K_ATP Channels and Pancreatic Islet Blood Flow in Anesthetized Rats: Increased Blood Flow Induced by Potassium Channel Openers

Male Sprague-Dawley rats, weighing 325–350 g, were obtained from a local breeding colony (Biomedical Center, Uppsala, Sweden) and were used in all experiments. All animals had free access to tap water and pelleted rat food. The use of animals was in accordance with international guidelines (NIH 85-23) and were approved by the local animal ethics committee at Uppsala University.

The rats were anesthetized with an intraperitoneal injection of thiobutabarbital sodium (130 mg/kg body wt, Inactin; Research Biochemicals International, Natick, MA) and placed on a heated operating table to maintain body temperature. Polyethylene catheters were inserted into the ascending aorta, via the right carotid artery, and into the left femoral artery and vein. The former catheter was connected to a pressure transducer (PDCR 75/1; Druck Ltd., Groby, U.K.), whereas the latter was used to infuse Ringer solution (5 ml · kg body wt⁻¹ · h⁻¹) to substitute for fluid losses. When the blood pressure had remained stable for at least 15 min an intravenous injection of diazoxide (Hypertstat; Schering-Plough AB, Stockholm, Sweden; 10, 20, or 30 mg/kg body wt) or the corresponding volume (2 ml/kg body wt) of saline was given. The blood flow measurements were then performed 20 min later.

The arterial blood perfusion of the whole pancreas, islets, duodenum, colon, adrenal glands, and kidneys was then measured with a microsphere technique (12). Briefly, a total of 1.5–2.0 × 10⁵ nonradioactive microspheres (EZ-Trac; Triton Microspheres, San Diego, CA), with a diameter of 10 μm, was injected via the catheter with its tip in the ascending aorta during 10 s. Starting 5 s before the microsphere injection, and continuing for a total of 60 s, an arterial blood sample was collected by free flow from the catheter in the femoral artery at a rate of ∼0.40 ml/min. The exact withdrawal rate was confirmed in each experiment by weighing the sample. Arterial blood was collected from the carotid catheter for determination of blood glucose and serum insulin concentrations as given below. The animals were then killed, and the pancreas and adrenal glands were removed in toto, blotted, weighed, and treated with a freeze-thawing technique, which visualized the pancreatic islets and microspheres (12). Approximately 100 mg each of the duodenum (around the papilla), colon (descending part), and left kidney (slice through the center) were also removed and treated in the same way. The number of microspheres in these samples were then counted in a microscope equipped with both bright and dark field illumination. The blood flow values were calculated according to the formula Q_org = Q_ref × N_org/N_ref, where Q_org is organ blood flow (ml/min), Q_ref is withdrawal rate of the reference sample, N_org is number of microspheres present in the organ, and N_ref is number of microspheres in the reference sample. Blood flow values based on the microsphere content of the adrenal glands were used to confirm that the microspheres were adequately mixed in the circulation. A difference <10% in the blood flow values was taken to indicate sufficient mixing.

The number of microspheres in the arterial reference sample was determined by sonicating the blood, then transferring samples to glass microfiber filters (pore size <0.2 μm), and then counting the number of microspheres in the same microscope as referred to above.

The surgical preparation was performed as given above. NNC 55-0118 (6-chloro-3-isopropylamino-4H-thieno[3,2-e]-1,2,4-thiadiazine 1,1-dioxide; Novo Nordisk A/S, Måløv, Denmark) was administered intravenously (3 mg/kg body wt) 20 min before the blood flow measurements. Control animals were given the same volume (2 ml/kg body wt) of vehicle alone. The blood flow measurements were performed in the same organs and with the microsphere technique referred to above.

The surgical preparation was performed as given above. Glipizide (Sigma, St. Louis, MO) was dissolved in a 50% solution of DMSO and then administered intravenously (25 μg/kg body wt) 45 min before the blood flow measurements. This time point was chosen since it was the earliest providing a significant decrease in blood glucose concentrations. Control animals were given the same volume (0.5 ml/kg body wt) of vehicle alone. The blood flow measurements were performed with the microsphere technique referred to above. However, only the blood perfusion to the whole pancreas, islets, and the adrenal glands was determined in these animals.

Blood glucose concentrations were measured with test reagent strips (Medisense; Svenska MediSense, Solna, Sweden). Serum immunoreactive insulin concentrations were measured with enzyme-linked immunosorbent assay (ELISA) (Rat Insulin ELISA; Mercodia AB, Uppsala, Sweden).

All values are given as means ± SE. Probabilities (P) of chance differences were calculated with Students unpaired t test or ANOVA with Bonferroni’s correction (SigmaStat; SSSPD, Erfart, Germany) A value of P < 0.05 was considered to be statistically significant.

As expected, diazoxide induced a marked, dose-dependent decrease in mean arterial blood pressure (Table 1). Blood glucose concentrations were increased to the same extent by all doses of diazoxide, whereas serum insulin concentrations were decreased only by the highest dose (30 mg/kg; Table 1). There was also a tendency to a decline in serum insulin when diazoxide was given at a dose of 20 mg/kg, although it did not attain statistical significance (P = 0.056).

Total pancreatic blood flow was increased by all doses of diazoxide, but no further increase was seen when the dose was raised from 20 to 30 mg/kg (Fig. 1). Islet blood flow was increased following administration of the two highest doses of diazoxide, with the most pronounced effect seen in animals receiving 20 mg/kg (Fig. 2). When islet blood flow was calculated as a fraction of total pancreatic islet blood flow, a decrease was seen when 10 or 30 mg/kg of diazoxide was given (Fig. 3). When 20 mg/kg diazoxide was given, the difference was not significant (P = 0.055). This indicates that diazoxide caused a redistribution of the blood perfusion within the pancreas in favor of the exocrine compartment.

The blood perfusion of the kidneys and adrenal glands was unaffected by diazoxide (Table 2). The highest dose of diazoxide induced an increase in both duodenal and colonic blood flow when compared with animals given 10 mg/kg, but not when compared with control rats (Table 2).

Because all doses of diazoxide decreased mean arterial blood pressure, we calculated the vascular conductance of the different organs in which blood flow measurements were made (Table 3). As can be seen, diazoxide at the lowest dose increased pancreatic vascular conductance. The value for adrenal vascular conductance showed an increase, whereas the values for the other organs were unchanged. The two highest doses of diazoxide, however, markedly increased the vascular conductance in all examined organs when compared with both control rats and those given 10 mg/kg diazoxide. No further increase in vascular conductance was seen when rats receiving 30 mg/kg diazoxide were compared with those given 20 mg/kg.

The drug did not affect blood glucose concentrations, but it did decrease serum insulin concentrations and mean arterial blood pressure in the anesthetized rats. No effects on duodenal, colonic, renal, or adrenal blood flow were seen after administration of NNC 55-0118. However, both total pancreatic and islet blood flow were increased after administration of NNC 55-0118. Fractional islet blood flow was not changed after administration of NNC 55-0118 (7.6 ± 0.7 vs. 8.9 ± 0.6%, respectively).

No effects on mean arterial blood pressure (104 ± 3 vs. 103 ± 1 mmHg in vehicle-treated versus glipizide-treated rats, respectively) were seen. Blood glucose concentrations were decreased 45 min after glipizide administration. Serum insulin concentrations were not measured in these animals due to a technical failure. No effects on total pancreatic blood flow were seen, whereas islet blood flow was decreased after glipizide administration. Consequently, fractional islet blood was also diminished.

K_ATP channels are composed of a pore-forming unit that is a member of the inwardly rectifying K⁺ channel family and constitutes either Kir6.1 or Kir6.2 (7). Normally, Kir6.1 is found in vascular smooth muscle cells (VSMCs), whereas Kir6.2 is located in β-cells (7). Recently Kir6.1 knockout mice were found to exhibit VSMCs, which have no detectable K_ATP currents and lack all vasodilator responses to K_ATP channel agonists (13). These animals showed spontaneous bouts of coronary vasospasm and ST-elevation, i.e., they reproduced key features of Prinzmetal angina (13). These findings would suggest that most VSMCs have Kir6.1 as their pore-forming unit. However, other studies have demonstrated that Kir6.2 also may be expressed in some muscle cells, and this confers different properties to VSMCs depending on whether their K_ATP channels contain Kir6.1 or Kir6.2 (14).

The potassium channel also consists of a regulatory subunit, the sulfonylurea receptor (SUR). This part of the channel complex exists in three forms: SUR1 found in β-cells and some neurons, SUR2A in cardiac and skeletal muscle, and SUR2B in VSMCs (7). In confirmation of this, recent experiments on SUR2 knockout mice demonstrated findings similar to those seen in Kir6.1 knockouts (i.e., a Prinzmetal phenotype and loss of K_ATP channels in VSMCs) (15).

The mechanism behind the blood flow increase presently seen after administration of diazoxide is likely to be an opening of VSMC K_ATP channels in pancreatic resistance vessels, i.e., presumably mainly arterioles. This would lead to a hyperpolarization of the muscle cell causing relaxation and vasodilatation (6,16,17). However, diazoxide is not selective for VSMC or β-cell K_ATP channels, but affects both. It should be kept in mind that there are no pharmacological or physiological studies that have determined the type of K_ATP channels present in pancreatic vasculature.

It has been suggested that diazoxide is selective for mitochondrial K_ATP channels in heart muscle (18–20). An effect on these channels has been suggested to mediate the protection conferred by this potassium channel opener against toxins in the heart (21,22) and the central nervous system (23–25). In β-cells, on the other hand, diazoxide has been suggested to act on K_ATP channels in both plasma membrane (26) and mitochondria (27). Indeed, the selectivity of diazoxide for mitochondrial K_ATP channels, irrespective of tissue, is likely to be dose dependent (28). The doses used in the present study are likely to affect channels on both of these cellular locations.

When interpreting the effects on blood flow by diazoxide, it should be noted that there was a marked decrease in mean arterial blood pressure at all doses given. Because organ blood flows other than those to the pancreas remained unchanged, the vascular conductivity was markedly increased in these organs. Such an increase was even more pronounced in the pancreas and presumably also in the islets, leading to an absolute increase in these organ blood flow values. The value for the islets is not included in the table, because we do not have any measurement of the weight of the islet organ, which precludes an exact calculation of vascular conductance. However, it can be estimated, assuming an islet volume of 1–2%, that it is increased to the same degree as that to the whole pancreas. Thus, it seems as if opening of the K_ATP channels induces most pronounced blood flow effects on islets and whole pancreas, despite the low blood pressure, which, nevertheless, is within the limits of autoregulation for the rat pancreas (29) and islets (30). Diazoxide had its most pronounced effects on total pancreatic blood perfusion, as manifested by the decreased fractional islet blood flow (i.e., a lowering of the fraction of blood diverted through the islets). This would suggest that arterioles in the exocrine parenchyma have more K_ATP channels or that the VSMCs in these arterioles are more sensitive to diazoxide. We would like to once again stress that there are no available anatomical or pharmacological studies on intrapancreatic K_ATP channel distribution.

In contrast to diazoxide, the newly described potassium channel opener NNC 55-0118 stimulated total pancreatic and islet blood flow to the same degree. No effects on the other measured regional blood flow values were seen. This is somewhat surprising, because NNC 55-0118 is selective for the Kir6.2/SUR1 K_ATP channel of the β-cell (11). Indeed, the effects on VMSCs are small and previous studies have demonstrated blood pressure effects similar to those in the present study (11). This opens the possibility that the effects of the K_ATP channel openers on islet blood perfusion are not primarily dependent on activation of VMSCs potassium channels (i.e., those primarily containing Kir6.1/Sur2B potassium channels). It can instead be that activation of the β-cell Kir6.2/SUR1 channel and/or that substances released from the islet cells themselves affect the islet arteriolar blood vessels and lead to the marked increase in blood flow. This notion would then explain why diazoxide increases the vascular conductance more in the pancreas and islets, since diazoxide would also affect the common Kir6.1-containing VSMC potassium channels, leading to the general increase in vascular conductance seen in all studied organs, including the whole pancreas. Furthermore, diazoxide would also affect the β-cell K_ATP channel, which could cause a further increase in islet blood flow (see below) and, perhaps, total pancreatic blood flow. Several previous studies have demonstrated the independent regulation of islet blood flow when compared with the exocrine pancreas (1,2).

In favor of such an interpretation is our finding after administration of glipizide, i.e., a closer of K_ATP channels, with a selectivity for the β-cell Kir6.2/SUR1 form. Glipizide leads to a decrease in islet blood flow but had no effects on total pancreatic blood flow. In line with the anatomical organization of islet afferent blood supply (see above), it can also be that a substance affecting the islets may release substances that can diffuse to the islet arterioles and directly affect VSMCs. Immunolocalization of K_ATP channels in rat islets have demonstrated that SUR1 and Kir6.2 are found not only in β-cells, but also in α-, δ-, and pancreatic polypeptide cells (31). This implies that any of these endocrine cells can be a source of a substance(s) that may affect the islet afferent arteriole. At present, the nature of such a mediator is unknown. Glucagon is worthy of mentioning in this context because it is a vasoactive hormone (32). Furthermore, the administered glipizide caused hypoglycemia, which is likely to have increased glucagon release. Preliminary findings with insulin-induced hypoglycemia also demonstrate a decreased pancreatic islet blood flow, and this response seems to be nervously mediated (P.-O. Carlsson, L.J., unpublished observations). Thus, to what extent the decreased islet blood flow seen after glipizide administration is due to an effect on K_ATP channels or is nervously mediated is at present unknown.

A piece of indirect evidence pointing toward an effect of K_ATP channels in the regulation of normal islet blood flow is the association between these potassium channels and adenosine. We have recently shown that adenosine, probably produced by β-cells, can stimulate islet blood flow during conditions associated with increased islet metabolism (3). Moreover, a coupling between K_ATP channels and adenosine effects on pancreatic vasculature has been suggested in experiments on the perfused rat pancreas (33). In this context, it has been reported that heart microvascular dilation in response to adenosine is enhanced at lower intraluminal pressures by selective activation of VSMC K_ATP channels (5). This should be compared with islets, where the capillary blood pressure is lower than in the exocrine parenchyma, but where the blood flow response to adenosine is more pronounced (3).

To what extent the increased blood glucose concentrations may affect the islet blood perfusion in the present study is of interest. It has previously been demonstrated that hyperglycemia may increase the blood perfusion of the islets (34), presumably mediated by nitric oxide (35). Diazoxide increased the values to ∼9 mmol/l, whereas NNC 55-0118 had no such effects. In view of the similarity on blood flow response but differences in blood glucose concentrations, it seems unlikely that the latter is responsible for the changes in islet blood flow. Furthermore, in a previous study, we observed that the blood glucose concentrations needed for this response to occur in anesthetized animals were higher than those seen in the present study (34). In general, however, the blood flow changes seen in the present study include effects mediated by K_ATP channels as well as more unspecific changes induced by altered blood pressure, blood glucose (see above), and serum insulin concentrations.

In an earlier study, the sulfonylurea tolbutamide was found to increase blood flow to the islets in rats in doses inducing hypoglycemia, but had no effects on total pancreatic blood flow (36). The reasons for the discrepancy between their findings and those in the present study are unknown. In the tolbutamide study, young rats, aged 5 weeks, were used, whereas we studied adult rats ∼15 weeks old. We have previously noted marked differences in the absolute values in islet blood flow depending on the age of the animals, and it is also likely that the blood flow regulation differs (37,38). It is also possible that there are differences in the actions of tolbutamide and glipizide, although this seems less likely. However, in view of previous findings when applying sulfonylureas, we consider a decrease in islet blood flow to be a logical consequence of administering these substances. Thus, depending on the dose, glibenclamide is known to reduce early and peak vasodilatation following reactive hyperemia in humans (39). Furthermore, glibenclamide (20 mg/kg) to rats increased vascular resistance and caused a systemic vasoconstriction (40). Therefore, it seems as if K_ATP channel closers would favor a vasoconstriction rather than a dilation. It may seem physiologically appropriate that insulin release induced by sulfonylureas would be associated with an increased blood perfusion, as suggested by Vetterlein et al. (36). However, in other contexts, we have been able to show that the very high basal blood perfusion of the islets allows for a substantial reduction in islet blood flow before any effects on insulin secretion can be observed (1).

In summary, our results suggest that K_ATP channels play an important role for regulation of total pancreatic and islet blood flow. To what extent effects on islet function are mediated by a modulated islet blood perfusion by substances affecting these channels merits further investigation.

Financial support was received from the Swedish Research Council (72X-109, 72X-8273), the Swedish Diabetes Association, the Swedish-American Diabetes Research Program funded by the Juvenile Diabetes Foundation and the Wallenberg Foundation, the EFSD/Novo Nordisk for Type 2 Diabetes Research Grant, the Novo Nordic Research Fund, and the Family Ernfors Fund.

The authors are grateful to Astrid Nordin for excellent technical assistance.