Modulation of β-Cell Ouabain-Sensitive 86Rb+ Influx (Na+/K+ Pump) by D-Glucose, Glibenclamide or Diazoxide

The activity of the β-cell Na+/K+ pump was studied by using ouabain-sensitive (lmM ouabain) 86Rb+ influx in β-cell-rich islets of Umeå-ob/ob mice as an indicator of the pump function. The present results show that the stimulatory effect of glucose on ouabain-sensitive 86Rb+ influx reached its approximate maximum at 5mM glucose. Pre-treatment of the islets with 20mM glucose for 60 min strongly reduced the glucose-induced stimulation of the Na+/K+ pump. Pre-treatment (60 or 180 min) of islets at 0mM glucose, on the other hand, did not affect the magnitude of the glucose-induced stimulation of 86Rb+ influx dunng the subsequent 5-min incubation. Glibenclamide stimulated the ouabain-sensitive 86Rb+ uptake in the same manner as glucose. The stimulatory effect, showed its apparent maximum at 0.5μM. Pre-treatment (60 min) of islets with 1μM glibenclamide did not reduce the subsequent stimulation of the ouabain-sensitive 86Rb+ influx. The stimulatory effect of glibenclamide and D-glucose were not .additive, suggesting that they may have the same mechanism of action. No direct effect of glibenclamide (0.01-1μM) was observed on the Na+/K+ ATPase activity in homogenates of islets. Diazoxide (0.4mM) inhibited the Na+/K+ pump. This effect was sustained even after 60 min of pre-treatment of islets with 0.4mM diazoxide. The stimulatory effect of glibenclamide and D-glucose were abolished by diazoxide. It is concluded that nutrient as well as non-nutrient insulin secretagogues activate the Na+/K+ pump, probably as part of the membrane repolarisation process.


INTRODUCTION
Stimulation of pancreatic fl-cells by D-glucose or sulphonylurea induces a complex electrical membrane activity, which has a pivot role in activating the secretory machinery. 11 The important role of K + permeability and activity of K + channels in the fl-cell plasma membrane is well established (for review see Ill). The ATPregulated K + channels dominate the resting conductance of the fl-cell membrane and the glucose-induced closing of these channels leads to the initial depolarisation up to the threshold for generation of an electrical slow wave complex. [2,3] This model requires the presence of an active Na +/K + pump. Early studies of labelled Rb + transmembrane transport *Corresponding author. Tel.: + 4690-786 51 09, Fax: + 4690-786 66 96, e-mail: janove.sehlin@histocel.umu.se showed that fl-cells accumulated 86Rb + and extruded 22Na + [4,5] in an ouabain-sensitive manner and indicated that neither glucose [4,6] nor sulphonylureas [7] inhibited the Na +/K + pump in pancreatic fl-cells. Without finding any direct effect of glucose on the electrogenic Na +/K + pump, electrophysiological studies indicated that activation of the Na +/K + pump could lead to a marked hyperpolarisation of the fl-cell membrane in the presence of 10 mM glucose and that this could be inhibited by ouabain. 8 Direct measurements of islet total content of Na + by integrating flame photometry have revealed that glucose lowers the islet Na + content E91 and measurements of cytoplasmic Na + by the fluorescent probe SBFI (sodium benzofuran isophtalate) indicated a glucose-induced decrease also in the free Na + activity. 10 These findings indirectly support the idea that the Na+/K + pump is activated during glucose stimulation of fl-cells. Some previous studies have suggested a direct or indirect inhibitory effect of insulin secretagogues, like glucose and sulphonylureas, on the Na +/K + ATPase activity. [11 13] In the present study we have further characterized the effect of D-glucose, a nutrient secretagogue, and the sulphonylurea glibenclamide, a non-nutrient secretagogue, on islet ouabain-sensitive 86Rb + influx, used as a marker for the Na +/K + pump in intact, fl-cell-rich islets of Umed-ob/ob mice.

Animals and Isolation of Islets
Adult, non-inbred, 7-8 months old, female ob/ob mice from the Umei colony (Ume-ob/ob), were used throughout the experiments. These animals have hyperplastic islets containing a high proportion of fl-cells (> 90%), [14] which makes it highly probable that the present data on isolated islets are representative of this cell type. All mice were fasted overnight but with free access to water, in order to facilitate the isolation of pancreatic islets. The pancreata were digested with collagenase to isolate individual islets as previously described. 151 The function of the isolated islets used in this study was monitored in separate control experiments by measuring the insulin released by islets at different glucose concentrations. Isolated islets used in the study showed an excellent secretory performance (data not shown).
The isolation medium used was a Krebs- ob/ob mice were transferred to a polypropylene micro test tube (Milian Instruments S.A., Geneva, Switzerland) and 60gl homogenisation medium and a glass bead were added. The islets were then homogenised by vibrating the test tube at a frequency of l kHz for 30s followed by a short centrifugation. The test tubes were then immediately frozen at -20C.

Assay of Na+/K + ATPase
The method for measuring the activity of Na +/ K + ATPase was adapted from Jorgensen and Skou. [16] The concentrations of Na + and K + were chosen to give a maximum rate of ouabainsensitive ATP hydrolysis (135mM Na + and 20mMK+). Previous studies on the cationic dependence of the Na +/K + ATPase of mouse pancreatic islets have shown that the enzyme activity at different Na +/K + ratios and the corresponding curve for Na + and K + activation of the fl-cell enzyme (Sandstr6m, Klaerke and Sehlin, unpublished data) are very similar to the data previously described for the Na +/K + ATPase in the kidney. [17'181 The activity of Na +/K + ATPase was measured as the difference in amount of inorganic phosphate released in the presence or absence of ouabain (1 mM). To permeabilize the membranes for measurements of Na +/K + ATPase from pancreatic fl-cells, the homogenate was preincubated with 75 gM DOC for 15 to 30 min at room temperature, according to Jorgensen and Skou. [17] Incubation was carried out in a histidine buffer with the following final composition Germany. All other chemicals were of analytical grade.

Effect of D-glucose
To estimate the activity of the Na +/K + pump in intact fl-cells, 86Rb + (K + marker) influx was studied in the absence or presence of I mM ouabain. Previous experiments have shown that 86Rb + uptake proceeds linearly for at least 5 min and I mM ouabain causes maximum inhibition. I21 Figure 1 shows that the stimulatory effect of glucose on the ouabain-sensitive portion of 86Rb + influx reached its approximate maximum at 5mM glucose (58%; n=14; P < 0.001). The basal rate of ouabain-sensitive 86Rb + influx was thus 63% of the maximum glucose-stimulated rate. The ouabain-resistant portion was significantly and maximally inhibited at 5mM glucose (37%; n=15; P <0.001).
Also the total 86Rb + influx was significantly reduced with an apparent maximum effect at 5mM glucose (10%; n=15; P <0.001). This is in line with the well established capacity of D-glucose to inhibit K + channels.
Pre-treatment with D-glucose In order to characterise the effects of glucose, the effect of pre-treatment of islets with the sugar on the 86Rb + influx was studied. Sixty min of pretreatment of islets with 20 mM glucose strongly reduced the glucose-induced stimulation of the ouabain-sensitive portion of 86Rb + influx (Fig. 2, left panel). No statistically significant stimulation above control level was found after this treatment (34% above control; P > 0.05; n 10) as compared with 74% above control (P < 0.001; n 10) when the islets were preincubated in the Ouabain-res. 25 Glucose concentration (mM) Ouabain-res.
Glucose concentration (mM) The terms 0G and 20G refer to the D-glucose concentration during the preincubation (60 or 180min) and incubation (5 min) periods respectively. For difference between a and b P < 0.001, between a and c P > 0.05 (n.s.), between b and c P < 0.01 and between d and e P < 0.05.

Deprivation
To investigate whether long-term fuel deprivation in vitro could affect the rate of ouabainsensitive 86Rb+ influx, we used 180min of preincubation of islets at 0mM glucose. The results show that this treatment did not affect the level of glucose-induced stimulation of 86Rb+ influx during the subsequent 5-min incubation (65%; P < 0.05; n 12) as compared with the same treatment for 60min (Fig. 2).
Neither the reduction of the ouabain-resistant portion (-29%; P < 0.001; n= 12) nor the total influx (-10%; P <0.005; n=12) was different compared to 60 min of treatment. It should be noted that the changes in different portions of the 86Rb + influx were similar, irrespective of the preincubation time (60 or 180 min) or glucose concentration during the preincubation time (0 or 3 mM).

Effect of Glibenclamide
The effect of the hypoglycaemic sulphonylurea, glibenclamide, on islet-ouabain-sensitive 86Rb + influx was studied by analysing the effect of a range of glibenclamide concentrations (0.1 to 10tM) in the absence or presence of I mM ouabain. Glibenclamide, like glucose, stimulated the Na +/K + pump (28%; P < 0.05; n =6). This stimulatory effect was evident already at 0.5 tM and did not further increased in magnitude with increasing concentration of glibenclamide (Tab. I). Glibenclamide    Islets were prepared as described in the materials and methods and then preincubated for 60 min at 37C in KRH medium containing 3 mM glucose but no glibenclamide. They were then incubated for 5 min at 37C in the absence of glucose but presence of different concentrations of glibenclamide. Data is expressed as mean values for primary data and difference from control 4-SEM for the number of experiments indicated in parentheses, n.s. denotes P > 0.05 for difference from control.    Homogenates of ob/ob-mouse islets were prepared and preincubated as described in the materials and methods and then incubated for 10 rain at 37C in the presence of different concentrations of glibenclamide. Data is expressed as mean values for primary data and difference from control 4-SEM for the number of experiments indicated in parentheses, n.s. denotes P > 0.05 for difference from control.
Earlier results showed that glibenclamide inhibited the activity of the Na +/K + ATPase in HIT-cells by as much as 40% at a drug concentration of 50 nM. [121 We studied the effect of glibenclamide on ouabain-sensitive ATP hydrolysis in islet homogenates by using concentrations of the drug ranging from 0 to 50 btM in order to characterise the proposed effect on Na +/K + ATPase of pancreatic fl-cells. Table IV shows that no significant effect of glibenclamide on the ATPase activity was observed at drug concentrations up to l btM, whereas 50tM caused a small reduction of probable statistical significance (Tab. IV).

Effect of Diazoxide
The effect of diazoxide on the rate of 86Rb+ influx was studied. Diazoxide at 0.4 mM inhibited the ouabain-sensitive 86Rb + influx by 38% (P 0.005; n= 15), while the ouabain-resistant portion was stimulated (62%; P 0.001; n 15) and the total influx was increased (31%;

DISCUSSION
The present results show that D-glucose, already at low concentrations (5 mM), causes maximum stimulation of the Na +/K + pump, as indicated by the ouabain-sensitive 86Rb+ influx. This finding is in accordance with previous results indicating that D-glucose lowers the islet Na + content in an ouabain-sensitive manner 9,21] and that this effect also appears to reach its maximum at 5mM glucose. [211 Also, the present results are in accord with the previous finding of glucose-induced increase in the rate of fractional outflow of 22Na + in islets, [22] although this effect was not abolished by ouabain. [22] The Na +/K + pump participates in the generation of the resting membrane potential in pancreatic fl-cells. 81 It has been suggested that as much as 75-80% of the basal energy production in the endocrine pancreas is consumed by the Na +/K + pump. [23] Our results show that the Na +/K + pump is active even in the absence of glucose when the fl-cells are resting. This basal activity amounts to more than 60% of the maximum Na +/K + pump activity seen in the presence of glucose. One explanation to consider for the glucose-induced stimulation of the Na +/K + pump in the present results could be that incubation in the absence of glucose causes a relative ATP deficiency and that the observed glucose effect simply represents a difference from starvation. If this was the case, one would expect that prolonging the treatment period in the absence of glucose from 5min to 60 or 180min should further aggravate this condition. However, when the islets were exposed to prolonged in vitro starvation (0mM glucose) we found no further change in the Na +/K + pump activity. The ATP supply in the fl-cells seemed to be enough to support the basal activity of the Na +/K + pump and to allow a seemingly normal glucose-stimulation of the pump after 180 min of in vitro starvation. It thus seems unlikely that the glucose effect is mainly due to increased metabolism-induced ATP production. It seems more likely that the depolarisation of the fl-cell by D-glucose stimulates the Na+/K + pump as part of the repolarisation of the cell. This idea is supported by the observations that the stimulatory glucose effect on the pump was reversed by diazoxide and that the effects of glucose and glibenclamide were not additive.
The present observation that pre-treatment of islets with a high glucose concentration (20 mM) decreases the subsequent glucose-induced stimulation of the Na +/K + pump is of interest, suggesting that this glucose effect is relatively transient. This may indicate that the Na +/K + pump is activated by glucose mainly during the acute, first phase of glucose stimulation. The apparent discrepancy between the present results and earlier reports, indicating that the glucose-induced lowering of the Na + content is not attributable to the activation of the Na +/K + pump, 91 could, at least in part, be explained by the transient appearance of the glucose effect, since the islets were pre-treated with the sugar for 60 min. Lg] It could be difficult to identify the remaining part of a transient glucose effect on the pump. Of course, glucose is likely to affect the fl-cell Na + content in several ways.
Sulphonylureas are thought to act mainly by stimulating the pancreatic fl-cells secretion of insulin. L241 It has been shown that these drugs affect several parameters in the islets, such as to reduce 86Rb+ permeability, [25'261 increase 36C1permeability, 271 increase 45Ca2 + uptake, [26,28,29] and depolarise the fl-cell membrane. [30 It has also been proposed that the sulphonylurea glibenclamide, directly inhibits the Na +/K + ATPase in fl-cells. [121 We were not able to detect any direct inhibitory effect of glibenclamide on the Na +/K + ATPase at 0.01-1.0 gM of the drug. In this concentration range glibenclamide stimulates the fl-cell secreting machinery. 311 The slight inhibitory effect of glibenclamide detected at the highest drug concentration is of questionable statistical significance and could be due to a direct interaction with the lipid membrane. [32] The present finding of a stimulatory effect of glibenclamide on the Na +/K + pump is in line with the idea of a role for the Na +/K + pump in electrical repolarisation. Sulphonylureas cause insulin secretion by closing the ATP-sensitive K+-channels and thereby depolarising the flcells. [33,34] The observation of similar effects of D-glucose and glibenclamide on the Na +/K + pump supports the idea that agents, which are causing electrical depolarisation of the fl-cells, also indirectly stimulate the Na +/K + pump as part of the repolarisation process. This idea is supported by the fact that diazoxide, an opener of ATP-sensitive K +-channels [34,35] and inhibitor of insulin secretion by hyperpolarising the fl-cells, [36] inhibited the basal Na +/K + pump activity. When glibenclamide and diazoxide were added in combination, the pump activity returned to control level. This could be explained by the opposite effects of these drugs on KTP channel activity and thereby on the membrane potential. The idea is also supported by previous studies indicating that diazoxide reversed the tolbutamide-induced increase in [Ca2+li while it did not affect the rise in [Ca2+]i caused by high K +. [37] Previous studies have shown that in contrast to the effect of D-glucose, tolbutamide did not lower the total islet Na + content [91 and indeed slightly increased the intracellular Na + activity, [11 which may partly be due to stimulated entry of Na + . [38] The present results show that both nutrient and non-nutrient insulin secretagogues, which act by depolarising the fl-cell membrane, also stimulate the ouabain-sensitive 86Rb+ influx.
The stimulation is abolished by diazoxide. We are unable to show any direct effects of Dglucose [391 or glibenclamide (this study) on the Na +/K + ATPase activity in islet homogenates.
We therefore suggest that the activation of the Na +/K + pump by those secretagogues occurs secondary to membrane depolarisation and that it may contribute to membrane repolarisation.