Leptin Effects on Pancreatic β-Cell Gene Expression and Function

The idea that leptin may directly act on pancreatic β-cells to control insulin secretion was derived from observations in genetically leptin-deficient ob/ob mice and leptin receptor–defective db/db mice. First, it was noted that defective leptin signaling in these mice strains leads to initial hyperinsulinemia in young animals, even before the development of the obese and diabetic phenotype (13–15). Second, this hyperinsulinemia was ameliorated by treatment of the ob/ob mice with recombinant leptin (2), suggesting a direct inhibitory action of leptin on insulin secretion from pancreatic β-cells.

Already at this stage, the idea was postulated that leptin may represent one part of a bidirectional feedback loop between adipose tissue and the pancreatic β-cell, with the other part consisting of insulin secreted from the pancreas and stimulating in turn leptin biosynthesis and secretion from adipose tissue. The adipogenic actions of insulin are well known, and stimulatory effects of insulin on leptin production in adipocytes were convincingly demonstrated (16,17).

Although the short isoform of the receptor (ObRa) is capable of signaling, the long isoform (ObRb) is currently believed to convey most of the physiological actions of leptin. As the first step to support the hypothesis of direct actions of leptin on pancreatic β-cells, the presence of ObRb on insulin-producing cells was shown. Messenger RNA for ObRb was found in high abundance in murine pancreatic islets but was not detected in the exocrine pancreas (18). Leptin receptor mRNA was detected in rat islets, even in greater abundance than that found in total brain, and was also detected in the pancreatic β-cell line βTC3 by RT-PCR. The presence of functional leptin receptors including ObRb on insulin-producing cells has since been confirmed in other studies and could also be demonstrated in β-cells derived from human pancreatic islets (19).

The main function of the pancreatic β-cell is the biosynthesis and adequate secretion of insulin to control blood glucose levels. This function is tightly regulated both by nutrients and hormonal modulators, such as enteroendocrine hormones (glucose-dependent insulinotropic polypeptide and glucagon-like peptide [GLP]-1).

The effects of leptin on insulin secretion have initially yielded conflicting results. In contrast to the pancreatic islets of leptin-deficient ob/ob mice, which exhibit a robust reduction of insulin secretion upon leptin stimulation (probably because of higher leptin sensitivity), chronically leptin-exposed pancreatic β-cells display variable results in different experimental conditions. In pancreatic β-cells, leptin exerts a biphasic dose response with respect to insulin secretion. In rat islets, 2 nmol/l leptin significantly and maximally suppressed insulin release, whereas high concentrations were almost ineffective (20,21). Incubation of pancreatic islets isolated from ob/ob mice with 6.25–100 nmol/l leptin for 1–2 h robustly inhibits insulin secretion, which was also demonstrated for the perfused pancreas of ob/ob mice (18,22). This effect of leptin on insulin release was smaller at higher glucose concentrations and could not be demonstrated when the enteroendocrine hormone GLP-1 was present (22,23), indicating that the more tonic inhibitory leptin effects on insulin secretion could be overcome by acute nutrient- and incretin-stimulated insulin release. Leptin also inhibits insulin release from ob/ob islets stimulated by activators of protein kinase C, such as acetylcholine or phorbol ester phorbol-12-myristate 13-acetate (23). In contrast to β-cells of leptin-deficient ob/ob mice, initial studies examining the effect of leptin on insulin secretion in normal rodents or cell lines yielded varying results. However, currently it is accepted that under physiological conditions leptin (1–20 nmol/l) significantly reduces insulin release from pancreatic β-cells. This has now been demonstrated by the majority of studies using perfused rat pancreas and isolated rat or mouse islets (19–41). Moreover, these observations derived from rodent β-cell lines or pancreatic islets are now confirmed in isolated human pancreatic islets (19,31,42,43), and there is recent evidence that a physiological increase in serum leptin levels significantly reduces insulin secretion in rats in vivo (44). The fact that leptin is able to inhibit insulin secretion from human islets at concentrations as low as 0.01 nmol/l implies that this regulatory loop between adipose tissue and the endocrine pancreas is also physiologically effective in humans.

The first step of insulin biosynthesis is proinsulin gene expression. It has been shown that the rate-limiting step determining steady-state proinsulin mRNA levels is via transcriptional regulation of the proinsulin gene promoter (45).

The majority of studies report on suppression of preproinsulin mRNA in pancreatic β-cells by leptin. Thereby, leptin has been shown to suppress preproinsulin mRNA expression in mouse βTC6 cells (31) in the rat pancreatic β-cell line INS-1 (38), in isolated primary rat islets (20,31), in ob/ob mouse islets, and in human islets (19). Intraperitoneal injection of ob/ob mice with 1 μg/g body wt murine leptin reduced steady-state levels of preproinsulin mRNA in isolated islets by 40% after 24 h (38). In studies of human islets, human leptin (6.25 nmol/l) evokes a time-dependent decrease in preproinsulin mRNA levels in the presence of 11.1 mmol/l glucose but not 5.6 mmol/l glucose (19). When the incretin hormone GLP-1, a known stimulator of the proinsulin gene promoter was used, leptin inhibited GLP-1–stimulated expression of preproinsulin mRNA in human islets at glucose concentrations of 5.6 and 11.1 mmol/l (38). In contrast, in INS-1 β-cells, leptin significantly reduced only preproinsulin mRNA expression stimulated by 25 mmol/l glucose, but not by lower glucose concentrations. These observations imply that the ability of leptin to reduce preproinsulin mRNA expression in pancreatic β-cells may depend on prior stimulation by incretin hormones or stimulatory ambient glucose concentrations. Further, the effects of leptin on steady-state preproinsulin mRNA levels in pancreatic β-cells were only observed after 16 h of incubation and were not seen at shorter incubation periods (38). This observation suggests that the effect of leptin on insulin biosynthesis may represent a more long-term character, and time kinetics imply that gene transcription may be necessary for this effect.

The effect of leptin on the transcriptional activity of the insulin gene promoter has been examined in a consecutive step. A dose of 6.25 nmol/l leptin inhibited a reporter vector expressing the luciferase gene under the control of 410 bp of the rat insulin I gene promoter in INS-1 cells at stimulatory glucose concentrations of 25 mmol/l but not at 5.6 mmol/l glucose (38). In contrast, the induction of transcriptional activity of the insulin promoter by additional stimulatory concentrations of 10 nmol/l GLP-1 at 11.1 mmol/l glucose was also inhibited by leptin. These findings suggest that stimulated insulin promoter activity by either GLP-1 or high glucose (25 mmol/l) represents a prerequisite for the inhibitory actions of leptin on insulin promoter activity (38).

Leptin signaling through ObR is intracellularly coupled with the janus kinase (JAK)–signal transducer and activator of transcription (STAT) pathway. Binding of leptin to the receptor (ObR) activates the receptor-associated kinase JAK2 via transphosphorylation and phosphorylates tyrosine residues on ObRb. Consecutively, transcription factors of the STAT family are recruited to the receptor and also phosphorylated. Phosphorylated STATs dimerize and translocate to the nucleus to regulate gene transcription. Also in pancreatic β-cells, we have found that repression of insulin promoter activity by leptin was associated with altered binding of the isoform STAT5b to specific DNA sequences within the promoter (38). STATs have mostly been shown to transcriptionally enhance gene expression. In contrast, we have demonstrated that leptin signaling, which activates several STATs in other tissues such as the hypothalamus (46,47), inhibits insulin biosynthesis via transcriptional repression of the proinsulin gene promoter (19,38). In a follow-up study, we sought to characterize this apparent contradiction at the molecular level.

Recently, a new family of molecules, called suppressor of cytokine signaling (SOCS), has been identified. This family of molecules is able to inhibit JAK-STAT signal transduction. These molecules contain a central src homology region 2 domain and a conserved COOH-terminal SOCS box. By their central SH2 domain, they bind directly to tyrosine-phosphorylated residues on the cytokine receptor–associated kinase JAK2. Expression of the SOCS proteins is induced by various cytokines, including interleukin-6, leukemia inhibitory factor, erythropoietin, and growth hormone in a tissue-specific manner, and once expressed, they in turn inhibit cytokine signaling via the JAK-STAT pathway in an intracellular negative feedback loop. Thus, SOCS molecules may function as cytokine inducible negative regulators of cytokine signaling. In the hypothalamus, it has been demonstrated that leptin induces expression of SOCS3 mRNA in areas where ObRb is expressed. Thus, the SOCS molecules may play an important role in the development of leptin resistance that is seen in obesity both in the central nervous system and the endocrine pancreatic β-cell. We examined leptin-mediated signal transduction and gene regulation in insulin-producing pancreatic β-cells more in detail (K. Laubner, T. Kieffer, S. Royer, J. Roller, F. Jacob, J.S., unpublished data).

In brief, by coimmunoprecipitation and Western blot analysis, we found that leptin stimulation (6.25 nmol/l) leads to recruitment and phosphorylation of STAT3 and STAT5b to the receptor-associated JAK2 in INS-1 β-cells. Further, we demonstrated time-dependent nuclear translocation of STAT3 and STAT5b in INS-1 pancreatic β-cells after leptin stimulation by fluorescence immunocytochemistry and transient transfection of expression vectors encoding the STAT5b green-fluorescent protein fusion protein. Moreover, leptin signaling in INS-1 cells and isolated pancreatic islets of ob/ob mice induces expression of SOCS3 at the transcriptional level in vitro and in vivo, as revealed by luciferase reporter gene assay and RT-PCR. In electrophoretic mobility shift assays using potential STAT-binding sites within the rat SOCS3 promoter, we found leptin-induced STAT3 DNA binding in INS-1 nuclear extracts. Whereas STAT5b alone transactivates the rat insulin 1 promoter in luciferase reporter gene assays in INS-1 cells, leptin-induced SOCS3 expression inhibits STAT5b-dependent transactivation of the rat insulin 1 promoter. In conclusion, we have acquired evidence that, in pancreatic β-cells, the gene regulatory effects of leptin not only involve transcriptional inhibition of the insulin gene promoter but also transcriptional activation of the promoter of the JAK-STAT inhibitory molecule SOCS3 by transcription factors of the STAT family, thereby constituting a regulatory feedback loop, which in turn inhibits leptin-dependent JAK-STAT activation. Leptin-induced SOCS3 expression may thus mediate the long-term inhibitory effect of leptin action in pancreatic β-cells that is seen in the context of endocrine signaling between adipose tissue and the endocrine pancreas within the adipoinsular axis (Fig. 1).

As described above, leptin exerts bifunctional gene regulatory effects on the proinsulin and SOCS3 genes in pancreatic β-cells. Thus, we hypothesized that additional genes may be regulated by leptin in insulin-producing cells. Using a subtractive PCR approach with cDNA pools derived from leptin and vehicle-treated INS-1 β-cells, we identified the catalytic subunit of protein phosphatase 1 (PP-1) as a leptin-regulated gene in pancreatic β-cells (P. Kühen, K. Laubner, J. Roller, F. Jakob, J.S., unpublished data). In brief, PP-1 has recently been characterized genetically as a candidate gene for type 2 diabetes (48,49). Further, serine/threonine phosphatase PP-1 is a key enzyme in the insulin-signaling cascade in muscle and liver. Through RT-PCR and fluorescence immunocytochemistry, we found coexpression of PP-1 and insulin in the pancreatic β-cell line INS-1 and in primary β-cells from human pancreatic islets. Moreover, RT-PCR and Northern and Western blot analysis demonstrated time-dependent inhibition of PP-1 mRNA and protein expression by leptin (100 nmol/l) within 48 h in pancreatic β-cells. Using a PP-1–specific assay, leptin reduced functional PP-1 enzyme activity by 64% in pancreatic β-cells, an effect that could be mimicked by the PP-1–specific inhibitors calyculin A and okadaic acid. Finally, leptin-, calyculin A–, or okadaic acid–mediated reduction of PP-1 activity leads to profound inhibition of both glucose- and glucagon-induced insulin secretion in pancreatic β-cells (Fig. 2).

Taken together, leptin exerts gene regulatory effects on multiple genes within insulin-producing pancreatic β-cells. Current knowledge, however, suggests that all gene regulatory effects concomitantly culminate to physiologically reduce insulin secretion and biosynthesis.

Insulin secretion from pancreatic β-cells relies in part on the activity of ATP-dependent K⁺ (K_ATP) channels. Closure (inactivation) of K_ATP channels in response to glucose or other insulin secretagogues depolarizes β-cells, resulting in the activation of voltage-dependent Ca²⁺ channels, a rise in cytosolic calcium concentration, and a rise in insulin secretion (50). Glucose-induced insulin secretion is further potentiated by hormone-mediated elevation of the intracellular second messengers cAMP/protein kinase A and phospholipase C/protein kinase C. Electrophysiological characterization of the effects of leptin on pancreatic β-cells revealed that leptin hyperpolarized the cell membrane, which led to inhibition of insulin secretion (22,27). This hyperpolarization has been demonstrated to be due to an increase in membrane conductance caused by the opening (activation) of K_ATP channels. Thus, the K_ATP channel is a molecular target of leptin in pancreatic β-cells for inhibition of insulin secretion. Consequently, activation of K_ATP channels in pancreatic β-cells by leptin reduces cytosolic calcium concentration, and this fall can be overcome by coincubation with 20 mmol/l glucose and GLP-1 (10 nmol/l). Although the molecular mechanism by which leptin activates K_ATP channels is not fully understood, studies indicate that phosphorylation or dephosphorylation of proteins may be involved (51), and treatment of insulinoma HIT-T15 cells with murine leptin results in a threefold activation of phosphatidylinositol (PI) 3-kinase (21). Phosphodiesterase 3B, which reduces the cellular content of cAMP, is also activated by leptin, and leptin (1–5 nmol/l) suppressed the elevation of cAMP induced by GLP-1 in HIT-T15 cells (5 nmol/ l) (21). These findings suggest that the inhibitory actions of leptin on insulin secretion are primarily mediated through the PI 3-kinase–dependent activation of phosphodiesterase 3B and a subsequent reduction of intracellular cAMP.

The inhibitory effects of leptin on preproinsulin gene expression, however, appear to be independent of the activation of K_ATP channels. The K_ATP channel opener diazoxide did not affect both leptin suppression of preproinsulin mRNA levels and inhibition of insulin promoter activity in INS-1 cells (38), indicating that gene regulatory effects of leptin use signal transduction pathways different from those that mediate the effect on insulin secretion.

The different functional, molecular, and gene regulatory effects of leptin involving insulin secretion (8,18–28,30,31,33–40,42–44,51–61), pancreatic β-cell gene expression (19,20,31,38,62,63), and signal transduction (21,22,27,38,51,54,64) are summarized in Table 1.

A growing body of evidence has been accumulated showing that the adipose tissue–derived hormone leptin directly acts on pancreatic β-cells in addition to its effects in the hypothalamus to reduce food intake and increase energy expenditure. At the cellular level, inhibitory effects of leptin on both insulin secretion and insulin biosynthesis, mainly represented by the inhibition of preproinsulin gene expression, have been demonstrated. In this context, leptin affects different signal transduction pathways and molecular targets.

Taken together, all effects of leptin in pancreatic β-cells that are known to date act concomitantly to exert a physiological long-term control of insulin secretion from the pancreatic β-cell, which adapts the amount of insulin secretion to the amount of body fat stores. It is important to note that this tonic restriction of insulin secretion by leptin generally does not seem to interfere with the short-term stimulatory actions of nutrients and hormones, such as glucose- and incretin-dependent insulin secretion.

Leptin seems to represent a signaling molecule from the adipose tissue to the endocrine pancreas to restrict insulin secretion according to the needs that are determined by body fat stores. As such, this effect establishes an “adipo-insular” transmission of signals from the adipocyte to the pancreatic β-cell. Vice versa, the lipogenic action of insulin has long been well known, and, consequently, it has been demonstrated that insulin stimulates both leptin biosynthesis and secretion from white adipose tissue. This establishes a classic endocrine adipo-insular feedback loop—the so-called “adipo-insular axis” (Fig. 3A). To date, most evidence to support this concept has come from experimental animal models and in vitro studies. A growing body of arguments, however, indicates that the adipo-insular axis also physiologically regulates insulin secretion and leptin production in humans (65). Leptin effects detected so far on pancreatic β-cells are generally delayed and involve more long-term cellular events. In contrast, acute stimulation of insulin secretion by nutrients and postprandially secreted hormones overcame most of the leptin effects on insulin secretion, but not on gene regulation, in the majority of studies. Thus, it is tempting to speculate that the inhibitory effect of leptin on insulin secretion within the adipo-insular axis is designed to adapt the set point of pancreatic insulin release to the body fat stores during the fasting state.

The adipo-insular axis may, however, play an important role during the development of type 2 diabetes in obese patients. During the development of this disease, initial hyperinsulinemia is believed to represent a simple compensatory response of the pancreatic β-cell to insulin resistance (66,67), and hyperglycemia is the consequence of pancreatic β-cell failure. However, in recent studies, it became obvious that hyperinsulinemia frequently precedes the development of insulin resistance, arguing at least in part for the existence of an early functional defect in insulin secretion and against obesity-induced insulin resistance fully explaining the development of type 2 diabetes.

In obese animal models and also in most obese patients, it has been shown that despite high levels of circulating leptin according to the increased fat mass, leptin seems to fail exerting its effect on the hypothalamus (i.e., reducing food intake and increasing energy expenditure). This observation has been termed “leptin resistance” and was attributed to several molecular alterations in postreceptor leptin signal transduction in the hypothalamus (68). Given the almost identical cellular targets and signal transduction pathways that are used by leptin in both the hypothalamus and the pancreatic β-cell, one may postulate the existence of leptin resistance also at the level of the pancreatic islet (8,69). Reduced leptin sensitivity of the pancreatic β-cell, however, leads to dysregulation of the adipo-insular axis, resulting in increased insulin release, which is then no longer under controlled repression by leptin. Thereby, leptin resistance at the level of the pancreatic β-cell may promote hyperinsulinemia in obese patients prone to developing type 2 diabetes (Fig. 3B). Elevated insulin concentrations promote both insulin resistance and further stimulation of leptin production and secretion from the adipose tissue, which may in turn enhance leptin resistance of the endocrine pancreas by further desensitizing leptin signal transduction pathways and constituting a vicious circle that promotes manifestation of type 2 diabetes in obese people (Fig. 3B).

Thus, identification of the molecular determinants of the adipo-insular axis and especially leptin resistance in pancreatic β-cells may provide novel targets for the development of therapeutic strategies. This process may prevent the early steps in the pathogenesis of type 2 diabetes and ultimately the manifestation of the disease in susceptible overweight patients.

The studies were in part supported by the Deutsche Forschungsgemeinschaft.

We thank K. Laubner and P. Kuehnen for providing results of ongoing studies and S. Nixdorf and J. Roller for expert technical assistance. We are grateful to Dr. R. Bretzel and Dr. M. Brendel, Islet Transplant Center, University of Giessen, Giessen, Germany, for providing the isolated human pancreatic islets. Fruitful collaboration in ongoing studies with Dr. T.J. Kieffer, University of British Columbia, Vancouver, Canada, is gratefully acknowledged.