Glucose regulation of insulin gene transcription factors
Pdx-1 is mainly localized at the nuclear periphery at low glucose concentrations (1–2mM) in β cells, and becomes phosphorylated and shuttles into the nucleus in response to concentrations stimulating insulin secretion. Many different signaling pathways have been shown to regulate the nucleo-cytoplasmic shuttling and transactivation potential of Pdx-1 under these conditions, including glycogen synthase kinase 3 (GSK3), p38/stress-activated protein kinase, phosphatidylinositol 3-kinase (PI3K), atypical protein kinase C isoforms, mitogen-activated protein kinase (MAPK), and Per-Arnt-Sim (PAS) kinase . Glucose also appears to regulate the interaction of Pdx-1 with various transcriptional coregulators. Thus, Pdx-1 is associated with the histone deacetylases HDAC-1 and HDAC-2 to downregulate insulin gene expression under low, non-insulin stimulating glucose concentrations . However, the HDAC-1/2 interaction is prevented at elevated glucose levels, which promotes association with the histone acetyltransferase p300, hyperacetylation of histone H4, and insulin gene transcription . In addition, SUMOylation of Pdx-1 increases its nuclear localization as well as its protein stability and is correlated with an increase in insulin promoter activity . Finally, Pdx-1 contains at least two sites of O-GlcNAcylation that increase its DNA binding activity [31,32].
In addition, glucose-induced phosphorylation of NeuroD1 regulates its nuclear localization and transactivation. However, precisely how the regulation is imposed is unknown. Treatment of pancreatic β cells with the MAPK/ERK kinase inhibitor PD98059 blocks nuclear NeuroD1 translocation, conditions that also impact Pdx-1 phosphorylation and insulin gene transcription . In addition, high glucose levels induce O-GlcNAcylation of NeuroD1 , which appears to be important for its translocation into the nucleus. Exactly how these distinct modifications control NeuroD1 activity in islet β cells in vivo still needs to be explored.
Increased MafA protein and DNA activity regulates glucose-dependent insulin gene transcription, with phosphorylation modulating insulin enhancer binding . In addition, the transactivation potential of MafA is potentiated by GSK3-mediated phosphorylation within the N-terminal region which allows recruitment of the P/CAF coactivator . However, precisely how phosphorylation regulates these properties of MafA remains to be determined.
Glucose regulation of insulin mRNA stability
In addition to its major effects on transcription of the insulin promoter, glucose markedly stabilizes preproinsulin mRNA. Indeed, the half-life of the message has been estimated to increase from 29 hours to 77 hours when switched from low to high glucose. Two elements located in the 3′-untranslated region of the mRNA molecule have been proposed as mediators of the glucose-stabilizing effect, the conserved UUGAA sequence and the pyrimidine-rich sequence (insPRS) . Stabilization appears to involve binding of a polypyrimidine tract-binding (PTB) protein to the insPRS, as binding was induced by glucose and prevented upon mutating the core PTB binding site.
Glucose regulation of the insulin gene in vivo
Variations in the amount of insulin mRNA at any given time represent the net effect of metabolic, hormonal, and neuronal stimuli on insulin gene transcription and mRNA stability. From the in vitro effects described earlier, it is predicted that increases in blood glucose should rapidly elevate preproinsulin mRNA levels in the endocrine pancreas (due to rapid stimulation of transcription), whereas a decrease in blood glucose would be followed by a slow decline in preproinsulin mRNA levels (due to reduced transcription and the long half-life of the message). Indeed, early studies in rats showed that 4 days of starvation are necessary to detect significant decreases in preproinsulin mRNA, which returned to basal values only 6 hours after refeeding  or 12 hours after glucose injection . The delay in the disappearance of the message upon fasting appears to be directly due to the stability of the mRNA, as insulin-induced hypoglycemia is followed by a rapid (2-hour) decrease in the level of precursors for insulin mRNA .
An additional level of complexity in the regulation of insulin gene expression in vivo is the potential interaction between β cells and non-β (i.e., α, δ, and PP) cells within the intact islet. Indeed, it is known that individual β cells behave differently from one another in terms of insulin secretion than in the context of the islet , a phenomenon which has also been demonstrated at the level of insulin gene transcription . Thus, the global response measured in an entire islet represents the integration of the individual responses. This fact should be kept in mind when interpreting experiments performed in clonal cell lines, which lack the level of regulation provided by islet architecture and the neighboring non-β cells.
Regulation of insulin gene expression by glucagon-like peptide-1 (GLP-1)
GLP-1 is an incretin hormone which strongly potentiates glucose-induced insulin secretion and is the target of a number of type 2 diabetes drugs . Most of the biologic effects of GLP-1 are mediated by its G protein-coupled receptor expressed on the β-cell surface . GLP-1 stimulates insulin gene expression by various mechanisms. First, it directly activates the cyclic-AMP (cAMP) response element (CRE) within the 5′-proximal control sequences of the insulin gene, by a mechanism which seems at least partly independent from protein kinase A activation . Second, it can augment the glucose-stimulated binding activity of Pdx-1 . Third, it can stimulate transcription of the PDX-1 gene (the promoter of which also contains a CRE) . Fourth, GLP-1 potentiates glucose-induced insulin gene transcription by activating NFAT (of which there are three binding sites on the rat insulin I promoter) via calcium/calmodulin-dependent protein phosphatase 2B (calcineurin) activation in response to a rise in intracellular calcium .
Physiologic inhibitors of insulin gene expression
Several physiologic inhibitors of insulin secretion also impair expression. Epinephrine and somatostatin, two hormones acting through G protein-coupled receptors and known inhibitors of insulin release , also inhibit the rate of insulin gene transcription . In addition, somatostatin decreases the stability of insulin mRNA . Glucagon, a key hormone in the counterregulatory response to hypoglycemia in vivo, stimulates expression of the inducible cAMP early repressor, which inhibits insulin gene transcription . Finally, the adipocyte-secreted hormone leptin decreases insulin gene expression . Thus, physiologic modulators of insulin secretion coordinately inhibit insulin gene expression, thereby ensuring that the long-term biosynthetic rate of insulin matches the secretory demand.
Short-term regulation of insulin gene expression by insulin
Whether insulin has a functionally relevant role in the regulation of insulin transcription and biosynthesis remains a debated issue [54,55]. Based on the observation that a rapid (within minutes) transcriptional response to glucose in insulin-secreting cells was mimicked by depolarizing agents and exogenous insulin and was suppressed by inhibiting insulin release or the PI3 kinase pathway, it was proposed that insulin acts in an autocrine manner to stimulate insulin gene transcription by binding to the insulin receptor . However, subsequent studies have failed to detect significant changes in preproinsulin mRNA levels upon short-term exposure to glucose and the physiologic relevance of an autocrine positive feedback on the β cell has been questioned .