers an increase in PP2A activity in human fibroblasts, an effect blocked by the p38 MAPK inhibitor SB203580 and the PP1/2A inhibitors okadaic acid and calyculin A. If a similar scenario is involved in insulin regulation of NET, we would expect that PP1/2A antagonists would also block insulin's regulation of NET. As seen in vas deferens, pretreatment of SK-N-SH cells with okadaic acid, but not the inactive analog norokadone, produced a significant reduction in basal NE transport (Fig.10). More importantly, okadaic acid (and not norokadone) abolished insulin's elevation of NE transport. A similar effect on insulin actions was seen with cells treated with calyculin A (Fig. 10). However, the PP1-selective inhibitor tautomycin and the PP2B inhibitor cyclosporin failed to alter either basal or insulin-stimulated NE transport (data not shown). These findings are consistent with a requirement for, and possible activation of PP2Ac, in the p38 MAPK pathway supporting NET regulation.
Insulin-Evoked Stimulation of NET Activity Is Not Accompanied by Changes in Cell Surface NET Abundance.
Changes in NE transport capacity could arise from an increase in the number of carrier molecules in the plasma membrane or an increase in the intrinsic activity of surface resident NETs, or both. To evaluate these possibilities, we used whole cell radioligand binding analyses with the NET-specific ligand [3H]nisoxetine and cell surface biotinylation approaches to document the impact of insulin treatments on NET surface expression. As previously described (Apparsundaram et al., 1998a), β-PMA treatment (1 μM; 30 min) produces a 25 ± 2% reduction in NE transport activity and a similar reduction of [3H]nisoxetine binding in intact cells (Fig. 11A). Similarly, genistein, wortmannin, and LY-294002 also produce a significant decrease in [3H]nisoxetine binding (Table 1), suggesting that constitutive tyrosine kinase and PI3K-linked pathways influence NET surface expression and provide for basal NE transport capacity. In contrast, under conditions where insulin causes a significant increase in NE uptake, no change in [3H]nisoxetine binding is evident in intact cells (Fig. 11A). Moreover, kinetic analysis of [3H]nisoxetine binding isotherms reveals no change in Kd andBmax of radioligand binding in intact cells (Table 2).
Although SK-N-SH cells have low levels of NETs, the recent availability of a sensitive monoclonal NET antibody permits the determination of NETs in this model by immunoblots. Using this approach, we assessed the effect of insulin and β-PMA on the population of hNETs accessible to the membrane-impermeant biotinylation reagent sulfo-NHS-biotin. In SK-N-SH cells, both ∼81- and ∼54-kDa species were detected in immunoblots of total extracts (Fig. 11B). The 81-kDa species constitutes approximately ∼75% of total hNET protein with approximately 40% of the total 81-kDa protein recovered in the plasma membrane fractions. Less than 5% of the total 54 kDa is found on the cell surface and thus we focused on the more mature form for our analyses. Consistent with estimates of changes in surface NETs with whole cell [3H]nisoxetine binding, β-PMA produces a 24 ± 2% decrease the level of NET protein recovered in biotinylated (surface) fractions (Fig. 11B). In a similar paradigm, insulin produces no change in the levels of biotinylated hNETs. We also obtained no change in cell surface nisoxetine binding and hNET biotinylation at 1 μM concentrations of insulin. Together with the results of surface radioligand binding, these data point to an activation of surface-resident NETs by insulin rather than an effect on transporter trafficking.
Discussion
IRs have been extensively studied as mediators of insulin's action in the periphery but also are found in the CNS (Raizada et al., 1988). With regard to the present study, insulin is known to alter stimulus-induced NE release and clearance of NE from extracellular space (Bhagat et al., 1981; Christensen, 1983; Shimosawa et al., 1992;Townsend et al., 1992). SK-N-SH cells possess IRs with a subunit composition similar to that reported for whole brain and neuronal tissue (Ota et al., 1988a; Raizada et al., 1988). In SK-N-SH cells, insulin binds to its receptors with aKd of ∼1 nM and stimulates 2-deoxyglucose uptake with an EC50 of 1 nM (Ota et al., 1988a).
IR is a membrane-bound tyrosine kinase whose activation leads to tyrosine autophosphorylation and the phosphorylation of IR-associated substrates. Consistent with this model, the tyrosine kinase inhibitor genistein, but not the inactive analog genistin, blocked the ability of insulin to enhance NE transport activity. One of the downstream consequences of IR activation is stimulation of the lipid kinase phosphatidylinositol-3-OH kinase or PI3K. Insulin-mediated activation of glucose transport parallels changes in PI3K activity (Shepherd et al., 1998). Two structurally distinct inhibitors of PI3K, wortmannin and LY-294002, are useful in evaluating the role of PI3K in insulin signal transduction (Shepherd et al., 1998). Wortmannin acts on the regulatory subunit (p85), whereas LY-294002 acts on the catalytic subunit (p110) to induce blockade of PI3K (Shepherd et al., 1998). We found that treatment of SK-N-SH cells with PI3K inhibitors, as with genistein, produced a significant reduction in NE transport, suggesting that constitutive tyrosine kinase and PI3K activity establish resting NE transport activity. Because there is no alteration in [3H]nisoxetine binding in total membrane fractions in the presence of these inhibitors, their effects are likely to occur via disruption of intracellular signaling mechanisms rather than through direct interaction with NETs. Wortmannin's actions in reducing NE transport in PC12 cells at micromolar concentrations has been suggested to occur via the inhibition of MAP kinase (Uchida et al., 1998). However, we observed inhibition of basal NET activity by nanomolar concentrations of wortmannin, suggesting a predominant involvement of PI3K. Consistent with this idea, PI3K is known to activate PKB/Akt and we document this activation as well in SK-N-SH cells after insulin treatments.
Evidence from biochemical (Apparsundaram et al., 1998b; Ramamoorthy and Blakely, 1999), radioligand binding (Apparsundaram et al., 1998a,b;Doolen and Zahniser, 2001) and immunofluorescence (Apparsundaram et al., 1998b) studies indicate that Na+/Cl−-coupled transporters undergo rapid changes in surface distribution in response to regulatory stimuli, involving in many cases PKC activation. The utility of [3H]nisoxetine binding for monitoring steady-state surface NETs has been previously described (Apparsundaram et al., 1998a) and was further confirmed by evaluating surface transporters in hNET stably transfected HEK-293 cells (Apparsundaram et al., 1998b) where sufficient NET expression permitted biotinylation analysis and direct documentation of NET internalization by confocal imaging. Through these measures, we observed that activation of PKC reduced surface density of hNETs in SK-N-SH cells and transporter transfected HEK-293. Similar to results in our previous studies with PKC activators, we observed that genistein, wortmannin, and LY-294002 reduced the density of NETs in the plasma membrane as measured by whole cell [3H]nisoxetine binding, suggesting that basal tyrosine kinase and PI3K activity dictate cell surface NET density. Consistent with these results, Doolen and Zahniser (2001) have recently reported a tyrosine kinase-mediated modulation of surface expression of DATs in Xenopus laevis oocytes. PI3K has also been implicated in regulation of the surface density of EAAT1 glutamate transporters (Davis et al., 1998). Finally, Law et al. (2000)have also reported a role for tyrosine kinases in the surface expression of GAT1 GABA transporters, involving in this case the tyrosine phosphorylation of GAT1 protein. We have found as yet no evidence for tyrosine phosphorylation of NETs after insulin activation (S. Apparsundaram and R. D. Blakely, unpublished findings), and thus we suspect that insulin's effects are mediated through multiple intermediate signaling partners downstream of PI3K.
Although altered trafficking of NET proteins likely underlies the effects of genistein, wortmannin, and LY-294002 on basal NE transport activity in SK-N-SH cells, we questioned whether insulin's action on NET was a case of modulated trafficking. In adipocytes and muscle, insulin induces the translocation of intracellular GLUT4 glucose transporter containing vesicles to the plasma membrane (Cheatham and Kahn, 1995). In rat hippocampal cells, insulin promotes recruitment of GABAA receptors to the neuronal membrane and enhances GABAA receptor-mediated inhibitory postsynaptic currents (Wan et al., 1997). The ability of insulin to increase Vmax of NE transport without altering Km for NE at first lead us to suspect that insulin's actions on NET was also a trafficking event. To our surprise, neither whole cell [3H]nisoxetine binding nor cell surface biotinylation efforts revealed changes in the surface density of transporter proteins and thus insulin's actions are more consistent with the enhancement of catalytic activity of NETs that are already inserted in the plasma membrane. Although less well defined than the translocation pathway, insulin has been shown also to increase glucose transport via pathways independent of GLUT4 trafficking (Sweeney et al., 1999). Notably, p38 MAPK has been implicated in GLUT4 activation (Sweeney et al., 1999) and we provide evidence that p38 MAPK mediates the stimulatory effect of insulin on NET activity in SK-N-SH cells. p38 MAPK mediates PI3K signaling in cardiac myocytes (Chun et al., 2000; Mockridge et al., 2000). Consistent with these data, inhibition of p38 MAPK phosphorylation by wortmannin and inhibition of PKB/Akt phosphorylation by p38 MAPK inhibitors in SK-N-SH cells indicate that p38 MAPK lies downstream of PI3K and upstream of PKB/Akt in the insulin-signaling pathway. Because blockade of p38 MAPK does not influence basal NET activity but abolishes insulin activation of NET, IRs may sustain p38 MAPK activation above that afforded by constitutive PI3K activity or distinct PI3K isoforms may participate. Multiple isoforms of PI3K, belonging to three different classes (Vanhaesebroeck and Waterfield, 1999), have been identified and could contribute differentially to basal and insulin-mediated regulation of NETs. Alternatively, PI3K activity may be required to sustain p38 MAPK activation arising through a separate pathway. In this regard, distinct MAPK pathways have been described involving ERK 1,2 (p42, p44), c-Jun NH2-terminal kinase, and p38 MAPK. Interestingly, a PI3K, PKB/Akt, and MAPK pathway also has been reported to mediate angiotensin II-evoked transcriptional regulation of NET expression in cultured rodent neurons (Yang and Raizada, 1999); however, in these studies a transcription-independent pathway for acute NET regulation by angiotensin II was also revealed.
The p38 MAPK pathway has been implicated in neuronal survival and apoptosis (Xia et al., 1995; Mao et al., 1999) but more recently has been found to play a role in more rapidly evoked synaptic plasticity. Postsynaptic p38 MAP kinase appears to mediate metabotropic glutamate receptor-dependent long-term depression via an as yet unidentified retrograde messenger to modulate glutamate release (Bolshakov et al., 2000). Conceivably, a diffusible messenger could be at work in our cultures as well to transfer insulin receptor activation of p38 MAP kinase to the up-regulation of NE transport in adjacent cells. Further studies are needed to evaluate of downstream products of p38 MAP kinase activation in NET regulation.
IRs activate bumetanide-sensitive Na+/K+/2Cl−cotransporters in alveolar cells via a Ca2+-dependent mechanism (Marunaka et al., 1999). In parallel, insulin alters intracellular [Ca2+] either by mobilization of intracellular Ca2+ stores, or by increasing Ca2+ influx (Ishida et al., 1996). In noradrenergic neurons of the myenteric plexus, insulin increases NE release by activating N- and L-type channels leading to an increase in Ca2+ influx (Cheng et al., 1997). We find that manipulations to limit extracellular Ca2+ influx via voltage-gated Ca2+ channels abolished insulin's action on NET and p38 MAPK activation. However, the site(s) of Ca2+ action is likely complex. We do not believe it involves PKC-dependent pathways as M3receptor activation and down-regulation of PKC via chronic phorbol ester treatments fail to impact insulin signaling to NETs. Moreover, PKC activation in SK-N-SH cells (and other systems) results in altered surface expression of NETs, a phenomenon that does not occur explain insulin's actions on NET. Finally, insulin does not appear to trigger a rise in cytosolic Ca2+ per se, unlike M3 receptor activators, and thus resting intracellular Ca2+ established through the basal activity of voltage-dependent Ca2+ channels is more likely permissive but not instructive in signal transduction leading to NET regulation. We did obtain evidence that one site of basal Ca2+ requirements is the activation of p38 MAPK itself.
How might insulin and activated p38 MAPK ultimately increase the intrinsic activity of NETs? Insulin stimulates the activity of Na+/K+-ATPase by enhancing the pump's dephosphorylation (Sweeney and Klip, 1998). The stimulatory effect of insulin on Na+/K+-ATPase activity occurs due to insulin-mediated activation of protein phosphatase PP1 (Ragolia et al., 1997). In a recent study with rat vas deferens, we reported that the catalytic subunit of protein phosphatase PP2A physically associates with NETs (Bauman et al., 2000). Additionally, the present study demonstrates that PP2A activity is required for insulin's actions to enhance NE transport. As with Na+ pump activation, insulin may sustain a dephosphorylated state of NETs through enhanced PP2A activation or association resulting in a change in the equilibrium between active and inactive conformations of NET protein resident at the cell surface (Fig. 12). Alternatively, insulin's actions on NET activation could be indirect, mediated via a Ca2+-dependent modulation of NET-interacting proteins that influence transporter catalytic efficiency. Recently, we have observed syntaxin 1A:NET complexes in transfected cells and native tissues (U. Sung, S. Apparsundaram, and R. D. Blakely, manuscript in preparation). Further analysis of mechanisms underlying insulin-mediated NET regulation may provide important insights of use in the therapeutic modulation of noradrenergic function and the reversal of compromised NET activity seen in pathophysiological states.
Acknowledgments
We acknowledge Andrea C. Cherrington, and Reena D. Duseja for assistance in transport assays; Qiao Han and Jackie Huller for expert technical assistance with aspects of cell culture; and Brian Wadzinski for discussions related to the engagement of protein phosphatases.
Footnotes
-
↵1 Current address: Department of Anatomy and Neurobiology, University of Kentucky Chandler Medical Center, Lexington, KY. 40536-0098.
-
This work was supported by National Institute of Neurological Disorders and Stroke award MH58921 and Diabetes Research Training Center pilot Grant DK20593 to R.D.B. and training grant T32 HL07323 to S.A.
- Abbreviations:
- NE
- norepinephrine
- NET
- norepinephrine transporter
- GABA
- γ-aminobutyric acid
- CNS
- central nervous system
- PKC
- protein kinase C
- IR
- insulin receptor
- PI3K
- phosphatidylinositol 3-OH kinase
- IGF-1
- insulin-like growth factor-1
- β-PMA
- β-phorbol-12-myristate-13-acetate
- BAPTA-AM
- 1,2-bis(o-amino-phenoxy)ethane-N,N,N′,N′-tetraacetic acid tetra(acetoxymethyl)ester
- fura-2 AM
- 1-[2-(5-carboxyoxazol-2-yl)-6-aminobenzofuran-5-oxy]-2-(2′-amino-5′-methylphenoxy)-ethane-N,N,N′,N′-tetraacetic acid acetoxy methylester
- PKB/Akt
- protein kinase B
- p38 MAPK
- p38 mitogen-activated protein kinase
- KRH
- Krebs-Ringer-HEPES
- PBS
- phosphate-buffered saline
- RIPA
- radioimmunoprecipitation assay
- PAGE
- polyacrylamide gel electrophoresis
- HBSS
- Hanks' balanced salt solution
- ANOVA
- analysis of variance
- [Ca2+]i
- intracellular calcium concentration
- ERK
- extracellular signal receptor-activated kinase
- PP2Ac
- catalytic subunit of protein phosphatase 2A
- PP1/PP2A
- protein phosphatase 1/2A
- Received April 24, 2001.
- Accepted August 9, 2001.
- The American Society for Pharmacology and Experimental Therapeutics