The Expression of Hormone-Sensitive Lipase in Clonal β-Cells and Rat Islets Is Induced by Long-Term Exposure to High Glucose

RPMI 1640 medium, fetal bovine serum, and streptomycin/penicillin were from Gibco BRL (Life Technologies, Gaithersburg, MD), and Hanks’ balanced salt solution (HBSS), glucose, and glucose analogs were from Sigma Chemical (St. Louis, MO). Collagenase P was from Boehringer Mannheim (Mannheim, Germany). [³²P]dCTP, Hybond-N+, Hybond-C extra, horseradish peroxidase–conjugated donkey anti-rabbit IgG secondary antibody, and enhanced chemiluminescence reagents were from Amersham Pharmacia (Uppsala, Sweden) Biotech; and RNAzol was from Nordic BioSite (Täby, Sweden); and ExpressHyb was from Clontech (Palo Alto, CA).

Male Sprague-Dawley rats that weighed ∼225 g were purchased from Møllegard A/S (Skensved, Denmark). The animals were maintained in a temperature-controlled room (22°C) on a 12-h light-dark cycle with lights on at 06:00. The study was approved by the Animal Ethics Committee at Lund University.

INS-1 cells (a gift from Professor Claes Wollheim, Geneva, Switzerland) were grown as previously described (18) in RPMI 1640 medium containing 11.1 mmol/l glucose supplemented with 10% heat-inactivated fetal bovine serum, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 50 μmol/l β-mercaptoethanol. The cells were cultured at 37°C in a humidified atmosphere (5% CO₂, 95% air). The cells were used at passage numbers 90–100. When cells reached 70–80% confluence, they were washed twice with phosphate-buffered saline (PBS) and then cultured in serum-free medium that contained different concentrations of glucose (or glucose analogs) for the indicated times.

Rat islets were isolated by the collagenase digestion technique. Briefly, 10 ml of cold HBSS supplemented with 1.0 mg/ml collagenase (Collagenase P, activity 1.52 units/mg) was injected into the bile duct, and the pancreas was filled retrogradely. The pancreas subsequently was removed and incubated for 24 min at 37°C. Islets were isolated after washing three times with HBSS. The islets were cultured overnight at 37°C in a humidified atmosphere (5% CO₂, 95% air) in RPMI 1640 medium supplemented with low or high concentrations of glucose (3.3 and 25 mmol/l, respectively), 10% heat-inactivated fetal bovine serum, 100 IU/ml penicillin, and 100 μg/ml streptomycin. After 16 h of incubation, the islets were washed twice with PBS and then lysed in SDS-PAGE sample buffer. Islet protein was measured using a commercial kit (BCA protein; Pierce, Rockford, IL).

After incubation of INS-1 cells and rat islets under different conditions, as indicated in the figure legends, cells were washed twice with PBS and then scraped with a rubber policeman into homogenization buffer (0.25 mol/l sucrose, 1 mmol/l dithioerythritol, 1 mmol/l EDTA, 10 μg/ml leupeptin, 10 μg/ml antipain, 1 μg/ml pepstatin A). Homogenization was performed by passing through a 27-gauge needle five times. The protein concentration was measured in the homogenates using the BCA-protein kit (Pierce). A total of 100 μg protein was separated by SDS-PAGE and electroblotted onto nitrocellulose membranes (Hybond-C extra). Western blot analysis was performed using an affinity-purified rabbit anti–rat HSL antibody, followed by a horseradish peroxidase–conjugated donkey anti-rabbit IgG secondary antibody and enhanced chemiluminescence reagents. The amount of HSL protein was quantified using a CCD camera (LAS 1000; Fuji, Tokyo, Japan).

Homogenates, prepared as for Western blot analysis, were assayed for diglyceride lipase activity using the mono-oleoyl-2-O-monooleylglycerol, a diacylglycerol analog, as substrate (19). One unit of enzyme activity corresponds to 1 μmol of fatty acid released per minute at 37°C. Immunoinhibition was performed by preincubating the cell homogenates with chicken anti–rat HSL antiserum (dilution 1:10) or with preimmune chicken serum (dilution 1:10) for 1 h at room temperature.

Glycerol release from INS-1 cells was measured in medium samples with an automatic luminometric kinetic assay as described previously (20).

INS-1 cells were lysed with 2 ml of RNAzol per 10-cm dish, and total cellular RNA was isolated according to the manufacturer’s instructions. Thirty micrograms of RNA was denatured and electrophoresed on a 1% agarose gel that contained 2.2 mol/l formaldehyde and then transferred to nylon membranes (Hybond-N+). The blots were probed with a full-length rat adipocyte HSL cDNA labeled with [³²P]dCTP using ExpressHyb according to the manufacturer’s instructions. Membranes were analyzed and quantified by digital imaging (Fujix Bas 2000; Fuji). For correction for differences in RNA loading on the gel, HSL mRNA levels were normalized to those of 18S rRNA, quantified through hybridization to a ³²P-end-labeled oligonucleotide (21).

The effect of glucose on HSL mRNA stability was determined by addition of actinomycin D, a potent inhibitor of RNA polymerase II, to the cell culture medium. After an overnight preincubation of INS-1 cells in serum-free RPMI 1640 with 11 mmol/l glucose, cells were switched to 25 or 3.3 mmol/l glucose and 4 μg/ml actinomycin D. Cells were harvested after 2, 4, and 10 h of incubation, and mRNA was extracted and quantified by Northern blot analysis (see above).

Values are expressed as means ± SE. Differences between individual data were compared with Student’s unpaired t test or by one-way analysis of variance, followed by Tukey’s test for individual comparisons. The critical probability used to determine significance was P < 0.05.

The effect of glucose on HSL expression in β-cells was studied by incubating INS-1 cells in 3.3 or 25 mmol/l glucose for 4, 8, 16, and 32 h. Western blot analysis of total cell homogenates revealed that after 4 and 8 h, there was no difference between high and low glucose, whereas after 16 and 32 h, the HSL protein was induced more than twofold (259 ± 20%) at high compared with low glucose concentrations (Fig. 1). The HSL activity, defined as the diglyceride lipase activity that was inhibited by neutralizing anti–rat HSL antibodies, showed the same degree of induction by high glucose concentrations (Table 1). Total diglyceride lipase activity, however, was induced much less, suggesting that lipase activity that is not accounted for by HSL is not increased by high glucose concentrations (Table 2). Preincubation with anti–rat HSL antibodies reduced the diglyceride lipase activity by 25%, which is similar to our previous findings (17).

To determine the dose dependence for the glucose-inducing effect on HSL expression, we incubated INS-1 cells with 1, 5, 10, 15, 20, 25, and 30 mmol/l glucose for 16 h. HSL expression was induced at all concentrations between 10 and 30 mmol/l, with maximum effects observed at 20–25 mmol/l (Fig. 2).

To examine the effect of high glucose concentration on HSL mRNA levels, we performed Northern blot analysis. Incubation of INS-1 cells for 16 h at 25 mmol/l glucose resulted in more than threefold higher levels of the 3.1-kb HSL mRNA found in these cells (17) compared with cells incubated at 3.3 mmol/l glucose (Fig. 3). No difference was observed for 18S rRNA levels, used to correct for differences in RNA loading on the gel. To determine whether induction of HSL mRNA levels can occur in the absence of transcription and whether glucose exposure affects the half-life of HSL mRNA, we incubated INS-1 cells at high (25 mmol/l) or low (3.3 mmol/l) glucose concentrations in the presence of 4 μg/ml actinomycin D. RNA was extracted at different time points and quantified by Northern blot analysis. These experiments demonstrated that actinomycin D inhibited the induction of HSL mRNA by high glucose and, furthermore, that the half-life of the HSL mRNA in INS-1 cells is short (<10 h) with no difference between high and low glucose conditions (data not shown). Consequently, glucose induction of HSL mRNA seems to not be the result of increased stability of HSL mRNA but rather the result of an increased transcription of the HSL gene.

To confirm in normal islet tissue the effect of glucose observed on HSL expression in INS-1 cells, we incubated freshly isolated rat islets for 16 h in RPMI 1640 medium supplemented with 10% serum and 3.3 or 25 mmol/l glucose. Western blot analysis showed a twofold induction of HSL protein at the higher glucose concentration (Fig. 4). Similar induction of HSL protein was seen in INS-1 cells when the incubation was performed in the presence of serum (data not shown). Two isoforms of HSL were found in lysates from rat islets, which is in agreement with an earlier study (17). Both isoforms of HSL were induced at high concentrations of glucose, although not to the same extent. At 3.3 mmol/l glucose, the two isoforms were expressed at equal levels, whereas at 25 mmol/l glucose, the level of expression of the 89-kDa isoform was 50% higher than that of the 84-kDa form.

To confirm that the induction of HSL by glucose correlated with an increased mobilization of triglyceride stores, the release of glycerol, which is an end product in lipolysis, was measured. After 16 h of incubation of INS-1 cells with 3.3 or 25 mmol/l glucose, medium samples were collected from each well and glycerol was measured. The accumulated glycerol was several-fold higher in cells that were incubated with high glucose compared with low glucose (Fig. 5). Cells that were incubated with high glucose concentrations released 17.5 ± 1.5 nmol glycerol/mg protein during the 16-h incubation, whereas cells that were incubated with 3.3 mmol/l glucose released only 1.0 ± 0.05 nmol glycerol/mg protein.

To gain insight into the mechanism of HSL induction by glucose, we investigated the effect of various sugars on the HSL expression. INS-1 cells were incubated in RPMI medium containing 3.3 mmol/l glucose and 21.7 mmol/l of one of various glucose analogs. Cells that were challenged with 25 mmol/l d-glucose showed a significant increase (223 ± 14%) in HSL protein expression (Fig. 6). Mannose, which is metabolized in β-cells and a potent insulin secretagogue, also increased the expression level of HSL (179 ± 9%), whereas fructose had no effect. 2-Deoxyglucose, which is phosphorylated by glucokinase but not metabolized further, was able to stimulate HSL expression significantly (130 ± 4%), albeit to a lower extent. Galactose and 3-orthomethylglucose, two sugars that are transported into the cell but not metabolized, had no effect on HSL expression. l-Glucose and mannitol did not induce HSL expression, showing that the effect of the sugars was not due to changes in osmolarity. Using Northern blot analysis, we investigated the effect of mannose and 2-deoxyglucose at the mRNA level. The result showed that mannose and 2-deoxyglucose induced HSL mRNA, mannose being as effective as glucose, whereas 2-deoxyglucose was less effective (data not shown). Taken together, the results suggest that glucose-6-phosphate is an important mediator of the glucose effect on HSL, although a contribution by additional mediators, generated through metabolization of glucose beyond the glucokinase step, cannot be excluded.

This study shows that long-term (≥16 h) exposure of INS-1 cells and islets to high glucose concentrations causes a concentration-dependent induction of HSL expression. The induction, which is paralleled by an increased lipolysis, is seen at the level of enzyme activity, protein, and as mRNA expression, suggesting that the effect is mediated via increased transcription of the HSL gene. This proposal is supported by the results of actinomycin D decay assays, which showed no significant effect of high glucose on HSL mRNA stability. Mannose and 2-deoxyglucose also induced HSL protein expression in the INS-1 cells, whereas other tested glucose analogs had no effect. The stimulation of HSL expression by 2-deoxyglucose was smaller than the effects seen with glucose and mannose, which may be due to the toxic effects of this analog. Consequently, glucose-6-phosphate may by the effector in stimulating HSL expression. However, the existence of additional mediators, generated through metabolization of glucose beyond the glucokinase step or generated as a result of early mitogenic events induced by glucose, i.e., activation of signal transduction pathways involving mitogen-activated protein kinase, protein kinase C, and Ca²⁺-dependent kinases, cannot be excluded and needs to be investigated further.

The regulation of HSL expression by glucose has been studied previously in adipocytes. With the murine adipocyte cell line 3T3-F442A as a model, deprivation of glucose was found to decrease HSL mRNA and HSL activity levels by a factor of 2–3 (22). In another study in which primary rat adipocytes were used as a model, glucose in combination with insulin was found to increase the levels of basal lipolysis as well as HSL protein expression (23). Glucose alone, however, had no effect. The reason for the discrepancy between the two studies is not known but could be attributed to the different experimental models. Glutamine has been shown not to be required for the glucose effect on induction of basal lipolysis in adipocytes, indicating that the hexosamine pathway is not involved in mediating the effect (23). Glucose analogs have not been tested in the adipocyte models, making it difficult to draw any conclusions about how the effect of glucose on HSL expression is mediated.

In agreement with what has been described previously, we observed two different HSL isoforms in isolated islets: one 84-kDa form and one 89-kDa form (17). INS-1 cells, however, express only the larger isoform. This may reflect either that β-cells express the 89-kDa isoform whereas other islet cell types express the 84-kDa form or that there is a difference between normal β-cells and INS-1 cells. Previous reverse transcriptase–polymerase chain reaction data indicated that the 84-kDa isoform in islets is identical to the main isoform expressed in adipocytes (17). In this study, we found that glucose induces both isoforms, although not to the same extent. Induction of both isoforms also was observed in adipocytes (23), suggesting that in both islets and adipocytes, there is a coordinated regulation of the two HSL isoforms by glucose. The possibility that the 84-kDa protein in islets represents a proteolytic fragment of the larger isoform cannot be excluded.

Because we found that incubation of cells or islets with high glucose for 16 h was required to induce HSL expression, whereas no effect of glucose was evident after 8 h of incubation, it is possible that the glucose effect on HSL is of major relevance during long-term exposure to high glucose. Chronically elevated glucose levels are known to alter β-cell secretory function, as evidenced in studies that showed sensitization to glucose with a markedly augmented secretion at low glucose concentrations and a lack of response to high glucose concentrations after prolonged exposure of β-cells to high glucose (24,25). The mechanisms underlying these adaptive changes of the β-cell to a hyperglycemic environment are only partly understood but may involve altered expression of a number of glucose-regulated genes, in particular glycolytic enzymes and other proteins involved in glucose metabolism. GLUT2 as well as downstream glycolytic enzymes, such as phosphofructokinase isozymes, glyceraldehyde-3-phosphate dehydrogenase, and l-pyruvate kinase, have been shown to be induced in the β-cell line INS-1 (26,27). Glucokinase activity has been found to be increased without altered mRNA levels (28). Insulin gene transcription and proinsulin mRNA translation increase in response to long-term exposure to high glucose levels (29). The massive deposition of glycogen has been proposed to contribute to the sensitization to glucose at low concentrations by channeling glucose to the glycolytic pathway (26). Long-term exposure of INS-1 cells to high glucose also has been found to induce lipogenic enzymes such as fatty acid synthase, acetyl-CoA carboxylase, and malic enzyme (13). Our present study suggests that the induction of HSL expression is part of the adaptive process to a long-term hyperglycemic environment. HSL was recently found to be expressed in β-cells and proposed to be involved in generating lipid-derived signals that are necessary for stimulus-secretion coupling (17). It is of interest that the dose-response relationship between glucose concentration and HSL expression (see Fig. 2) displays a marked glucose effect within the range of hyperglycemic levels, which correspond to the levels seen in patients with type 2 diabetes. It is suggested that the adaptive response to long-term glucose exposure in HSL expression evolves for the generation of lipid signals needed for the augmentation of insulin secretion to normalize the hyperglycemia. The nature of the lipid-derived signals is unknown, but among suggested candidates are malonyl CoA, long-chain acyl CoA, and diglycerides. Long-chain acyl CoA thereby might be the effector signal, augmenting the exocytosis of insulin, whereas malonyl CoA is more of a regulatory signaling molecule, because it inhibits carnitine palmitoyltransferase 1, leading to inhibition of fatty acid oxidation with a resulting accumulation of long-chain acyl CoA in the cytosol (8,30). Other candidate lipid-derived signals are fatty acid derivatives that act as ligands for the peroxisomal proliferating activated receptors, which may regulate genes of importance for the β-cell adaptation to hyperglycemia.

We found evidence of increased lipolytic activity after exposure of β-cells to a high glucose concentration because glycerol release was stimulated under such conditions. Thus, increased HSL expression and activity may contribute to an increased flux of all potential lipid-signaling molecules, through hydrolysis of the increased triglyceride stores. Noteworthy, fatty acid oxidation is decreased as β-cells are exposed to high concentrations of glucose (13), which is in agreement with the notion that simultaneous increase in triglyceride deposition and lipolysis mainly serves the purpose of providing an increased flux of lipid-signaling molecules. The concerted induction of lipogenic and lipolytic enzymes thus is triggered by the increased glycolytic flux and may be part of the mechanism behind the glucose sensitization and altered secretory function observed upon prolonged exposure of β-cells to high glucose concentrations. It may be speculated that the β-cell adaptation to long-term glucose exposure fails in type 2 diabetes, initiating the progressive islet dysfunction. If so, then pharmacological means to augment and support the glucose-induced HSL expression might be a new avenue for preservation of islet function in type 2 diabetes.

This work was supported by the Swedish Medical Research Council (project nos. 11284 [C.H.], 1034 [P.A.], and 6834 [B.A.]), The Swedish Diabetes Association, Center of Excellence Grant from the Juvenile Diabetes Foundation (United States), the Knut and Alice Wallenberg Foundation (Sweden), and the following foundations: Novo Nordisk, Albert Påhlsson, Thelma Zoega, and Crafoord.

We thank Sara Larsson and Lena Kvist for excellent technical assistance and Professor Claes Wollheim for providing INS-1 cells.