Mutations in the promoter of adenylyl cyclase (AC)-III gene ...

Rapid Publication Mutations in the Promoter of Adenylyl Cyclase (AC)-III Gene, Overexpression of AC-III mRNA, and Enhanced cAMP Generation in Islets From the Spontaneously Diabetic GK Rat Model of Type 2 Diabetes Samy M. Abdel-Halim, Amel Guenifi, Bing He, Bei Yang, Maha Mustafa, Bo Höjeberg, Jan Hillert, Moiz Bakhiet, and Suad Efendi´c

Glucose-induced insulin release is decreased in the spontaneously diabetic GK rat, a nonobese rodent model of type 2 diabetes. Forskolin restores the impaired insulin release in both the isolated perfused pancreas and isolated islets from these rats (AbdelHalim et al., Diabetes 45:934–940, 1996). We demonstrate here that the insulinotropic effect of forskolin in the GK rat is due to increased generation of cAMP and that it is associated with overexpression of adenylyl cyclase (AC)-III mRNA and gene mutations. The AC-III mRNA overexpression was demonstrated by in situ hybridization using oligonucleotide probes binding to different regions of the rat AC-III mRNA. It was associated with the presence of two point mutations identified at positions –28 bp (A → G) and –358 bp (A → C) of the promoter region of the AC-III gene and was demonstrable in both GK rat islets and peripheral blood cells. Transfection of COS cells with a luciferase reporter gene system revealed up to 25-fold increased promoter activity of GK AC-III promoter when compared with normal rat promoter (P < 0.0001). In conclusion, forskolin restores the impaired insulin release in islets of the GK rat through enhanced cAMP generation. This is linked to overexpression of AC-III mRNA in GK islets due to two functional point mutations in the promoter region of the AC-III gene. Diabetes 47:498–504, 1998

From the Endocrine and Diabetes Unit (S.M.A.-H., A.G., S.E.), Department of Molecular Medicine, Karolinska Hospital; and the Division of Neurology (B.He, B.Y., M.M., B.Hö., J.H., M.B.), Huddinge University Hospital, Karolinska Institute, Stockholm, Sweden. Address correspondence and reprint requests to Dr. Samy M. AbdelHalim, Endocrine and Diabetes Unit, Department of Molecular Medicine, Karolinska Hospital, L1:02, S-171 76 Stockholm, Sweden. E-mail: [email protected]. Received for publication 6 August 1997 and accepted in revised form 17 December 1997. AC, adenylyl cyclase; [Ca 2+] i, intracellular calcium concentration; CRE, cAMP response element; PCR, polymerase chain reaction. 498

A

central defect in type 2 diabetes is an impaired insulin response to glucose stimulation. Such a defect is also found in the GK (Goto-Kakizaki) rat, a lean spontaneous animal model of type 2 diabetes (1,2). The defective insulin release by the islets of the GK rat is due to impaired stimulus-secretion coupling of the glucose signal rather than decreased insulin availability (3) or b-cell density (4). Glucose-induced insulin release is also impaired in these islets when the KATP channels are inactivated by diazoxide and the membranes depolarized by high K+ (5), suggesting that one or more signals directly coupling glucose metabolism to the exocytosis of insulin are impaired. Stimulation of insulin release by forskolin restores the impaired insulin response in isolated islets and perfused pancreas of the GK rat (5). Additionally, it enhances to a larger extent insulin release at substimulatory concentration of glucose in GK rats compared with control rats. Such insulinotropic effects of forskolin in GK rats are elicited through a KATP channel–independent pathway, suggesting a direct effect on exocytosis of insulin (5). Forskolin is known to induce its effects by strong stimulation of adenylyl cyclase (AC) and cAMP generation. To date, nine AC isoenzymes are classified, of which five are Ca2+-regulated (AC-I, AC-III, ACV, AC-VI, and AC-VIII) (6,7). The AC isoforms expressed in the pancreatic islet are unknown. Of interest in this regard is that AC activity in pancreatic islets was shown to be Ca2+ dependent (8). Based on this background, the current study was primarily designed to determine the molecular mechanisms mediating the enhanced insulinotropic response of forskolin in GK islets. RESEARCH DESIGN AND METHODS Animals. Male GK rats were from our colony at Karolinska Hospital, Stockholm, Sweden, and control Wistar rats were from a local breeder (B&K Universal, Sollentuna, Sweden). Both groups were ~3 months old and had similar body weights (316 ± 9 [n = 19] for GK rats vs. 309 ± 9 g [n = 19] for control rats). Blood glucose concentration just before pancreas isolation was significantly elevated in GK rats (11.3 ± 1.4 [n = 15] in GK rats vs. 4.4 ± 0.1 mmol/l [n = 15] in control rats, DIABETES, VOL. 47, MARCH 1998

S.M. ABDEL-HALIM AND ASSOCIATES

P < 0.0001). All rats were fed ad libitum with free access to water and were kept in rooms with alternating 12-h periods of light and darkness. Isolation of pancreatic islets and insulin release in batch incubation studies. Islets were isolated from the rats by digestion with collagenase (BoehringerMannheim, Mannheim, Germany). The islets were cultured for 48 h in RPMI-1640 supplemented with 11 mmol/l glucose and 10% (vol/vol) fetal calf serum. The islets were then maintained in Krebs-Ringer bicarbonate buffer supplemented with 20 g/l bovine plasma albumin (Sigma, St. Louis, MO) with 3.3 mmol/l glucose for 45 min. The batches of 20 islets were incubated with 3.3 or 16.7 mmol/l glucose alone and with 5 µmol/l forskolin (Sigma) at both glucose concentrations for 45 min. After that period, 900 µl of the effluent was removed and frozen at –20°C to be subsequently analyzed for insulin content by radioimmunoassay of insulin (9). cAMP measurements. cAMP was determined in the same batches of 20 islets in which insulin release was measured. After removal of 900 µl for determination of insulin, the medium was replaced with 300 µl ice-cold acetate buffer containing 0.5 mmol/l isobutylmethylxanthine. The islets were then immersed in boiling water for 4 min and left to cool at 4°C. The islets were sedimented by brief centrifugation, sonicated, and subsequently frozen at –20°C. cAMP content was measured using a commercial kit NEK-033 (Du Pont-NEN, Boston, MA). Detection of AC-III mRNA expression by in situ hybridization. After incubation of the islets for 45 min with the two concentrations of glucose with and without forskolin, the islets were transferred onto restricted areas of microscope slides (ProbeOn, Fisher Scientific, Pittsburgh, PA) and dried at 55°C for 30 min. In situ hybridization was performed as described before (10). A mixture of four oligonucleotide probes, ~48 bases long, binding to different regions of rat AC-III mRNA were used to increase the sensitivity of the method. In a second series of experiments, a mixture of two probes (probes 1 and 3, Table 1) were performed. The AC-III mRNA sequence (11) was obtained from GenBank and the oligonucleotide probes were selected by the MacVector system (Table 1). A probe identical to the AC-III mRNA was used as a control probe in parallel slides to assure the specificity of the hybridization signals. After emulsion autoradiography, islet cells expressing mRNA for AC were examined in dark-field microscopy and presented at 3200 magnification. Using sense oligonucleotide as a control probe produced a uniformly weak background signal without revealing any positive cells. Direct polymerase chain reaction sequencing of promoter region of ACIII gene in isolated islets and peripheral blood. Islets and blood were collected from six control and six GK rats. The islets and blood from two GK rats or Wistar rats were pooled. Additionally, blood was also collected from another nondiabetic rat strain, the DA rat. DNA from islets and leukocytes in peripheral blood were extracted by standard methods (12). Polymerase chain reaction (PCR) primers were designed by Oligo 5.0 (NBI, Plymouth, MN) according to the database from GenBank (acc. no. S64908) (13). A forward primer corresponding to positions –568 to –549 bp relative to the transcription start site was linked with a 18-bp –21M13 sequence (resulting in primer sequence TGTAAAACGACGGCCAGTTC TTGAGCTGCCTCCCAAAG). The reverse primer corresponded to positions +263 to +282 bp relative to the transcription start site (GTTCAGCATCCGTGGTCGCA). PCR was carried out in the Gene-AMP 9600 cycler (Perkin-Elmer, Norwalk, CT) for 35 cycles of 96°C for 10 s, 60°C for 1 min, and 72°C for 1 min. A dye primer cycle sequencing kit with Ampli Taq FS enzyme (Perkin-Elmer) was applied to sequencing of PCR fragments using the labeled standard –21M13 sequencing primer. Sequencing reactions were resolved on an ABI-377 DNA automated sequencer. All processes followed the manufacturer’s instructions. Transfection of reporter gene with the mutant AC-III promoter. A luciferase reporter vector, pGL3-enhancer (Promega Biotec, Madison, WI), was used to compare the function of the GK and wild-type AC-III promoters after transfection to COS cells. The 580-bp sequences of the AC-III promoter regions of GK and Wistar rats were amplified by PCR using primers containing 59 Sac I and HindIII restriction sites (forward primer TTTTGAG CTCTCTTGAGCTGCCTCCCAAAG [as described above], reverse primer AATTAAGCTTTGGAAACGCCGAGTAGGTGG [corresponding to positions +12 to –8 relative to the transcription start site]), allowing insertion into the pGL3enhancer vector. After cloning into Escherichia coli, the cloned inserts were confirmed by sequencing. Recombinant vectors were transfected into 5 3 105 COS cells (cultured in Dulbecco’s medium containing 1 g/l glucose and 10% fetal bovine serum) using Transfectam Reagent (Promega) parallel with pGL3 control vector and pGL3 enhancer vector without insertion as positive and negative controls. Each transfection was performed in five parallel experiments. Luciferase activity in transfected COS cells was assessed after a 48-h culture in 20 µl COS cells extract mixed with 100 µl luciferase assay reagent using a scintillation counter. As expected, the positive control pGL3 vector used for monitoring transfection efficiency showed the highest promoter activity (2,257 ± 299 light intensity units), whereas the activity of the pGL3 vector without insert did not differ from background (nontransfected COS cells) activity (6 ± 1 and 5 ± 1 light intensity units, respectively). DIABETES, VOL. 47, MARCH 1998

FIG. 1. Insulin secretion and cAMP generation. Effect of 5 µmol/l forskolin in the presence of 3.3 or 16.7 mmol/l glucose on concurrent insulin secretion (A) and cAMP generation (B) from control (h, n = 7–8) and GK (j, n = 7–8) rat islets. Results are expressed as means ± SE. F, forskolin, G, glucose.

In another set of experiments, luciferase vectors carrying the mutant or the normal AC-III promoter fragments were cotransfected with pCAT3 control vectors (Promega) as the internal control plasmid for normalization of transfection efficiency. In these experiments 2 µg luciferase vector and 2 µg pCAT3 control vector mixed with 8 µl Transfectam (Promega) were cotransfected in each cell preparation (2.5 3 105 COS cells in six-well plates). Cotransfections were performed in at least three cell preparations. Groups of cotransfected cells were incubated for 48 h with and without 5 µmol/l forskolin. After 48 h, the cells were harvested using 200 µl reporter lysis buffer (Promega) and divided into two parts for estimation of luciferase and CAT activities. To determine CAT activity, the cell extract was heated at 60°C for 10 min and frozen at –70°C until assay. The activity of luciferase was determined as described above. CAT assay was performed using 14C-labeled chloramphenicol (Du Pont-NEN) and n-butyryl coenzyme A, following the CAT assay kit description (Promega). The background CAT activity (nontransfected COS cells) was 7,407 ± 39 cpm (n = 2), while CAT activity was almost identically increased in all the cell groups whether being cotransfected with vectors carrying the normal or GK mutant promoter fragment in the absence or the presence of 5 µmol/l forskolin (147,719 ± 1,957 cpm [control promoter, n = 3], 146,520 ± 1,381 cpm [control promoter plus forskolin, n = 3], 148,083 ± 1,934 cpm [GK promoter, n = 3], and 149,801 ± 388 cpm [GK promoter plus forskolin, n = 3]). Data analysis. All data are expressed as means ± SE. Statistical tests were performed with SigmaStat for Windows Version 1.0 (Jandel Scientific Software, Erkrath, Germany). Tests of significance of difference were performed using the t test for unpaired data and the Mann-Whitney rank-sum test.

RESULTS

Insulin release and cAMP generation. Insulin release (Fig. 1A) was decreased in GK rat islets at 3.3 mmol/l (0.02 ± 0.004 in GK vs. 0.04 ± 0.003 pmol/islet in controls, P < 0.002) and 16.7 mmol/l (0.04 ± 0.004 in GK vs. 0.15 ± 0.015 499

AC-III GENE PROMOTER, cAMPLEVELS, AND INSULIN RELEASE IN GK RAT ISLETS

TABLE 1 Probes used for detection of rat AC-III mRNA expression by in situ hybridization Probe

Bases

1

994–1041 (R&C) CTCCCTCAGCAGCTGCATCCCTTCTAGCTCGTCTTGCTGCTGCTGGGC

2

1486–1515 (R&C) GTGGTCCTCCCGGTAGTCAGGCAGGCCGCAGATGCAGTAGTAACAGTC

3

2227–2274 (R&C) ACAGCTGAAGGCAGCCCCACTCTGCTTCTCCTTCTCCACCGAGTAGCG

4

3018–3065 (R&C) ACATGCTCTGGCAACATGTTGGTGACCAAGGCCTCGTTCCACCGGCGC

5

3018–3065 (Control) GCGCCGGTGGAACGAGGCCTTGGTCACCAACATGTTGCCAGAGCATGT

From GenBank acc. no. M55075. R&C, reverse and complemented.

pmol/islet in controls, P < 0.0001) glucose. Stimulation with forskolin increased insulin release by approximately fourfold in control islets at 3.3 mmol/l (0.04 ± 0.003 without forskolin vs. 0.16 ± 0.026 pmol/islet with forskolin, P < 0.0001) and 16.7 mmol/l glucose (0.15 ± 0.015 without forskolin vs. 0.63 ± 0.068 pmol/islet with forskolin, P < 0.0001). In confirmation of our previous findings (5), forskolin induced a strong 9-fold increase in insulin release in GK islets at 3.3 mmol/l (0.02 ± 0.004 without forskolin vs. 0.22 ± 0.017 pmol/islet with forskolin, P < 0.0001) and an 18-fold increase at 16.7 mmol/l glucose (0.04 ± 0.004 without forskolin vs. 0.66 ± 0.073 pmol/islet with forskolin, P < 0.0001). Hence, in the presence of forskolin, insulin responses were similar in control and GK islets at 3.3 and 16.7 mmol/l glucose, but the relative increase in response to forskolin was much more pronounced in GK islets, similar to our findings with forskolin in the isolated perfused pancreas (5). cAMP levels (Fig. 1B) in GK islets were similar to those of control islets at 3.3 mmol/l glucose (31.2 ± 3.7 in GK vs. 24.0 ± 1.6 fmol/islet in controls, NS) and were significantly higher at 16.7 mmol/l glucose (38.9 ± 4.8 in GK vs. 18.1 ± 0.8 fmol/islet in controls, P < 0.0001). In control islets, forskolin induced a sixfold increase in cAMP levels at 3.3 mmol/l glucose (24.0 ± 1.6 without forskolin vs. 147.9 ± 13.8 fmol/islet with forskolin, P < 0.0001) and an approximately fivefold increase in nucleotide level at 16.7 mmol/l glucose (18.1 ± 0.8 without forskolin vs. 84.9 ± 9.8 with forskolin, P < 0.0001). In GK islets, forskolin induced even higher cAMP levels at both 3.3 mmol/l glucose (31.2 ± 3.7 without forskolin vs. 279.0 ± 14.3 fmol/islet with forskolin, P < 0.0001) and 16.7 mmol/l glucose (38.9 ± 4.8 without forskolin vs. 242.8 ± 13.9 fmol/islet with forskolin, P < 0.0001). Thus, forskolin elicited an enhanced cAMP response in GK islets compared with control islets at both 3.3 mmol/l (279 ± 14 in GK vs. 148 ± 14 fmol/islet in control, P < 0.0001) and 16.7 mmol/l glucose (243 ± 14 in GK vs. 85 ± 10 fmol/islet in controls, P < 0.0001). Islet AC-III mRNA expression by in situ hybridization. mRNA expression of AC-III was determined in islets under the above conditions in which insulin and cAMP generation were studied. In a first series of experiments using a mixture of four probes (Probes 1–4, Table 1) to increase the sensitivity, an intense expression of AC-III mRNA already at 3.3 500

mmol/l glucose was noted in GK islets compared with control islets. A similar picture was noted at 16.7 mmol/l glucose. Forskolin clearly enhanced expression of AC-III mRNA in control islets both at 3.3 and 16.7 mmol/l glucose, mirroring the rises in cAMP and insulin release at similar experimental conditions. Overexpression of AC-III mRNA was, however, already marked in GK islets at 3.3 or 16.7 mmol/l glucose, rendering it difficult to discern a further rise in expression level in response to forskolin at either low or high glucose stimulation. To decrease the enhanced basal mRNA signals in GK islets to delineate the effects of forskolin, a set of experiments (Fig. 2) in which a mixture of only two probes (probes 1 and 3, Table 1) was performed. mRNA expression was clearly reduced compared with the series using a mixture of four probes. However, the enhanced ACIII mRNA overexpression in GK islets was clearly visible at glucose levels, and the addition of forskolin further induced mRNA expression severalfold higher in the GK islets compared with the control islets (Fig. 2B, D, F, and G). Sequencing of the promoter region of AC-III gene in GK and control rats. The promoter region of the AC-III gene was selected for mutation screen by DNA sequencing. A genomic DNA fragment of 825 bp was sequenced including 563 bp in the 59 flanking region. Obtained sequences from GK (n = 6), Wistar (n = 6), and DA (n = 1) rats were identical to the published sequence (13) except for two point mutations at positions –28 bp (A → G) and –358 bp (A → C) (Fig. 3) relative to the transcription starting site of the AC-III gene in islets and peripheral blood of GK rats. In the electrophoretogram, clear monophasic peaks of the two point mutations were obtained, indicating their homozygous character. No base changes were found in the mRNA transcription sequence from positions 1 to 262. It is also interesting to consider the possibility of autoregulation by cAMP of AC-III in view of the presence of a cAMP response element (CRE) motif in the AC-III promoter. A classical CRE was described to be constituted of the motif 59TGACGTCA-39 or similar composition (14). We confirm the presence of a motif with the composition 59-TGACGTC-39, which more than likely is representing a CRE. The position of this motif is –244 to –238 bp relative to the transcription start site. Interestingly, this CRE motif lies between the two described mutations in the GK rat AC-III promoter and closer DIABETES, VOL. 47, MARCH 1998


FIG. 2. AC-III mRNA expression by in situ hybridization. AC-III mRNA expression determined by in situ hybridization in single islets in the presence of 3.3 mmol/l glucose (A: control, B: GK rat); 3.3 mmol/l glucose and 5 µmol/l forskolin (C: control, D: GK rat); 16.7 mmol/l glucose (E: control, F: GK rat); 16.7 mmol/l glucose and 5 µmol/l forskolin (G: control, H: GK rat). Photomicrographs were not taken at the same magnification. F, forskolin; G, glucose. DIABETES, VOL. 47, MARCH 1998

501


FIG. 3. Schematic structure of the promoter region of the AC-III gene. (j), exons; ( ), transcription start site. Two point mutations upstream from the transcription start site were observed at positions –28 (A → G) and –358 (A → C) only in GK islets and peripheral blood cells.

to the –358-bp mutation. This report, to our knowledge, is the first to describe such a finding, and accordingly, nothing is yet known about its regulation. Experiments specially designed to test this motif and the possible effects of the two described mutations on its regulation need to be addressed in future studies. Transfection analysis of AC-III gene promoter mutations. The presence of novel mutations in the AC-III gene promoter in GK rats prompted us to assess their potential functional importance. Indirect evidence of a regulatory effect of the mutant promoter was obtained by using a luciferase reporter gene system. Constructs containing the two novel mutations from GK rat or wild-type promoter sequences were inserted in a pGL3-enhancer vector. An approximately sixfold higher luciferase activity was obtained from vectors carrying the mutant promoter (1,875 ± 187 light intensity units) compared with vector carrying the wild-type promoter (305 ± 61 light intensity units, P < 0.0001 versus vectors carrying mutant promoter) (Fig. 4A). In another set of experiments, COS cells were cotransfected with pCAT-3 control vectors together with vectors carrying the GK or normal promoter (Fig. 4B and C). An ~24fold enhanced luciferase activity was obtained from vectors carrying the mutant GK promoter (3,660 ± 250 light intensity units) compared with vectors carrying the normal promoter (153 ± 57 light intensity units, P < 0.0002). Addition of forskolin (Fig. 4C) did not enhance these responses, and ~26-fold increased luciferase activity was still elicited from vectors carrying the GK promoter (2,441 ± 253 light intensity units) than vectors carrying normal promoter (95 ± 31 light intensity units, P < 0.0008). The current results may indicate that the effects of forskolin on the AC-III message are probably mediated by a posttranscriptional mechanism.

([Ca2+]i), suggesting that they are mediated through a direct interaction with the exocytotic machinery in the b-cell (5). It has been suggested that at least 80% of the effects of cAMP generation on insulin release in the pancreatic b-cell are mediated through direct interaction with the exocytotic machinery, with the remainder through rises in [Ca2+]i (21). In 1979, Valverde et al. (8) described an AC activity in islets that is Ca2+ stimulated. Of the nine AC isoenzymes thus classified, five are Ca2+ regulated (AC-I, AC-III, AC-V, AC-VI, and AC-VIII) (6,7). We chose to focus on AC-III in examining the molecular mechanisms behind the enhanced forskolin effects on insulin release and cAMP generation in GK islets for the following reasons. First, significantly more cAMP generation in response to forskolin has been reported in HEK-293 cells, a human kidney cell line, when transfected with the AC-III than that with the AC-I gene (22), demonstrating that stimulation of AC-III allows for considerable modulation of intracellular cAMP concentration (11). Second, regular Ca2+ oscillations in response to stimulation with forskolin have been elicited when HEK-293 cells are transfected with the AC-III gene, but not with the AC-I gene, indicating that expression of AC-III is necessary for the elicitation of regular Ca2+ oscillations in the cell (22). In analogy with HEK-293 cells, forskolin is known to induce Ca2+ oscillations and periodic insulin release from

DISCUSSION

The question of glucose-induced cAMP generation from pancreatic islets is controversial, with reports showing that glucose induces a rise (15,16) or has no effect (17–19) on nucleotide levels. In the present study, glucose failed to induce significant elevations of cAMP levels in control or GK islets. This indicates that significant rises in cAMP generation may not be crucial for glucose-induced insulin release in normal islets (17,18) and supports the view that a basal permissive rate of cAMP formation may be a requirement for maintenance of the normal response to glucose stimulation (20). In contrast, increased generation of cAMP appears to play a key role in the transduction of the forskolin signal to the insulin-releasing machinery. Thus, forskolin significantly enhanced insulin release and cAMP generation at a substimulatory glucose concentration in controls and, to an even greater extent, GK islets. The insulinotropic effects of forskolin were also obtained through a KATP channel–independent pathway and are not associated with an increase in intracellular Ca2+ levels 502

FIG. 4. Transfection analysis of the two mutations in the promoter region of AC-III gene in the GK rat. Light intensity reflecting luciferase activity produced by COS cells transfected with vectors carrying a fragment of the promoter region of AC-III gene from a control rat (h) or the corresponding fragment containing the two mutations from a GK rat (j) only (A) or cotransfected with pCAT-3 control vectors (B and C) are shown. Effect of 5 µmol/l forskolin is marked (C). Results are expressed as light intensity unit means ± SE. DIABETES, VOL. 47, MARCH 1998


pancreatic islets (18), which suggested the possible expression of AC-III in islets. Taken together, increased activity of AC-III in GK islets then appeared to be a candidate to explain the markedly enhanced cAMP generation in response to forskolin in GK islets. Therefore, we investigated the possibility that alterations in the expression of AC-III mRNA may explain the enhanced cAMP generation in GK islets. These experiments were performed under the same experimental and stimulatory conditions for which we studied concurrent cAMP generation and insulin release. The finding of a forskolin-induced expression of AC-III mRNA in islets is, to our knowledge, without precedent. Interestingly, we found not only overexpression of AC-III mRNA in GK islets but also two functional mutations in the promoter of the AC-III gene. The possibility that other AC isoforms are expressed in the pancreatic islet and contribute to cAMP generation by forskolin cannot be dismissed. The isoforms of AC expressed in the pancreatic b-cell are unknown, and the current report is the first to describe the expression of an AC isoform in the pancreatic islet. While the currently described functional mutations in the AC-III promoter and overexpression of AC-III mRNA in the GK rat can account for the enhanced effects of forskolin on cAMP generation and insulin release in GK islets, it is unclear whether these changes in AC-III gene constitute an important mechanism behind the impaired insulin response to glucose in the GK rat. In this context, it is of interest that the regulation of AC-III is distinct from all other AC isoenzymes characterized thus far in that it is stimulated by hormones in vivo and inhibited by rises in [Ca2+]i (6,23). The expression of ACIII in islets can thus allow for the presence of a closed negative-feedback loop between cAMP and Ca2+ that can result in stable oscillations of both intracellular messengers (24). Accordingly, if overexpression of AC-III induces even small alterations in the generation of cAMP, this may disturb the delicate modulatory signals exerted by both intracellular messengers, leading to loss of regular oscillations of Ca2+ and insulin in the b-cell. This hypothesis is supported by the finding of high basal [Ca2+]i levels in GK islets (5,25). In the b-cell, the cloned a1-subunits of L-type voltage-dependent calcium channels have been shown to have a number of activated cAMP-dependent protein kinase phosphorylation sites (26–28). Thus, increased cAMP levels, by enhancing phosphorylation of Ca2+ channels, could be responsible for the high basal [Ca2+]i level in the GK islet. Expression of AC-III gene was first demonstrated in the olfactory tissue (11) and later in the human cell line HEK293, bovine brain, spinal cord, adrenal medulla, adrenal cortex, heart atrium, aorta, lung, and retina (29). A similar distribution of the mutant AC-III gene, leading to increased cAMP generation, could have many additional pathophysiological implications for the diabetic state in the GK rat and possibly in humans. An important extension of the current findings will be to determine whether similar alterations in the AC-III gene are present in patients with type 2 diabetes. The AC-III gene was mapped to chromosome 2 in humans (30,31) but not cloned. In conclusion, the current study confirms our findings of decreased insulin response to glucose in GK islets, which is restored by forskolin stimulation (5). The current results extend this observation by demonstrating that the restorative effect of forskolin on insulin release in GK islets can be DIABETES, VOL. 47, MARCH 1998

accounted for by increased cAMP generation rather than altered sensitivity to the nucleotide. Furthermore, two functional point mutations in the promoter region of the AC-III gene were found in the pancreatic islets and peripheral blood of GK rat, which resulted in overexpression of AC-III mRNA in GK islets. To our knowledge, the current findings are the first to describe mutational changes in an AC gene resulting in overexpression of its mRNA associated with a pathological state. ACKNOWLEDGMENTS

This work was supported by the Swedish Society of Medicine, the Medical Research Council in Sweden (K97-19X-0003433C), and the Research Funds of the Karolinska Institute, Stockholm, Sweden. We are most grateful to Professor Hans Link, Division of Neurology, Huddinge University Hospital, Karolinska Institute, for critical review of this report and to Professor Jean-Claude Henquin, Université de Louvain, Brussels, Belgium, for technical information on the cAMP radioimmunoassay. The skillful technical assistance of Anita Gustafsson, Division of Neurology, and Yvonne Strömberg, Department of Molecular Medicine, Karolinska Institute, is greatly acknowledged. Parts of this work were presented in abstract form at the 32nd Annual Meeting of the European Association for the Study of Diabetes, Vienna, Austria, 1–5 September 1996. REFERENCES 1. Abdel-Halim SM, Guenifi A, Luthman H, Grill V, Efendic´ S, Östenson C-G: Impact of diabetic inheritance on glucose tolerance and insulin secretion in spontaneously diabetic GK-Wistar rats. Diabetes 43:281–288, 1994 2. Abdel-Halim SM, Östenson C-G, Andersson A, Jansson L, Efendic´ S: A defective stimulus-secretion coupling rather than glucotoxicity mediates the impaired insulin secretion in the mildly diabetic F1 hybrids of GK-Wistar rats. Diabetes 44:1280–1284, 1995 3. Abdel-Halim SM, Guenifi A, Efendic´ S, Östenson C-G: Both somatostatin and insulin responses to glucose are impaired in the perfused pancreas of the spontaneously diabetic GK (Goto-Kakizaki) rat. Acta Physiol Scand 148:219–226, 1993 4.Guenifi A, Abdel-Halim SM, Höög A, Falkmer S, Östenson C-G: Preserved bcell density in the endocrine pancreas of young, spontaneously diabetic GotoKakizaki (GK) rats. Pancreas 10:148–153, 1995 5. Abdel-Halim SM, Guenifi A, Khan A, Larsson O, Berggren P-O, Östenson C-G, Efendic´ S: Impaired coupling of glucose signal to the exocytotic machinery in diabetic GK rats: a defect ameliorated by cAMP. Diabetes 45:934–940, 1996 6. Wayman GA, Impey S, Storm DR: Ca2+ inhibition of type III adenylyl cyclase in vivo. J Biol Chem 270:21480–21486, 1995 7. Antoni FA: Calcium regulation of adenylyl cyclase: relevance for endocrine control. Trends Endocrinol Metab 8:7–14, 1997 8. Valverde I, Vandermeers A, Anjaneyulu R, Malaisse WJ: Calmodulin activation of adenylate cyclase in pancreatic islets. Science 206:225–227, 1979 9. Herbert V, Lau KS, Gottlieb CW, Bleicher SJ: Coated charcoal immunoassay of insulin. J Clin Endocrinol Metab 25:1375–1384, 1965 10. Dagerlind Å, Friberg K, Bean AJ, Hökfelt T: Sensitive mRNA detection using unfixed tissue: combined radioactive and non-radioactive in situ hybridization histochemistry. Histochemistry 98:39–49, 1992 11. Bakalyar HA, Reed RR: Identification of a specialized adenylyl cyclase that may mediate odorant detection. Science 250:1403–1406, 1990 12. Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A Laboratory Man ual. Nolan C, Ed. New York, Cold Spring Harbor Laboratory Press, 1989 13. Wang MM, Tsai RY, Schrader KA, Reed RR: Genes encoding components of the olfactory signal transduction cascade contain a DNA binding site that may direct neuronal expression. Mol Cell Biol 13:5805–5813, 1993 14. Meyer TE, Waeber G, Lin J, Beckmann W, Habener JF: The promoter of the gene encoding 39,59-cyclic adenosine monophosphate (cAMP) response element binding protein contains cAMP response elements: evidence for positive autoregulation of gene transcription. Endocrinology 132:770–780, 1993 15. Grill V, Cerasi E: Activation by glucose of adenyl cyclase in pancreatic islets of the rat. FEBS Lett 33:311–314, 1973 503


16. Grill V, Cerasi E: Stimulation by D-glucose of cyclic adenosine 39:59monophosphate accumulation and insulin release in isolated pancreatic islets of the rat. J Biol Chem 249:4196–4201, 1974 17. Sharp GWG: The adenylate cyclase-cyclic AMP system in islets of Langerhans and its role in the control of insulin release. Diabetologia 16:287–296, 1979 18. Henquin J-C, Meissner HP: The ionic, electrical, and secretory effects of endogenous cyclic adenosine monophosphate in mouse pancreatic B cells: studies with forskolin. Endocrinology 115:1125–1134, 1984 19. Gembal M, Detimary P, Gilon P, Gao Z-Y, Henquin J-C: Mechanisms by which glucose can control insulin release independently from its action on adenosine triphosphate-sensitive K+ channels in mouse B cells. J Clin Invest 91:871–880, 1993 20. Malaisse WJ, Malaisse-Lagae F: The role of cyclic AMP in insulin release. Expe rientia 40:1068–1075, 1984 21. Ämmälä C, Ashcroft FM, Rorsman P: Calcium-independent potentiation of insulin release by cyclic AMP in single b-cells. Nature 363:356–358, 1993 22. Wayman GA, Hinds TR, Storm DR: Hormone stimulation of type III adenylyl cyclase induces Ca2+ oscillations in HEK-293 cells. J Biol Chem 270:24108–24115, 1995 23. Wei J, Wayman G, Storm DR: Phosphorylation and inhibition of type III adenylyl cyclase by calmodulin-dependent protein kinase II in vivo. J Biol Chem 271:24231–24235, 1996 24. Cooper DMF, Mons N, Karpen JW: Adenylyl cyclases and the interaction

504

between calcium and cAMP signaling. Nature 374:421–424, 1995 25. Kato S, Ishida H, Tsuura Y, Tsuji K, Nishimura M, Horie M, Taminato T, Ikehara S, Odaka H, Ikeda H, Okada Y, Seino Y: Alterations in basal and glucose-stimulated voltage-dependent Ca2+ channel activities in pancreatic b cells of noninsulin-dependent diabetes mellitus GK rats. J Clin Invest 97:2417–2425, 1996 26. Seino S, Chen L, Seino M, Blondel O, Takeda J, Johnson JH, Bell GI: Cloning of the a1 subunit of a voltage-dependent calcium channel expressed in pancreatic b cells. Proc Natl Acad Sci USA 89:584–588, 1992 27. Yaney GC, Wheeler MB, Wei X, Perez-Reyes E, Birnbaumer L, Boyd AE III, Moss LG: Cloning of a novel a1-subunit of the voltage-dependent calcium channel from the b-cells. Mol Endocrinol 6:2143–2152, 1992 28. Leiser M, Fleischer N: cAMP-dependent phosphorylation of the cardiac-type a1 subunit of the voltage-dependent Ca2+ channel in a murine pancreatic bcell line. Diabetes 45:1412–1418, 1996 29. Xia Z, Choi E-J, Wang F, Storm DR: The type III calcium/calmodulin-sensitive adenylyl cyclase is not specific to olfactory sensory neurons. Neurosci Lett 144:169–173, 1992 30. Haber N, Stengel D, Defer N, Roeckel N, Mattei MG, Hanoune J: Chromosomal mapping of human adenylyl cyclase genes type III, type V, and type IV. Hum Genet 94:69–73, 1994 31. Gaudin C, Homcy CJ, Ishikawa Y: Mammalian adenylyl cyclase family members are randomly located on different chromosomes. Hum Genet 94:527–529, 1994

DIABETES, VOL. 47, MARCH 1998