mTOR is a serine/threonine kinase that is conserved through evolution and is essential for cell growth and proliferation.¹¹ mTOR constitutes the catalytic unit of two distinct complexes, mTORC1 and 2,^12,13 which both contain mTOR, mLST8 (mammalian lethal with SEC13 protein)/GβL (G-protein β-protein subunit-like) and Deptor (DEP domain-containing mTOR-interacting protein) (Fig. 1). mTORC1 constitutes the rapamycin-sensitive arm of mTOR signaling and contains Raptor (regulatory associated protein of mTOR) and PRAS40 (proline-rich AKT substrate of 40 kDa).^14,15 mTORC2 is insensitive to rapamycin and also contains Rictor (rapamycin in-sensitive companion of mTOR), mSIN1 (mammalian stress-activated MAPK-interacting protein 1) and Protor 1/2 (protein observed with rictor 1/2).^12,13 mTORC2 phosphorylates AKT on serine 473, suggesting that this pathway could be indirectly linked to proliferation.¹⁶ mTORC1 activity is modulated by multiple upstream signals, which include growth factors, amino acids, cytokines, energy and cellular stressors. mTORC1 is negatively regulated by TSC1 and TSC2 and the small G protein RHEB (Ras homolog enriched in brain) (Fig. 1). The major function of TSC1 is to stabilize TSC2 and prevent its ubiquitin-mediated degradation.^17,18 TSC2 phosphorylation and inactivation by AKT and ERK, among others, releases the inhibition of RHEB, leading to activation of mTOR.^19-24 In contrast, phosphorylation and activation of TSC2 by AMPK and GSK3β inhibits mTOR signaling.^23,25 Although amino acids are important regulators of mTORC1 activity, the mechanisms responsible for this activation are not as well-characterized as the regulation by insulin or growth factors. RAG-GTPases have been identified as important regulators of this process by transporting mTORC1 to specific cellular compartments and facilitating the ability of RHEB-GTP to activate mTORC1.^26,27 Ste20-related kinase, MAP4K3 and the class III PI3K, mVPS34, have also been shown to link amino acid sensing to mTORC1.^28-30 Stress and hypoxia act to limit growth and protein synthesis by mechanisms that are incompletely understood but involve HIF-1 and REDD1.^31,32 Finally, reduction in ATP levels in conditions of energy stress suppresses mTORC1 signaling via several mechanisms, including both AMPK-dependent and independent pathways.^23,33,34

Activation of mTORC1 signaling leads to translation initiation, which includes the synthesis of secreted proteins. mTORC1 controls growth (cell size) and proliferation (cell number) by modulating mRNA translation through phosphorylation of 4E-BPs and the S6 kinases (Fig. 1).^35-38 The mTORC1-S6K arm regulates multiple biological processes, including cell growth/size, gene transcription, cell survival, synaptic plasticity and adipocyte differentiation.^39,40 S6K promotes mRNA translation initiation and other steps of protein synthesis by phosphorylating an array of substrates, including ribosomal S6 protein, a component of the 40S ribosome, and eIF4B,^41,42 PDCD4 (programmed cell death 4), eEF2K,⁴³ SKAR (S6K1 Aly/REF-like substrate),⁴⁴ CCTβ,⁴⁵ and CBP80.⁴⁵ However, chronic activation of S6K negatively regulates IRS1, therefore inactivating the PI3K/AKT signaling pathway.^1-4,47 Phosphorylation of the 4E-BPs triggers the release of eIF4E, initiating cap-dependent translation.^48,49 This arm of mTORC1 signaling has been linked to induction of cell proliferation independent of the mTORC1-S6K1 arm.⁵⁰ Together these substrates downstream of mTORC1 play an integral role in mRNA translation initiation and progression, thus controlling the rate of protein synthesis.

Rapamycin treatment has been shown to have mixed effects on insulin secretion, depending on the experimental conditions.^56-59 The role of mTORC1 in β-cell proliferation using rapamycin has been explored in multiple studies, which have resulted in five important observations. (1) Rapamycin treatment blocks β-cell expansion, cell size and proliferation induced by an activation of AKT in β cells.⁶⁰ (2) Inhibition of β-cell proliferation by rapamycin results from alterations in cyclin D2 and D3 levels and Cdk4 activity. (3) Rapamycin modulates cyclin D2 synthesis and stability, suggesting that mTORC1 regulates post-transcriptional modifications of cell cycle components in β cells. (4) Rapamycin also ameliorates β-cell expansion in a model of insulin resistance, suggesting that mTORC1 coordinates β-cell adaptation to hyperglycemia and type 2 diabetes.⁹ (5) Rapamycin treatment results in reduced proliferation in β cells in pregnant mice and causes anti-proliferative effects in transplanted rat β cells in vivo.^61,62 Together, these studies suggest a role of mTORC1 in the regulation of β-cell mass and cell cycle and imply that mTORC1 can mediate adaptation of β cells to insulin resistance. However, it is important to note that in vivo studies designed to reveal the importance of mTORC1 using rapamycin are less than ideal because of modulation of insulin sensitivity.^63,64 In addition, it is now accepted that a major limitation of rapamycin for in vivo and in vitro studies is the inhibition of mTORC1 activity toward only a subset of mTORC1 substrates.^8,9,65-67 Therefore, inhibiting mTORC1 signaling in vivo by genetic deletion of Raptor in β cells will provide new insights into the role of this pathway in nutrient- and growth factor-dependent regulation of β-cell proliferation.

The role of mTORC1 in β cells has been explored using animal models with gain of mTORC1 function. Mice with conditional deletion of Tsc2 in β cells (βTsc2^-/-) exhibit lower glucose levels, hyperinsulinemia and improved glucose tolerance. These changes are explained by increases in β-cell mass, proliferation and cell size.⁷ A separate study demonstrated that conditional Tsc2 deletion in β cells exhibited a similar phenotype, but these mice developed diabetes and β-cell failure after 40 weeks.⁸ The differences in the phenotype between these reports are most likely explained by different genetic backgrounds and the RIP-Cre line (hypothalamic expression) used. Further confirmation of the effect of mTORC1 activation in increased β-cell mass comes from mice overexpressing Rheb in β cells.⁶⁸ While the effect on cell size has been regularly described in these models, it is important to note that the effect of mTORC1 on β-cell proliferation has not been consistently observed in these studies. It is possible that different experimental conditions could, in part, explain these differences. However, one important conclusion derived from these studies is that more research is needed to determine the mechanisms and signaling components downstream of mTORC1 that control β-cell mass, proliferation and function.

The TSC1/TSC2 complex acts as a functional inhibitor of mTOR. Studies in human, mice, flies and yeast strongly suggest that their gene products are interdependent, and that these proteins function primarily as a complex.^18,69-71 The role of Tsc1 in β cells was evaluated by deletion of Tsc1 specifically in β cells using RIP2-Cre mice.⁷² These mice presented a complex phenotype, with older mice displaying hyperphagia, obesity and diabetes due to the activation of mTorc1 in RIP2-Cre-expressing hypothalamic neurons.⁷³ However, evaluation of the β-cell phenotype prior to the onset of obesity and insulin resistance in 4-week-old mice showed that Tsc1-deficient mice exhibit augmented β-cell hypertrophy, increased insulin synthesis and release and consequent hyperinsulinemia. Interestingly, this biphasic phenotype observed is similar to that observed in mice with conditional deletion of Tsc2 in β cells using the same RIP2-Cre line.⁸ Therefore, it remains unclear if the disparate β-cell phenotypes observed following conditional deletion of Tsc1 vs. to Tsc2 are consequences of β-cell specific functions of these proteins or alterations resulting from the Cre line used.

To further examine the consequences of conditional deletion of Tsc1 in β cells, we crossed Tsc1 floxed mice with a RIP-Cre line that displays little hypothalamic expression (βTsc1^-/-).⁷⁴ Absence of Tsc1 expression in islets from βTsc1^-/- mice was demonstrated by immunostaining and immunoblotting (data not shown). Evaluation of glucose and insulin levels in 5- to 6-month-old βTsc1^-/- mice revealed lower random glucose and hyperinsulinemia (Fig. 2A and B). βTsc1^-/- mice also displayed improved glucose tolerance when compared with control (Fig. 2C), and this phenotype was also observed in older mice (10–12 mo, data not shown). In a parallel experiment, we set out to determine if concomitant deletion of Tsc2 in a Tsc1-deleted background has additional effects in β cells. βTsc1^-/-/Tsc2^‑/- (RIP-Cre Tsc1^f/f Tsc2^f/f) mice showed lower basal glucose levels and hyperinsulinemia compared with control mice, but the levels were not significantly different than those observed in βTsc1^-/- mice (Fig. 2A and B). Interestingly, glucose tolerance was also improved in βTsc1^-/-/Tsc2^-/- mice but not significantly different from βTsc1^-/- mice (Fig. 2C). Evaluation of β-cell morphometry demonstrated that βTsc1^-/- mice had increased in β-cell mass that was not significantly different than that observed in βTsc1^-/-/Tsc2^-/- mice (Fig. 2A and B). The results of these experiments demonstrate that deletion of Tsc1 in β cells has a positive impact on β-cell mass expansion, insulin levels and glucose tolerance that persists with aging. Moreover, these data suggest that deletion of both Tsc1 and Tsc2 does not have an additive effect on glucose metabolism and β-cell mass.

The members of the 4E-BP family include three paralogs (4E-BP 1, 2 and 3) with different tissue expression profiles. 4E-BP proteins are repressors of the translation initiation factor 4E (eIF4E) and therefore inhibit protein translation. Phosphorylation of 4E-BPs at multiple sites by mTORC1 leads to dissociation of eIF4E from 4E-BP, allowing elF4E binding to eIF4F and eIF4G. This complex formation promotes the translational machinery of mRNAs with high 5'-UTR secondary structures, such as those that encode ribosomal proteins, elongation factors and other proteins involved in the assembly and function of the translational machinery.^38,77,78 This pathway regulates translation of mRNAs that encode important proteins for many biological processes, including proliferation. Our most current knowledge on the role of these proteins has been based on experiments using 4E-BP1-deficient cells or mice. The combined disruption of 4e-bp1 and 4e-bp2 in mice increased their sensitivity to diet-induced obesity by accelerated adipogenesis. These animals displayed increased insulin resistance associated with increased S6K activity and impaired Irs2/Akt signaling in peripheral tissues. Unfortunately, the specific β-cell phenotype of these animals was not analyzed.⁵¹ Recent studies have also revealed an important role for 4E-BP in protection against endoplasmic reticulum (ER) stress in β cells. Expression of 4e-bp1 was increased by direct transcriptional activation of the Eif4e-bp1 gene in islets under ER stress in several mouse models of diabetes.⁷⁹ Most importantly, islets from 4e-bp1-null mutant mice were more susceptible to ER stress-induced apoptosis, suggesting that 4E-BP1 could be a survival factor for β cells. In vitro studies have demonstrated that growth factors, such as insulin, and nutrients, such as amino acids and glucose, induce phosphorylation of 4E-BP1 in islets and insulinoma cells.^80-83 However, little is known about the importance of different 4E-BPs and the function of these proteins on regulation of β-cell proliferation, size, survival, mass and function. In summary, studies to determine the importance of 4E-BP1 in β cells have been limited by the alteration on insulin sensitivity observed in the global knockout model. Our understanding of the role of 4E-BP2 in different tissues is limited, but this protein is abundant in the brain and is required for learning and memory, indicating that in tissues it has functions independent of 4E-BP1.⁸⁴

The ribosomal protein S6 kinase (S6K) is described as a regulator of cell growth, protein translation and proliferation.⁸⁵ The activity of S6K1 is regulated by both mTORC1 and PDK1, and several of its downstream targets have been identified. S6K1 and S6K2 regulate ribosomal protein S6, the eukaryotic initiation factor 4B (eIF4B), SKAR and eukaryotic elongation factor 2 kinase (eEF2K).⁸⁶ S6K1- and S6K2-mediated phosphorylation is required for full activation and subsequent induction of the 40S ribosomal protein S6, which can then induce cell growth and proliferation. The importance of S6K signaling in β cells has been assessed in genetically modified mouse models. S6K1-deficient mice are viable and fertile and only present mild phenotypes during development because of a compensatory increase in levels of S6K2.^87,88 Additionally, these mice display glucose intolerance and hypoinsulinemia with impaired insulin secretion.⁸⁷ The β cells display a reduced size and decreased transcription of insulin, demonstrating the importance of S6K1 in regulating glucose homeostasis as well as cell growth. Interestingly, S6K1-deficient mice are resistant to age and high-fat diet-induced obesity and remain insulin-sensitive due to the loss of the negative feedback loop from S6K on Irs1 and Irs2. Together, these results suggest an in vivo role for S6K in desensitizing tissues to insulin,^1,3,47 although the significance of this feedback regulation on IRS1 has recently been questioned.⁸⁹ In contrast, S6K2-deficient mice are phenotypically similar to the control mice, suggesting that S6K1 might be more important than S6K2 for glucose homeostasis.⁹⁰ Recent findings also demonstrated that S6K is important for insulinoma formation induced by activation of AKT signaling, implicating this kinase in regulation of β-cell cell cycle progression.⁹¹ Little is known about the signaling events downstream of S6K. The importance of ribosomal S6 protein was assessed using knock-in mice containing alanine substitutions at all five phosphorylatable serine residues of their S6 protein.⁹² These mice exhibit impaired glucose tolerance, lower insulin levels and increased insulin sensitivity. The pancreatic insulin content was reduced in these mice but was not associated with alterations in β-cell mass, implying an effect of S6 protein on insulin synthesis. The similarity of this phenotype with that of S6K-deficient mice suggests that ribosomal S6 protein is a critical substrate in relating metabolic signals from S6K.

To study the role of S6K activation in pancreatic β cells, we developed transgenic mice overexpressing a constitutively active form of S6K in β cells (S6KCA^RIP).⁷⁶ Activation of S6K signaling in these mice improves insulin secretion in the absence of changes in β-cell mass. These findings revealed for the first time that activation of S6K induces insulin resistance in β cells by feedback inhibition of IRS signaling. Importantly, these studies also demonstrated that S6K regulates IRS2 levels, a major determinant of β-cell proliferation and survival. The β-cell insulin resistance induced by this mechanism had major negative effects on cell cycle progression by modulating levels of Cdk2, p27 and p16. Another major effect of insulin resistance in this model was increased apoptosis due to decreased survival signals from AKT. The balance between increased cell cycle entry with slow progression and enhanced apoptosis resulted in lack of β-cell mass expansion. Furthermore, these observations underscore the importance of the S6K-mediated negative feedback on IRS1/2 signaling in this phenotype and suggest that one of the major consequences of chronic exposure of β cells to nutrient overload is the development of impaired IRS signaling. In addition, these studies show that S6K activation recapitulates the cell size but not the proliferative phenotype of models with activation of mTORC1 signaling. These findings serve to elucidate some of the abnormalities observed in adaptive responses of β cells to nutrient excess and tentatively explain some of the mechanisms involved in glucose toxicity. Together, these data suggest that the negative feedback of S6K on IRS signaling can be a major modulator of β-cell mass and function in vivo.

Paraformaldehyde-fixed pancreatic tissues were embedded in paraffin using standard techniques. Sections were deparaffinized, rehydrated and incubated with blocking solution. Sections were incubated overnight at 4°C with antibodies against insulin (Linco Research), followed by secondary antibodies conjugated to FITC (Jackson Immunoresearch). DAPI-containing mounting media (Vector Laboratories) was added to the coverslips. β-cell mass assessment was performed by point counting morphometry from five insulin-stained sections (5 μm) separated by 200 μm using NIH Image Pro software (Media Cybernetics, Inc.).

Fasting glucose levels were measured after an overnight fasting using an AccuChek II glucometer (Roche Diagnostics). Fed glucose levels were obtained at the same time in the morning of each day. Glucose tolerance test was performed by intraperitoneal delivery of 2 g/kg glucose to mice after 12 h fasting. Blood glucose was monitored for 2 h after glucose delivery. Plasma insulin levels were measured using mouse insulin ELISA kit (ALPCO).

All data are represented as the mean ± SEM from at least three independent experiments unless otherwise indicated. Data were analyzed by Student t-test or ANOVA followed by post hoc analysis where appropriate. Results were considered statistically significant when the p-value was less than 0.05 (Fig. 3).