E2F transcription factors are often depicted as the ultimate effectors of the G1/S transition of the cell cycle. When bound to DNA, they exist either as free E2F/DP heterodimers, or associated in larger complexes containing members of the retinoblastoma family (pRB, p107, p130) and members of the cyclin/cdk protein families. Cyclin-dependent kinases (CDK) define a family of serine/threonine protein kinases that phosphorylate a number of substrates mainly implicated in cell cycle progression and transcription. Association of E2Fs with proteins of the pRB family facilitates active repression through recruitment of histone deacetylases 4 and lysine/arginine methyl-transferases 5 . Subsequent phosphorylation of the retinoblastoma protein by the cyclin/CDK complexes results in the release of the E2F transcription factors and activation of the transcription of genes required for progression through G1 into the S phase of the cell cycle. CDKs ultimately translate external signaling into transcriptional response, which is the final step of the regulatory cascade 6 . This makes the CDK-pRB-E2Fs potential candidates for the control of the crosstalk between proliferative stimuli and metabolic response. E2Fs functions are not restricted to the control of cell proliferation and it is now accepted that E2Fs participate in cellular processes beyond the cell cycle 7 . This has been extensively demonstrated through the characterization of E2fs-null mouse models 8 and the number of genes identified as E2F targets encoding factors involved in DNA replication, mitosis, DNA repair, apoptosis, differentiation, and development (reviewed in 9 ). Of particular interest is the observation that E2Fs, pRB, and cdk4 could be involved in the regulation of a metabolic network, including the control of adipose tissue function 10 11 .

Proliferation of β-cells is a key mechanism to maintain postnatal β-cell mass 12 13 , and is the primary mechanism for β-cell regeneration 14 15 16 . It is now clear that pancreatic β-cells are able to replicate, albeit the origin of newly formed islets remains controversial. However, the mechanisms controlling β-cell replication unequivocally require cell cycle regulation. Gene inactivation studies of cell cycle regulators underscore the important role for these proteins in maintaining the β-cell population. Together with our previous observation that E2f1 -/- mice have decreased β-cell mass, due to decreased proliferation 17 , other cell cycle regulators have been implicated in this process (for review, see 18 . Similar to E2f1-/-, Cyclin D1+/-; Cyclin D2-/- 19 , Cyclin D2 -/- 15 , and Cdk4 -/- 20 mice have decreased β-cell mass. Conversely, mice with gene inactivation of the cell cycle inhibitors p27 21 22 , p16^INK4A 23 have increased islet mass. Interestingly, genetic rescue of a mutant Cdk4 insensitive to INK4 inhibition (CDK4^R24C) in a Cdk4-null background results in normal β-cell mass, suggesting a key role for CDK4 in controlling proliferation of β-cell 24 .

We previously evaluated the effects of E2F1 on glucose homeostasis. E2f1 -/- mice show an overall reduction in pancreatic size, as the result of impaired postnatal pancreatic growth 17 . E2F1 is not involved in embryonic β-cell development, since E2F1 is expressed at the final step of pancreas development, as demonstrated by in situ hybridizations. Because of the disproportionate small pancreas E2f1 -/- β-cells secrete insufficient amounts of insulin in response to a glucose load, resulting in glucose intolerance. However, despite this decreased insulin secretion, these animals were protected for the development of diabetes because of a dramatic increase in insulin sensitivity in other tissues, mainly adipose tissue 25 . In contrast, E2f1/E2f2 double KO mice are diabetic, showing exocrine pancreatic insufficiency, indicating that E2fs gene dosage and/or tissue specificity are critical for glucose homeostasis and pancreatic function 26 27 . These observations suggest however that E2F1 and E2F2 might cooperate to regulate β-cell mass and functions through cell-autonomous 3 17 and non-cell autonomous 26 28 effects. Surprisingly, we observed that E2F1 was highly expressed in non-proliferating adult pancreatic β-cells 17 , suggesting that besides its impact on β-cell number, E2F1 could also play a role in pancreatic β-cell function and insulin secretion.

Pancreatic β-cell K_ATP channels are composed by the heteromeric association of the sulfonylurea receptor 1 (SUR1) with the K⁺ ATP channel inward rectifier (KIR6.2). These channels play a critical role in the regulation of glucose-induced insulin secretion by controlling membrane polarization 29 30 . The observation that Kir6.2 -/- mice display defective insulin secretion and increased insulin sensitivity is interesting 31 , since it strongly resembles a particular phenotype of E2f1-/- mice 17 . Global gene expression analysis of E2f1+/+ and -/- pancreas revealed that the mRNA and protein levels of Kir6.2 were decreased. Most importantly, we demonstrated that Kir6.2 is a bona fide E2F1 target gene, as observed by chromatin immunoprecipitation performed on pancreas and isolated islet tissues. Most importantly, rescue of Kir6.2 expression in E2f1 -/- isolated islets restored glucose-stimulated insulin secretion, confirming that E2F1 exerts cell-autonomous effects on insulin secretion through its transcriptional control of Kir6.2 expression 3 . Furthermore, since Kir6.2 transcription is positively regulated by glucose, we studied the impact of glucose and glucose-stimulated insulin in the control of CDK4 activity and subsequent E2F1-dependent transcription. Our data demonstrated that glucose-stimulated insulin secretion triggers CDK4-dependent pRb phosphorylation in the β-cell, and subsequent transcriptional activation of the Kir6.2 gene by E2F1/DP-1 heterodimers. Finally, chemical invalidation of CDK4 activity, through the use of a specific inhibitor, resulted in glucose intolerance and impaired glucose-stimulated insulin secretion in mice, confirming a role for the CDK4-pRB-E2F1 pathway in insulin secretion process. Altogether, our data provide evidence that the CDK4-pRB-E2F1 pathway directly contributes to insulin secretion through the regulation of Kir6.2 expression in pancreatic β-cells both in vitro and in vivo. Our results reveal a novel feed-back control mechanism of the insulin response.