Metabolic Role of PTEN in Insulin Signaling and Resistance

Abstract

Phosphatase and tensin homolog (PTEN) is most prominently known for its function in tumorigenesis. However, a metabolic role of PTEN is emerging as a result of its altered expression in type 2 diabetes (T2D), which results in impaired insulin signaling and promotion of insulin resistance during the pathogenesis of T2D. PTEN functions in regulating insulin signaling across different organs have been identified. Through the use of a variety of models, such as tissue-specific knockout (KO) mice and in vitro cell cultures, PTEN's role in regulating insulin action has been elucidated across many cell types. Herein, we will review the recent advancements in the understanding of PTEN's metabolic functions in each of the tissues and cell types that contribute to regulating systemic insulin sensitivity and discuss how PTEN may represent a promising therapeutic strategy for treatment or prevention of T2D.

THE ROLE OF PTEN IN INSULIN SIGNALING: OVERVIEW

PTEN is a critical negative regulator of insulin signaling through its role in the dephosphorylation of phosphatidylinositol 3,4,5-triphosphate (PIP₃) to phosphatidylinositol 4,5-bisphosphate (PIP₂) in the phosphoinositide 3-kinase (PI3K) pathway (Fig. 1; Nakashima et al. 2000; Simpson et al. 2001). Traditionally, PTEN has been defined as a tumor suppressor, as it regulates proliferation, cell growth, and survival to prevent tumor formation (Stambolic et al. 1998). However, because the PI3K pathway is a major signaling network activated in response to insulin; PTEN dysregulation has been implicated in the regulation of insulin signaling and glucose homeostasis (Nakashima et al. 2000; Simpson et al. 2001).

Figure 1.

Schematic of insulin signaling pathway. The insulin pathway is activated by the binding of insulin to the insulin receptor. Binding of insulin to its receptor will recruit insulin receptor substrate (IRS) to phosphorylate the tyrosine kinase (TK) domains of the insulin receptor. The activated IRS will then phosphorylate and activate PI3K. Activated phosphoinositide 3-kinase (PI3K) will facilitate the phosphorylation of PIP₂ to PIP₃. PTEN acts as a negative inhibitor of the phosphorylation of PIP₂ to PIP₃ as it promotes the phosphorylation of PIP₃ back to PIP₂. PIP₃ will activate downstream PDK that promotes GLUT4 translocation to the plasma membrane of skeletal muscles. Alternatively, IRS phosphorylates and activates the GRB2/SHC/SOS pathway, which promotes the phosphorylation of Ras-GDP to Ras-GTP. Ras-GTP ultimately activates ERK1/2, which mediates GLUT4 transcription and cellular proliferation.

Insulin signaling is initiated with insulin binding to the insulin receptor (IR), resulting in canonical activation of PI3K, an enzyme that phosphorylates membrane-bound PIP₂ to PIP₃ (Whitman et al. 1985; Carpenter et al. 1990; Hawkins et al. 1992; Stephens et al. 1993). PIP₃ in turn acts as a docking site for pleckstrin homology (PH) domain-containing proteins, which then coordinate the activation of numerous downstream proteins including AKT/PKB, and which then phosphorylates a wide array of downstream targets that initiate the anabolic actions of insulin (Alessi et al. 1997; Franke 1997; Currie et al. 1999). PTEN, as a negative regulator of this pathway, dephosphorylates PIP₃ to PIP₂ to effectively inhibit the effects of PI3K signaling in response to insulin. Accordingly, loss-of-function mutations in PTEN can result in enhanced insulin signaling in different organs leading to protection against insulin resistance, a key pathogenic process in type 2 diabetes (T2D) (Grinder-Hansen et al. 2016). Thus, PTEN may be an attractive target for therapeutic intervention in T2D.

CLINICAL IMPLICATIONS OF PTEN IN INSULIN RESISTANCE AND T2D

T2D, as defined by hyperglycemia and insulin resistance, has reached a global epidemic with the World Health Organization reporting that more than 422 million people worldwide are affected by this disease (WHO 2016). Emerging clinical evidence from both individual and population studies implicate PTEN in the regulation of insulin resistance in addition to its well-known role as a tumor suppressor (Pal et al. 2012). For example, PTEN haploinsufficiency as a result of a loss-of-function mutation in Cowden's syndrome is associated with increased insulin sensitivity in addition to increased risk of tumor formation (Iida et al. 2000). Indeed, individuals with PTEN haploinsufficiency show decreased baseline insulin and glucose levels, suggesting improved insulin sensitivity. Furthermore, these individuals have increased AKT phosphorylation in response to insulin in adipose tissue (Pal et al. 2012).

Conversely, studies have reported that polymorphisms in PTEN are associated with increased insulin resistance or metabolic syndrome, a major risk for the onset of T2D. For example, substitution of cytosine to guanine in position 9 in the 5′-untranslated region in PTEN is commonly found in Japanese individuals with T2D (Ishihara et al. 2003). Indeed, individuals with this polymorphism have increased PTEN protein levels, decreased AKT phosphorylation, and increased insulin resistance when compared with healthy controls (Ishihara et al. 2003). Furthermore, studies in individuals with metabolic syndrome reveal decreased PTEN promoter hypermethylation, likely associated with PTEN silencing (Hashemi et al. 2013), further supporting that PTEN down-regulation has a protective effect in T2D.

The most instrumental experimental tools that have helped elucidate the metabolic functions of PTEN in vivo have been the Pten knockout (KO) mouse models (Table 1). Whereas whole-body Pten KO leads to embryonic lethality in mice (Di Cristofano et al. 1998), mice with Pten-deficiency specifically in metabolic tissues are viable and allow the analysis of tissue-specific contributions of PTEN signaling in altering insulin action and glucose homeostasis (Fig. 2). The majority of mice featuring tissue-specific KO for Pten in metabolic organs (pancreas, liver, muscle) share several features of improved metabolic functions, with increased insulin sensitivity and glucose tolerance (Stiles et al. 2004; Wijesekara et al. 2005; Nguyen et al. 2006; Tong et al. 2009; Wang et al. 2010, 2014). Herein, we will review the role of PTEN in each of the peripheral and central tissues and cell types, and its impact on metabolism with a focus on tissue-specific KO mice used to assess insulin signaling in vivo (Fig. 3).

Table 1.

Mouse studies on different tissue-specific knockout (KO) of PTEN and their various effects on metabolic phenotypes^a

Tissue	Mice model	Diet fed	Phenotypeb	References
Pancreatic α cell	Glucagon (Glu-Cre)	NCD/HFD	▴IS, ▾fasting BG, ▴GT, ▾HS, ∼BW, ∼AD, maintained α- and β-cell architecture, no tumor formation	Wang et al. 2015b
Pancreatic β cell	Rat insulin promoter (RIP-Cre)	NCD/HFD	▴IS, ▾fasting BG, ▴GT, ▴BW, ▴β-cell mass and size in islets, architecture of β cell maintained, ▴insulin secretion	Nguyen et al. 2006; Stiles et al. 2006; Wang et al. 2010
	RIP-Cre- LEPdb/db	NCD/HFD	∼IS, ∼fasting BG, ∼GT, ▴BW, maintained β-cell architecture, ▴β-cell function	Wang et al. 2010
	Rip-Cre-inducible Myc (Myc-Tam)	NCD	▴BG, ▴islet mass, no protection against diabetes induced by c-myc activation, proliferation, and apoptosis of β-cell islet, no tumor formation	Radziszewska et al. 2009
	Pdx1-Cre/Pdx1-Cre-ROSA		▾fasting BG, ▾fed SI, ∼plasma CHO, ∼plasma TG, ▴IS, ▴GT, enlarged pancreatic islets, ▴pancreatic acinar proliferation	Tong et al. 2009
Skeletal muscle–soleus and extensor digitorum longus	Muscle creatinine kinase (Mck-Cre)	NCD/HFD	▴IS, ▾fasting BG, ▴GT, ▾BW, ▾fasting PI, KO did not develop spontaneous hyperglycemia, β-cell hypertrophy, and hyperinsulinemia, ∼adverse tumor development	Wijesekara et al. 2005
	Glucosoamine-induced insulin-resistant rat skeletal muscle	N/A	▾Insulin-stimulated GU, ▴GU, ▴PTEN expression Metformin: ▴Insulin-stimulated GU, ▴GLUT4 translocation, ▾glucosamine-induced cellular apoptosis	Wang et al. 2015a
Brain	Rip-Cre hypothalamus–partial deletion	NCD	▴IS, ▴GT, ▾SI, ▾BG, ▾BW, ▴circulating IGF-1, small-body-phenotype	Choi et al. 2008
	Adenovirus- mediated PTEN KO in rat mediobasal hypothalamus (MBH)	NCD/HFD	▾FI, ▾WG, ▾BW, decreased FI blunted by HFD, ▴IS, ▾BG, ▾glucose perfusion, ▴fasting PI, ▴AKT phosphorylation during HFD	Sumita et al. 2014
	Adenovirus-mediated PTEN overexpression in rat MBH	NCD/HFD	▴WG, ▴FI, ▾insulin-induced decreases in BG, ▾GIR to maintain euglycemia, ▴lipogenesis genes	Sumita et al. 2014
	PTEN and LEP gene (Obrb) KO in mice arcuate nucleus POMC neurons	NCD	▾BW, ▴IS, ▴conversion of WAT to BAT, ▴skeletal muscle GU, ▴RMR, ▴O2 consumption	Plum et al. 2007
	PTEN KO in ventromedial hypothalamus (VMH) SF-1 neurons	NCD/HFD	▴BW, ▴weight gain, ▴weight gain similar to wild-type in HFD	Klöckener et al. 2011
	PTEN KO in hypothalamic POMC neurons	NCD/HFD	▴BW, ▴BL, hyperphagia, ▾POMC neuronal projections, ▴LEP resistance, ▴hyperpolarization of POMC neurons via insulin	Plum et al. 2007
Adipose tissue	WAT-specific binding protein (aP2-Cre)	NCD	▴IS, ▴GT, ▾FBG, ▾PI, ∼APN, ∼AD, ∼BW, ∼ serum TG, ▾circulating and adipose resistin	Kurlawalla-Martinez et al. 2005
	Inducible adipocyte-specific PTEN KO	Doxycycline diet, HFD	▴AN, ▴BW, ▾SI, ▴IS, ▴GT, ▾fasting BG, ▴β3-adrenergic-stimulated lipolysis, ▾TNF-α in subcutaneous adipose tissue, ▾movement, ▾fasting SI, ▴AD, ▾HS, ∼adipocyte size or number, healthy hyperplasia, hyperglycemic at rest, ▾macrophage infiltration	Morley et al. 2015
	In vitro 3T3-L1 adipocytes PTEN overexpression	N/A	▾Insulin-stimulated PI3K signaling, ▾AKT activation, ▾P70S6K activation, ▾GLUT4 translocation, ▾GU	Nakashima et al. 2000
	BAT PTEN overexpression (PTENTg)	NCD/HFD	▾FBG, ▾fasting PI, ▴IS, ▾HS, ▴BW under HFD, FI normalized by body weight, ▴WI, ▴RMR, ▾LEP, ▾adipocyte size, ▾CHO, ▾tumor formation, ▴mouse life, ▾DNA damage	Ortega-Molina et al. 2012
Liver	Alb-Cre	NCD/HFD	▴HS, ▾AD, ▴IS, ▴GT, ▾FBG, ▾serum LEP, ▾fasting PI, ▾gluconeogenesis, ▾fasting PG, ▴glycogen synthesis, ▴liver FA synthesis, ▴lipogenesis enzymes	Stiles et al. 2004
	Alb-Cre	NCD	▾HS, ▴liver TG, ▴liver CHO, ▴IS, ▾SI, ▾BG, ▴adipogenic and lipogenic genes, ▴HD, ▴liver adenomas, ▴liver BrdU proliferation	Horie et al. 2004
	Alb-Cre—with focus on brain metabolism	NCD/HFD	▴GT, ▾PKB, ▴brain GU, ▴LTP, ▴synaptic plasticity, ▴brain IS	Patil et al. 2018
	Alb-Cre—with focus on peripheral insulin signaling	NCD	▴HS, ▾gluconeogenesis, ▾GO, ▴glycolysis, ▴IS, ▴GT, ▾lipid storage in fat deposits, ▴skeletal muscle insulin-stimulated GU, ▴browning of mesenteric WAT, ▴liver FGF21 secretion	Peyrou et al. 2015
	Antisense-mediated oligonucleotide PTEN KO	NCD	▾ob/ob weight gain, ▾db/db weight gain, hyperglycemia, ▴IS, ▴AKT expression, ▾CHO, ▾serum TG, ▾SI, ▾BG	Butler et al. 2002
Thyroid	TPO-cre	NCD	▾TCA cycle genes expression, ▾mitochondrial genes expression	Antico Arciuch et al. 2013

(HFD) high-fat diet, (NCD) normal chow diet, (IS) insulin sensitivity, (FBG) fasting blood glucose, (GT) glucose tolerance, (HS) hepatic steatosis, (BW) body weight, (AD) adiposity, (BG) blood glucose, (LEP) leptin, (SI) serum insulin, (TG) triglyceride, (CHO) cholesterol, (GU) glucose uptake, (GLUT4) glucose transporter 4, (IGF-1) insulin growth factor 1, (FI) food intake, (WG) weight gain, (PI) plasma insulin, (AKT) protein kinase B, (GIR) glucose infusion rate, (POMC) proopiomelanocortin, (WAT) white adipose tissue, (BAT) brown adipose tissue, (O₂) oxygen, (BL) body length, (APN) adiponectin, (TNF-α) tumor necrosis factor α, (PI3K) phosphoinositide 3-kinase, (p70S6K) ribosomal protein S6 kinase β-1, (WI) water intake, (RMR) resting metabolic rate, (FA) fatty acid, (PG) plasma glucose, (PKB) plasma ketone bodies, (HD) hepatic damage, (LTP) long term potentiation, (GO) glucose output, (FGF21) fibroblast growth factor 21, (ob/ob) leptin gene KO mice, (db/db) leptin receptor KO mice.

^aSymbols: (▴) increased, (▾) decreased, (∼) no change.

^bChanges in phenotype under HFD unless experiments were performed in normal chow diet.

Figure 2.

Schematic of metabolic altering effects of PTEN in various tissue. Black arrows indicate the generation of tissue-specific in vivo knockouts of PTEN, whereas the green arrow indicates the promotion of global events found commonly across in each tissue-specific PTEN knockout. The metabolic parameters altered in the specific organ are depicted in each of the organs represented in each box including the skeletal muscle, thyroid, liver, white adipose tissue, brain, and pancreatic islets.

Figure 3.

Model of insulin signaling across the various metabolic organs in the body. Each box highlights the unique signaling components of insulin in the liver, pancreatic β cell, and muscle and adipose tissue, respectively. Shown in red is PTEN's role in regulating the various components of insulin signaling in the respective metabolic organs and potential functions that are regulated by insulin signaling.

PANCREATIC ISLETS

The pancreatic islets α and β cells synthesize and secrete key metabolic hormones, glucagon and insulin, respectively, that maintain plasma glucose homeostasis. These functions are often compromised during the pathogenesis of T2D as a result of α- and β-cell dysfunction (Unger et al. 1970; Butler et al. 2003; Ueki et al. 2006). Insulin-mediated PI3K signaling plays a key role in maintaining α- and β-cell function, and PTEN is a critical negative regulator in the maintenance of pancreatic islet's structural and functional integrity (Wang et al. 2010, 2015b). PTEN has been shown to have an important biological role in regulating α- and β-cell growth, as well as glucagon and insulin secretion through the regulation of D-type cyclins, p27, mammalian target of rapamycin (mTOR), and glycogen synthase kinase 3β (GSK3B), and by indirectly suppressing pancreatic and duodenal homeobox (PDX-1) expression via mTOR inhibition (Holland et al. 2002; Wang et al. 2010; Yang et al. 2014; Alejandro et al. 2017; Blandino-Rosano et al. 2017; Bozadjieva et al. 2017).

Mechanistically, PTEN disruption in β cells leads to increased AKT phosphorylation, which promotes activation of the mTOR pathway, more specifically of the rapamycin-sensitive mTORC1 complex, which is composed of mTOR, Raptor, GβL, and DEPTOR. The effect of AKT on mTORC1 is primarily mediated by its inhibitory phosphorylation of TSC2, which impairs the ability of the TSC1/TSC2 complex to inhibit the small G-protein Rheb, which is essential to activate mTOR (Huang and Manning 2009). In addition, in HEK293T cells, AKT has been shown to activate mTORC1 by phosphorylating proline-rich Akt substrate of 40 kDa (PRAS40) on Thr246 to sequester it from mTORC1, thus activating mTORC1 (Vander Haar et al. 2007). Alternatively, in HEK293T cells, AKT can modulate mTORC1 activity by regulating IκB kinase α (IKKα) to reduce the affinity between Raptor and mTORC1, which inhibits the mTORC1 complex activity (Dan et al. 2014). Indeed, studies using β-cell-specific Rptor KO mice (βraKO), generated using the rat Ins2 promoter (RIP2), revealed β-cell-specific disruption of mTORC1 resulting in β-cell loss, impaired insulin secretion, hyperglycemia, and premature diabetes (Blandino-Rosano et al. 2017). The decreased activation of mTORC1 affects two critical downstream mTORC1 targets and effectors: 4E-BPs that control downstream β-cell proliferation, and S6K, which modulates β-cell autophagy, size, and apoptosis (Blandino-Rosano et al. 2017). Furthermore, reactivation of eIF4E in βraKO mice by deleting Eif4ebp2 restored the expression of carboxypeptidase E (CPE), an essential enzyme in the folding of proinsulin (Davidson and Hutton 1987), indicating that the 4E-BP2/eIF4E/CPE axis is the primary mechanism of action for mTOR's regulatory actions on insulin secretion (Blandino-Rosano et al. 2017).

Disruption of PTEN resulting in activated AKT also leads to GSK3β Ser9 phosphorylation leading to its deactivation (Cross et al. 1995). This in turn can reduce PDX-1 phosphorylation resulting in prevention of PDX-1 breakdown, ultimately leading to increased β-cell mass and function (Holland et al. 2002). PTEN has also been identified as a critical factor in maintaining β-cell structure postnatally (Yang et al. 2014). Studies in mice using tamoxifen-induced Pten deletion under the rat Ins2 gene promoter (RIP) revealed decreased expression of cell-cycle inhibitor p27 and p16^INK4a at 10 months of age (Yang et al. 2014), suggesting that PTEN may increase β-cell proliferation even later in age through bypassing the replicative checkpoints (Yang et al. 2014). These studies highlight the important negative regulatory role for PTEN in maintaining β-cell function.

Pancreatic β-cell PTEN mRNA has also been identified as a critical target of many microRNAs (miRNAs) to regulate β-cell proliferation and function. One such miRNA is miR-494, which has been shown to bind and inhibit β-cell PTEN mRNA. As such, knockdown of miR-494 resulted in impaired insulin secretion, increased apoptosis, and decreased proliferation of INS-1 cells. siRNA-mediated down-regulation of PTEN reversed the effects of miR-494 knockdown on insulin secretion, cell proliferation, and apoptosis of pancreatic β cells (He et al. 2017). Furthermore, women with gestational diabetes have decreased levels of miR-494 that correlated with increased β-cell apoptosis and compromised insulin secretion (He et al. 2017). Another miRNA that has been shown to regulate β-cell proliferation is miR-132. In the MIN6 mouse insulinoma cell line, overexpression of miR-132 decreased PTEN levels, resulting in increased cell survival and proliferation (Mziaut et al. 2017).

Collectively, down-regulation of PTEN leads to enhanced β-cell growth and proliferation while maintaining secretory function. Importantly, the enhanced growth does not occur concomitantly with dedifferentiation or tumorigenesis. In fact, even with concomitant activation of a potent oncogene, c-Myc, Pten deletion leads to apoptosis rather than tumorigenesis (Radziszewska et al. 2009).

PTEN disruption in α cells improves glucose homeostasis and protects against both obesity and diabetes through suppression of glucagon secretion. Mice with α-cell-specific Pten KO under the control of glucagon promoter (GluCre⁺-Pten^fl/fl) showed decreased glucagon secretion and improved insulin sensitivity (Wang et al. 2015b). Although these mice did not show changes in α-cell mass, the pancreatic α-cell line αTC1 revealed increased proliferation in pancreatic α cells on treatment with insulin and this was ablated after blocking the glucagon receptor, suggesting that glucagon itself may promote α-cell proliferation (Liu et al. 2011). Notably, GluCre⁺-Pten^fl/fl mice display improved glucose tolerance with suppression in glucagon secretion, supporting a function for PTEN in promoting insulin action in α cells through glucagon suppression (Wang et al. 2015b). GluCre⁺-Pten^fl/fl mice displayed no changes in the ratio of α-cell to β-cell area with similar pancreatic islet mass to their wild-type (WT) controls (Wang et al. 2015b), which suggests that PTEN disruption in α cells can maintain the structural integrity in α and β cells, which is often disrupted in T2D (Henquin and Rahier 2011).

Increased PI3K and mTORC1 signaling can up-regulate FoxA2, a key developmental transcriptional factor that regulates glucagon secretion and synthesis through regulating key glucagon secretory genes such as SUR1 that regulates the K_ATP channel Kir6.2, which is essential for glucagon secretion (Heddad Masson et al. 2014; Bozadjieva et al. 2017). In support of this, α-cell-specific mTORC1 KO under the glucagon promoter revealed decreased glucagon secretion with down-regulation of FOXA2 secretion, indicating that mTORC1 can regulate glucagon secretion through FOXA2 (Bozadjieva et al. 2017). Overall, these results show that disrupting PTEN in either α or β cells leads to improved glucose homeostasis through independent mechanisms.

LIVER

The liver is a major site of insulin action that facilitates suppression of gluconeogenesis, as well as promotion of lipid and glycogen synthesis (Valverde et al. 2003). Defective hepatic PI3K signaling is a major factor in the pathogenesis of T2D because it is associated with disrupted hepatic insulin response and increased hepatic glucose production (HGP) (Miyake et al. 2002; Titchenell et al. 2015). Consistent with the well-known negative regulatory role of PTEN in the PI3K pathway, PTEN has been shown to regulate HGP, triglyceride production, and hepatic steatosis. Liver-specific PTEN KO mouse models under the albumin promoter (AlbCre⁺-Pten^fl/fl) show enhanced hepatic insulin signaling and improved insulin sensitivity compared with their WT counterparts when challenged with a high-fat diet (HFD) (Stiles et al. 2004).

One important target of PTEN in the liver is the AKT-GSK3β pathway in which activated AKT on PTEN deletion inactivates GSK-β by phosphorylating it on Serine 9 (Sutherland et al. 1993; Cross et al. 1995). Reduced GSK-3β activity as a consequence of Ser 9 phosphorylation leads to decreased inhibitory phosphorylation of glycogen synthase, resulting in increased hepatic glycogen storage (Fiol et al. 1988). Similar to these in vitro results, Alb-cre⁺Pten KO mice show elevated hepatic phospho-GSK-3β with increased hepatic glycogen content. PTEN loss, through increasing AKT activity, can also promote phosphorylation of Thr-32, Ser-253, or Ser-315 of Forkhead box protein O1 (FOXO-1) (Brunet et al. 1999; Rena et al. 1999). This in turn leads to FOXO-1 sequestration by 14-3-3, leading to nuclear export and degradation (Brunet et al. 1999). Inactivation of FOXO-1 induces transcription of adipogenic genes, including perixosome proliferator-activated receptor γ (PPAR-γ) (Dowell et al. 2003), and inhibits transcription of the gluconeogenic genes phosphoenolpyruvate carboxykinase (PEPCK) and glucose 6-phosphatase (G6Pase) (O-Sullivan et al. 2015), leading to suppression of HGP (Brunet et al. 1999). Indeed, Albcre⁺-Pten^fl/fl mice showed decreased expression of G6Pase and PEPCK (Stiles et al. 2004). Moreover, AlbCre⁺-Pten^fl/fl mice have dramatically increased hepatic steatosis, supporting PTEN's input into regulating hepatic lipid synthesis (O-Sullivan et al. 2015). This is likely attributed to PPAR-γ-mediated FOXO-1 inhibition, similar to liver-specific PPAR-γ KO protecting mice from HFD-induced hepatic steatosis (Morán-Salvador et al. 2011). Furthermore, PTEN, through the mTORC1 pathway, can regulate hepatic glucose and lipid metabolism by activating the adipogenic sterol regulatory element-binding 1 (SREBP-1c) in an insulin-dependent manner (Foretz et al. 1999a,b). In this pathway, mTORC1 activates its downstream effector SREBP-1c, which in turn activates its response elements SREa and SREb to bind and activate the liver glucose kinase (GK) promoter in an insulin-dependent manner (Kim et al. 2004). Glucokinase phosphorylates glucose to glucose 6-phosphate, which activates glycogen synthase and results in increased hepatic glycogen storage (Seoane et al. 1996; Foretz et al. 1999a; Kim et al. 2004). Interestingly, hepatic PTEN deletion may have effects in distant metabolic tissues such as muscle and adipose tissue through increased production of fibroblast growth factor 21 (FGF21) (Peyrou et al. 2015), resulting in beiging of adipose tissue and increasing muscle insulin sensitivity (Huang et al. 2017).

Numerous miRNAs can exert their regulatory effects on hepatic PTEN. Notably, mmu-miR-152-3p (miR-152) (Wang et al. 2016), miR-19α (Dou et al. 2015), and miR-20a-5p (Fang et al. 2016) have been shown to promote hepatic glycogenesis through negatively regulating PTEN (mRNA) expression. miR-19α and miR-152 can directly bind to PTEN mRNA and thereby counteract interleukin 6 (IL-6)-induced hepatic insulin resistance (Dou et al. 2015; Wang et al. 2016). In contrast, miR-20a-5p binds to TP63 mRNA, which in turn can indirectly down-regulate PTEN through inhibition of p53 in NCTC1469 cells resulting in increased glycogenesis and AKT activation (Fang et al. 2016). Indeed, p53 was previously reported to promote PTEN transcription (Stambolic et al. 2001). Overall, these studies highlight that hepatic PTEN can be regulated by multiple miRNAs as another potential therapeutic strategy for T2D.

SKELETAL MUSCLE

The skeletal muscle is a key tissue in regulating glucose homeostasis, acting as a major site of glucose uptake from circulation in response to insulin. This is primarily mediated through activation and translocation of GLUT4 to the plasma membrane (Suzuki and Kono 1980; Birnbaum 1989; Frevert and Kahn 1997). In particular, PTEN in the skeletal muscle regulates glucose uptake and glycogen synthesis by regulating glut-4 expression. PTEN disruption promotes increased AKT activity, which stimulates its effector TBC1 domain family member 4 (AS160) (Eguez et al. 2005). Increased AS160 activity in turn promotes increased vesicular trafficking of GLUT4 to the plasma membrane resulting in increased glucose uptake (Sano et al. 2003). In support of this, muscle-specific PTEN KO under the muscle creatine kinase promoter (MCKCre⁺-PTEN^fl/fl) improves glucose uptake in the soleus muscle, supporting enhanced insulin action in this muscle group and resulting in improved glucose tolerance and systemic insulin sensitivity (Wijesekara et al. 2005).

PTEN in the skeletal muscle can also regulate glycogen synthesis through enhancing the GSK-3 pathway (Moxham et al. 1996). Indeed, muscle biopsies from type 2 diabetic patients show insulin resistance and elevated GSK-3β activity, suggesting that glycogen synthesis may be impaired because of disruption of glycogen synthase activity (Nikoulina et al. 2000). Indeed, MCKCre⁺-PTEN^fl/fl mice had more glycogen storage in muscle compared with WT controls (Wijesekara et al. 2005). Overall, these studies support that PTEN in the skeletal muscle is an important regulator for insulin-mediated glucose uptake and glycogen synthesis. Importantly, in this tissue, PTEN deletion does not appear to promote an increase in muscle mass or tumorigenesis.

ADIPOSE TISSUE

Adipose tissue was classically defined as an organ for fat storage; however, it is now also recognized as a dynamic endocrine organ that secretes many metabolic hormones known as adipokines, such as adiponectin, leptin, and resistin (Kershaw and Flier 2004). Adipose tissue can be classified into multiple types, including white adipose tissue (WAT), brown adipose tissue (BAT), and beige or brite adipose tissue (WAT that has features of BAT within WAT depots). WAT primarily serves as an energy storage site for lipid droplets, whereas BAT predominately serves as a thermal regulator by burning fuel products for thermogenesis and is characterized by its high expression of mitochondrial uncoupling protein-1 (UCP-1) (Ridley et al. 1986). All adipose tissue depots are highly regulated by insulin in modulating lipid synthesis, glucose uptake, thermal regulation, and metabolism, primarily through insulin-mediated PI3K signaling (Blüher et al. 2002).

PTEN can regulate adipocyte biology using similar molecular mechanisms as in hepatocytes. For example, PTEN can regulate adipogenesis via AKT-mediated inhibition of FOXO-1, which in turn leads to activation of PPAR-γ (Fan et al. 2009). On the other hand, PPAR-γ can transcriptionally regulate PTEN by directly binding to its promoter site (Teresi et al. 2006). Indeed, PPAR-γ stimulation using agonists on 3T3-L1 adipocytes led to down-regulation of PTEN resulting in increased glucose uptake (Shen et al. 2006; Kim et al. 2007; Fan et al. 2009). PTEN can also regulate adipogenesis in WAT through the activating mTORC1/SREBP-1c pathway (Porstmann et al. 2008) leading to fatty acid synthase and lipoprotein lipase activity, which promotes differentiation of adipocyte precursor cells resulting in increased adipogenesis (Kim and Spiegelman 1996, 1). In line with these findings, inducible adipocyte PTEN (Aip)KO mice show improved insulin signaling with increased whole-body adiposity with increased body mass suggesting that disruption of WAT PTEN can enhance insulin signaling to promote increased PPAR-γ and SREBP-1c activity (Morley et al. 2015). Furthermore, HFD-fed AipKO mice show changes in adipocyte endocrine function with fat pads isolated from AipKO mice showing decreased tumor necrosis factor α (TNF-α) and increased adipose and circulating adiponectin levels (Morley et al. 2015). Overall, PTEN deletion in WAT led to attenuation in inflammatory changes with endocrine function that promote enhanced insulin sensitivity (Morley et al. 2015). Similar to AiPKO mice, PTEN KO mice under the control of adipocyte protein 2 (aP2) (also known as fatty acid binding protein 4) promoter show improved WAT insulin sensitivity (Kurlawalla-Martinez et al. 2005). However, in contrast to AipKO, aP2 PTEN KO mice show decreased adiposity and body weight compared with WT controls (Kurlawalla-Martinez et al. 2005; Morley et al. 2015). This difference is likely because of widespread expression of aP2 beyond adipose tissue (Martens et al. 2010).

Contrary to the well-known role of WAT in insulin signaling, the metabolic effects of BAT were not appreciated in adulthood until the recent rediscovery of active BAT in humans (Nedergaard et al. 2007; Cypess et al. 2009; van Marken Lichtenbelt et al. 2009; Virtanen et al. 2009). In diet-induced obesity rodent models, decreased BAT insulin signaling and function was observed, suggesting the importance of insulin signaling in BAT (Shimizu et al. 2014; Roberts-Toler et al. 2015). Interestingly, PTEN overexpression leads to increased UCP-1 by up-regulating PGC-1α (Ridley et al. 1986; Ortega-Molina et al. 2012). Furthermore, mice overexpressing PTEN in BAT (Pten^tg) showed decreased AKT signaling with increased PGC-1α, FOXO-1, and UCP-1 levels in BAT. In contrast with effects of PTEN in WAT, Pten^tg mice showed a paradoxical improvement in insulin sensitivity, suggesting that BAT PTEN may have an opposing role in whole-body insulin sensitivity (Ortega-Molina et al. 2012). The findings from PTEN^tg mice are similar to those in BAT-specific DJ-1 KO mice under the myf5 promoter, which also show improved insulin sensitivity (Wu et al. 2017). Indeed, BAT isolated from DJ-1 KO mice have elevated PTEN, UCP-1, and FOXO-1 levels, with decreased AKT signaling. Mechanistically, DJ-1 can increase ubiquitination of PTEN through mind bomb-2 (Mib2), an E2 ubiquitin kinase that was found to be autophosphorylated by DJ-1 (Wu et al. 2017). Overall, PTEN may have opposing roles in BAT and WAT in regulating metabolism, which warrants further studies.

BRAIN

The brain is critical in regulation of metabolism. In particular, the hypothalamus plays a central role in the regulation of food intake and energy homeostasis (Niswender et al. 2003; Stanley et al. 2005). Within the hypothalamus, different types of neurons each contribute uniquely in regulating glucose homeostasis, appetite, and satiety (Grossman, 1975). In particular, there are two groups of neurons that regulate satiety, the appetite-stimulating agouti-related peptide (AGRP) and appetite-suppressing proopiomelanocortin (POMC) neurons (Balthasar et al. 2005; Gropp et al. 2005).

PTEN shows a unique regulatory role in metabolism in different hypothalamic neurons. POMC-specific PTEN KO mice show increased hyperpolarization of POMC neurons leading to uncontrolled food intake and obesity (Plum et al. 2006). POMC neurons isolated from POMC-specific PTEN KO mice show increased ATP-sensitive potassium channel activity (K-ATP), which was inhibited by the PI3K inhibitor LY294002 (Plum et al. 2006). This suggests that PTEN, through modulating PIP₃, controls K-ATP channel hyperpolarization. PTEN can also regulate HGP through the vagus output, as K-ATP channels in the hypothalamus can regulate HGP (Pocai et al. 2005) by suppressing the hepatic vagus nerve efferents (Pocai et al. 2005; Kimura et al. 2016). The inhibition of vagal output can affect hepatic Kupffer cells by promoting the down-regulation of α-7 nicotinic receptors, ultimately resulting in the activation of IL-6 signaling (Kimura et al. 2016). Subsequently, this promotes hepatic STAT3 activation, which will transcriptionally down-modulate G6Pase and PEPCK, resulting in decreased HGP (Inoue et al. 2006; Ramadoss et al. 2009).

Furthermore, PTEN has been reported to regulate neuronal insulin signaling through inhibiting downstream focal-adhesin kinase (FAK) (Gupta and Dey 2012) through dephosphorylating FAK at Tyr397 (Tamura 1998). With PTEN disruption, increased FAK activity can subsequently activate downstream extracellular signal-regulated kinase 1/2 (ERK1/2) through phosphorylation and abrogating insulin signaling (Gupta and Dey 2012). This is in contrast to FAK in nonneuronal cells, where it acts as a positive regulator of insulin signaling (Bisht et al. 2007; Gupta et al. 2012). Similarly, nonneuronal ERK can promote cell growth, whereas neuronal ERK promotes neuronal demise (Subramaniam et al. 2004) through opposing actions of insulin (van der Heide et al. 2003). Activated ERK abrogates insulin signaling through phosphorylating the serine residues of IRS-1 to inhibit its activation following insulin signaling (Gual et al. 2005; Gupta and Dey 2012). Overall, different neuronal-specific PTEN KO models show that PTEN can regulate neuronal insulin signaling through uniquely modulating FAK-ERK1/2 pathways.

Last, hypothalamic PTEN has also been shown to regulate body size and peripheral insulin sensitivity, potentially through activation of the peripheral nervous system (Choi et al. 2008; Wang et al. 2014). Mice with Pten deletion under the control of RIP2 promoter (RIPcre-Pten KO) displayed improved peripheral insulin sensitivity with increased insulin signaling in the hypothalamus and macrophage polarization skewing toward an M2-like anti-inflammatory state in the periphery. Interestingly, these protective phenotypes were abolished following vagotomy in RIPcre-Pten KO mice, suggesting that PTEN-deficient RIPcre neurons at least partially mediates its actions through the autonomic nervous system to regulate peripheral insulin sensitivity (Wang et al. 2014).

THYROID

The effects of Pten deletion on the metabolic response of thyroid follicular cells are quite remarkable. Thyrocyte-specific deletion of Pten leads to the development of hyperplastic thyroid glands from birth (Yeager et al. 2007). Strikingly, these Pten^−/− nontransformed thyroid epithelial cells increase their glycolytic metabolic flux, as indicated by a dramatic increase in lactate production observed in mutant cells, and drastically reduce TCA cycle function. This metabolic reprogramming of Pten^−/− thyroid epithelial cells is critical for proliferation and is achieved through a mechanism involving the AKT-mediated phosphorylation of AMPK on Ser485, and its consequent inactivation despite AMP/ATP and ADP/ATP ratios that would normally stimulate AMPK activity. In turn, AMPK inhibition leads to the coordinated down-regulation of the RNA expression of TCA cycle and OXPHOS genes, leading to dysfunctional mitochondria and reduced ability to perform respiratory metabolism (Antico Arciuch et al. 2013). Thus, in the thyroid, PTEN controls the balance between OXPHOS and glycolysis by regulating AMPK activity, and the shift toward glycolysis is essential to induce increased proliferation. Interestingly, a cascade with several similarities is also observed in Caenorhabditis elegans, where daf-18/PTEN represses insulin-mediated neuroblast proliferation through a pathway that includes aak-2/AMPK as a downstream effector (Zheng et al. 2018).

CONCLUDING REMARKS

In this review, we have discussed the negative regulatory function of PTEN in insulin signaling during physiological and T2D conditions. The function of PTEN in each tissue is context-dependent as PTEN regulates a variety of metabolic functions including insulin sensitivity, glucose homeostasis, energy balance, adiposity, and alterations in metabolic hormone levels. Because of the complexity of the regulatory role of PTEN in insulin action, understanding the precise role of tissue-specific PTEN is challenging. Thus, the different animal models and examining the essential metabolic effects of PTEN in the various specific tissues and cell types have uncovered complex physiologic mechanisms and potential therapeutic opportunities for inhibiting PTEN to overcome insulin resistance and T2D (Table 1). These studies have strengthened our understanding of the molecular pathways regulated by PTEN in the pathogenesis of insulin resistance and T2D, which will facilitate therapeutic strategies. Further clinical studies examining the impact of PTEN disruption in insulin resistance and during the pathogenesis of T2D may be challenging given the potential ensuing oncogeneses. Future strategies in tissue-specific targeting of PTEN in specific metabolic tissues without affecting proliferative, cancer prone tissues and further refining molecular targets that address metabolic without the oncogenic effects could provide promising new therapeutic opportunities to overcome insulin resistance and T2D.