This Article Abstract Full Text (PDF) All Versions of this Article: 21/11/2775    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yu, J. Articles by Barker, C. J. Search for Related Content PubMed PubMed Citation Articles by Yu, J. Articles by Barker, C. J.

An Autocrine Insulin Feedback Loop Maintains Pancreatic ß-Cell 3-Phosphorylated Inositol Lipids

The Rolf Luft Research Center for Diabetes and Endocrinology, Karolinska Institutet, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden

Address all correspondence and requests for reprints to: Per-Olof Berggren, The Rolf Luft Research Center for Diabetes and Endocrinology, Karolinska Institute, Karolinska University Hospital Solna, SE-171 76 Stockholm, Sweden. E-mail: per-olof.berggren{at}ki.se.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Phosphatidylinositol 3-kinases (PI3Ks) have a central role in pancreatic ß-cell function. Downstream events include the regulation of K_ATP channel activity, insulin secretion, gene transcription, and cell survival. Fewer data are available on the 3-phosphorylated inositol lipids (3-PIs) that are the primary products of these kinases. We characterized these PI3K products in insulin-secreting HIT T15 cells and were able to demonstrate, for the first time the presence of phosphatidylinositol 3,5-bisphosphate [PtdIns(3,5)P₂]. We then showed that glucose can significantly increase PtdIns(3,4,5)P₃, PtdIns(3,4)P₂, and notably PtdIns(3,5)P₂. We investigated the mechanism(s) whereby these molecules are generated under both basal and glucose-stimulated conditions. We postulated that insulin exocytosis could drive the rises in 3-PIs. In our experimental system, we could detect a rise in insulin secretion within 1 min of glucose stimulation, thus allowing the possibility that early rises in 3-PIs are regulated by secreted insulin. This was confirmed because blockade of the ß-cell insulin receptor completely abrogated the glucose-mediated increase of all three lipids, driving their concentrations below basal levels. Using primary pancreatic islets and either blockade of the insulin receptor or antibodies to insulin, we verified that basal insulin secretion is responsible for the maintenance of 3-PIs. Therefore, autocrine insulin signaling, a feature compromised in diabetes, is essential to up-regulate both basal and glucose-stimulated levels of a vital family of second messengers that preserve and drive pancreatic ß-cell function.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
SIGNAL TRANSDUCTION VIA phosphatidylinositol 3-kinases (PI3Ks) is of critical importance in pancreatic ß-cell function. A growing number of key processes are controlled by these pathways (1), including the regulation of ion flux (2, 3, 4), insulin secretion (2, 3, 4, 5), and gene transcription (6, 7, 8, 9, 10). However, there is surprisingly little information on the initial lipid products of PI3Ks and how they respond to glucose stimulation. Because each lipid can be produced by a different PI3K isoform (11) and has specific protein targets (11), the identification and response of these lipids under different physiological conditions is important for a full understanding of pathways downstream of PI3Ks in ß-cell signal transduction. Furthermore, not all actions of PI3Ks are mediated by their lipid kinase activity (12, 13), making it important to ascertain when there is a genuine rise in the lipid products.

In the family of 3-phosphorylated inositol lipids (3-PIs), four species have been identified in eukaryotic cells, namely phosphatidylinositol 3-phosphate (PtdIns3P), phosphatidylinositol 3,4-bisphosphate [PtdIns(3,4)P₂], PtdIns(3,5)P₂, and phosphatidylinositol 3,4,5-trisphosphate [PtdIns(3,4,5)P₃, or PIP₃]. They are generated by phosphorylation of PtdIns, PtdIns 4-phosphate (PtdIns4P), and PtdIns(4,5)P₂ on the 3-hydroxyl group of the inositol ring by PI3Ks, although a substantial proportion of PtdIns(3,4)P₂ is likely to be derived from PtdIns(3,4,5)P₃ dephosphorylation (11). PtdIns(3,5)P₂ is formed by phosphorylation of PtdIns3P by PI5-kinases, probably PIKfyve (phosphoinositide kinase with a specificity for the five position containing a fyve finger) (14, 15).

With the exception of PtdIns(3,5)P₂, these lipids have been described in ß-cells labeled with either [³H]myo-inositol or [³²P]orthophosphate (16). Their response to glucose alone has not been reported, but raised concentrations of PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃ did follow the combined application of high glucose and carbachol (16). However, no mechanism was offered for these increases. We hypothesized that an autocrine insulin feedback loop could be the mechanistic explanation for this rise in these 3-phosphorylated lipids. However, to test this hypothesis, there was a need for a detailed characterization of these inositides due to a discrepancy in the published information on PtdIns(3,4,5)P₃ in ß-cells (16, 17).

We now show an essential role for secreted insulin in the production of PtdIns(3,4,5)P₃, PtdIns(3,4)P₂, and PtdIns(3,5)P₂ under both basal and glucose-stimulated conditions.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Characterization of the 3-PIs in Insulin-Secreting Cells
To test our hypothesis that an autocrine insulin feedback loop was behind the 3-PI signaling, we first needed to establish which lipids were present in our ß-cells. In this initial investigation, our first aim was to reexamine the basis of the high basal PtdIns(3,4,5)P₃ reported in a previous study (16), because a subsequent study using a PtdIns(3,4,5)P₃ mass assay did not confirm these data (17). We examined the basal levels of PtdIns(3,4,5)P₃ by labeling HIT T15 cells with [³H]myo-inositol and extracted the lipids using protocols designed to maximize the recovery of PtdIns(3,4,5)P₃. These extracts were then deacylated and the resulting glycerophosphate derivatives separated on HPLC. Our initial experiments seemed to confirm the observations of the previous study (16), indicating that a high basal level of PtdIns(3,4,5)P₃ was present in insulin-secreting cells, for example see Fig. 1A. However, the peak that was presumed to be GroPIns(3,4,5)P₃, the deacylation product of PtdIns(3,4,5)P₃, in the HPLC profile consisted of one major peak with a significant shoulder, indicating that at least two different compounds were present. A similar profile was observed in the original report (16). The peak was checked with [³H]Ins(1,4,5)P₃ and [³H]Ins(1,3,4,5)P₄ standards by splitting an additional sample into two and running half with and half without the standards (Fig. 1, A and B). We found that running standards in separate runs was an unreliable method of assigning peak identity, even if a mock extraction was added to the standards. Clearly, Ins(1,4,5)P₃ contributed to the peak A (Fig. 1B), presumed to be GroPIns(3,4,5)P₃, if compared with the same peak from the other half of the sample eluted without the standards (Fig. 1A). Ins(1,3,4,5)P₄ eluted after the putative PtdIns(3,4,5)P₃ peak and was denoted peak B. These data demonstrate that our putative PtdIns(3,4,5)P₃ peak was likely to contain Ins(1,4,5)P₃ as well as PtdIns(3,4,5)P₃ and possibly other unknown compounds. We repeated the chromatography of the deacylated lipids using a gradient with greater resolving power, Fig. 1, C and D. The original peak A was then separated into three peaks, denoted A, A1, and A2 (Fig. 1C). Again samples were split into two and run with or without standards, Fig. 1, C and D, respectively. This showed that peak A coeluted with Ins(1,4,5)P₃, and peaks A1 and A2 ran between Ins(1,4,5)P₃ and Ins(1,3,4,5)P₄ (peak B), which eluted last. It was unclear whether peak A1 or A2 corresponded to GroPIns(3,4,5)P₃. To resolve this, a ³²P-labeled GroPIns(3,4,5)P₃ internal standard was prepared and added to the deacylated extract and run on HPLC (see Fig. 2A). The peak A2 was identified as GroPIns(3,4,5)P₃, the deacylated product that corresponds to PtdIns(3,4,5)P₃. The origin of the unidentified peak A1 is discussed below. In the early characterization of deacylated PtdIns(3,4,5)P₃ of Stephens et al. (18), three similar peaks are also apparent on the chromatogram. An explanation for the reason why Stephens et al. (18) and our current study could resolve these peaks, but the early report could not, may lie with the HPLC columns used. The original PtdIns(3,4,5)P₃ study in insulin-secreting cells used a Partisil SAX column (16), which has half the resolving power of the Partisphere SAX column used by Stephens et al. (18) and ourselves. In summary, our data demonstrate that the putative PtdIns(3,4,5)P₃ comprises only about 0.1% of the amount of the PtdIns(4,5)P₂ peak. This contrasts with the previous studies that reported a figure of about 4%.

View larger version (9K): [in this window] [in a new window]   Fig. 1. HPLC Profiles Showing the GroPIns(3,4,5)P3 Region of the Chromatogram Deacylated lipid samples were split into two halves and eluted by using a previously described gradient (19 ). A, Elution of half the sample without standards. B, Elution of the second half spiked with [3H]-labeled standards of Ins(1,4,5)P3 and Ins(1,3,4,5)P4. The Ins(1,4,5)P3 standard contributed to the major peak A with Ins(1,3,4,5)P4 eluting later and denoted peak B. Two other experiments gave similar results. Similar samples were also split and separated by HPLC on a revised gradient with greater resolving power (see Materials and Methods). C, The original peak A in panels A and B was separated into three peaks, peak A, A1, and A2. D, The second half of the sample was spiked with [3H]-labeled standards of Ins(1,4,5)P3 and Ins(1,3,4,5)P4. The Ins(1,4,5)P3 standard contributed to peak A. Peaks A1 and A2 remained to be identified. Peak B was the standard of Ins(1,3,4,5)P4. E, Standard [3H]PtdIns(4, 5)P2 was added to an equivalent dish of unlabeled cells as described for A–D and subjected to extraction and deacylation. The sample was then run using the improved gradient. Both peak A, corresponding to Ins(1,4,5)P3, and peak A1 (see Discussion) were detected. Two other experiments gave similar results.

View larger version (18K): [in this window] [in a new window]   Fig. 2. HPLC Separation of [3H]-Labeled and Deacylated 3-Phosphorylated Lipids with [32P]-Labeled Internal Standards Deacylated inositol lipids were separated exactly as described in Fig. 1, C and D. Internal [32P]-labeled standards were prepared (see Materials and Methods) and included in the HPLC runs. A, The GroPIns(3,4,5)P3 region of the chromatogram showing the coelution with genuine [32P]GroPIns(3,4,5)P3. B, The GroPInsP and GroPInsP2 regions of the chromatogram showing the coelution of deacylated species with 32P-labeled standards of GroPIns3P, GroPIns(3,5)P2, and GroPIns(3,4)P2. , Deacylated [3H]inositol lipid extract; , [32P]-labeled internal standards. This trace is typical of two others.

To identify the remaining 3-phosphorylated lipids, the deacylated forms of [³²P]-labeled PtdIns3P, PtdIns(3,4)P₂, and PtdIns(3,5)P₂ internal standards were prepared and then run on HPLC with deacylated [³H]myo-inositol-labeled cell extracts. Figure 2B shows the coelution of peaks corresponding to all three of these lipids. Of particular importance is the discovery of PtdIns(3,5)P₂ in pancreatic ß-cells.

Glucose Alone Can Significantly Increase PtdIns(3,5)P₂, PtdIns(3,4)P₂, and PtdIns(3,4,5)P₃
A previous study has demonstrated that a combination of high concentrations of carbachol (0.5 mM) and glucose (27 mM) results in an increase in PtdIns(3,4,5)P₃ (16). To investigate whether PtdIns(3,4,5)P₃ and other 3-phosphorylated lipids are changed with glucose stimulation alone, and that this is sufficient to establish an insulin feedback loop, we prelabeled the HIT T15 cells with [³H]myo-inositol for 48 h. They were then preincubated in a Krebs buffer containing low glucose (0.1 mM) and then stimulated with 10 mM glucose alone for 1 and 5 min, respectively, and analyzed by classical HPLC. Simultaneously, we also used various agents to address the mechanism of the glucose-induced rise in 3-PIs. The rationale for the particular agents used is detailed in the next section. The medium was also taken to assess insulin secretion, again as detailed below. The data on the glucose stimulation showed significant increases in PtdIns(3,4,5)P₃ (see Fig. 4, C and G) and PtdIns(3,4)P₂ (see Fig. 4, B and F) after 1 min glucose stimulation, with a continued significant rise after 5 min. There was only a significant increase in PtdIns(3,5)P₂ (see Fig. 4, A and E) after 5 min. There was no significant change in PtdIns3P (see Fig. 4, D and H).

View larger version (14K): [in this window] [in a new window]   Fig. 4. Effects of Nimodipine and HNMPA-(AM)3 on Glucose-Induced 3-PI Generation The cells were preincubated for 30 min in Krebs buffer containing 0.1 mM glucose and in the presence of 10 µM nimodipine or 100 µM HNMPA-(AM)3 or equal amount of dimethylsulfoxide (vehicle) for control samples and then stimulated by 10 mM glucose for 1 (A–D) or 5 (E–H) min. The bars represent the average of four independent experiments each carried out in triplicate, normalized to the time-matched control as 100%. Statistical analysis was performed as described in Fig. 3. *, P < 0.05; **, P < 0.01; ***, P < 0.001. G, Cells were preincubated with dimethylsulfoxide alone and stimulated with 10 mM glucose; G+N, cells were preincubated with nimodipine and then stimulated with 10 mM glucose; G+H, cells were preincubated with HNMPA-(AM)3 and stimulated with 10 mM glucose. In the statistical analysis, we also tested the significance of the glucose-stimulated rise above control, and this is presented here to avoid confusion with the graph. The glucose-stimulated increase of PtdIns(3,4)P2 was significant at 1 and 5 min (P < 0.001 and P < 0.01, respectively). The PtdIns(3,4,5)P3 increase was significant at 1 and 5 min (P < 0.01 and P < 0.001, respectively). The increase in PtdIns(3,5)P2, was significant only after 5 min (P < 0.05), and there was no statistically significant change in PtdIns3P at either 1 or 5 min. NS, Not significant.

Figure 3 shows the insulin secretion data from our experiments. After 1 min glucose stimulation, the level of insulin secretion is doubled (Fig. 3A) and reaches a 4-fold increase by 5 min (Fig. 3B). Thus, increased insulin secretion could explain the early rise in 3-PIs. After 1 min glucose stimulation, blockade of the L-type channel by nimodipine does not significantly decrease the rise in insulin secretion. However, after 5 min glucose stimulation, the secretion of insulin is strongly dependent on Ca²⁺ influx through the L-type Ca²⁺ channel. Our data indicate that some insulin secretion can even occur if the channel is blocked, but with time, there is an increased dependency on insulin secretion triggered by the influx of Ca²⁺. The fact that insulin signaling can still be enhanced in the presence of the Ca²⁺-channel blocker nimodipine is important to bear in mind when interpreting the lipid data in the presence of this Ca²⁺-channel blocker. Preincubating the cells with the cell-permeant insulin receptor tyrosine kinase inhibitor (hydroxy-2-naphthalenylmethyl) phosphonic acid acetoxy methyl ester [HNMPA-(AM)₃] (20, 21), before glucose stimulation, either enhances, at 1 min, or inhibits, at 5 min, glucose-dependent insulin secretion, suggesting the involvement of both a negative and a positive insulin feedback loop, dependent on the time of stimulation. Whether there is an increased or decreased secretion in response to HNMPA treatment is immaterial with respect to changed 3-PI levels, because any feedback through the insulin receptor is blocked by the inhibitor.

View larger version (22K): [in this window] [in a new window]   Fig. 3. Effects of the L-Type Ca2+-Channel Blocker Nimodipine and the Insulin Receptor Blocker HNMPA-(AM)3 on Glucose-Stimulated Insulin Secretion The samples described here come from the supernatant of the same experimental dishes that were used for lipid analysis (see Fig. 4). The samples were stimulated with 10 mM glucose for 1 (A) or 5 (B) min. G, Cells were preincubated only with dimethylsulfoxide and stimulated by 10 mM glucose; G+N, cells were preincubated with 10 µM nimodipine and stimulated by 10 mM glucose; G+H, cells were preincubated with 100 µM HNMPA-(AM)3 and stimulated by 10 mM glucose. The supernatants were analyzed for insulin content using a standard RIA. The bars represent the average of four experiments, control being 100%. Statistical analysis was performed by using one-way ANOVA and followed by Bonferroni’s multiple comparison test (GraphPad, Prism 3.03). *, P < 0.05; **, P < 0.01; ***, P < 0.001.

An interesting and very important piece of information was that basal secretion acts to maintain the levels of 3-phosphorylated lipids in insulin-releasing cells and thus presumably constitutes an important physiological signal in the insulin secretory process. Because our current studies were carried out exclusively on a ß-cell line, to reinforce the validity of this novel observation, we decided to examine the effect of basal insulin secretion on the 3-PIs in primary pancreatic ß-cells (Fig. 5). To be confident that the effects we were seeing came specifically from ß-cells, we used islets from an ob/ob mouse variant that is normo-insulinemic and normo-glycemic but whose islets consist of more than 90% ß-cells. We used two approaches to curtail the effect of basal insulin secretion. First, to make a more direct comparison with the HIT cell data, we used the insulin receptor blocker HNMPA. We also used antibodies to insulin to mop up the secreted insulin, a technique that has been used effectively to study the importance of insulin feedback signaling (22, 23) and that removes any doubt about the use of pharmacological tools like HNMPA. Islets were prelabeled with [³H]inositol and then preincubated at 3 mM glucose in the presence of either HNMPA or vehicle. HNMPA did not significantly change insulin secretion when compared with time-matched control incubations (data not shown). Separate experiments were carried out with insulin antibodies or IgG. Figure 5, A and B, shows that the level of all 3-PIs is reduced with respect to their controls by about 25%. In HNMPA-treated cells, this reduction in 3-PIs was significantly different compared with controls in all three lipids measured. However, in antibody-treated cells, the drop was significant only for PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃. Because access of the antibodies to intra-islet insulin may be restricted, the magnitude of the drop of 3-PIs by antibody treatment may underestimate the dependency of the lipids on basal insulin secretion. Thus, basal insulin secretion is critical for the maintenance of 3-PIs in primary ß-cells as well as in insulin-secreting cell lines.

View larger version (16K): [in this window] [in a new window]   Fig. 5. Basal Insulin Secretion Maintains 3-PIs in Primary Mouse ß-Cells A, Mouse ß-cells were labeled, washed, and preincubated in 3 mM glucose for 30 min, as described in Materials and Methods, in the presence of either 100 µM HNMPA or vehicle (dimethylsulfoxide). B, In separate experiments, insulin antibodies (to porcine insulin) or the appropriate IgG (0.8 mg/ml) were applied. Reactions were terminated and lipids extracted as described in Materials and Methods. The data are expressed as a percentage of the IgG control for each respective lipid and are the average ± SEM of three separate preparations of islets. *, P < 0.05; **, P < 0.01.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

The first identification of PtdIns(3,5)P₂ in insulin-secreting cells may come as no surprise because this lipid has now been found in a wide variety of other eukaryotic cells (14, 15) and is formed by the phosphorylation of PtdIns3P by PI5-kinases (11, 25) or PIKfyve (14, 15). However, its suggested roles in vesicle transport (14, 15, 26, 27, 28, 29, 30, 31, 32, 33) and in cell stress responses (14, 26, 32, 33) will impact on essential aspects of ß-cell physiology, e.g. exocytosis of insulin and ß-cell preservation during conditions of stress and thus maybe serve as a new target in ß-cell pathology.

We were able to demonstrate, using glucose stimulation alone, a significant increase in all 3-PIs with the exception of PtdIns3P. This is in contrast to a previous report in which glucose’s unique contribution to increases in 3-PI was masked by the simultaneous application of both high concentrations of glucose and carbachol (16). Our data suggest not only that PtdIns(3,4,5)P₃ but also PtdIns(3,4)P₂ and notably PtdIns(3,5)P₂ concentrations can be raised by glucose stimulation alone. This indicates that the PI3Ks (class I and II) as well as PI5K and/or PIKfyve (11, 14, 25) are likely to be involved in the acute ß-cell responses to glucose.

Although there is much circumstantial, indirect evidence in the literature that 3-PIs form part of the autocrine feedback loops mediated by insulin secretion in pancreatic ß-cells (1), direct addition of insulin to unstimulated ß-cells has (17), or has not (16), resulted in increases in 3-PIs. Our new data not only support our hypothesis that an insulin feedback loop drives these increases, but directly show that this insulin feedback is actually essential for the glucose-stimulated rise in 3-PIs in insulin-secreting cells. This is true for both those lipids that have been previously studied in ß-cells such as PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃ as well as PtdIns(3,5)P₂, the latter described here for the first time. A striking observation in our system is that even within 1 min of glucose stimulation, there is a rise in PtdIns(3,4,5)P₃ and PtdIns(3,4)P₂ that is insulin dependent and follows an early rise in insulin concentration. This suggests that the insulin feedback loop can operate comparatively rapidly and within the timeframe of the initial ß-cell stimulus-secretion coupling events

The mechanistic link between insulin signaling and the generation of PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃ is well established (34); however, only recently has a possible direct link between insulin signaling and PtdIns(3,5)P₂ been described. PIKfyve, an enzyme responsible for PtdIns(3,5)P₂ synthesis, is activated by insulin (15, 35) via a protein kinase B/Akt-directed phosphorylation (36). This suggests that PtdIns(3,5)P₂ production is secondary to PI3K signaling. Because PtdIns(3,5)P₂ levels, unlike those of PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃, are only modestly increased by glucose stimulation, it is likely that the main actions of PtdIns(3,5)P₂ are mediated under basal conditions. In contrast, PtdIns(3,4)P₂ and PtdIns(3,4,5)P₃ are acutely responsive to increased glucose concentration and the concomitant increase in secreted insulin.

One assumption we have made is that the effect of insulin is mediated by its own receptor, but in addition, insulin can generate the formation of 3-PIs via the IGF-I receptor, which is also found on ß-cells (37). Although HNMPA was specifically devised to block insulin receptor signaling (20, 21), it is a pharmacological agent, and one cannot discount the role of the IGF-I on these grounds. However, when we examined the basal concentrations of insulin in control or glucose-stimulated HIT cells, they were 0.5 nM (basal) and 1.7 nM (stimulated), respectively. At these concentrations, we would anticipate little contribution of IGF-I-mediated vs. insulin receptor-mediated signaling.

When using pharmacological tools or ß-cell lines, a degree of caution needs to be exercised when interpreting results. In regard to the pharmacological agents, we detected no gross effect of these agents on cell viability during the short time course of our experiments. However, side effects on the glucose response not directly related to actions on insulin secretion may occur. For example, nimodipine can inhibit glucose transport (38) and thus may impair glucose sensing. Depression of insulin signaling can also down-regulate glucokinase gene expression (10), although this latter effect would not be expected to have any impact during the timescale of our experiments. However, data obtained by these inhibitors is supported by non-pharmacological tools discussed in the next section, suggesting that our results do reflect the physiological situation. Criticism can also be leveled at our use of cell lines. Nonetheless, an independent study published as we were submitting this paper strongly supports our data. This study used an alternative approach to monitor PtdIns(3,4,5)P₃ alone (39). Using a fluorescent PtdIns(3,4,5)P₃ binding reporter construct, these authors clearly demonstrated in another ß-cell system the dependency of glucose-mediated increases in PtdIns(3,4,5)P₃ on insulin feedback Thus, this is not just a HIT cell phenomenon. However, an unexpected and exciting outcome of our studies in the HIT T15 ß-cell line was the dependency of 3-PIs on basal insulin secretion. Given the novelty of this finding, it was paramount that we verified it in a primary ß-cell system.

Our data from primary mouse ß-cells islets, either using insulin receptor inhibition by HNMPA or by removal of secreted insulin by antibodies, also showed reduced 3-PI levels under basal conditions. This indeed suggests that this phenomenon is not an artifact of either a model cell line or a pharmacological tool. Of course, in the intact islet, insulin signaling is more complex because basal insulin secretion may influence other islet cells, e.g. -cells, leading to paracrine effects. Under low-glucose conditions, -cells themselves are triggered to secrete hormones such as glucagon that can further modify ß-cell function. However, we would not anticipate insulin-provoked increases in -cell 3-PIs to contribute much to our overall data, because in our particular mouse model, -cells represent only about 5% of the total cell population.

It has been reported that basal insulin secretion is important for the subsequent ability of the cell to release insulin after glucose stimulation in primary islets (23). Our study now indicates at least one key signaling pathway that will be compromised by removal of basal insulin feedback. Thus, important signaling is taking place under both stimulated and basal conditions. Now that it is appreciated that the ß-cell itself has functional insulin receptors (6, 7, 8, 9, 10), it is possible that the insulin-resistance phenomenon in type 2 diabetes can be enacted in this cell. If this occurs, it will affect both the acute responses to increased insulin secretion and also influence longer-term responses to basal insulin secretion, which is nonetheless important for acute ß-cell exocytosis (23) and probably viability. It is tempting to speculate that in the classic insulin-sensitive tissues, basal plasma insulin performs a hitherto unrecognized role in the maintenance of basic phosphoinositide signaling. If this is the case, the impact of insulin resistance may be more pervasive than previously appreciated.

In conclusion, we have demonstrated that glucose-dependent production of the key 3-PIs, PtdIns(3,4,5)P₃, PtdIns(3,4)P₂, and PtdIns(3,5)P₂ is mediated by secreted insulin and that even insulin secreted under basal conditions is important in maintaining the level of these lipid signals. Hence, our data for the first time show that the autocrine insulin feedback loop drives a constantly active phosphoinositide signaling platform that maintains both ß-cell exocytotic capacity and survival.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Cell Culture and Radiolabeling
HIT T15 cells (passages 78–80) were routinely maintained in RPMI 1640 medium with 10% fetal bovine serum, glutamine (2 mM), and a penicillin (100 IU/ml)/streptomycin (100 mg/ml) cocktail (Invitrogen AB, Stockholm, Sweden). For the experiments, the cells were cultured in 92-mm-diameter petri dishes with a modified medium RPMI 1640, 1640-M, for 5 d. The modified medium consisted of RPMI 1640 that was glucose and inositol free (Invitrogen). This base medium was then supplemented with the following additives: 0.4 mM MgSO₄, 50 µM inositol, and 5.5 mM glucose with glutamine and penicillin/streptomycin added at the same concentrations as the standard medium above. Fetal bovine serum was dialyzed using a 1000-molecular-weight cutoff Spectro Por tubing (Spectrum Europe BV, DG Breda, The Netherlands) and then added to the modified medium at a concentration of 10%. The cells were labeled with 10 µCi/ml [2-³H]myo-inositol for 48 h before experiments were undertaken.

Experimental Conditions and Lipid Analysis
For the characterization and identification of 3-PIs under basal conditions, HIT T15 cells were grown in [³H]myo-inositol for 48 h and then incubated in a Krebs buffer supplemented with 0.05% BSA, 0.1 mM glucose, 50 µM inositol, and 2 µM CaCl₂ for 30 min in a water bath at 37 C. Extraction and deacylation of the lipids were carried out as described below. For glucose-stimulation experiments, cells were preincubated with or without nimodipine and HNMPA-(AM)₃ in a similar Krebs buffer as above for 30 min and then stimulated with 10 mM glucose for either 1 or 5 min. Cells were quenched with 1 ml ice-cold 1 M HCl. A 10-µl aliquot of a lipid carrier (Sigma phosphoinositide mix, 25 mg/ml; Sigma Aldrich Sweden AB, Stockholm, Sweden) was immediately added and the plate left to stand on ice for 20 min. An established lipid extraction protocol to quantitatively extract lipids, especially for PtdIns(3,4,5)P₃, was used (40). Throughout extraction, only siliconized glass or plastic-ware and pipette tips were used. Cells were then harvested by use of a cell scraper, and the plates washed with 2.73 ml of a solution combined from 0.6 ml 1 M HCl, 5 mM tetrabutyl ammonium sulfate, and 2.13 ml methanol. Chloroform (4.27 ml) was added to split the phase. The mixture was vortexed and the phases separated by centrifugation. The lower phase containing the inositol lipids was transferred to tubes already containing 1.43 ml of the synthetic upper phase. The phases were mixed and centrifuged, and the lower phase was removed into clean tubes. Both the initial upper phase and the synthetic upper phase (left after removing the first synthetic lower-phase wash) were sequentially reextracted with 2.23 ml of the synthetic lower phase, mixed, and separated centrifugally. This final lower phase was combined with the originally washed lower phase, the tube filled with N₂, and the lipid extract stored at –20 C. To determine PtdIns(3,4,5)P₃, the lipid extract was dried under N₂ and deacylated using methylamine, exactly as described by Anderson et al. (40). The products were stored at –20 C until separated by HPLC.

To be able to detect 3-PIs in the relatively little material that could be generated from primary ob/ob mouse islets, we modified the above lipid-labeling protocol. Islets, 300 per condition, were labeled in a RPMI 1640 medium containing only 2 µM inositol and 100 µCi/ml [³H]inositol. They were then washed with the Krebs buffer described above, this time containing 3 mM glucose, and then preincubated in this same media for 30 min in the presence of 0.8 mg/ml of either insulin antibodies (to porcine insulin) or the appropriate IgG. Antibodies were dialyzed before use to remove azide and other contaminants. Reactions were terminated, and lipids extracted as described above. Note that pancreatic islets were prepared from mice in accordance with local ethical guidelines.

Preparation of [³²P]-Labeled Lipid Standards
To verify the identities of 3-phosphorylated lipids, [³²P]-labeled lipids were prepared as internal standards. The [³²P]-labeled PtdIns3P, PtdIns(3,4)P₂, PtdIns(3,4,5)P₃, and PtdIns(3,5)P₂ standards were generated by phosphorylating a mixture of the cold inositol lipids PtdIns, PtdIns4P, PtdIns(4,5)P₂ (Sigma Aldrich Sweden), or PtdIns5P (Echelon Research Laboratories, Salt Lake City, UT) with an equal proportion of phosphatidylserine and recombinant PI3K (Alexis Biochemicals, KELAB, Göteborg, Sweden) together with [-³²P]ATP (Amersham, Pharmacia Biotech, Uppsala, Sweden). The mixtures were incubated at 37 C for 1 h. Then the lipids were extracted, dried, deacylated, and stored as described above. Small aliquots of the standards were added to the samples and run on HPLC.

HPLC
The separation of deacylated inositol lipids was performed by HPLC on a 25-cm Whatman Partisphere-SAX column (Laserchrom, Rochester, UK). In initial experiments (Fig. 1, A and B), we used a published method (41). For all remaining studies, an improved gradient was used. The gradient was generated from deionized H₂O (buffer A) and 1.0 M (NH₄)₂HPO₃ adjusted to pH 3.8 with H₃PO₄ (buffer B). The gradient was as follows: 0 min, 0% B; 5 min, 0% B; 60 min, 15% B; 80 min, 15% B; 88 min, 20% B; 108 min 30% B; and 123 min, 100% B. Radioactivity was determined by the addition of Packard Ultima Flo AP scintillant and counted on a Packard CA 2000 scintillation counter (both from CIAB, Stockholm, Sweden).

Insulin Secretion Analysis
Insulin secretion was measured in the same dishes as those used to determine the inositol lipids. A 5-ml aliquot, from a total of 8 ml, was removed from each dish and was stored at –20 C. Insulin release was measured by RIA (42). Radioactivity was counted by a Packard -counter.

	ACKNOWLEDGMENTS

 
We thank Drs. B. and I. B. Leibiger for helpful discussions.

	FOOTNOTES

 
This work was supported by grants from the Karolinska Institute, Novo Nordisk Foundation, the Swedish Research Council, the Swedish Diabetes Association, the Juvenile Diabetes Research Foundation International, European Foundation for the Study of Diabetes (EFSD), EuroDia (LSHM-CT-2006-518153), The Family Erling-Persson Foundation, Åke Wibergs Foundation, and Berth von Kantzow’s Foundation.

Disclosure Statement: The authors have nothing to disclose.

First Published Online July 24, 2007

Abbreviations: HNMPA-(AM)₃, (Hydroxy-2-naphthalenylmethyl) phosphonic acid acetoxy methyl ester; 3-PI, 3-phosphorylated inositol lipid; PI3K, phosphatidylinositol 3-kinase; PtdIns3P, phosphatidylinositol 3-phosphate; PtdIns(3,4)P₂, phosphatidylinositol 3,4-bisphosphate; PtdIns(3,4,5)P₃, phosphatidylinositol 3,4,5-trisphosphate.

Received for publication November 13, 2006. Accepted for publication July 20, 2007.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 

1. Barker CJ, Leibiger IB, Leibiger B, Berggren PO 2002 Phosphorylated inositol compounds in ß-cell stimulus-response coupling. Am J Physiol Endocrinol Metab 283:E1113–E1122
2. Aspinwall CA, Qian WJ, Roper MG, Kulkarni RN, Kahn CR, Kennedy RT 2000 Roles of insulin receptor substrate-1, phosphatidylinositol 3-kinase, and release of intracellular Ca²⁺ stores in insulin-stimulated insulin secretion in ß-cells. J Biol Chem 275:22331–22338[Abstract/Free Full Text]
3. Khan FA, Goforth PB, Zhang M, Satin LS 2001 Insulin activates ATP-sensitive K⁺ channels in pancreatic ß-cells through a phosphatidylinositol 3-kinase-dependent pathway. Diabetes 50:2192–2198[Abstract/Free Full Text]
4. Larsson O, Barker CJ, Berggren PO 2000 Phosphatidylinositol 4,5-bisphosphate and ATP-sensitive potassium channel regulation: a word of caution. Diabetes 49:1409–1412[Abstract]
5. Zawalich WS, Zawalich KC 2000 A link between insulin resistance and hyperinsulinemia: inhibitors of phosphatidylinositol 3-kinase augment glucose-induced insulin secretion from islets of lean, but not obese, rats. Endocrinology 141:3287–3295[Abstract/Free Full Text]
6. Macfarlane WM, Smith SB, James RF, Clifton AD, Doza YN, Cohen P, Docherty K 1997 The p38/reactivating kinase mitogen-activated protein kinase cascade mediates the activation of the transcription factor insulin upstream factor 1 and insulin gene transcription by high glucose in pancreatic ß-cells. J Biol Chem 272:20936–20944[Abstract/Free Full Text]
7. Leibiger IB, Leibiger B, Moede T, Berggren PO 1998 Exocytosis of insulin promotes insulin gene transcription via the insulin receptor/PI-3 kinase/p70 s6 kinase and CaM kinase pathways. Mol Cell 1:933–938[CrossRef][Medline]
8. da Silva Xavier G, Varadi A, Ainscow EK, Rutter GA 2000 Regulation of gene expression by glucose in pancreatic ß-cells (MIN6) via insulin secretion and activation of phosphatidylinositol 3'-kinase. J Biol Chem 275:36269–36277[Abstract/Free Full Text]
9. Leibiger B, Moede T, Schwarz T, Brown GR, Kohler M, Leibiger IB, Berggren PO 1998 Short-term regulation of insulin gene transcription by glucose. Proc Natl Acad Sci USA 95:9307–9312[Abstract/Free Full Text]
10. Leibiger B, Leibiger IB, Moede T, Kemper S, Kulkarni RN, Kahn CR, de Vargas LM, Berggren PO 2001 Selective insulin signaling through A and B insulin receptors regulates transcription of insulin and glucokinase genes in pancreatic ß-cells. Mol Cell 7:559–570[CrossRef][Medline]
11. Vanhaesebroeck B, Leevers SJ, Ahmadi K, Timms J, Katso R, Driscoll PC, Woscholski R, Parker PJ, Waterfield MD 2001 Synthesis and function of 3-phosphorylated inositol lipids. Annu Rev Biochem 70:535–602[CrossRef][Medline]
12. Bondeva T, Pirola L, Bulgarelli-Leva G, Rubio I, Wetzker R, Wymann MP 1998 Bifurcation of lipid and protein kinase signals of PI3K to the protein kinases PKB and MAPK. Science 282:293–296[Abstract/Free Full Text]
13. Pirola L, Zvelebil MJ, Bulgarelli-Leva G, Van Obberghen E, Waterfield MD, Wymann MP 2001 Activation loop sequences confer substrate specificity to phosphoinositide 3-kinase alpha (PI3K). Functions of lipid kinase-deficient PI3K in signaling. J Biol Chem 276:21544–21554[Abstract/Free Full Text]
14. Michell RH, Heath VL, Lemmon MA, Dove SK 2006 Phosphatidylinositol 3,5-bisphosphate: metabolism and cellular functions. Trends Biochem Sci 31:52–63[CrossRef][Medline]
15. Shisheva A 2001 PIKfyve: the road to PtdIns 5-P and PtdIns 3,5-P(2). Cell Biol Int 25:1201–1206[CrossRef][Medline]
16. Alter CA, Wolf BA 1995 Identification of phosphatidylinositol 3,4,5-trisphosphate in pancreatic islets and insulin-secreting ß-cells. Biochem Biophys Res Commun 208:190–197[CrossRef][Medline]
17. Harvey J, McKay NG, Walker KS, Van der Kaay J, Downes CP, Ashford ML 2000 Essential role of phosphoinositide 3-kinase in leptin-induced K(ATP) channel activation in the rat CRI-G1 insulinoma cell line. J Biol Chem 275:4660–4669[Abstract/Free Full Text]
18. Stephens LR, Hughes KT, Irvine RF 1991 Pathway of phosphatidylinositol(3,4,5)-trisphosphate synthesis in activated neutrophils. Nature 351:33–39[CrossRef][Medline]
19. Berggren PO, Larsson O 1994 Ca²⁺ and pancreatic B-cell function. Biochem Soc Trans 22:12–18[Medline]
20. Saperstein R, Vicario PP, Strout HV, Brady E, Slater EE, Greenlee WJ, Ondeyka DL, Patchett AA, Hangauer DG 1989 Design of a selective insulin receptor tyrosine kinase inhibitor and its effect on glucose uptake and metabolism in intact cells. Biochemistry 28:5694–5701[CrossRef][Medline]
21. Baltensperger K, Lewis RE, Woon CW, Vissavajjhala P, Ross AH, Czech MP 1992 Catalysis of serine and tyrosine autophosphorylation by the human insulin receptor. Proc Natl Acad Sci USA 89:7885–7889[Abstract/Free Full Text]
22. Leibiger B, Wahlander K, Berggren PO, Leibiger IB 2000 Glucose-stimulated insulin biosynthesis depends on insulin-stimulated insulin gene transcription. J Biol Chem 275:30153–30156[Abstract/Free Full Text]
23. Srivastava S, Goren HJ 2000 Insulin constitutively secreted by ß-cells is necessary for glucose-stimulated insulin secretion Diabetes 52:2049–2056
24. Krugmann S, Anderson KE, Ridley SH, Risso N, McGregor A, Coadwell J, Davidson K, Eguinoa A, Ellson CD, Lipp P, Manifava M, Ktistakis N, Painter G, Thuring JW, Cooper MA, Lim ZY, Holmes AB, Dove SK, Michell RH, Grewal A, Nazarian A, Erdjument-Bromage H, Tempst P, Stephens LR, Hawkins PT 2002 Identification of ARAP3, a novel PI3K effector regulating both Arf and Rho GTPases, by selective capture on phosphoinositide affinity matrices. Mol Cell 9:95–108[CrossRef][Medline]
25. Tolias KF, Rameh LE, Ishihara H, Shibasaki Y, Chen J, Prestwich GD, Cantley LC, Carpenter CL 1998 Type I phosphatidylinositol-4-phosphate 5-kinases synthesize the novel lipids phosphatidylinositol 3,5-bisphosphate and phosphatidylinositol 5-phosphate. J Biol Chem 273:18040–18046[Abstract/Free Full Text]
26. Dove SK, Cooke FT, Douglas MR, Sayers LG, Parker PJ, Michell RH 1997 Osmotic stress activates phosphatidylinositol-3,5-bisphosphate synthesis. Nature 390:187–192[CrossRef][Medline]
27. Cooke FT, Dove SK, McEwen RK, Painter G, Holmes AB, Hall MN, Michell RH, Parker PJ 1998 The stress-activated phosphatidylinositol 3-phosphate 5-kinase Fab1p is essential for vacuole function in S. cerevisiae. Curr Biol 8:1219–1222[CrossRef][Medline]
28. Sbrissa D, Ikonomov OC, Shisheva A 1999 PIKfyve, a mammalian ortholog of yeast Fab1p lipid kinase, synthesizes 5-phosphoinositides. Effect of insulin. J Biol Chem 274:21589–21597[Abstract/Free Full Text]
29. McEwen RK, Dove SK, Cooke FT, Painter GF, Holmes AB, Shisheva A, Ohya Y, Parker PJ, Michell RH 1999 Complementation analysis in PtdInsP kinase-deficient yeast mutants demonstrates that Schizosaccharomyces pombe and murine Fab1p homologues are phosphatidylinositol 3-phosphate 5-kinases. J Biol Chem 274:33905–33912[Abstract/Free Full Text]
30. Ikonomov OC, Sbrissa D, Mlak K, Kanzaki M, Pessin J, Shisheva A 2002 Functional dissection of lipid and protein kinase signals of PIKfyve reveals the role of PtdIns 3,5-P2 production for endomembrane integrity. J Biol Chem 277:9206–9211[Abstract/Free Full Text]
31. Dove SK, McEwen RK, Mayes A, Hughes DC, Beggs JD, Michell RH 2002 Vac14 controls PtdIns(3,5)P(2) synthesis and Fab1-dependent protein trafficking to the multivesicular body. Curr Biol 12:885–893[CrossRef][Medline]
32. Jones DR, Gonzalez-Garcia A, Diez E, Martinez-A C, Carrera AC, Merida I 1999 The identification of phosphatidylinositol 3,5-bisphosphate in T-lymphocytes and its regulation by interleukin-2. J Biol Chem 274:18407–18413[Abstract/Free Full Text]
33. Bonangelino CJ, Nau JJ, Duex JE, Brinkman M, Wurmser AE, Gary JD, Emr SD, Weisman LS 2002 Osmotic stress-induced increase of phosphatidylinositol 3,5-bisphosphate requires Vac14p, an activator of the lipid kinase Fab1p. J Cell Biol 156:1015–1028[Abstract/Free Full Text]
34. Shepherd PR, Withers DJ, Siddle K 1998 Phosphoinositide 3-kinase: the key switch mechanism in insulin signalling. Biochem J 333:471–490[Medline]
35. Ikonomov OC, Sbrissa D, Mlak K, Shisheva A 2002 Requirement for PIKfyve enzymatic activity in acute and long-term insulin cellular effects. Endocrinology 143:4742–4754[Abstract/Free Full Text]
36. Berwick DC, Dell GC, Welsh GI, Heesom KJ, Hers I, Fletcher LM, Cooke FT, Tavare JM 2004 Protein kinase B phosphorylation of PIKfyve regulates the trafficking of GLUT4 vesicles. J Cell Sci 117:5985–5993[Abstract/Free Full Text]
37. Kulkarni RN 2002 Receptors for insulin and insulin-like growth factor-1 and insulin receptor substrate-1 mediate pathways that regulate islet function. Biochem Soc Trans 30:317–322[CrossRef][Medline]
38. Ardizzone TD, Lu XH, Dwyer DS 2002 Calcium-independent inhibition of glucose transport in PC-12 and L6 cells by calcium channel antagonists. Am J Physiol Cell Physiol 283:C579–C586
39. Idevall Hagren O, Tengholm A 2006 Glucose and insulin synergistically activate PI3-kinase to trigger oscillations of phosphatidylinositol-3,4,5-trisphosphate in ß-cells. J Biol Chem 281:39121–39127[Abstract/Free Full Text]
40. Anderson KE, Stephens LR, Hawkins PT 1999 Phosphoinositide 3-kinases. In: Milligan G, ed. Signal transduction: a practical approach. 2nd ed. Oxford, UK: Oxford University Press; 283–300
41. Brearley CA, Hanke DE 1997 Assaying inositol phospholipid turnover in plant cells. In: Shears, SB, ed. Signaling by inositides: a practical approach. Oxford, UK: Oxford University Press; 1–32
42. Efanov AM, Zaitsev SV, Berggren PO 1997 Inositol hexakisphosphate stimulates non-Ca²⁺-mediated and primes Ca²⁺-mediated exocytosis of insulin by activation of protein kinase C. Proc Natl Acad Sci USA 94:4435–4439[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) All Versions of this Article: 21/11/2775    most recent Author Manuscript (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Yu, J. Articles by Barker, C. J. Search for Related Content PubMed PubMed Citation Articles by Yu, J. Articles by Barker, C. J.