Different Regulated Expression of the Tyrosine Phosphatase-Like Proteins IA-2 and Phogrin by Glucose and Insulin in Pancreatic Islets: Relationship to Development of Insulin Secretory Responses in Early Life

Mouse monoclonal antibodies 76F and 3C12 against the cytoplasmic domain of IA-2 were generated as previously described (14). Rabbit antibody to the α subunit of the insulin receptor (type A and B) was from Biogenesis (Poole, U.K.). Polyclonal guinea pig anti-serum to phogrin (16) was a kind gift of J. Hutton (Barbara Davis Center, Denver, CO), and guinea pig anti-insulin antiserum was kindly provided by P. Jones (King’s College, London).

Rat islets were isolated from Wistar rats of different ages (aged 1, 5, and 10 days and 5-week-old adults). Animals were kept according to the guidelines of the U.K. home office and killed by cervical dislocation. Islets from juvenile rats (aged 1, 5, or 10 days) were isolated by collagenase digestion (Roche Diagnostics, Lewes, U.K.) and enrichment on a discontinuous Percoll gradient (17). Adult rat islets were isolated by collagenase digestion after intraductal perfusion of the pancreas with collagenase P and DNAse and enriched on a three-step discontinous BSA gradient (18). Single pancreatic β-cells (> 95% pure) and non−β-cells were purified from trypsin-dissociated adult rat islet cells by autofluorescence-activated cell sorting (19).

Human islets were isolated from pancreata of cadaveric organ donors by modifications of the automated method of Ricordi (20,21). Approved informed consent was obtained from the relatives of organ donors. Isolated human islet preparations (60–80% purity) were stained with dithizone before selection under the dissection microscope to > 95% purity. Human islets were exposed to dithizone for <5 min, conditions that have no effect on β-cell function (22).

After enrichment, islets (200 islets per condition) were cultured in 60-mm culture dishes in RPMI medium. Isolated pancreatic β-cells and non-β-cells were reaggregated (19) and incubated at 1–3 × 10⁵ cells per dish in HAM’s F10 medium for 9 days. The medium contained 3, 6, 10, or 20 mmol/l glucose in combination with agents indicated in the text or figures. In some experiments, insulin signaling was blocked with an antibody to the insulin receptor. At the concentration used, the antibody inhibited insulin-stimulated phosphorylation of mitogen-activated protein kinase in Hep-G2 cells but was without effect on IGF-1-stimulated phosphorylation (data not shown). After culture, islets were analyzed for IA-2, phogrin, and insulin protein and mRNA and for insulin secretion in response to secretagogues. Accumulated insulin in the culture medium was measured by enzyme-linked immunosorbent assay (ELISA).

Isolated islets were washed in HEPES-buffered saline and prepared for Western blotting with monoclonal antibodies to IA-2 as previously described (14). The integrated optical density of bands representing IA-2 was measured by densitometry and compared with a standard curve of purified recombinant IA-2 (23) run on each blot. The interassay variability of IA-2 protein measurements in 10 independent blots was <20%, and the intra-assay variation was <8%.

IA-2 biosynthesis was assessed by ³⁵S-methionine/cysteine incorporation and detection by immunoprecipitation and autoradiography. Isolated adult rat islets (200 per condition) were incubated for 2 h at 37°C in 100 μl HEPES-buffered Hanks’ balanced salt solution (HBSS) containing 1 mg/ml BSA, 6 mmol/l glucose, 10% methionine, cysteine-deficient amino acid mix (Amersham-Pharmacia Biotech, Amersham, U.K.), and 4 MBq EasyTag Express protein labeling mix (NEN, Hounslow, U.K.) in the presence or absence of 10 nmol/l insulin. Islets were lysed (14) and immunoprecipitated with monoclonal IA-2 antibody 76F or an irrelevant control antibody. Immune complexes were isolated on protein A Sepharose and analyzed by SDS-PAGE and autoradiography.

IA-2 degradation was studied in pulse-chase experiments. Islet proteins were labeled for 1 h at 37°C in 100 μl HEPES-buffered HBSS containing 1 mg/ml BSA, 20 mmol/l glucose, 10% methionine, cysteine-deficient amino acid mix, and 12 MBq EasyTag Express protein labeling mix. Islets were either collected immediately or after a 12-h chase period in RPMI medium containing 6 mmol/l glucose with or without 10 nmol/l insulin. Radiolabeled islet IA-2 was detected by immunoprecipitation and autoradiography as described above.

RNA from 20 islets per condition was isolated using the High Pure RNA Isolation Kit (Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. A sample of 500 ng total RNA was transcribed with poly d(T)₁₅-primers to cDNA using Expand Reverse Transcriptase (Roche Diagnostics). The cDNA was purified using the High Pure PCR Product Purification Kit (Roche Diagnostics).

Expression of IA-2, phogrin, and insulin mRNA was analyzed by quantitative real-time RT-PCR using the FastStart DNA Master SybrGreen I Kit in the LightCycler system (Roche Diagnostics). To generate control templates for the quantification, we cloned rat β-actin (478–819 bp), IA-2 (1642–2463), phogrin (1923–2048), and insulin (4243–4538) cDNA fragments from INS-1 rat insulinoma cells by RT-PCR into the pGEM-T Easy vector (Promega, Mannheim, Germany) (15). Sequencing confirmed 100% identity to published sequences. Control plasmids were linearized by restriction digestion with ApaI, PstI, or EcoRI and purified using the High Pure PCR Product Purification Kit (Roche Diagnostics).

PCRs were run in glass capillaries (Roche Diagnostics) in a 20-μl volume containing 2 μl FastStart DNA SybrGreen I mix, 3 mmol/l MgCl₂, 0.5 μmol/l of each primer, and 2 μl control template (standard) or cDNA. PCR conditions were 10 min denaturing at 95°C followed by 40 cycles of 15 s at 95°C, 5 s annealing, and 10 s at 72°C. To exclude amplification of traces of genomic DNA, intron-spanning primers were chosen. Primer pairs and annealing temperatures for each mRNA species were as follows: β-actin 342-bp fragment: 5′-ACCCACACTGTGCCCATCTA-3′ and 5′ -GCCACAGGATTCCATACCCA-3′ (58°C); IA-2 114-bp fragment: 5′ -TGCGCTCATTGCTGCTTACTCTG-3′ and 5′ -GGCGCTCCTTATCCCGTTGTTT-3′ (63°C); phogrin 149-bp fragment: 5′-CCTGCATCCTGGCCGTTCTCCTG-3′ and 3′ -ACGTTGGCGGCATAGCTCCTGGTA-5′ (68°C); insulin-1 164-bp fragment: 5′-ACCCAAGTCCCGTCGTGAAGT-3′ and 5′ -CCAGTTGGTAGAGGGAGCAGATG-3′ (61°C).

Amplification products were detected after each run by measurement of fluorescence intensity at 530 nm. A melting curve analysis was performed to confirm the specificity of the amplified PCR product. A set of four standards was used for quantification of each gene product run in parallel with the cDNA samples under identical PCR conditions. Standard curves were generated by serial dilutions of the control templates ranging from 10⁷ to 10⁴ copies (β-actin), 10⁶ to 10³ copies (IA-2 and phogrin), and 10⁸ to 10⁵ copies (insulin). The crossing cycle numbers of the logarithmic linear PCR phase of the standards were plotted against the logarithm of their concentrations. From the standard curves, cDNA copy numbers were interpolated by the LightCycler data analysis software using the Fit Points method as described in the manufacturer’s instructions. β-actin was used to normalize the probes to equal mRNA/cDNA levels. Data were calculated as copy number IA-2, phogrin, or insulin mRNA relative to β-actin.

The ability of freshly isolated or cultured islets to secrete insulin was assessed over a 2-h period in a static assay. Cultured islets of average size were selected under the stereomicroscope, washed, transferred into 96-well plates, and incubated for 2 h at 37°C in Krebs-balanced salt solution containing BSA (1 mg/ml) and 2 or 20 mmol/l glucose. Ten samples of one islet each were run in parallel for every condition. After 2 h, the insulin content of the supernatant was quantified by ELISA.

Insulin secretion was expressed relative to the DNA content of islets in each sample. After incubation, islets were sonicated in 100 μl PBS containing 2 mmol/l EDTA. After the addition of Hoechst 33258 solution, fluorescence of the samples was determined (excitation 360 nm, emission 460 nm) and compared with a standard curve of DNA.

Insulin in islet extracts and cell supernatants was measured by competitive ELISA. Elisa plates (Maxisorb; Nunc, Roskilde, Denmark) were coated with guinea pig anti-insulin antibody, blocked with BSA (5 mg/ml in Tris-buffered saline [TBS]/BSA), and subsequently incubated for 2 h with the insulin-containing sample and biotinylated insulin in TBS/BSA. Biotinylated insulin captured on the plate was detected with Streptavidin-alkaline phosphatase and developed with p-nitrophenyl phosphate. All samples were analyzed in duplicate and insulin concentration determined from a standard curve of rat insulin included on each plate. The detection range of the assay, defined as the limits within which the coefficient of variation between 10 independent assays was <20%, was 20–3,500 pg.

The significance of differences between groups was analyzed by Student’s t test or ANOVA, as appropriate.

We previously described a progressive increase with age in the islet IA-2 detected by immunohistochemistry on pancreas tissue sections of rats during the first 2 weeks of life (5). We set out to confirm these findings at the molecular level and to extend our observations to the closely related protein, phogrin (IA-2β). IA-2 and phogrin contents of extracts of isolated islets from rats of different ages (1, 5, 10 days and adult) were analyzed by Western blotting (Fig. 1A). Both IA-2 and phogrin resolved as two to three major bands of ∼66,000 Mr (Fig. 1A), which represent different post-translationally modified forms of the proteins (8,24). IA-2 was detected at very low levels in islets from 1-day-old rats, but increased progressively with age during the first 10 days of life. In contrast, phogrin levels were already high at 1 day of age and showed little variation in the older animals (Fig. 1B). These results extend the previous in situ hybridization studies indicating that phogrin is already expressed at the protein level in islets at birth and confirm that stable high expression of IA-2 is not detected until days or weeks after birth.

Islet IA-2 content was compared with acute insulin secretory responses to 20 mmol/l glucose of islets from rats of different ages to determine to what extent increases in IA-2 content are associated with changes in islet function. IA-2 levels within the islet were quantified by densitometry of Western blots and comparison with known amounts of recombinant protein. Both IA-2 content (Fig. 1C) and insulin secretion in response to 20 mmol/l glucose (Fig. 1D) progressively increased during the first 10 days of life (P < 0.0001). In contrast to changes in IA-2 content and insulin secretory responses, the islet insulin content was not influenced by age (Fig. 1C).

To understand causes of increased islet IA-2 after birth, we investigated factors that regulate IA-2 protein expression in neonatal rat islets in vitro. Glucose is a major stimulator of β-cell function and has been shown to upregulate a number of secretory granule proteins in isolated pancreatic islets (5). We therefore studied the influence of glucose on IA-2 content of islets from 5-day-old and adult rats cultured between 4 h and 10 days. Glucose stimulated IA-2 protein levels of islets from 5-day-old rats, with maximal effect after 48 h incubation at 10 mmol/l glucose (Fig. 2). Stimulatory effects of glucose were also observed in adult rat islets, although IA-2 content of freshly isolated islets was higher and stimulation less marked than in neonatal islets (Fig. 2). After sorting of dissociated islets into pancreatic β-cells and non-β-cells, there was no significant difference in IA-2 levels between the freshly sorted cell populations, but after culture, IA-2 levels were increased in the β-cells at 20 mmol/l glucose (P < 0.05) (Fig. 2).

Glucose-mediated increases in islet IA-2 content could represent direct effects of glucose or be secondary to autocrine action of secreted insulin. Stimulators and potentiators of secretion, such as tolbutamide or agents that increase cAMP (IBMX and forskolin) or activate protein kinase C (phorbol myristic acid [PMA]), all increased islet IA-2 content at 6 mmol/l glucose (P < 0.0001) (Fig. 3). This stimulation was accompanied by an increase in insulin secretion into the medium (Fig. 3). Exogenous insulin (3–16 nmol/l) increased IA-2 protein expression when added to medium containing 6 mmol/l glucose (P < 0.0001) (Fig. 3). IGF-1 also increased IA-2 content, an effect that was not inhibited by addition of an insulin receptor antibody. Glucose (10 mmol/l)-induced increases in islet IA-2 levels were abolished by blocking insulin binding to its receptor with an insulin receptor antibody. Diazoxide-mediated inhibition of glucose-induced insulin secretion blocked glucose-induced increases in IA-2 expression, an effect that could be partially reversed by exogenous insulin (Fig. 3).

In contrast to the results for IA-2, levels of phogrin protein in islets from 5-day-old rats were not altered by glucose, diazoxide, or tolbutamide (Fig. 4A). In adult human islets, similar effects of glucose, insulin, and insulin receptor antibodies on IA-2 protein levels could be observed (P < 0.05) (Fig. 4B), confirming that the observations in rat islets are relevant to human physiology.

Insulin-regulated expression of two islet proteins, glucokinase and insulin, occurs at the level of gene transcription (3,25, 26), and evidence for transcriptional regulation of IA-2 by agents that increase cAMP in β -cell lines has been presented (15). To understand the contribution of transcriptional regulation of IA-2 by glucose, insulin, and other agents, we compared IA-2 protein and mRNA levels in islets from 5-day-old rats using similar conditions as described above. Glucose (10 mmol/l) and IBMX both stimulated IA-2 protein and mRNA by two- to threefold (P < 0.005) (Fig. 5), and similar effects were seen with forskolin and PMA (data not shown). In contrast, tolbutamide and exogenous insulin increased IA-2 protein levels in the absence of equivalent increases in mRNA (Fig. 5). Anti-insulin receptor antibodies and diazoxide blocked glucose-stimulated increases in IA-2 protein levels without diminishing the IA-2 mRNA. These data show that insulin exerts a strong influence on the regulation of IA-2 protein levels but not of IA-2 mRNA. IA-2 mRNA levels are regulated by glucose and substances that increase intracellular cAMP. The phogrin mRNA copy number was 10-fold lower than for IA-2 but was not significantly influenced by glucose, forskolin, or IBMX (Fig. 6).

Metabolic labeling experiments were performed to determine whether post-transcriptional regulation of IA-2 levels by insulin occurs at the level of protein biosynthesis or degradation. Islets were incubated for 2 h in the presence of ³⁵S-methionine and radiolabeled IA-2 immunoprecipitated with a specific antibody and analyzed by SDS-PAGE and autoradiography. Newly synthesized radiolabeled IA-2 was almost undetectable on incubation of islets at 6 mmol/l glucose, but synthesis was stimulated in the presence of insulin (Fig. 7A). Pulse-chase experiments indicated that degradation of IA-2 was slow, and insulin did not alter the intensity of 66,000-Mr IA-2 bands after a 12-h chase (Fig. 7B).

Insulin secretion in response to glucose and other secretagogues is low throughout fetal life in humans and rodents and these responses increase after birth (27–30). Poor secretory responses appear not to be a consequence of inadequate insulin content of β-cells. Low levels of expression or activity of glucose transporters and metabolic enzymes may contribute (31,32), but other components of the secretory machinery may also be deficient in fetal and early neonatal life. The present study demonstrates that development of regulated insulin secretion after birth is accompanied by an increase in islet levels of IA-2, a tyrosine phosphatase-like protein that is proposed to mediate interactions of secretory granules with the islet cytoskeleton (12). The age-dependent increase in IA-2 levels is specific; similar increases were not detected for the closely related granule protein phogrin or for insulin.

Islets from neonatal rats could be stimulated to express adult levels of IA-2 by culture with stimulatory concentrations of glucose. Islets from adult rats also show glucose-stimulated increases in IA-2, although these were less marked than in immature rats. Glucose effects are seen in isolated β-cells from adult rats, whereas IA-2 content did not change on glucose treatment of non-β-cells. IA-2 is thus a “glucose response protein” in β-cells (33) and may represent one of the components of islet secretory granules previously shown to be increased by glucose (8). However, the major effect of glucose appears to be indirect and mediated by local effects of secreted insulin. Thus, blocking insulin secretion with diazoxide, or inhibiting insulin action on the islet insulin receptor antibodies, can abolish the stimulation of islet IA-2 levels by glucose. Other stimulators of insulin secretion, such as IBMX, PMA, or tolbutamide, upregulate IA-2 protein. Insulin itself increases islet IA-2 levels and reverses the inhibitory effects of diazoxide. These results identify insulin as a major regulator of IA-2 expression and the dominant mediator of the stimulatory effects of glucose on islet IA-2 levels in vitro.

Tyrosine phosphorylation of the insulin receptor and its substrates IRS-1 and IRS-2 are early events in insulin signaling, and phosphorylation of all three signaling molecules has been demonstrated in β-cells in response to both insulin and glucose (34,35). The IRS family of proteins are also substrates for the IGF-1 receptor. IGF-1 was also found to stimulate IA-2 expression, an effect that was not inhibited by insulin receptor antibodies, and is therefore likely to be mediated by its own receptor. These observations are consistent with IA-2 expression being regulated via signaling pathways common to both insulin and IGF receptors and raise the possibility that both hormones could influence islet IA-2 expression in vivo. A predominant effect of autocrine regulation of IA-2 by secreted insulin is suggested by the strong inhibitory effects on IA-2 expression of insulin receptor antibodies in vitro and the observation that IA-2 expression within islets is induced during the period in early life when insulin secretory responses develop.

Glucose has been shown to stimulate the biosynthesis of a number of proteins in the pancreatic β-cell (5,33), and, like IA-2, the regulation of these is in part dependent on the local action of secreted insulin. Thus, insulin stimulates the expression of both insulin and glucokinase, with regulation occurring at the level of transcription (3). To determine whether the increases in IA-2 protein observed in this study were also the result of transcriptional control, the influence of glucose, insulin, and insulin secretagogues on IA-2 mRNA was determined by quantitative RT-PCR. Glucose, agents that elevate cellular cAMP levels, or agents that activate protein kinase C each increased islet IA-2 mRNA levels. Increased IA-2 expression in response to elevated cAMP is consistent with the presence of a cAMP response element in the 5′-untranslated region of the IA-2 gene (36). These results confirm and extend previous observations in the Ins-1 cell line (15). However, insulin- and tolbutamide-stimulated IA-2 protein levels were not accompanied by an increase in IA-2 mRNA over the time course studied. Furthermore, antibodies to the insulin receptor did not inhibit glucose-induced increases in IA-2 mRNA. The study therefore identifies two levels of regulation of IA-2 expression: the transcriptional regulation already identified (15), which is mediated by cAMP and protein kinase C, and post-transcriptional regulation through insulin receptor signaling pathways. Because mRNA levels were determined at a single relatively late time point, we cannot exclude that insulin also exerts earlier transient effects on IA-2 gene transcription.

Insulin may stimulate islet IA-2 protein levels by increasing protein translation or by inhibiting protein degradation. Glucose has been shown to stimulate the biosynthesis of a number of islet secretory granule proteins by activating translation (5), an effect that may be dependent on regulation of the initiation phase of mRNA translation by secreted insulin (37). Regulation of IA-2 expression has been studied in rat pituitary cells (38), and increased transcription and translation could only partially account for the accumulation of IA-2 seen by treatment of cells with a combination of estradiol, insulin, and epidermal growth factor. Our own metabolic labeling studies with isolated rat islets indicated that regulation of IA-2 expression by insulin was largely at the level of protein biosynthesis. Pulse-chase experiments indicated that IA-2 degradation is slow at low glucose concentrations and not affected by insulin. Degradation or processing of IA-2 may therefore represent a level of control only under conditions in which exocytosis is stimulated (12).

Phogrin is closely related to IA-2, having a similar structure, tissue distribution, and subcellular localization. However, we found that phogrin levels were not increased by agents that modify β-cell function, indicating that secretory granule membrane components are subject to different modes of regulation. Similarly, there was no age dependence of islet levels of phogrin. Mature levels of phogrin were already detected in the neonate, consistent with in situ hybridization studies on phogrin (described as PTP-NP) in which phogrin mRNA was detected throughout fetal pancreas development (13). Differences in timing of expression of these proteins may have significance in type 1 diabetes, a disease in which both IA-2 and phogrin are targets of autoimmunity. Autoantibody reactivity to phogrin largely occurs as a result of cross-reactivity with the homologous regions on IA-2 (39), and different levels of expression and mechanisms of regulation, particularly in fetal and neonatal periods important for induction of immunological self-tolerance, may contribute to the higher susceptibility of IA-2 to autoimmunity. Understanding differences in regulation of IA-2 and phogrin is important to elucidate the roles of these PTP-like molecules in islet function and to determine whether disrupted insulin regulation of IA-2 can contribute to β-cell defects in type 2 diabetes.

This study was supported by grants from the European Union (TMR network CT 970142) and a Medical Research Council Training Fellowship to G.A.R.

The authors thank Drs. Peter Jones, Shanta Persaud, and John Hutton for gifts of antibodies.