The impact of modest but prolonged (3 h) exposure to high physiological glucose concentrations and hyperkalemia on the insulin secretory and phospholipase C (PLC) responses of rat pancreatic islets was determined. In acute studies, glucose (5–20 mM) caused a dose-dependent increase in secretion with maximal release rates 25-fold above basal secretion. When measured after 3 h of exposure to 5–10 mM glucose, subsequent stimulation of islets with 10–20 mM glucose during a dynamic perifusion resulted in dose-dependent decrements in secretion and PLC activation. Acute hyperkalemia (15–30 mM) stimulated calcium-dependent increases in both insulin secretion and PLC activation; however, prolonged hyperkalemia resulted in a biochemical and secretory lesion similar to that induced by sustained modest hyperglycemia. Glucose- (8 mM) desensitized islets retained significant sensitivity to stimulation by either carbachol or glucagon-like peptide-1. These findings emphasize the vulnerability of the β-cell to even moderate sustained hyperglycemia and provide a biochemical rationale for achieving tight glucose control in diabetic patients. They also suggest that PLC activation plays a critically important role in the physiological regulation of glucose-induced secretion and in the desensitization of release that follows chronic hyperglycemia or hyperkalemia.
when challenged in vivo or in vitro with increasing concentrations of glucose, a number of β-cell-signaling pathways are activated, culminating in dose-dependent increases in insulin secretion. The proportionate activation of exocytosis to meet the challenge of hyperglycemia depends upon the tight coupling between glucose metabolism and the activation of a number of established second messenger systems (23, 26, 27, 50). In vivo, the secretory response to glucose is amplified by a number of diverse potentiators, including glucagon-like peptide-1 (GLP-1) and vagally derived acetylcholine. We reported recently (37) that sustained hyperglycemia, 3 h of exposure to 20 mM glucose, resulted in parallel impairments in both insulin secretion and PLC activation. Since the secretory and biochemical lesions induced by high glucose were faithfully reproduced by cholinergic stimulation, we suggested that PLC was intimately involved in both processes. The results from these studies were also in accord with previous findings made by us (42, 45–47) using islets desensitized with tolbutamide, glucosamine, monomethylsuccinate, or forskolin. Prior sustained exposure to any of these compounds results in a deterioration of glucose-induced secretion that is paralleled by impaired glucose-induced activation of PLC.
Not addressed in these prior experiments, but an area replete with important clinical overtones, is the potential vulnerability of the β-cell to modest but sustained hyperglycemia. This information is particularly important at this time because of human studies demonstrating the profound impact of even modest elevations in the ambient glucose level on insulin secretion and the progression to type 2 diabetes (17, 33). This information is provided in the present series of studies. The results emphasize not only the exquisite sensitivity of the β-cell to glucose, but also the vulnerability of the secretory process to this fuel.
MATERIALS AND METHODS
The detailed methodologies employed to assess insulin output from collagenase-isolated islets have been previously described (48, 49). Male Sprague-Dawley rats were purchased from Charles River and weighed 325–425 g when studied. All animals were treated in a manner that complied with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. The animals were fed ad libitum. After Nembutal (pentobarbital sodium, 50 mg/kg; Abbott, North Chicago, IL)-induced anesthesia, islets were isolated by collagenase digestion and handpicked, using a glass loop pipette, under a stereomicroscope into Krebs-Ringer bicarbonate (KRB) supplemented with 3 mM glucose. They were free of exocrine contamination.
After isolation, islets were loaded onto nylon filters (Sefar America, Kansas City, MO). Some islets were immediately perifused, whereas others were subjected to a 3-h incubation with various agonists prior to perifusion (see below). All islets were perifused with KRB buffer at a flow rate of 1 ml/min for 30 min in the presence of 3–5 mM glucose to establish basal and stable insulin secretory rates. After this 30-min stabilization period, they were perifused with the appropriate agonist or agonist combinations, as indicated in the figure legends and in results. Perifusate solutions were gassed with 95% O2-5% CO2 and maintained at 37°C. Insulin released into the medium was measured by radioimmunoassay (1).
After being loaded onto nylon filters, other groups of islets were incubated for 3 h. The filter, with islets attached, was placed in a small glass vial. KRB solution (400 μl), supplemented with the 5–10 mM glucose, was then added. In some studies, 15 mM KCl was added together with 5 mM glucose. The vials were capped with a rubber stopper and then gently aerated for 10 s with 95% O2-5% CO2 and maintained at 37°C. After 90 min, they were aerated again for 10 s. After the 3 h, they were perifused as described above and in the figure legends.
Groups of 18–26 islets were loaded onto nylon filters and incubated for 3 h in a myo-[2-3H]inositol-containing KRB solution made up as follows: 10 μCi of myo-[2-3H]inositol (specific activity 16–23 Ci/mmol) were placed in a 10 × 75 mm culture tube. To this aliquot of tracer, 255 μl of warmed (to 37°C) and oxygenated KRB medium supplemented with 5 mM glucose, 5 mM glucose plus 15 mM KCl, or 8 mM glucose were added. After being mixed, 240 μl of this solution were gently added to the vial with islets. The vial was capped with a rubber stopper, gassed for 10 s with 95% O2-5% CO2, and incubated at 37°C. After 90 min, the vials were again gently oxygenated. After the labeling period, the islets were washed with 5 ml of fresh KRB and perifused. Samples (200 μl) were analyzed every 2 min for [3H]inositol radioactivity from minute 28 to minute 70 of the perifusion. Fractional efflux rates were calculated as described previously (11, 38, 41).
Inositol phosphate studies.
Groups of 18–26 islets were loaded onto nylon filters and incubated for 3 h in a myo-[2-3H]inositol-containing KRB solution made up as follows: 10 μCi of myo-[2-3H]inositol (specific activity 16–23 Ci/mmol) were placed in a 10 × 75 mm culture tube. To this aliquot of tracer 255 μl of warmed (to 37°C) and oxygenated KRB medium supplemented with 5–10 mM glucose were added. After being mixed, 240 μl of this solution were gently added to the vial with islets. The vial was capped with a rubber stopper, gassed for 10 s with 95% O2-5% CO2, and placed in a metabolic shaker at 37°C. After 90 min, the vials were again gently oxygenated. After the labeling period, the islets were washed with 5 ml of fresh KRB and used for inositol phosphate (IP) measurements (see below).
After being washed to remove free labeled inositol, the islets on nylon filters were placed in small glass vials. Added gently to the vial were 400 μl of warmed (to 37°C) KRB supplemented with 10 mM LiCl to prevent IP degradation and 20 mM glucose. The vials were capped, and after 30 min, the generation of IPs was stopped by adding 400 μl of 20% perchloric acid. Total IPs formed were then measured using Dowex columns as described previously (4, 39)
Hanks’ solution was used for the islet isolation. The perifusion medium consisted of 115 mM NaCl, 5 mM KCl, 2.2 mM CaCl2, 1 mM MgCl2, 24 mM NaHCO3, and 0.17 g/dl bovine serum albumin. The 125I-labeled insulin for the insulin assay and the [3H]inositol were purchased from PerkinElmer Life Sciences (Boston, MA). Bovine serum albumin (RIA grade), fatty acid-free albumin, glucose, isradipine, carbachol, GLP-1, and the salts used to make the Hanks’ solution and perifusion medium were purchased from Sigma (St. Louis, MO). Rat insulin standard (lot no. 615-ZS-157) was the generous gift of Dr. Gerald Gold, Eli Lilly (Indianapolis, IN). Collagenase (type P) was obtained from Roche (Indianapolis, IN).
Statistical significance was determined using Student’s t-test for unpaired data or analysis of variance. A P value ≤0.05 was taken as significant. Values presented in the figures and results represent means ± SE of at least three observations.
Acute physiological hyperglycemia and insulin secretion from perifused rat islets.
Experiments from numerous laboratories, including our own, often use 15–27 mM glucose to stimulate the islet. However, these studies have no physiological correlate, since these degrees of hyperglycemia rarely, if ever, exist in vivo. Since one primary goal of these experiments was to establish how sustained physiological elevations in the ambient glucose level influence subsequent β-cell responses, we first confirmed the sensitivity of our preparation to physiological increments in glucose. As shown in Fig. 1, modest elevations in the ambient glucose level from 3 to 10 mM concentrations that encompass physiological excursions usually observed in vivo markedly increased insulin secretory rates. When measured 35–40 min after the onset of stimulation and compared with 3 mM glucose (23 ± 4 pg·islet−1·min−1, n = 6), release rates increased 3.5-, 5.0-, 8.7-, and 13.3-fold with 6, 7, 8, and 10 mM glucose, respectively. Although not depicted in this figure, 20 mM glucose stimulation resulted in an ∼25-fold increase in secretion rates above those noted with 3 mM glucose alone. For example, insulin release rates during the final 5 min of stimulation with 20 mM glucose averaged 584 ± 14 pg·islet−1·min−1 (n = 13). These findings agree well with results obtained using the perfused rat pancreas (16, 20, 29) or human islets studied in vivo (14, 15, 36) and attest to the physiological integrity of the collagenase-isolated islets used in these studies.
Additional studies using freshly isolated islets demonstrated that 15–30 mM KCl, in the presence of 5 mM glucose, evokes a brisk secretory response that was most dramatic during the initial min of stimulation and declined thereafter (Fig. 2). Like the secretory response to high glucose (22, 41), release to the cation was vulnerable to inhibition by the calcium channel antagonist isradipine (80 nM).
Effects of sustained physiological hyperglycemia and hyperkalemia on β-cell secretory responses.
After a 3-h incubation with 5 mM glucose, islets exhibited brisk secretory responses to 20 mM glucose (Fig. 3). However, if the glucose level used during the 3-h incubation was increased from 7 to 10 mM, an entirely different secretory picture emerged. Although similar secretory responses to 20 mM glucose were seen from 6 mM glucose-preincubated islets, release from 7 to 10 mM glucose-preincubated islets was dramatically reduced. For example, after only a 3-h exposure period to 7 mM glucose, peak secretory responses to subsequent 20 mM glucose stimulation decreased from 385 ± 45 pg·islet−1·min−1 (n = 16) to 135 ± 12 pg·islet−1·min−1 (n = 6), a decline of ∼65–70%.
The alteration in 20 mM glucose-induced insulin secretion induced by prior 3-h exposure to 10 mM glucose could not be attributed to the potential presence of contaminating fatty acids in the BSA usually employed in these studies. Desensitization by 10 mM glucose was readily induced using 3-h exposure to 10 mM glucose using fatty acid-free BSA. For example, secretion rates measured 35–40 min after the onset of 20 mM glucose stimulation averaged 376 ± 34 pg·islet−1·min−1 (n = 8) from 5 mM glucose-pretreated islets but decreased to 94 ± 14 pg·islet−1·min−1 (n = 8) from 10 mM glucose-pretreated islets exposed to fatty acid-free albumin during the 3 h prior to the perifusion.
A similar reduction in 20 mM glucose-induced secretion was also evident if islets were incubated for 3 h with 15 mM KCl in the presence of 5 mM glucose (Fig. 4). Similar to its inhibitory effect on 15 mM KCl-induced secretion, the inclusion of the calcium channel antagonist isradipine (80 nM) significantly reduced the desensitizing effect of hyperkalemia on 20 mM glucose-induced release.
Effects of sustained physiological hyperglycemia and hyperkalemia on β-cell PLC responses.
Our previous study (37), where islets desensitized by 10–20 mM glucose or the cholinergic agonist carbachol was used, demonstrated that the secretory lesion caused by these compounds could not be attributed to any effect on insulin content or altered glucose metabolism. However, a defect in PLC activation was observed in desensitized islets (37). The impact of modest but sustained physiological hyperglycemia or hyperkalemia of this signal transduction pathway was investigated next. Prior to stimulation with 20 mM glucose, the efflux of label was slightly higher in the groups of islets exposed to 15 mM potassium or 8 mM glucose during the labeling period (Fig. 5). This in all likelihood reflects the dephosphorylation of IPs formed during the labeling period, their subsequent hydrolysis, and efflux of [3H]inositol from the β-cell. Of particular significance, islets chronically exposed to 8 mM glucose or to 5 mM glucose plus 15 mM KCl exhibit a marked impairment in 20 mM glucose-induced [3H]inositol efflux compared with islets incubated with 5 mM glucose.
The efflux of [3H]inositol is a convenient, albeit indirect, approach to monitor both the activation of PLC and insulin secretion from the same group of islets. This methodology has been employed by a number of investigators, including us (2, 28, 34, 40, 44). To strengthen the association between defective glucose-induced PLC activation and impaired secretion, we directly measured the generation of labeled inositol phosphates in control and desensitized islets. The data are given in Table 1. After a 3-h labeling period in 5 mM glucose, subsequent stimulation of these islets with 20 mM glucose resulted in a marked accumulation of IPs. This response was significantly blunted, in a dose-dependent manner, if islets were labeled in the presence of 8–10 mM glucose and then stimulated with 20 mM glucose.
Restorative effects of carbachol or GLP-1 on desensitized islets.
In the preceding series of experiments where islets desensitized by modest increments in the ambient glucose concentration were used, we employed 20 mM glucose to assess the severity of the lesion in the secretory process. In the final series of experiments, two additional issues were addressed: 1) what is the impact of modest but sustained hyperglycemia (8 mM glucose) on the secretory response to 10 mM glucose, and 2) what is the impact of several physiological and pharmacological potentiators of secretion on release from these islets?
As shown in Fig. 6, prior 3-h exposure to 8 mM glucose virtually abolished the subsequent insulin secretory response to 10 mM glucose alone. Also depicted in this figure are the secretory responses of 8 mM glucose-desensitized islets to the further addition of the cholinergic agonist carbachol (10 μM) or the incretin factor GLP-1 (100 nM) to 10 mM glucose-stimulated islets. In each case a significant restorative effect on release was noted with the addition of the cholinergic agonist carbachol or the incretin factor GLP-1.
Long-term islet perifusions recapitulate observations made using static incubations.
It might be reasonably argued that the impairment in insulin secretion induced by modest hyperglycemia may be an artifact of the static incubation protocol and may not necessarily reflect secretory events occurring during a dynamic perifusion. To address this important issue, additional studies were performed. After isolation, islets were perifused for 30 min with 5 mM glucose. For an additional 3 h, they were stimulated with 8 or 10 mM glucose (Fig. 7). In response to these higher glucose levels, a rapid biphasic secretory response was evoked. Maximal secretory rates, achieved after ∼40–60 min of stimulation, peaked at ∼400 pg·islet−1·min−1. As the perifusion progressed, however, rates of secretion declined and after 3 h of stimulation were reduced >70–75% compared with peak secretion rates measured during the first hour of stimulation. For example, during the final 10-min collection period, the secretory response to 8 or 10 mM glucose averaged 89 ± 16 or 112 ± 17 pg·islet−1·min−1to these glucose levels respectively. These secretory findings are reminiscent of those previously reported by Bolaffi and coworkers (8–10) and Curry (12). Insulin content measured at the termination of the perifusion averaged 184 ± 14 (n = 4) and 172 ± 18 (n = 5) ng/islet in the 8 and 10 mM glucose-stimulated islets respectively.
On the basis of previous studies (8–10, 21, 43), it is clear that sustained hyperglycemia, even modest elevations within the high physiological range, induces a significant functional alteration in the ability of the pancreatic β-cell to secrete insulin. A central issue replete with important clinical overtones is the level of glucose necessary to induce the secretory anomaly in the β-cell and the identification of the signaling system that is so vulnerable to hyperglycemia. The potential significance of these issues has been highlighted by two recent in vivo observations. First, a progressive loss of first-phase secretion to glucose is noted with even modest elevations of the fasting glucose within the high physiological range (17). Second, it has been demonstrated (33) that higher fasting plasma glucose levels within the physiological range are an independent risk factor for the future development of type 2 diabetes. Thus determining the nature of the biochemical lesion responsible for hyperglycemia-induced desensitization and possible means to reverse it assume potential clinical significance.
A necessary prerequisite to establish the identity of the biochemical defect that culminates in the secretory lesion characteristic of type 2 diabetes is the employment of a system that retains the appropriate insulin secretory response characteristics. For example, it has been demonstrated numerous times (3, 12, 16, 19, 25, 29, 30) using the perfused rat pancreas preparation that, in response to sustained hyperglycemia, a brisk biphasic secretory response from the β-cell occurs. The magnitude of the secretory response to high glucose is invariably more than 20-fold above prestimulatory rates of insulin release. As shown in Fig. 1, our collagenase-isolated islets retain a similar degree of responsiveness to high glucose, adding credence to the significance of the biochemical findings observed. Unfortunately, the vitality of the β-cell’s response to glucose is an often overlooked issue, particularly in the many studies involving tumoral cell lines that usually display quite aberrant responses compared with results obtained using the perfused pancreas preparation or the type of islets used herein.
The activation of PLC by glucose or other agonists is not without potential drawbacks, however. The sustained activation of PLC by a large number of compounds desensitizes islets to subsequent stimulation (37, 43, 47, 50). In the present report, modest levels of glucose within the high physiological range were employed to induce desensitization. We observed that sustained exposure to glucose levels as low as 7 mM induced a significant secretory lesion, a finding reported previously by Bolaffi et al. (9). Altered secretory responsiveness induced by modest hyperglycemia (8 mM glucose) was also paralleled by a defect in glucose-induced PLC activation. As demonstrated in other studies (37), the reduction in insulin secretion caused by elevated glucose levels cannot be attributable to any alteration in islet insulin stores or in the ability of the islet to metabolize glucose. Moreover, and as shown here, it cannot be attributed to the potential presence of fatty acids in the albumin used. Of course, we cannot exclude the possibility that in vivo fatty acids may play some role, although this remains controversial. The most logical interpretation of the present data, and one consistent with a number of other reports from this lab, is that irrelevant of the agonist used to desensitize islets, the biochemical lesion responsible for defective secretion is the impaired activation of PLC (42, 43, 45, 51).
Our conclusion that impaired activation of PLC is responsible, at least in part, for glucose toxicity has been reinforced by our prior studies with the cholinergic agonist carbachol (37). Sustained exposure to this muscarinic agonist impairs both glucose-induced PLC activation and secretion. We considered it prudent to explore this concept further by using hyperkalemia to desensitize the β-cell. Because of its dramatic effect on the resting membrane potential, many investigators have employed hyperkalemia to stimulate secretion. Like glucose, increases in the potassium level bathing them results in a rapid secretory response (6, 18). Unlike glucose, however, potassium has no known effect on the metabolism of the β-cell. By setting a new Nernst equilibrium potential, hyperkalemia depolarizes the β-cell membrane, an electrical event that causes the opening of voltage-gated calcium channels. The resulting increase in calcium activates a number of calcium-dependent response elements, including PLC. While initially increasing the release of insulin, the sustained activation of this enzyme results in desensitization of the secretory response, a finding first reported by Björklund and Grill (7). An obligatory role for the calcium-dependent activation of PLC in the multiple effects that potassium exerts on the β-cell is supported by several findings. First, hyperkalemia activates PLC (5, 6, 35) and stimulates insulin secretion (2). Second, preventing the migration of calcium into the β-cell with calcium channel antagonists reduces the acute insulin stimulatory effect of the cation. Third, potassium-induced PLC activation is also impaired in the absence of calcium (5). Finally, as shown herein, sustained exposure to 15 mM potassium culminates in desensitization to glucose stimulation. The simplest interpretation of these data, and one in accord with the action of a variety of agonists, including glucose and carbachol, on the β-cell is that, whereas short term, acute activation of PLC participates in the stimulation of secretion, its sustained activation culminates in a decisive secretory impairment.
In the final set of experiments, we explored the potential impact of several potentiators of secretion on glucose-desensitized islets. The glucose-dependent contribution of these agonists to the secretory response is an elegant physiological arrangement. By amplifying the stimulatory effect of glucose on the β-cell, they not only reduce the duration and magnitude of glycemia necessary to evoke a secretory response, they also insure that the secretory response will be commensurate to satisfy insulin requirements. In islets desensitized by prior 3-h exposure to 8 mM glucose (144 mg%), subsequent stimulation with 10 mM glucose failed to evoke secretion. However, inclusion of carbachol or GLP-1 resulted in a significant restoration of release. Thus, even in a β-cell refractory to the primary stimulatory effect of glucose alone, significant sensitivity to other physiological agonists is retained. In the case of carbachol, the restorative effect on secretion is most likely mediated by a PLC isozyme distinct from the one activated by glucose (24). The increased generation of phosphoinositide-derived second messengers by the cholinergic agonist compensates for the inability of glucose to activate PLC. GLP-1, on the other hand, by increasing islet cAMP levels in islets, exerts a similar sensitizing effect on the exocytotic apparatus (13, 23, 41). This elegant supplementary stimulatory arrangement insures that a severe lesion in glucose recognition does not culminate in the complete absence of an in vivo insulin secretory response.
With the realization that even modest increases in the glucose level bathing them may have adverse consequences on β-cell performance, the glucose level diagnostic of diabetes has been reduced from 140 to 126 mg%. However, the nature of the biochemical lesion induced by such a modest elevation of glucose, although recognized as being clinically important, has not been identified. The present studies may represent a potentially important step toward identifying the lesion induced by hyperglycemia. Modest but sustained elevation of glucose over the physiological range of 7–10 mM for as short as 3 h results in significant decreases in the ability of a subsequent glucose stimulus to increase insulin secretion. The cause of the aberrant release profile appears to reside in the inability of glucose to activate PLC. Our working hypothesis is that, although acute and temporally appropriate increases in PLC activation participate in a positive fashion to the evocation of secretion, chronic activation of the enzyme results in negative feedback regulation that impairs the capacity of glucose to subsequently activate the same enzyme. One possibility is that the phosphorylation of key amino acid residues of PLC is responsible for these observations (31, 32). Alternatively, considering the calcium dependency of the enzyme, reduced or limited availability of the divalent cation may play an important regulatory role as well.
In conclusion, the studies reported herein provide a potential biochemical explanation for the clinical observation that even modest but sustained increases in plasma glucose are deleterious to the β-cell. Our position is that a critically important enzyme, PLC, is involved. It is interesting to also point out that, although β-cell rest with diazoxide has been suggested as a potential therapy in diabetes, this compound has also been shown to protect islet PLC from the adverse impact of hyperglycemia as well (37). Attention might now be focused on the potential therapeutic usefulness of the observations made in this report.
These studies were supported by National Institutes of Health Grant no. 41230 to W. S. Zawalich
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by American Physiological Society