Activation of the Sodium Pump Blocks the Growth Hormone-Induced Increase in Cytosolic Free Calcium in Rat Adipocytes¹

Abstract

GH promptly increases cytosolic free calcium ([Ca²⁺]_i) in freshly isolated rat adipocytes. Adipocytes deprived of GH for 3 h or longer are incapable of increasing [Ca²⁺]_i in response to GH, but instead respond in an insulin-like manner. Insulin blocks the GH-induced increase in [Ca²⁺]_i in GH-replete cells and stimulates the sodium pump (i.e. Na⁺/K⁺-ATPase), thereby hyperpolarizing the cell membrane. Blockade of the Na⁺/K⁺-ATPase with 100 μm ouabain reversed these effects of insulin and enabled GH to increase[ Ca²⁺]_i in GH-deprived adipocytes. Both insulin and GH activated the sodium pump in GH-deprived adipocytes, as indicated by increased uptake of ⁸⁶Rb⁺. Decreasing availability of intracellular Na⁺ by blockade of Na⁺/K⁺/2Cl⁻ symporters or Na⁺/H⁺ antiporters abolished the effects of both hormones on ⁸⁶Rb⁺ uptake and enabled both GH and insulin to increase [Ca²⁺]_i in GH-deprived adipocytes. The data suggest that hormonal stimulation of Na⁺/K⁺-ATPase activity interferes with activation of voltage-sensitive calcium channels by either membrane hyperpolarization or some unknown interaction between the sodium pump and calcium channels.

THE ACUTE EFFECTS of GH on rat adipocytes are complex and strongly influenced by the time elapsed from the previous exposure to the hormone. When added to adipocytes that have been deprived of GH for 3 h or longer, GH produces effects that are remarkably similar to those of insulin and include increased glucose uptake and lipogenesis (1, 2). These transient insulin-like effects of GH appear to follow from activation of the cytosolic tyrosine kinase JAK2 (3), resulting in tyrosine phosphorylation of the insulin receptor substrates (IRS-1 and IRS-2) (4–7) and activation of phosphatidylinositol-3-kinase (6). When added to adipocytes less than 3 h after prior exposure to the hormone either in vivo (freshly isolated cells) or in vitro (GH pretreated), GH increases the intracellular free calcium concentration ([Ca²⁺]_i) 2- to 3-fold (8, 9) as a result of influx of calcium through voltage-sensitive channels (10, 11), and no insulin-like response is seen (9, 12). GH produces no such change in[ Ca²⁺]_i in the GH-deprived adipocytes that express an insulin-like response (8, 9, 11). In contrast to the insulin-like response, activation of calcium influx appears to be independent of JAK2, at least in Chinese hamster ovary (CHO) cells engineered to express GH receptors (13), and proceeds through a mechanism that appears to require activation of a phosphatidylcholine phospholipase C and a calcium-independent isoform of protein kinase C (10). Neither the substrate for protein kinase C nor the manner in which the GH receptor signals to phospholipase C is known.

Compelling evidence suggests that the[ Ca²⁺]_i is a critical determinant of the ability of adipocytes to respond to GH with an increase in glucose metabolism (12). Treatment of GH-deprived cells with the calcium ionophore A23187 in the presence of 1.5 mm calcium markedly decreased or prevented expression of the insulin-like response, and conversely, incubation in calcium-free medium or pharmacological blockade of calcium channels allowed expression of the insulin-like effect in freshly isolated adipocytes that would otherwise be unresponsive. These results indicate that although not expressed, the underlying capacity to respond to GH in an insulin-like manner is intact in freshly isolated and GH-pretreated adipocytes as well as in GH-deprived cells and suggest further that the failure of GH to activate calcium channels in GH-deprived cells allows them to express an insulin-like response.

To gain insight into the basis for the failure of GH-deprived adipocytes to increase[ Ca²⁺]_i when treated with GH, we revisited the finding that insulin blocks the GH-induced rise in[ Ca²⁺]_i in freshly isolated or GH-pretreated adipocytes (8). Abundant evidence indicates that insulin rapidly causes adipocyte plasma membranes to hyperpolarize (14–17), probably by stimulating the electrogenic Na⁺/K⁺-ATPase in the plasma membrane (18–20). Hyperpolarization would be expected to oppose the activation of voltage-sensitive calcium channels and the consequent influx of Ca²⁺. The present studies were undertaken to examine the possibility that among its insulin-like actions, GH stimulates Na⁺/K⁺-ATPase and thereby prevents the activation of voltage-sensitive calcium channels. This report describes apparent relationships between GH or insulin stimulation of Na⁺/K⁺-ATPase and[ Ca²⁺]_i in rat adipocytes.

Materials and Methods

Animals and cells

Normal 160- to 200-g male rats of the Charles River Laboratories, Inc. CD strain (Charles River Laboratories, Inc., Kingston, NY) were used in all experiments in accordance with protocols approved by the University of Massachusetts Medical School animal care and use committee. Husbandry conditions and preparation and handling of epididymal and perirenal adipocytes were described previously (10). Adipocytes that were studied as soon as possible after isolation are referred to as freshly isolated cells. These cells are unresponsive to the insulin-like actions of GH (1, 2). Adipocytes that were studied after incubation without hormone for at least 3 h are referred to as GH-deprived cells. These cells are responsive to the insulin-like effects of GH (1, 2). To control for nonspecific changes that might occur during the 3-h in vitro incubation and yet have cells that are unresponsive to the insulin-like actions of GH, 100 ng/ml recombinant methionyl human GH (hGH) was added for the first hour of incubation. These cells, which are referred to as GH-pretreated cells, were then washed and incubated for an additional 2 h without hormone.

Measurement of [Ca²⁺]_i

Adipocytes were loaded for 30 min with the fluorescent Ca²⁺ indicator, fura-2/AM (Molecular Probes, Inc., Eugene, OR), and[ Ca²⁺]_i was measured in individual cells according to procedures described previously (8). Briefly, adipocytes suspended from a coverslip in a temperature-controlled chamber (37 C) were perifused with Krebs-Ringer bicarbonate buffer that contained 1 mg/ml glucose and 0.1% (wt/vol) BSA (Metrix, fraction IV, Reheis Chemical Co., Phoenix, AZ) at a flow rate of 1 ml/min. The perifusion chamber was mounted on the stage of an inverted microscope (Diaphot, Nikon, Melville, NY), and the adipocytes were sequentially illuminated through a ×10 UV lens (Nikon, glycerin NA 0.5) with 3-nsec pulses of light (337 or 380 nm) delivered every 33 msec through a bifurcated quartz fiber extending from a nitrogen laser and a tunable dye laser (Laser Science, Cambridge, MA) to the epiport of the microscope. Images were recorded with a CCD Camera (FTM 800, Philips Components, Slatersville, RI) and processed using an NEC Power Mate computer. To measure responses in multiple members of a population of cells, sequential observations of 10–15 sec each were made in 30–50 randomly selected cells over a 9-min period. These cells were perifused with GH, insulin, and/or other test substances for 30–60 sec before and throughout the scanning interval. Mean values for[ Ca²⁺]_i obtained in the first minute of scanning did not differ significantly from values obtained in the fifth or ninth minutes of scanning, indicating that measured values remained constant during the 9-min period of observation.

Rubidium uptake

The activity of the Na⁺/K⁺-ATPase was estimated in intact cells by measuring the rate of ouabain-sensitive uptake of ⁸⁶Rb⁺ according to the procedure of Resh et al. (19). Briefly, adipocytes were suspended at a dilution of 1:12 (vol/vol; ∼10⁶ cells/ml) in Krebs-Ringer phosphate buffer that contained 4% BSA and were incubated for 15 min in the absence or presence of 500 ng/ml hGH, 250 μU/ml insulin, and/or 100 μm ouabain (Calbiochem, Los Angeles, CA) before addition of ⁸⁶RbCl (NEN Life Science Products, Boston, MA) to give a final concentration of 5 μCi/ml. At the times indicated, triplicate 200-μl aliquots of cell suspension were transferred to 400-μl snap cap polyethylene tubes that contained 100μ l dinonylphthallate oil (MCB Manufacturing Chemists, Cincinnati, OH) and were centrifuged for 1 min in a Beckman Coulter, Inc. (Palo Alto, CA), microcentrifuge as described by Gliemann et al. (21). The tubes were sliced through the oil layer, and the upper portions containing the cells were transferred to liquid scintillation vials. The cells were solubilized in 500 μl 5% SDS and counted after adding 5 ml Optifluor (Packard Instruments, Meriden CT). The data are expressed as moles of ⁸⁶Rb⁺(K⁺) taken up per 10⁶ cells.

Reagents

hGH was a gift from Genentech, Inc. (South San Francisco, CA). sn-1,2-Dioctanoyl glycerol (DOG), 5-(N-methyl-N-isobutyl) amiloride (MIA), and nimodipine were obtained from Research Biochemicals International (Natick, MA). Ouabain was purchased from Sigma (St. Louis, MO), and bumetanide was purchased from Calbiochem (Los Angeles, CA). All reagents were of the highest purity available.

Statistics

[Ca²⁺]_i measurements for each experimental condition were made in at least three independent cell populations. For analytic purposes, each population of adipocytes prepared from the pooled epididymal and perirenal fat from one or two rats was treated as a single observation. At least eight independent cell populations consisting of cells pooled from at least four rats per population were used for study of ⁸⁶Rb⁺ (K⁺) uptake for each experimental condition. Statistical significance was evaluated by paired analysis using Student’s t test and applying the Bonferroni adjustment to correct for additive type I errors when multiple comparisons were made (22).

Results and Discussion

As reported previously (8, 10), GH more than doubled[ Ca²⁺]_i when added to GH-pretreated adipocytes (Fig. 1A), and the increase was evident within 1 min. Insulin alone had no effect on[ Ca²⁺]_i, but completely nullified the effect of GH when added simultaneously with GH. DOG, an active diacylglyceride, like GH, produced a more than 2-fold increase in [Ca²⁺]_i, and its effects were also completely blocked by insulin, suggesting that the inhibitory effect of insulin on the action of GH is exerted distal to the activation of phospholipase C and the formation of diacylglycerol. To evaluate the possibility that the inhibition produced by insulin might result from stimulation of Na⁺/K⁺-ATPase and the resulting membrane hyperpolarization, the foregoing measurements were repeated in the presence of 100 μm ouabain, a concentration high enough to inhibit both isoforms of the Na⁺/K⁺-ATPase known to be present in rat adipocytes (23). Although ouabain alone had no effect on[ Ca²⁺]_i, in its presence insulin failed to prevent the increase in[ Ca²⁺]_i caused by either GH or DOG (Fig. 1B).

Figure 1

Insulin antagonizes the effects of GH and DOG on[ Ca²⁺]_i in GH-pretreated adipocytes incubated in the absence (A), but not in the presence (B) of 100 μm ouabain. GH (500 ng/ml), insulin (100 μU/ml), or DOG (50μ m) was added alone or in the indicated combinations to fura-2-loaded adipocytes just before measurement of[ Ca²⁺]_i. Each bar indicates the mean ± sem of [Ca²⁺]_i in at least 3 cell populations, each including 30–50 adipocytes. *, P < 0.01 compared with control.

Because GH mimics most of the actions of insulin in GH-deprived adipocytes (24), we explored the possibility that it might also stimulate Na⁺/K⁺-ATPase and therefore oppose any effects on[ Ca²⁺]_i. If so, GH, like insulin, should prevent the increase in[ Ca²⁺]_i caused by DOG in GH-deprived cells, and this inhibitory effect should be prevented by ouabain. In accord with previous observations (8, 9), GH failed to increase [Ca²⁺]_i when added to GH-deprived adipocytes (Fig. 2A), although other cells from the same populations doubled their[ Ca²⁺]_i when treated with DOG. When added together with DOG, GH prevented the increase in[ Ca²⁺]_i, suggesting that some consequence of GH action might have antagonized the activation of L-type calcium channels in these cells. When GH was added to GH-deprived adipocytes in the presence of 100 μm ouabain,[ Ca²⁺]_i increased nearly 3-fold (Fig. 2B), indicating that the signaling pathway leading to calcium channel activation is intact in these cells, but calcium influx is blocked by a ouabain-sensitive reaction. No further increase was seen when GH and DOG were added simultaneously to ouabain-treated adipocytes. The sensitivity of the increase in[ Ca²⁺]_i to blockade by 100 nm nimodipine (Fig. 2) is consistent with earlier evidence that the increase in[ Ca²⁺]_i depends on activation of L-type calcium channels (10). These observations are consistent with the possibility that GH, like insulin, activates Na⁺/K⁺-ATPase in GH-deprived adipocytes. It is noteworthy that under both basal and stimulated conditions,[ Ca²⁺]_i was significantly lower in the GH-deprived adipocytes (p<0.05) than in GH-pretreated adipocytes (Fig. 2vs.Fig. 1), in accord with previous observations (9, 10) and presumably as a consequence of the effects of GH pretreatment on the abundance of L-type calcium channels in the plasma membrane (25).

Figure 2

GH antagonizes the effects of DOG on[ Ca²⁺]_i in GH-deprived adipocytes incubated in the absence (A), but not in the presence (B), of 100μ m ouabain. GH (500 ng/ml), insulin (100 μU/ml), DOG (50 μm), or nimodipine (100 nm) was added alone or in the indicated combinations to fura-2-loaded adipocytes just before measurement of [Ca²⁺]_i. Each bar indicates the mean ± sem of[ Ca²⁺]_i for at least 3 cell populations, each including 30–50 adipocytes. *, P < 0.01 compared with control.

To assess the effects of GH on the activity of the Na⁺/K⁺-ATPase in intact adipocytes, uptake of ⁸⁶Rb⁺ (which serves as a marker for K⁺) was measured in the absence and presence of 100 μm ouabain. Because the insulin-like response appears to be fully developed by 15 min after the addition of GH and persists for at least 30 min thereafter (26), adipocytes were preincubated with 500 ng/ml hGH or 250 μU/ml insulin for 15 min before addition of ⁸⁶Rb⁺, and uptake of ⁸⁶Rb⁺(K⁺) was measured at various times thereafter. In the absence of hormones, ⁸⁶Rb⁺(K⁺) uptake progressed in a curvilinear fashion for at least 30 min, as described by previous investigators (19), but the rate of uptake by the GH-deprived cells was about 25% greater than that in freshly isolated or GH-pretreated adipocytes (Fig. 3). The effect of GH deprivation was most evident for the ouabain-sensitive component of ⁸⁶Rb⁺(K⁺) uptake, but ouabain-insensitive uptake of ⁸⁶Rb⁺(K⁺) was also somewhat increased in these cells. The explanation for these findings is unknown.

Figure 3

Effects of prior exposure to GH on the rate of uptake of ⁸⁶Rb⁺(K⁺) by rat adipocytes. Freshly isolated, GH-deprived, or GH-pretreated adipocytes were incubated with 5 μCi/ml ⁸⁶RbCl for 30 min in the absence or presence of 100 μm ouabain. Ouabain-sensitive uptake is defined as the difference between ⁸⁶Rb⁺(K⁺) taken up in the absence or presence of ouabain. Each bar represents the mean ± sem for 8 (freshly isolated and GH-deprived) or 14 (GH-pretreated) adipocyte populations. Triplicate aliquots of cells were analyzed for each measurement. *, P < 0.02; †, P < 0.01 (compared with freshly isolated and GH-pretreated cells). §, P < 0.05 (compared with freshly isolated cells).

Insulin stimulated ⁸⁶Rb⁺(K⁺) uptake to a similar extent in all three adipocyte populations (Fig. 4), and this response continued at an apparently linear rate (r² = 0.94) for at least 30 min in freshly isolated adipocytes (Fig. 4A). In both GH-deprived and GH-pretreated adipocytes, the effect of insulin appeared to taper off by about 10 min after the addition of ⁸⁶Rb⁺, but remained in evidence for at least 30 min (Fig. 4, B and C). GH had no effect on ⁸⁶Rb⁺(K⁺) uptake in freshly isolated adipocytes (Fig. 4A) or in adipocytes that had been pretreated with GH in the first hour of incubation (Fig. 4C). In GH-deprived adipocytes, however, GH produced a small, but transient, stimulation of ⁸⁶Rb⁺(K⁺) uptake that was significant (P < 0.01) 5 min after addition of ⁸⁶Rb⁺. Because ⁸⁶Rb⁺ was not added until 15 min after GH, it is possible that increased ⁸⁶Rb⁺ (K⁺) uptake actually lasted for as long as 20 min.

Figure 4

Effects of insulin or GH on ⁸⁶Rb⁺(K⁺) uptake by adipocytes. Freshly isolated (A), GH-deprived (B), or GH-pretreated (C) cells were preincubated for 15 min with insulin (250 μU/ml) or GH (500 ng/ml) in the absence or presence of 100 μm ouabain before the addition of 5 μCi/ml ⁸⁶RbCl. Triplicate aliquots of cells were analyzed at the indicated time points. Ouabain-sensitive uptake is defined as the difference between ⁸⁶Rb⁺(K⁺) taken up in the absence or presence of 100 μm ouabain. Each point represents the mean and sem of the difference in ouabain-sensitive ⁸⁶Rb⁺(K⁺) uptake attributable to GH or insulin for eight experiments with freshly isolated or GH-pretreated cells (A and C) or 14 experiments with GH-deprived cells (B). Virtually identical results were obtained when total ⁸⁶Rb⁺(K⁺) uptake in the absence of hormone was subtracted from that taken up in the presence of hormone. The effects of insulin were significant (P< 0.01) at all time points in all three cell groups, except for the 5 min point in the GH-pretreated cells (P < 0.05). The only significant (P < 0.01) effect of GH was seen at the 5 min point in the GH-deprived cells (B).

The increases in [Ca²⁺]_i shown in Fig. 2 were recorded during the first 9 min after the addition of GH and were fully developed within 1–2 min. For the stimulation of Na⁺/K⁺-ATPase to be relevant with respect to regulation of Ca²⁺ influx, it, too, must therefore occur within 1–2 min after the addition of GH. Unfortunately, measurement of ⁸⁶Rb⁺(K⁺) uptake is too slow and too imprecise to assess enzyme activity in so short a time. When ⁸⁶Rb⁺ was added immediately after GH, ouabain-sensitive ⁸⁶Rb⁺(K⁺) uptake was increased by 2.46 ± 0.66 nmol/million cells (n = 8; P < 0.01) 5 min later and remained at 1.66 ± 0.37 nmol/million cells above the control level at 10 min (P< 0.01). These values are quite similar to those shown in Fig. 4B and suggest that the increase in Na⁺/K⁺-ATPase activity occurs within the early minutes after the addition of GH.

Although the findings that GH and insulin stimulate Na⁺/K⁺-ATPase activity and that ouabain reverses the inhibitory effect of insulin and GH on Ca²⁺ influx in GH-deprived adipocytes are consistent with our hypothesis, the effect of ouabain on Ca²⁺ could nevertheless be unrelated to its effect on the sodium pump. We therefore determined whether other agents that decrease Na⁺/K⁺-ATPase activity might have the same effects as ouabain on Ca²⁺ influx. Basal and stimulated activity of the Na⁺/K⁺-ATPase are limited by the intracellular concentration of Na⁺, which, in rat adipocytes, appears to be well below the concentration that produces half-saturation of the catalytic α-subunits of the enzyme (26, 27). Sustaining the augmented activity of the enzyme, as shown in Fig. 4, requires accelerated influx to replenish intracellular Na⁺. Sargeant et al. (28) reported that the increase in ⁸⁶Rb⁺(K⁺) uptake caused by insulin in mouse 3T3-L1 adipocytes was blocked by bumetanide, a specific inhibitor of the Na⁺/K⁺/2Cl⁻ symporter. Sodium is also taken up in exchange for H⁺, and insulin stimulates the activity of this antiporter in rat adipocytes (29). We therefore examined the effects of bumetanide and MIA, a specific inhibitor of the Na⁺/H⁺ antiporter, on the insulin- and GH-dependent stimulation of Na⁺/K⁺-ATPase in GH-deprived adipocytes.

In accord with the results shown in Fig. 4, both hormones increased ⁸⁶Rb⁺(K⁺) uptake when measured 5 min after the addition of ⁸⁶Rb⁺ (Fig. 5). About 60% of the total ⁸⁶Rb⁺(K⁺) uptake was attributable to the activity of Na⁺/K⁺-ATPase, as judged by its sensitivity to ouabain (Fig. 3), which also obliterated the entire hormone-dependent increase in ⁸⁶Rb⁺(K⁺) uptake. Blockade of the Na⁺/K⁺/2Cl⁻ symporter with bumetanide abolished the effects of both GH and insulin and decreased ⁸⁶Rb⁺(K⁺) uptake by approximately 40%, or nearly two thirds of the ouabain-sensitive component (data not shown), indicating that the Na⁺/K⁺/2Cl⁻ symporter provides an important route of Na⁺ influx. Very similar results were seen when the Na⁺/H⁺ antiporter was blocked. MIA abolished the effects of GH and insulin (Fig. 5) and decreased ⁸⁶Rb⁺(K⁺) uptake by about 30% (data not shown), suggesting that the Na⁺/H⁺ antiporter also contributes significantly to the influx of Na⁺. Whether insulin (or GH) activates the Na⁺/K⁺-ATPase indirectly by activating either of these transporters, as suggested in other studies (28, 29), cannot be determined from these data. Curiously, the inhibitory effects of bumetanide and MIA on ⁸⁶Rb⁺(K⁺) were not additive, and blockade of both transporters simultaneously produced no greater inhibition than that seen with bumetanide alone (data not shown), suggesting that Na⁺ entry through other routes, perhaps by exchange for Ca²⁺ (30) or through a Na⁺ channel, may accelerate at low intracellular sodium concentrations. The combination of bumetanide and ouabain decreased ⁸⁶Rb⁺(K⁺) uptake approximately 7% more than ouabain alone (P < 0.05), and addition of MIA produced no further decrease (data not shown). The approximately 25% of ⁸⁶Rb⁺(K⁺) uptake that is insensitive to these inhibitors presumably reflects traffic through potassium channels or nonspecific cation channels.

Figure 5

Effects of ouabain (100 μm), bumetanide (100 μm), and MIA (25 μm) on ⁸⁶Rb⁺(K⁺) uptake in response to insulin or GH in GH-deprived adipocytes. The cells were incubated with insulin (250 μU/ml) or GH (500 ng/ml) for 15 min in the absence or presence of ouabain, bumetanide, or MIA before ⁸⁶Rb⁺ was added. ⁸⁶Rb⁺(K⁺) uptake was measured 5 min later. Each bar represents the mean increase in ⁸⁶Rb⁺(K⁺) uptake ± sem produced by insulin or GH for 12 or more experiments.* , P < 0.001 compared with zero, as determined by paired analysis.

We next compared bumetanide and MIA with ouabain for possible effects on [Ca²⁺]_i in GH-deprived adipocytes after treatment with GH or insulin (Fig. 6). Even though ouabain, bumetanide, and MIA differ in their effects on intracellular concentrations of H⁺, K⁺, and Cl⁻, virtually identical results with respect to[ Ca²⁺]_i were obtained with all three agents. GH and insulin, neither of which produced any change in [Ca²⁺]_i when added alone to GH-deprived adipocytes, more than doubled[ Ca²⁺]_i when added in the presence of bumetanide, MIA, or ouabain (Fig. 6, A–C). Identical effects of GH and insulin were seen when GH-deprived adipocytes were incubated in Na⁺-free medium (Fig. 6D). Thus inhibition of Na⁺/K⁺-ATPase, either with ouabain or by depriving it of substrate (i.e. Na⁺), enabled not only GH, but also insulin, to increase [Ca²⁺]_i in GH-deprived adipocytes. These results thus provide strong support for the idea that increased activity of Na⁺/K⁺-ATPase limits the ability of GH or insulin to activate L-type calcium channels. The precise relationship between pump activity and channel activation is not known. Preliminary data indicate that GH causes the plasma membranes of GH-deprived cells to hyperpolarize by 2–3 mV (Yamaguchi, H., and H. M. Goodman, unpublished data), but whether this change is large enough or takes place fast enough to block channel activation remains to be demonstrated.

Figure 6

Effects of decreased Na⁺/K⁺-ATPase activity on[ Ca²⁺]_i after treatment of GH-deprived adipocytes with 500 ng/ml GH or 100 μU/ml insulin. Each bar represents the mean ± sem for 3 (B–D) or 4 (A) independent cell populations, each involving measurements in 30–50 cells. Bumetanide (100 μm), MIA (25 μm), and ouabain (100 μm) were added 1 min before insulin or GH. The cells were transferred to Na⁺-free medium (i.e. Na⁺ was replaced with choline) 1 min before the addition of insulin or GH.* , P < 0.05 compared either to[ Ca²⁺]_i in adipocytes incubated in medium of the same composition but without hormone, or to[ Ca²⁺]_i in adipocytes incubated with the corresponding hormone but without inhibitors.

The findings that insulin increases[ Ca²⁺]_i in the presence of ouabain or other inhibitors of Na⁺/K⁺-ATPase in GH-deprived adipocytes (Fig. 6) contrast sharply with data shown in Fig. 1, in which insulin produced no increase in[ Ca²⁺]_i in GH-pretreated adipocytes even when ouabain was present. These effects of GH pretreatment have been confirmed in three additional experiments in which GH-deprived and GH-pretreated adipocytes from the same rats were compared (data not shown). Thus, it is likely that some GH-dependent event limits the ability of insulin to signal Ca²⁺ entry. Neither the mechanism for such an action nor its physiological rationale is readily apparent. We recently found that insulin activates L-type calcium channels in GH-deprived adipocytes that were preincubated with the phosphatase inhibitor, okadaic acid (31), but the effects of okadaic acid were seen only after a 2-h lag period and appeared to involve transcriptional events. It is unclear how, or if, the effects of okadaic acid relate to the findings described here.

The present findings indicate that GH-deprived adipocytes retain the requisite enzymatic apparatus for increasing[ Ca²⁺]_i in response to GH, but the increase is blocked by the simultaneous activation of the sodium pump. Activation of Na⁺/K⁺-ATPase by GH is not seen in GH-pretreated adipocytes, and hence, the increase in[ Ca²⁺]_i is expressed. GH-deprived adipocytes also respond to GH with an increase in glucose metabolism (1, 2). The ability of adipocytes to exhibit an insulin-like response to GH with respect to glucose is accompanied by increased activity of Na⁺/K⁺-ATPase and is inversely related to expression of an increase in[ Ca²⁺]_i. Blockade of the GH-induced calcium influx with verapamil permits expression of the insulin-like increase in glucose metabolism in GH-pretreated adipocytes (12), indicating that the increase in[ Ca²⁺]_i is necessary for suppression of the increase in glucose metabolism. The present finding that stimulation of the sodium pump prevents the increase in[ Ca²⁺]_i strongly suggests that the activity of the Na⁺/K⁺-ATPase may play a pivotal role in determining the nature of the response to GH. Just how GH stimulates the activity of this enzyme is unknown.

A hypothetical model representing these early actions of GH in fat cells is shown in Fig. 7. According to this model, the dimeric GH receptor complex entrains at least two simultaneous events. Activation of JAK2, which leads to tyrosine phosphorylation of IRS-1 and -2 and the Stat (signal transducer and activator of transcription) proteins, produces the insulin-like effects (32) and genomically mediated events (3), as indicated on the left of the figure. Stimulation of phospholipase C by some still unknown transducer results in activation of L-type calcium channels and calcium influx (10), as shown on the right of the figure. In the GH-deprived cells, the influx of calcium is blocked by the insulin-like stimulation of the Na⁺/K⁺-ATPase, and glucose metabolism is also increased. In freshly isolated or GH-pretreated cells, activation of Na⁺/K⁺-ATPase appears to be blocked by some as yet unknown, short-lived protein(s) whose synthesis is GH dependent (9), and calcium influx proceeds unchecked. The resulting increase in[ Ca²⁺]_i is required to prevent the increase in glucose metabolism (12) by some as yet unknown mechanism. GH is normally secreted in pulses every 3–4 h (33). If the present findings are indicative of in vivo events, it is likely that residual effects of each preceding secretory episode interfere with activation of Na⁺/K⁺-ATPase by each succeeding pulse of GH secretion and thereby maintain a state of insensitivity to the insulin-like effects of GH.

Figure 7

Schematic representation of the pathways of GH action in adipocytes. MAPK, Mitogen-activated protein kinase; ??, unknown transducing protein; PLC, phospholipase C; PKC, protein kinase C. Solid arrows indicate stimulation; dashed arrows indicate inhibition (see text).

¹

This work was supported by NIDDK Grant DK-19392. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health. A preliminary report of these findings was presented in a poster at the 81^st Annual Meeting of The Endocrine Society, San Diego, California, June 1999.

²

Supported by NIH Training Grant DKO-7302.

³

Deceased February 9, 1999.