Effect of Growth Hormone on Insulin Signaling

Abstract

Growth hormone (GH) is a pleiotropic hormone that coordinates an array of physiological processes, including effects on bone, muscle, and fat, ultimately resulting in growth. Metabolically, GH promotes anabolic action in most tissues except adipose, where its catabolic action causes the breakdown of stored triglycerides into free fatty acids (FFA). GH antagonizes insulin action via various molecular pathways. Chronic GH secretion suppresses the anti-lipolytic action of insulin and increases FFA flux into the systemic circulation; thus, promoting lipotoxicity, which causes pathophysiological problems, including insulin resistance. In this review, we will provide an update on GH-stimulated adipose lipolysis and its consequences on insulin signaling in liver, skeletal muscle, and adipose tissue. Furthermore, we will discuss the mechanisms that contribute to the diabetogenic action of GH.

Introduction

GH research and its use in clinics to cure growth disorders span nearly a century. Elegant work by Evans and Long in 1922 demonstrated that intraperitoneal injection of a fresh bovine anterior hypophyseal extract accelerated growth in rats (1). Later, studies by Houssay and colleagues in 1930s and early 1940s established that hypophyseal extracts or crude anterior pituitary extracts accelerated growth but also promoted diabetes in dogs and cats (2–5). In 1942, Houssay demonstrated that these pituitary extracts induced histological changes in many tissues and damaged pancreatic Langerhans islets, indicating their diabetogenic potential (6). After successful isolation of GH from the anterior pituitary by Li and colleagues in 1944 (7), a series of studies by multiple groups reported that GH has the potential to induce growth, diabetes, and hyperglycemia in animals (7–11). Rabinowitz et al in 1965 showed in humans that GH promotes diabetogenic action by antagonizing insulin action in muscle and adipose via promoting the release of free fatty acids (FFAs) from the adipose tissue (12). In 1963, human growth hormone (HGH) isolated from pituitary glands of cadavers was approved for clinical use to treat children who were unusually short due to pituitary disorders resulting in GH deficiency (D). In 1985, the U.S. Food and Drug Administration (FDA) approved recombinant (r) HGH to be used for clinical and research purposes (13).

The pleiotropic actions of GH involve multiple organs and physiological systems. GH exerts many metabolic effects that persist throughout life (14, 15). Kastin et al. showed that lipid mobilizing substances present in the hypothalamus and pituitary of mammals are involved in the central neuronal control of obesity and hypothesized that GH-releasing hormone (GHRH) might influence lipolysis secondarily by regulating pituitary hormones (e.g., GH) release (16). Growth promoting effects of GH on cartilage and bone are most notable, especially during the adolescent years (17). In addition to promoting growth, GH exerts profound anabolic actions during postnatal skeletal development, in-part through stimulating insulin-like growth factor-1(IGF-1) (17). GH induces skeletal IGF-1 synthesis, sulfate uptake, osteoblast hypertrophy, proliferation of the prechondrocytes, widening of the epiphyseal plate, stimulation of cartilage growth, mineralization and bone remodeling (18, 19). Multiple studies have shown that GH induces growth in two steps: GH directly activates growth hormone receptor (GHR) to stimulate IGF1 synthesis and then IGF-1 acts on target cells to exert some of the physiological effect of GH (20–22); thus GH acts directly and indirectly (via IGF-1) on cells of various tissues. This ‘dual effector’ theory of GH action was first postulated by H. Green several decades ago (23). To exert their effects, GH and IGF-1 bind to their cognate receptors [GHR and IGF-1 receptor (IGF-1R)] respectively and activate an array of intracellular signaling systems resulting in the regulation of many genes that mediate cellular differentiation, growth, and other physiological functions (24). In adults, GH induces a rapid fat loss by stimulating lipolysis while inhibiting the lipogenic actions of insulin(25). In addition to JAK-STAT, GH action is also propagated via linkage of GHRs to several cellular effector systems, including Src/SHC, MAPK, PI3K-Akt, and PKC. Differential activation of these pathways under varied physiological conditions determines the anabolic versus catabolic actions of GH (26–32). Alterations in GH or GHR action is associated with Laron Syndrome (LS) and Acromegaly in humans.

Recent studies in human adipocytes have shown that GH suppresses the expression of lipid droplet associated protein, fat specific protein 27 (FSP27) and destabilizes Peroxisome Proliferator-Activated Receptor Gamma (PPARγ), the master transcriptional regulator of adipogenesis (33–36). This derangement stimulates excessive adipose tissue lipolysis resulting in free fatty acids flux in the systemic circulation. GH-mediated deregulation of fat metabolism, therefore, promotes the progression of insulin resistance, a prime suspect for the development of Type 2 diabetes, cardiovascular and metabolic diseases (37–40). Furthermore, GH induces insulin resistance in skeletal muscle and liver via multiple mechanisms to promote its diabetogenic actions (41–44). Based on recent in vivo and in vitro studies, this review will provide a concise summary of the molecular mechanisms underlying GH-mediated development of insulin resistance, a pre-requisite for the progression of diabetogenic actions of GH and metabolic diseases.

GH indolence in Dwarf mouse models enhances insulin sensitivity

George Snell first reported dwarfism in mice in 1929 (45). All dwarf mouse models harboring natural spontaneous homozygous mutations (Pit1 mutation in Snell dwarf, Prop-1 mutation in Ames dwarf, a missense mutation in the Growth hormone-releasing hormone in Little mice or targeted gene deletion of Growth hormone receptor gene in Laron dwarf mice) are insulin sensitive, long-lived and metabolically healthy (46–52). Interestingly, all these mice share one specific genetic component - impaired GH production/signaling/action. Since GH and insulin are counterregulatory hormones, therefore, reduced insulin secretion, low fasting insulin and glucose levels in conjunction with enhanced insulin sensitivity are the major contributing factors for the improved metabolic health and extended longevity in these dwarf mice (46, 50, 51, 53–55). Houssay demonstrated that pituitary insufficiency produces extreme insulin sensitivity in multiple animal models (cat, dog, toad, rabbit, monkey and man) which can be reversed by administration of anterior-pituitary extracts, making GH a prime suspect for its anti-insulinemic effects (56). Since Snell and Ames dwarf mice lack multiple genetic factors, namely growth hormone (GH), prolactin (PRL), and thyroid-stimulating hormone (TSH), the enhanced insulin-sensitive phenotype could not be precisely attributed to a single factor. However, in 1997 targeted disruption of the GH receptor (R) by Kopchick’s laboratory resulted in a mouse model of Laron Syndrome (LS) discussed below (52). These mice are fat and short in stature with low levels of IGF-1 and high levels of GH. Since they lack a functional GHR, they are GH resistant. They also have reduced insulin secretion, display reduced plasma glucose levels, but exhibit increased insulin sensitivity (50, 54). They are also resistant to the development of type 2 diabetes and cancer and are long lived. If fact, they are the longest-lived laboratory mouse (57, 58). This GHR−/− mouse displays a phenotype consistent with patients with LS who possess homozygous inactivation mutations of the GHR gene. These patients are short in stature and ‘chubby’ with low IGF-1, high GH and, again, are GH resistant. Remarkably, they were found to be resistant to diabetes and cancer (59, 60). Therefore, the GHR−/− mouse has helped to precisely pinpoint that GH signaling antagonizes insulin action and disruption of GH action results in low IGF-1 levels as observed in the dwarf mouse models described above. Together, the lack of GH’s direct and indirect effects is a key to their improved metabolic health and very important in their extended longevity. Interestingly, the antagonistic effects of GH and insulin have been conserved during the evolutionary process and are prevalent in many organisms, ranging from yeast to worms and mice to humans (61).

Laron syndrome versus Acromegaly and Insulin sensitivity

Laron syndrome (LS) was first described by Dr. Zvi Laron in 1966. It is a rare genetic disorder inherited in an autosomal recessive manner that is caused by inactivating mutations in the GHR gene and leads to severely depressed IGF-1 levels. LS patients are characterized by dwarfism, facial phenotype, obesity and hypogenitalism (62, 63) with elevated serum GH, but low IGF1 levels, and are GH insensitive or resistant (64, 65). Interestingly, despite having obesity, LS patients are, metabolically healthy, and their risk of developing insulin resistance, cancer, and cardiovascular complications compared to their unaffected relatives is extremely low (59, 60, 66–70). Furthermore, Dr. Jamie Guevara-Aguirre’s group reported that an Ecuadorian cohort of LS patients has significantly reduced systemic insulin concentration and a very low HOMA-IR (homeostatic model assessment–insulin resistance) index, indicating higher insulin sensitivity, and the absence of diabetes (59). Hence, anti-insulinemic actions of GH in the GHR−/− mouse mimics that in LS patients.

On the other hand, acromegaly is rare disorder that is characterized by a concomitant increase in both GH and IGF-1. Physiologically, nutrient intake, and glucose suppress GH secretion in healthy subjects but not in patients with acromegaly (71). Chronically elevated levels of GH in these patients induce lipolysis, and hepatic/peripheral insulin resistance (72–74) followed by a compensatory increase in insulin secretion to maintain euglycemia (75). Insulin resistance in patients with acromegaly is associated with reduced body fat mass (76, 77). Besides promoting insulin resistance, elevated levels of FFAs can induce corresponding lipotoxicity that can promote beta-cell exhaustion (72, 78–81). Therefore, these patients, despite having less fat and more lean mass, if left untreated can develop diabetes, cardiovascular, and other metabolic problems (82–85). The first line of treatment of these patients is removal of the GH-secreting pituitary tumors via surgery. If successful, the patients are cured. If remnants of tumor persist, medical treatment is required to decrease GH and/or IGF-1 to normal, age/sex adjusted normal levels at which point insulin sensitivity improves (86).

Interestingly, dwarf mouse models of GH inactivity (Snell, Ames, LS mouse) mimic LS patients and serve an excellent model to study the pathophysiology of GH and its actions in humans. However, recently developed GHR deficient pig model also parallel hallmarks of human LS patients pathophysiology (such as GH insensitivity, high GH and low circulating IGF1 levels, increased adiposity, reduced postnatal body and organ growth) and therefore, offer an appealing model for treatment trials and to nail down the mechanistic studies of human diseases (87). Furthermore, porcine embryonic development resembles that of humans (88), and the pig is closer to human anatomically, physiologically (89) and genetically (90) than the mouse. Finally, the porcine model could bridge the gap between proof-of-concept and discrepancies between rodent and human subject studies (87).

Dawn Phenomenon: A tug of war between GH and Insulin action

Allen Kowarski showed that Type 1 diabetic subjects display a “dawn phenomenon” defined by periodic episodes of hyperglycemia occurring in the early morning hours between 5:00 AM and 8:00 AM (91). This phenomenon reflects an early-morning increase in insulin requirement of patients with insulin-dependent diabetes mellitus (IDDM) as well as Type 2 diabetes patients (92), an effect observed in all age groups but prominently in elderly diabetic subjects (93). In healthy subjects, however, blood glucose and insulin levels remain steady throughout the night, with a slight surge in insulin secretion before dawn to suppress hepatic gluconeogenesis and prevent GH-induced hyperglycemia. The observed increase in insulin in the early morning hours despite stable levels of glucose indicates a temporally increased insulin need in nondiabetic individuals similar to that found in individuals with diabetes (94). Diabetic patients are negatively affected by the dawn phenomenon because of the dysregulated hepatic gluconeogenesis and inability of pancreatic beta-cells to produce compensatory insulin secretion (94, 95). The underlying cause for these hyperglycemic episodes remained a mystery until Gerich’s lab showed that increased plasma glucose levels significantly correlated with peak plasma GH concentration (96).

Interestingly, when nocturnal surges in GH secretion were prevented by infusion of somatostatin, plasma glucose levels and hepatic glucose production were stabilized, demonstrating that nocturnal surges in GH secretion were the prime culprit for dawn phenomenon (96). Further work by Rosenberg’s group showed that GH-induced increase in FFA level contributes to the dawn phenomenon and suppression of sleep-induced GH secretion abolished early morning hyperglycemia (97). Furthermore, GH pulses associated with an increase in endogenous glucose production and decreased rates of peripheral glucose uptake were entirely reversed by pharmacological inhibition of lipolysis with acipimox (98). Therefore, the GH-driven decrease in insulin sensitivity was predominantly due to GH’s effect on adipose tissue lipolysis (99, 100). To our knowledge, the Dawn phenomenon in mice has not been reported, probably due to their nocturnal feeding behavior and unlimited food accessibility.

Growth hormone, hyperinsulinemia and insulin resistance

In healthy human subjects, insulin is released in low-amplitude pulses and rapidly removed from the circulation, possibly to avoid chronic insulin exposure (101, 102). Studies in rats and human subjects in whom pancreatic insulin output was suppressed show that pulsatile delivery of exogenous insulin confers better glycemic control and inhibition of lipolysis over continuous insulin infusion (103, 104). During the development of insulin resistance, the body usually responds by increasing serum insulin concentration to maintain insulin signaling. However, a sustained compensatory increase in insulin concentration causes hyperinsulinemia and dampens insulin response, further making the insulin-resistant state worse (105–108). Chronic insulin exposure causes insulin receptor down-regulation and defects in the glucose transport system per se (109). Hyperinsulinemia results in diminished levels of tyrosine and serine phosphorylation of the insulin receptor. Prolonged insulin exposure alters the insulin receptor tyrosine kinase domain and or proteins interacting with this domain in a way that sets the insulin refractory state in the cells (110). Patients with acromegaly display chronically elevated levels of GH and are insulin resistant (111). Due to the antagonistic action of GH on insulin (GH promotes lipolysis while insulin promotes lipogenesis per se), chronically elevated levels of GH in patients with acromegalic are mimicked by a compensatory increase in insulin secretion leading to hyperinsulinemia and therefore, predisposing these patients to insulin resistant/diabetic state (75, 86, 111). Compensatory hyperinsulinemia in acromegalic patients impairs insulin receptor autophosphorylation significantly (110). Insulin binding studies in erythrocytes of the acromegalic patients showed a decreased binding due to a reduction in the receptor number per cell but with no alterations in the affinity of insulin to its receptor (111) and a greater decrease in insulin binding was observed in acromegalic patients with diabetes (112). Similarly, the total insulin receptor concentration per cell was decreased in proportion to the hyperinsulinemia in circulating monocytes, i.e. the receptor concentration was inversely related to the basal level of insulin, similar to that found in patients with obesity, diabetes, and insulin-secreting tumors (113).

Growth hormone and skeletal muscle insulin resistance

GH regulates myofiber number and promotes skeletal muscle development (114). Like adipose tissue, skeletal muscle also stores triglycerides, and to-date, three muscle-specific GHR knockout mouse lines have been generated to study GH function in skeletal muscle. These studies are extensively discussed in a recent review by List et al (58). Muscle-specific deletion of GHR or IGF-1R in mice leads to defective skeletal muscle development and performance (114). Furthermore, mice with skeletal muscle-specific deletion of GHR develop metabolic features, including elevated blood glucose and TG levels, marked peripheral adiposity, insulin resistance, glucose intolerance, and display increased inhibitory phosphorylation of IRS-1 (114). On the contrary, two subsequent studies contradict these findings and show that muscle specific GHRKO mouse lines (mGHRKO and MuGHRKO) show marked health improvements (115). Vijayakumar et al. have shown that skeletal muscle-specific loss of GHR signaling reduced liver and muscle triglyceride content, improved insulin sensitivity, and protected mice against high-fat diet-induced metabolic deterioration (116). List et al. show that male MuGHRKO mice have reduced levels of glucose, insulin, c-peptide, and enhanced glucose tolerance (115), while mGHRKO mice on a high-fat diet have decreased adiposity, inflammation, muscle and hepatic triglyceride content increased carbohydrate utilization, and enhanced insulin sensitivity (116). Furthermore, GH action on muscle had minimal effect on muscle strength or endurance, and disruption of GHR in muscle increased lifespan in male MuGHRKO mice (115). Therefore, it seems that the genetic background of the mouse and the promoter used to drive Cre expression in these mouse lines may be responsible for the contradictory results in these studies. Conversely, transgenic mice overexpressing GH show reduced IR levels and its tyrosine phosphorylation, reduced IRS-1 tyrosine phosphorylation, and defective PI3-kinase activation upon insulin stimulation (42). Similarly, rats treated with GH also show reduced IRS phosphorylation and impaired activation of skeletal muscle insulin signaling due to reduced PI3K activity (42, 117, 118). Furthermore, increased levels of p85α in transgenic mice overexpressing GH or human placental GH, suppress insulin sensitivity in skeletal muscle (42, 119). Interestingly, deletion of a single allele of p85α restored PI3-kinase activity and improved skeletal muscle insulin sensitivity in these mice, suggesting that GH-mediated insulin resistance in skeletal muscle is promoted via its effect on PI3-kinase activity (41).

Although GH replacement therapy (GHRT) in growth hormone deficient (GHD) patients offers clinical benefits, it may deteriorate glucose metabolism in muscles (120), impairs insulin-stimulated glucose turnover into glycolytic flux and glycogen synthesis/glucose storage (121), and promote insulin resistance (122). However, GH-induced acute skeletal muscle insulin resistance can be reversed by pharmacological inhibition of lipolysis or insulin action (123–126). GH infusion does not alter Akt (a critical mediator of the insulin signaling pathway that triggers insulin-stimulated glucose uptake via GLUT4(123)) protein expression or GLUT4 protein expression in skeletal muscles of healthy subjects (123), obese (127, 128) or GHRT patients (122). Recent evidence suggests that elevated GH levels as observed under fasting/starvation or upon GH infusion suppress pyruvate dehydrogenase (PDH) activity in muscles (74, 129). This decrease in PDH activity is the result of the metabolic shift from glucose to fatty acid as a primary source of fuel resulting in insulin resistance without altering insulin signaling in human subjects (74, 129). Several studies have shown that FFAs inhibit insulin-stimulated glucose uptake secondary to intramyocellular accumulation of di-acyl glycerol (DAGs) and ceramides, which directly impair insulin signaling (39, 130). Interestingly, GH-induced insulin resistance, despite its dependence on lipolysis, does not affect insulin signaling per se but decreases skeletal muscle glucose uptake (129). Overall, GH exerts skeletal muscle insulin resistance via multiple mechanisms, including increased p85α protein expression (in rodents), reduced IR levels and IRS-1 phosphorylation, decreased PDH activity, and preference for a shift in metabolic fuel from glucose to lipids.

Growth hormone and hepatic insulin resistance

Impaired ability of insulin to suppress hepatic glucose production and adipose lipolysis causes increased hepatic acetyl CoA and links adipose lipolysis to hepatic insulin resistance associated with obesity and T2D (131). Multiple mechanisms contribute to GH-induced hepatic insulin resistance. GH-induced adipose lipolysis alters glucose and lipid oxidation (33, 44), antagonizes insulin action in the liver (44), elicits hepatic gluconeogenesis, lowers peripheral glucose uptake and impairs liver insulin sensitivity in healthy subjects (127, 132, 133). Conversely, low levels of GH associate with hepatic lipid accumulation (134) and GH replacement has been suggested to resolve the fatty liver condition in diet-induced obese rodents and in GH-deficient patients (135). Phosphatase and tensin homolog (PTEN) is a widely known negative regulator of insulin/PI3K signaling (136). Chronic GH treatment upregulates PTEN expression to undermine hepatic insulin signaling (137). GH is thought to increase gluconeogenesis (74) but chronic GH exposure impairs gluconeogenesis (138, 139) and reduces the number of insulin receptors (IR) significantly without altering IR kinase activity in rat liver (140). In agreement with this study, the transgenic mice overexpressing GH also show reduced hepatic IR levels with 3–5 times increase in compensatory phosphorylation of basal IR and IRS-1 and PI3-kinase activity that is refractory/insensitive to further stimulation by insulin (43). Additionally, Akt expression, a downstream target of PI3K, is significantly increased in the liver of GH overexpressing mice to compensate for the defects in the upstream signaling cascade. However, Akt phosphorylation (a measure of insulin sensitivity) was severely diminished, thus resulting in reduced glucose uptake (141). Livers of Ames Dwarf mice and GHR knockout mice on the other hand have impaired GH action/signaling and therefore, exhibit increased IR levels, its cognate phosphorylation and are sensitive to insulin stimulation (142, 143). Furthermore, visceral fat transplant from GHRKO mice into normal mice increases phosphorylated insulin receptor levels and enhances hepatic insulin sensitivity (144). Similarly, blocking GH action/signaling via administration of human GH receptor antagonist, pegvisomant, shows reduction in plasma FFA levels, suppression of hepatic glucose production and improved insulin sensitivity in patients with acromegaly (145) and obese individuals (146). These studies point to the role of cross-organ talk and extrahepatic factors (GH-mediated FFAs and or other unidentified molecular entities of adipogenic origin) in the development of hepatic insulin resistance. Overall, these findings suggest that GH-mediated suppression of insulin’s antilipolytic ability is the prime cause of hepatic insulin resistance (147, 148).

Metabolic effects of GH in Adipose tissue

GH is required for the differentiation of adipocytes and contributes differentially to fat depot development at different anatomical locations (82). Since GH levels are inversely related to adipose tissue mass in both mice and humans, one would expect that GH-mediated reduction of adipose tissue would have health benefits. However, mounting evidence suggests that despite reducing fat mass, GH deteriorates metabolic health in mice and humans (82). Adipose tissue displays significant depot differences with respect to hormonal responsiveness and metabolic activity (149). For example, GHR−/− mice have increased subcutaneous adipose tissue with decreased fibrosis and enhanced insulin sensitivity (33, 150, 151). This may be due to the relative reduction in visceral fat which shows an enhanced lipolysis due to higher β-adrenergic response and reduced α-adrenergic response (152–154). GH increases lipolysis in adipocytes partly through the beta-adrenergic response (155). WAT, in response to appropriate stimuli such as cold exposure and β-adrenergic agonist, undergoes browning wherein it acts as an inducible thermogenic adipocyte (beige adipocytes). It increases UCP1 expression and energy production to mimic brown adipose tissue (BAT) function (156–159). Therefore, beige adipocytes exhibit UCP1-dependent thermogenesis, and their mitochondria use lipids or carbohydrates as substrates for energy production, similar to brown adipocytes (160). Also, exercise stimulates adrenergic response by potentiating AMP-activated protein kinase (AMPK) to induce PGC1a phosphorylation and activation resulting in mitochondrial biogenesis (161). Chronic exercise, therefore, induces browning of subcutaneous WAT and decreases body weight due to loss of fat mass (162). Interestingly, AMPK subunits and beiging markers such as uncoupling protein 1 (UCP1) and PR domain containing 16 (PRDM16) are upregulated in WAT of transgenic mice overexpressing GH, but downregulated in growth hormone receptor knockout mouse and GH-STAT5 inactivated (GHR-391 mouse) (163). As expected, GHRKO mice and GHR-391 mice are unable to respond to adaptive thermogenesis and other beiging agents (such as ADRB3 agonist treatment and FGF21 infusion), suggesting that in addition to fasting-induced adipose lipolysis, GH also integrates adrenergic input in WAT to promote the formation of beige adipocytes via STAT5 activation (163).

GH contributes to the growth and development during the adolescent years and the GH/IGF-1 axis acts as a prime mediator of insulin resistance across the entire spectrum of normal puberty (164). Furthermore, increased insulin dosage is anticipated with the onset of puberty in patients with insulin-dependent diabetes, and GH exacerbates this process (165–168). Low doses of GH-treatment in combination with diet restrictions reduce visceral adipose tissue mass, LDL cholesterol levels, triglycerides, free fatty acids, and improve insulin sensitivity and muscle mass but predisposes individuals to insulin resistance (169, 170). Chronic elevation in GH levels in patients with acromegaly suppresses the anti-lipolytic action of insulin, tipping the balance in favor of fat catabolism resulting in lipotoxicity, which promotes insulin resistance and related metabolic diseases (146, 171).

Patients with elevated levels of GH embody clinical manifestations of the diabetogenic effects of GH since they display increased rate of insulin resistance, hyperinsulinemia, and type 2 diabetes (72, 172, 173). Similarly, patients with acromegaly have elevated level of GH, and therefore, have reduced body fat and increased lean body mass but still develop insulin resistance (74, 174–177). Transgenic mice overexpressing GH mimic patients with acromegaly and have less total body fat mass and higher muscle mass than littermate controls but develop insulin resistance (178–180). Furthermore, GH administration in GHD children is associated with a redistribution of adipose tissue from an abdominal anatomical location to peripheral location, a significant reduction in abdominal adipocyte size, a significant reduction in overall basal rates of lipogenesis and a reduction in the anti-lipolytic actions of insulin (181). On the contrary, decreased GH action increases visceral as well as total adiposity as observed in LS patients (182) and mouse models of reduced GH action such as Ames dwarf mice (183) and large animal models (87). The decreased adiposity could be attributed to GH-mediated lipolysis and, to a lesser degree, in a preferential increase in intermuscular adipose tissue.

GH exerts anabolic effects through stimulation of IGF-I, insulin, and FFAs under fed conditions. Fasting however, increases GH levels in the systemic circulation, decreases insulin sensitivity, and blunts JAK-STAT pathway activation (184–187). Interestingly, the nocturnal increase in FFA during sleep is absent in GH-deficient patients, strongly linking lipolysis to GH (188). Similar to fasting, exercise also induces lipolysis due to adrenergic signaling (189). Elevated levels of GH observed during fasting reduce anti-lipolytic gene G0 switch 2 (G0S2) mRNA and protein expression but increase adipose triglyceride lipase (ATGL, a primary lipase in adipose tissue), protein activity with a concomitant rise in circulating fatty acids (190). However, both ATGL and G0S2 remain unchanged during exercise, suggesting that the ATGL-G0S2 complex is an important long-term regulator of lipolysis under physiological conditions (such as fasting) in humans (191). Recently, we have shown that GH downregulates expression of the negative regulators of lipolysis, namely, FSP27 and G0S2 (34, 36). Forced expression of FSP27 however, suppresses GH-induced adipose lipolysis and restores insulin sensitivity (34, 36). As stated above, studies in human subjects and mouse models show that GH affects adipose insulin sensitivity via multiple mechanisms including increased stimulation of adipose lipolysis (192, 193), expression of p85α in mice (194), increased activity of hormone-sensitive lipase (HSL) and ATGL in adipose tissue (36, 195–198), reduced insulin receptor expression (139), reduced adipose tissue lipoprotein lipase (LPL) activity, and impaired expression of FSP27 and G0S2 (34, 36) discussed below and shown in Figure 1. Overall, these studies provide a solid foundation and framework and support that GH-mediated lipolysis is one of the critical regulators of insulin action in both healthy subjects and diabetic patients.

Figure 1. Intracellular model of GH signaling under normal and pathological conditions in Human adipocytes:

Upon stimulation by GH binding, GHR activates Jak2. Activated Jak2 phosphorylates GHR on tyrosine residues which in turn recruit members of the STAT family of transcription factors. Additionally, Src kinase and Shc adaptor proteins are also recruited to the GHR to activate a series of signaling pathways. STATs homodimerize or heterodimerize and translocate to the nucleus to regulate an array of genes under physiological conditions. Coordinated action of STATs and other signaling pathways leads to a balanced state of physiological lipolysis to provide energy for growth, development and metabolism under normal and fasting/starvation conditions.

However, chronic levels of elevated GH (as seen in patients with acromegaly) leads to activation of the Ras-Raf-MEK-ERK pathway which dominates other signaling pathways. As a result, phopshorylated ERK translocates to nucleus and phosphorylates PPARγ. Phopshorylated PPARγ is targeted for degradation leading to downregulation of anti-lipolytic genes such as FSP27 and G0S2. Sustained levels of systemic FFAs as a result of deregulated lipolysis cause lipotoxicity and pave the way for the development of insulin resistance, diabetes and related metabolic diseases.

Transcriptional Regulation of Lipid droplet (LD) associated Proteins and mechanisms of GH-induced lipolysis

Fatty acids are stored as triglycerides in LDs in adipocytes. LDs contain various proteins on their surface (199), which are involved in lipogenesis, lipolysis, or maintaining lipid homeostasis. LD-associated proteins regulating lipolysis such as ATGL, G0S2, PLIN1, and FSP27, also known as Cell Death-Inducing DFFA-like Effector Protein C (CIDEC) are all direct transcriptional targets of master adipogenesis regulator PPARγ (200–202). PPARγ can sense and interpret fatty acid signals (203, 204). Therefore, alterations in the intracellular concentration of FFAs directly impact PPARγ-dependent gene regulation (205). Recent studies have highlighted the role of FSP27 as a GH target to induce lipolysis and insulin resistance (33–36). FSP27 is highly expressed in WAT and regulates lipid droplet dynamics and fat metabolism (206–212). Puri’s lab has recently described its major role in adipocyte lipolysis and insulin sensitivity in human adipocytes (206, 207).

Recent studies in human subjects, primary human adipocytes and 3T3-L1 adipocytes have shown that GH treatment downregulates FSP27 and G0S2 expression, resulting in LD fragmentation and systemic flux of FFAs (34, 36). Insulin decrease lipolysis via regulating FSP27 expression. FSP27 interacts with ATGL to regulate its lipolytic action in human adipocytes (206). Also, upon insulin stimulation, FSP27 recruits EGR1 at the ATGL promoter to directly inhibit ATGL transcription and its lipolytic capacity, suggesting that FSP27 acts at multiple levels to tightly regulate ATGL-mediated lipolysis (213). Mechanistically, GH binding to its receptor activates multiple signaling pathways, including the Ras-Raf-MEK-ERK1,2 pathway (26, 214–216). MEK1-mediated phosphorylation of PPARγ leads to its nuclear export and degradation (217, 218). GH-mediated activation of MEK/ERK signaling downregulates FSP27, destabilizes PPARγ, and increases HSL phosphorylation to promote lipolysis in human adipocytes (36). Forced expression of FSP27 or blocking MEK/ERK signaling stabilized PPARγ in the nucleus, improved insulin signaling and restored GH-induced lipolysis to normal levels (34, 36). Counterintuitively, GH administration also resulted in FSP27 mRNA upregulation suggestive of an additional transcription regulation mechanism for FSP27 expression (34). Such a mechanism is physiologically obligatory to optimally regulate lipolysis and maintain adipose tissue homeostasis. GH also signals through STAT5, which can homodimerize or heterodimerize with other STAT family members and translocate to the nucleus in MEK and PPARγ independent manner to regulate gene expression and mediate GH effects (26). Interestingly, inhibiting STAT5 phosphorylation with a subsequent GH stimulation nearly abolishes FSP27 mRNA expression in 3T3-L1 adipocytes (34). These results suggest that GH-mediated STAT5 phosphorylation upregulates FSP27 expression under physiological conditions to keep GH stimulated MEK/ERK signaling under check and maintain an optimal level of lipolysis. Furthermore, FSP27 is significantly reduced in the perigonadal fat and subcutaneous fat of STAT5^ΔN/ΔN-mutant mice, which express hypomorphic forms of both Stat5a and Stat5b (34). These studies provide molecular insights as to how GH-mediated activation of MEK/ERK causes destabilization of the transcription factor PPARγ resulting in the downregulation of FSP27, G0S2, and other lipid-droplet associated genes while simultaneously phosphorylating STAT5 to regulate LD homeostasis and prevent excessive lipolysis.

Concluding remarks

GH and insulin both have mitogenic and metabolic effects (Figure 2). However, whether GH and insulin actions will synergize to promote mitogenic or antagonize to promote metabolic effects, depends on the duration of exposure to the respective hormone. Multiple lines of evidence suggest that impaired adipose tissue homeostasis plays a significant role in maintaining whole-body insulin sensitivity, and GH-mediated adipose tissue lipolysis is one of the most critical regulators of insulin resistance in both healthy subjects and diabetic patients.

Figure 2. Synergistic and antagonistic effects of growth hormone and insulin signaling.

Acute GH and insulin exposure act synergistically (as shown in green) to promote normal growth, metabolism, differentiation, and development. However, chronic GH exposure antagonizes insulin action and vice-versa (shown by red arrows) to promote insulin resistance in multiple organs (Liver, skeletal muscle, and adipose tissue) via multiple mechanisms. Dysregulated GH and insulin action, under pathological conditions alter gene/protein expression and their enzymatic activity in insulin sensitive organs to promote insulin resistance that leads to the development of Type 2 diabetes and related metabolic diseases.