ReviewThe Impact of Adipose Tissueon Insulin Resistance inAcromegalyNicoleta Cristina Olarescu1,2,* and Jens Bollerslev1,2

Adipose tissue (AT) is recognized as key contributor to the systemic insulinresistance and overt diabetes seen in metabolic syndrome. Acromegaly is adisease characterized by excessive secretion of growth hormone (GH) andinsulin-like growth factor I (IGF-I). GH is known both for its action on ATand for its detrimental effect on glucose metabolism and insulin signaling. Inactive acromegaly, while body fat deports are diminished, insulin resistance isincreased. Early studies have demonstrated defects in insulin action, both at thehepatic and extrahepatic (i.e., muscle and fat) levels, in active disease. Thisreview discusses recent data suggesting that AT inflammation, altered ATdistribution, and impaired adipogenesis are potential mechanisms contributingto systemic insulin resistance in acromegaly.

Acromegaly and AT HomeostasisAcromegaly is a rare disease characterized by excessive secretion of GH, commonly caused bya pituitary adenoma, which results in constitutively increased levels of IGF-I. In acromegalyprolonged exposure to GH and IGF-I causes systemic manifestations, including insulin resis-tance and diabetes mellitus. In the general population, insulin resistance and diabetes are mostoften associated with increased body fat, a hallmark of metabolic syndrome [1]. However, inactive acromegaly, despite a favorable body composition (decreased body fat and increasedmuscle mass), patients often present with insulin resistance and overt diabetes [2–4]. Themechanisms involved in the appearance of insulin resistance/diabetes in acromegaly are poorlyunderstood, but previous studies indicate that the liver and skeletal muscles are likely tocontribute. GH is a counter-regulatory hormone that antagonizes the effects of insulin, whereasIGF-I is an insulin sensitizer [5,6]. Thus, GH and IGF-I exert opposite effects on insulin homeo-stasis, and the clinical relevance is orchestrated by the delicate balance between the effects ofthese two hormones on target tissues.

Adipose tissue has recently been established as a central player in energy and glucosemetabolism, and different mechanisms have been proposed to explain the role of AT in insulinresistance [7–10]. However, a question arises as to whether these mechanisms, primaryidentified in studies of obesity and/or diabetes, are applicable in a disease as acromegalywhere AT mass is markedly decreased. Because GH receptors are abundant in fat, AT isresponsive to GH stimulation [11–14]. In fact, the most prominent metabolic effect of GH is amarked increase in lipolysis and free fatty acid (FFA) levels [5].

The purpose of the present review is to highlight the evidence regarding the specific role of AT oninsulin resistance in acromegaly, in the context of the recently described advances on the role ofAT inflammation, the release of FFAs, ectopic fat deposition, and the alteration of AT distribution.Fat is one of the most affected tissues in active acromegaly, and understanding the mechanisms

TrendsDespite decreased body fat andincreased muscle mass, insulin resis-tance, and overt diabetes are presentin acromegaly. GH antagonizes insulinaction and promotes insulin resistance.

Early evidence supports a role for liverand skeletal muscles, but later studiesunderline AT as one of the contributorsto the development of insulin resis-tance in acromegaly.

AT inflammation, lipolysis and releaseof free fatty acids, ectopic or altered fatdeposition and distribution, andimpaired adipogenesis are the mainmechanisms to explain AT-driven insu-lin resistance in active acromegaly.

Treatment modalities impact differentlyon glucose metabolism, but not on ATdeposition that is predominantly in thetrunk and visceral depots.

An individualized approach for themanagement of diabetes based onthe mechanisms of AT-driven insulinresistance would be desirable in thefuture.

1Section of Specialized Endocrinology,Department of Endocrinology, OsloUniversity Hospital, Rikshospitalet,Oslo, Norway2Faculty of Medicine, University ofOslo, Norway

*Correspondence: nicola@ous-hf.no(N.C. Olarescu).

226 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 http://dx.doi.org/10.1016/j.tem.2016.02.005

© 2016 Elsevier Ltd. All rights reserved.

by which AT contributes to the systemic insulin resistance might lead to novel potentialtherapeutic strategies for diabetes in these patients. Moreover, the hyperinsulinemia presentin the early stages of insulin resistance potentiates the liver sensitivity to GH and leads to higherIGF-I production. Thus finding specific targets, such as AT, to reduce the insulin resistance inactive disease may also lead to improved disease control.

The Role of AT on Systemic Insulin ResistanceInsulin increases glucose transport and breakdown in fat and skeletal muscles, and stimulatesglycogen synthesis and suppresses glucose production in the liver (Figure 1). Insulin resistancein skeletal muscles and liver, and b cell failure, represent the core pathophysiologic defects intype 2 and secondary diabetes mellitus [9]. The insulin-dependent glucose disposal in AT isrelatively small compared to skeletal muscles which account for 75% [6,15]; however, as ametabolic integrator and a key endocrine organ, AT interacts with other insulin-responsivetissues and contributes to impaired insulin action via a variety of mechanisms including secretionof an altered adipokine profile, release of FFAs and lipotoxicity, local and systemic inflammation,and ectopic/altered fat accumulation or impaired adipogenesis [8–10,16].

AT-Driven Systemic InflammationThe association between systemic low-grade, chronic inflammation (metaflammation) andinsulin resistance is widely accepted [17]. As a metabolically active tissue, AT secretes an arrayof inflammatory factors (e.g., TNF-/, MCP1, IL-6, IL-8, IL-1b, RBP4) that crosstalk with othercells in an autocrine, endocrine, and paracrine manner [7,8]. In addition, other mechanisms such

↓Glucose uptake↓Lipogenesis↑Lipolysis

Growth hormone

↓Glucose uptake↓Glucose oxida�on?↑Lipid deposi�on?↑Lipid oxida�on?

↑Glucose uptake↑Lipogenesis↓Lipolysis

Insulin

↑Glucose uptake↑Glycogen synthesis

↑Glucose cycling↑Glycogenolysis↑Gluconeogenesis↓De novo lipogenesis↓Lipid uptake↑Ketogenesis?

↑Glycogen synthesis↓Glycogenolysis↓Gluconeogenesis↑De novo lipogenesis

Figure 1. The Metabolic Effects of Growth Hormone (GH) and Insulin on Carbohydrate and Lipid Metabolism. GH antagonizes the action of insulin oncarbohydrate and lipid metabolism. The overall effect is to increase endogenous glucose production and to reduce glucose disposal. The effects on carbohydrate (darkblue) and lipid (light blue) metabolism are presented. Symbols: ", increase; #, decrease.

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 227

as adipocyte senescence, necrosis, and death, as well as increased extracellular matrixdeposition, may potentiate the inflammatory state of AT [18].

The activation of inflammatory pathways (i.e., JNK and IKKb) within metabolic tissuespotentiates insulin resistance via serine phosphorylation of IRS proteins and inhibition ofinsulin signaling [19]. Furthermore, inflammatory cytokines attract immune cells such as M1macrophages to local tissues as well as systemically, further exacerbating insulin resistancein liver, muscle, and brain. FFAs and other lipids released through lipolysis have also beenshown to stimulate inflammation in pancreatic islet b cells, adipocytes, and macrophages[20–24].

The Role of Adipokines in Insulin ResistanceAT synthesizes and releases a variety of adipokines that improve (e.g., adiponectin, omentin-1,vaspin, chemerin, FGF21) or impede (e.g., RBP4, resistin, TNF/) insulin sensitivity [25]. Inaddition, the adipokines have other diverse roles, such as influencing feeding behavior, hepaticfibrosis, cardiovascular function, and vascular tone. Altered adipokine production and/orsecretion owing to AT dysfunction contributes to the pathogenesis of systemic insulinresistance.

Ectopic Lipid-Induced Insulin ResistanceDefective insulin signaling in AT leads to increased lipolysis and deficient lipid storage, thuselevating circulating FFAs and promoting extra-adipose lipid deposition. Recent studies presentcompelling evidence that ectopic lipid accumulation in liver and skeletal muscles is involved in thepathogenesis of insulin resistance [26]. The specific signaling pathways are slowly being eluci-dated. An early hypothesis postulated that lipids inhibit key glycolytic enzymes in muscle, thusleading to impaired insulin-stimulated glucose use. However, recent work has shown that theintracellular accumulation of FFAs metabolites (i.e., diacylglycerols) in muscle and liver inducesthe activation of a novel protein kinase C pathway with subsequent impairments in insulin signaling[15,26,27].

Importance of AT DistributionNumerous studies have shown that central (abdominal) obesity has a particularly adverseeffect on insulin sensitivity [28,29]. In humans, the subcutaneous (SAT, situated under theepidermis) and intra-abdominal (e.g., visceral-VAT, omental, and retroperitoneal) adiposetissue depots represent the main compartments for fat storage. Lately, other fat depots(e.g., intramuscular, pericardial, and bone marrow) have been demonstrated to have a rolein glucose metabolism; however, many questions remain unanswered and this is an area ofactive investigation [30].

There is evidence suggesting that SAT is not directly implicated in the etiology of insulinresistance or may even have a protective role by functioning as a sink for excess circulatingfat [7,28,31]. Conversely, VAT is associated with an elevated risk for diabetes, hypertension,atherosclerosis, dyslipidemia, and cancer. Currently, there are two main hypotheses to explainthe negative impact of VAT on glucose tolerance, insulin resistance, and lipid metabolism. First,‘the portal hypothesis’ holds that increased lipolytic activity in VAT and consequent higherdelivery of FFAs through the portal vein to the liver negatively affects liver function [32]. Second,VAT produces more proinflammatory cytokines and less anti-inflammatory adiponectin, thuscontributing to systemic low-grade inflammation [33]. There are regional differences between thevarious fat compartments that contribute to their differential functions. In this regard SAT differsfrom VAT with respect to the characteristics of adipocyte precursors, including replication,adipogenesis, apoptosis and senescence, adipocyte morphology, inflammatory cell infiltration,vascularization, and innervation [31].

228 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4

Mechanisms Underlying Insulin Resistance in AcromegalyActive acromegaly is associated with insulin resistance and diabetes. While the mechanisms toexplain the development of insulin resistance in acromegaly are not yet clearly defined, there isevidence suggesting that excess GH directly impairs insulin signaling in responsive cells anddiminishes glucose uptake, promotes lipolysis and FFAs release, and increases hepatic glucoseproduction. Other putative GH-mediated mechanisms that alter insulin sensitivity might involvethe production of inflammatory adipokines causing local and systemic inflammation, decreasedadipogenesis and lipogenesis, and the deposition of ectopic lipids in insulin-sensitive tissuessuch as skeletal muscles.

Historical Evidence from Clamp StudiesEarly clamp studies showed that patients with acromegaly required a lower glucose infusion rateto maintain euglycemia, had impaired suppression of glucose production at physiological insulinconcentrations, and lower glucose utilization than healthy controls, supporting a defect in bothhepatic and extrahepatic (i.e., muscle and fat) insulin action in active disease [34]. At the hepaticlevel, even in the presence of normal glucose tolerance, the chronic GH excess is accompaniedby an increase in glucose cycling (whereby phosphorylated glucose is not stored as glycogen oroxidized, but is instead rapidly dephosphorylated and released back into the bloodstream) ratherthan by glucose production [35]. Furthermore, it has been shown that, both in the fasting stateand during glucose load, there was an impairment of muscle glucose uptake and non-oxidativeglucose utilization (i.e., glycogen synthesis), but no disturbance of glucose oxidation [36].

The first evidence that AT is resistant to insulin action in acromegaly was identified using an invitro approach of culturing isolated mature adipocytes from patients with acromegaly andcontrols [37]. The authors demonstrated that, in the fasting state, the effect of insulin on glucoseutilization was almost completely abolished while insulin binding was somewhat reduced,whereas the antilipolytic effect of the hormone remained intact. Furthermore, after oral glucoseingestion, AT resistance to insulin was further enhanced and also involved insulin-inducedantilipolysis. These findings indicated a post-binding defect in insulin-induced AT glucoseoxidation in acromegaly, while the antilipolytic effect of insulin was not altered in the fastingstate [37].

GH Lipolytic/Antilipogenic Effect in AT and the Release of FFAsThe primary effect of GH in the basal state is to promote lipid mobilization and oxidation, and toinhibit triglyceride accumulation in AT, effects that are contrary to the anabolic actions of insulin(Figure 1). GH stimulates directly lipolysis in humans, but the specific mechanisms are not entirelyclear. However, results from animal studies indicate that the lipolytic effect of GH involvesincreased responsiveness to b-adrenergic signaling leading to increased adenylyl cyclaseactivation. This model has been supported by human studies showing that the in vivo lipolyticeffect of GH is counteracted by acipimox, an adenylyl cyclase inhibitor [38]. Interestingly,stimulation with GH does not increase lipolysis in explants of human AT or isolated humanadipocytes, suggesting that innervation is needed for GH to promote lipolysis [39].

GH inhibits lipogenesis by attenuating the expression of fatty acid synthase in AT and inhibitinglipoprotein lipase activity, the main enzyme involved in the intravascular release of FFAs fromtriacylglycerol [40,41].

Although there is no doubt that acute, limited, administration of GH to healthy subjects increasesplasma FFAs [42], there are fewer reports showing elevated FFA concentrations in a setting ofconstant increased GH secretion, as in acromegaly [43,44]. However, a significant decreaseof circulating FAAs accompanied by an improvement of insulin sensitivity is registered aftertreatment for acromegaly, supporting that the release of FFAs due to the lipolytic effect of GH

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 229

might negatively impact on systemic insulin sensitivity [45]. The proposed mechanism by whichthe rise in FFAs induced by GH in human subjects alters insulin sensitivity in skeletal muscles is viainactivation of the pyruvate dehydrogenase complex (involved into transforming pyruvate intoacetyl-CoA, that further is used in the citric acid cycle), thus impairing glucose oxidation in themitochondria [46]. However, ectopic lipid deposition in skeletal muscles and the liver may representanother mechanism, but this remains to be demonstrated in acromegaly. It is of interest that aprevious study from the early 1990s did not observe any difference in circulating FFA levelsbetween patients with active acromegaly and controls, suggesting that insulin resistance inacromegaly may be present even in the absence of increased circulating FFA levels [36].

AT Distribution and Ectopic Fat Accumulation in AcromegalyFat mass is decreased and lean mass is increased in patients with active acromegaly versusmatched controls, and treatment results in an increase of fat mass and a decrease in lean masswithout altering total body weight [4,47–52]. Albeit decreased, the AT mass is positivelyassociated with insulin resistance, suggesting that AT possesses increased metabolic activityin active acromegaly [4,47–49]. Although both SAT and VAT are decreased in active disease, fatdeposition is predominantly in the trunk and visceral depots following treatment, and patientswith less fat accumulation in the trunk have a better reduction in insulin resistance [49,52].

It could be tempting to hypothesize that ectopic fat deposition in the liver and muscle may representone of the mechanisms to explain the insulin resistance seen in acromegaly. However, there isevidence against hepatic lipid accumulation as one of the underlying mechanisms; studies haveshown that GH signaling limits hepatic lipid accumulation in mice [53,54], liver lipid content issignificantly lower in patients versus healthy subjects [55], and treatment with pegvisomant (a GHreceptor blocker) leads to increased hepatic lipid accumulation [56]. Regarding muscle, inter-muscular AT (located between muscle groups and beneath the muscle fascia) is increased inacromegaly [50], whereas intramyocellular lipid content is not changed [52]. It remains to bedetermined if increased intermuscular AT has any role in insulin resistance in acromegaly.

The Role of GH in Adipocyte DifferentiationAs mentioned previously, failure of adipocyte precursors to differentiate results in an impairedcapacity of AT to expand, thus causing ectopic fat deposition and insulin resistance. Previousstudies investigating the role of GH on pre-adipocyte differentiation were inconclusive, particu-larly because the effects seen in cell lines differed from observations in primary cells [57,58].Moreover, the existence of non-adipose cells (e.g., fibroblasts, endothelial cells, macrophages)in the stromal vascular fraction (SVF), that are known to impact on adipogenesis [59], weakensthe interpretation of the results. An additional observation is that adipocytes secrete IGF-I uponGH stimulation, and thus the individual effects of these two hormones could not be separated.Culturing SVF cells isolated from human SAT in the presence of GH led to a reduced number ofdifferentiating cells as well as suppressed expression and enzymatic activity of glycerol-3-phosphate dehydrogenase, corresponding to impaired adipogenesis [58]. Similar results wereobserved in SVF cells from epididymal depots of rats, SAT of cows, and porcine perirenaldepots. Recently, it was shown that lack of GH receptor signaling increases adipogenesis in apure population of mouse AT mesenchymal stem cells (MSCs) [57]. Thus, impaired adipo-genesis in the presence of excess GH might represent a novel mechanism to further explain GH-induced insulin resistance.

AT Inflammation and Adipokine Release in AcromegalyChronic, low-grade, non-resolving inflammation in metabolically active tissues is a hallmark ofinsulin resistance [24]. Studies to assess the inflammatory status of AT in patients withacromegaly are lacking. Moreover, the majority of in vitro data have been obtained from miceand results described in human primary adipose cells culture are limited [60]. It was shown that

230 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4

GH increases gene expression of nicotinamide phosphoribosyltransferase (NAMPT)/visfatin,monocyte chemotactic protein 1 (MCP1), vascular endothelial growth factor-A (VEGF-A) andinterleukin-6 (IL-6) in human adipocytes of subcutaneous and visceral origin [4,48]. Thesecytokines may then recruit various inflammatory cells and promote angiogenesis, leading toa more ‘inflamed’ fat. A recent study corroborated this idea and demonstrated that SAT andmesenteric AT depots in mice with excess GH action showed greater macrophage andregulatory T cell infiltration [61]. Taken together, these observations propose a model in whichacromegaly is accompanied by an increased AT inflammatory phenotype (Figure 2, Key Figure).

A variety of adipokines have been measured in active acromegaly [4,47–49,62–64]. Adiponectinrelease has been shown to be decreased, stay unchanged, or even increased [47,63,64]. Thediscrepancies are difficult to reconcile, but a recent study showed that the murine adiponectingene promoter has a STAT5 binding site that is occupied following GH stimulation and leads todecreased transcriptional activity [65].

Key Figure

Model of Adipose Tissue (AT) Appearance in Active Acromegaly

Healthy adipose �ssue

↔ Recruitment of inflammatory cells

↔ Angiogenesis

↔ Inflamma�on

↔ Extracellular matrix deposi�on

↔ Adipogenesis

An�-inflammatory adipokines(e.g., adiponec�n)

Local/systemic insulin sensi�vity

Adipose �ssue in acromegaly

Inflammatory adipokines(e.g., NAMPT, VEGF-A, MCP1, IL-6)

Local/systemic insulin resistance

↑Recruitment of inflammatory cells

↑Angiogenesis

↓Adipogenesis

Adipocyte

Lymphocyte

Fibroblast

Dendri�c cell

Blood Vessel

Collagen

Macrophage

MSC

Key:

↑Extracellular matrix deposi�on?

↑Inflamma�on

Figure 2. In healthy AT there is a balanced between adipocytes, inflammatory cells, and blood vessels, whereas in active acromegaly under growth hormone (GH)stimulation the adipocytes increase the secretion of inflammatory adipokines (e.g., NAMPT, VEGF-A, MCP1, IL-6) that recruit inflammatory cells and promoteangiogenesis, finally leading to a more ‘inflamed fat’ that mediates both local and systemic insulin resistance. In addition, impaired adipogenesis and a possibleincrease of extracellular matrix deposition will likely contribute.

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 231

Leptin was found to be decreased in active acromegaly, reflecting the decreased amount of AT[47], while resistin does not appear to be altered [60]. NAMPT/visfatin is a serum adipokine that isincreased in active acromegaly [4,66]. This is of interest because NAMPT seems to be involved inthe switch from glycolysis to fatty acid oxidation via NAD+ production and the activation of twoNAD-dependent deacetylases, SIRT1 and SIRT6 [67]. If this mechanism also operates in skeletalmuscles or adipocytes of humans, it might explain the well-described GH ‘sparing of glucose atthe expense of oxidizing fatty acids’ effect [5]. Another observation to consider is that NAMPT,again through production of NAD+, is linked to the regulation of circadian clock genes, which areinvolved in the regulation of metabolism and energy homeostasis [68]. Thus, a disturbedcircadian clock as a result of increased NAMPT might contribute to disturbed glucose andenergy homeostasis in acromegaly.

Vaspin and omentin, two visceral adipokines, were recently assessed in acromegaly, before andafter treatment [49]. Both adipokines were modulated by glucose and insulin levels, and thuswere not valid estimators of VAT mass. Moreover, because vaspin showed a positive associationwith GH at baseline, and because the decrease in vaspin levels during treatment stronglycorrelated with improvement of insulin resistance, it is likely that serum vaspin represents abiomarker reflecting disease activity as well as insulin resistance.

Crosstalk Between GH and Insulin SignalingGH binds to its preformed GH receptor dimer and activates it, leading to signal transductionthrough activation of the Janus-family tyrosine kinase-2 (JAK2) and Src family kinases (Figure 3)[69]. These events recruit and/or activate a variety of signaling pathways, including STATs, SHC,and IRS proteins, that finally lead to transcription of target genes. The pathway activation isreversed by the action of specific tyrosine phosphatases and receptor degradation, assisted byinduced SOCS proteins [70,71].

Insulin binds to and activates the insulin receptor, which belongs to a subfamily of receptortyrosine kinases including the IGF-I receptor. Among the intracellular substrates of the insulin/IGF-I receptors are IRS proteins, SHC, and CAP/Cbl (Figure 3) [6]. Two major pathways of theinsulin IRS signaling cascade are the Ras/ERK and PI3K pathways. In insulin target cells, such asadipose and myocytes, insulin-stimulated glucose transport is mostly mediated by PI3K activity,but other signals seem to be required (e.g., CAP/Cbl pathway).

The GH and insulin signaling pathways share some downstream intracellular factors (Figure 3).In the short term, GH produces insulin-like effects elicited by stimulation of insulin signaling viaJAK2/IRS, JAK2/SHC/Ras, and Src/PLCg/Ras. By contrast, increased expression of SOCSgenes has been shown to negatively regulate both GH and insulin signaling [72]. Studies in ATof transgenic mice expressing bovine GH and 3T3-L1 adipocytes showed that chronic GHexposure leads to inhibition of PI3K activity via increased p85/ expression [73,74]. However,a study performed in healthy subjects showed evidence against a direct effect of GH on PI3Kactivity in skeletal muscles and fat biopsies [12,75]. These discrepant results may reflectspecies differences or technical issues. Because GH and insulin have some commonintracellular substrates, such as IRS and SHC, competition for these substrates mightreciprocally perturb their signaling in the state when both hormones are simultaneouslyincreased [70].

Treatment Modalities in AcromegalyThe Impact on Glucose MetabolismControlling the biochemical abnormalities seen in acromegaly is associated with improvedglucose metabolism, but different treatment modalities may have a differential impact on glucosehomeostasis in view of their particular mechanisms of action. In this regard, somatostatin

232 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4

analogs (SA) (i.e., octreotide, lanreotide, and pasireotide) act not only centrally by inhibiting GHrelease, and consecutively decreasing insulin resistance, but also in the periphery on the keyorgans for glucose metabolism. For example, in the pancreas, SAs decrease insulin and to alower extent glucagon secretion, thus potentially impairing glycemic control [76]. Indeed, severalstudies have shown a detrimental effect of SA treatment on glucose metabolism in patients withactive acromegaly, but other studies failed to show any difference in insulin resistance [77,78].This suggests that the balance between the positive effects mediated by the decrease in GHlevels, and the negative effects on pancreatic b cells, determines whether treatment of acro-megaly with SAs worsens glucose metabolism or not.

Glucosetransport

Insulin/IGF-Ireceptors

Glucose transportlipid synthesis

Prolifera�ondifferen�a�on

Proteinsynthesis

Glycogenesis

Transcrip�on oftarget genes

Growth andmetabolism

GHreceptor

IRS SHC

SrcJAK2

SOCS

SHC

CAP

PI3Kp85

aPKCAkt

Ras

GSK-3β

PLCγ

mTOR

Cbl

STAT1

STAT3

STAT5

MEK

ERK

p110

Figure 3. Crosstalk Between Growth Hormone (GH) and Insulin Signaling Pathways. GH binds to its preformed receptor dimer and activates it, leading tosignal transduction through activation of JAK2 and Src [69–71]. These events recruit and/or activate a variety of signaling pathways, including STATs, SHC, and IRSproteins (green lines). Among the intracellular substrates of the insulin/IGF-I receptors are IRS proteins, SHC, and CAP/Cbl. Two major pathways of IRS signaling cascadeare Ras/ERK and PI3K [6] (blue lines). The GH and insulin signaling pathways share some downstream intracellular factors (thick lines). In the short term, GH producesinsulin-like effects elicited by stimulation of insulin signaling via JAK2/IRS, JAK2/SHC/Ras, and Src/PLCg/Ras. Increased expression of SOCS genes has been shown tonegatively regulate both GH and insulin signaling [72] (red lines). Studies in mice showed that chronic GH exposure leads to inhibition of PI3K activity via increased p85/expression [73,74] (red broken line), but this mechanism was not present in humans [12,75]. GH and insulin have some common intracellular substrates, (e.g., IRS andSHC), and competition for these substrates might reciprocally alter their signaling. There are differences between various cell types depending on the relative expressionof signaling components and duration of GH and insulin simultaneous elevation. Abbreviations: aPKC, atypical protein kinase C; ERK, extracellular signal-regulated kinase(also known as mitogen-activated protein kinase); GSK-3b, glycogen synthase kinase-3b; IRS, insulin receptor substrate; JAK2, Janus kinase 2; MEK, dual specificitymitogen-activated protein kinase kinase 2; mTOR, mammalian target of rapamycin; PI3K, phosphoinositide 3-kinase; PLCg, phospholipase Cg; RAP, Ras-related protein;SHC, SH2-domain containing transforming protein; SOCS, suppressor of cytokine signaling; Src, proto-oncogene protein tyrosine kinase Src; STAT, signal transducerand activator of transcription.

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 233

Insulin increases GH hepatic sensitivity. This was primarily described in patients with type 1diabetes. These patients presented with GH resistance (i.e., high GH levels and low IGF-I levels)that was reversed by intraportal insulin administration, but not by subcutaneous administration,thus clearly showing that hepatic insulin levels are positively correlated with GH effects in the liver[79]. Later, it was demonstrated that insulin increases the biosynthesis and cell surfacetranslocation of GH receptors, that potentially translates into increased GH binding and signal-ing, and higher IGF-I production [80]. These results imply that, in addition to the possibleinhibitory effect on GH release from pituitary adenomas, SAs diminish hepatic IGF-I productionalso by decreasing pancreatic insulin secretion. The latter mechanism is more prevalent in thecase of pasireotide (a SA multireceptor ligand with affinity for somatostatin receptors 1, 2, 3, and5) owing to an additional effect on somatostatin receptor 5 which is known to have an importantrole in mediating pancreatic insulin secretion [81].

It is generally accepted that the GH receptor blocker, pegvisomant, has beneficial effects onglycemic control in active acromegaly. By blocking GH receptors competitively, pegvisomantimproves peripheral and hepatic insulin sensitivity [45,82]. Interestingly, the effect of pegvisom-ant is tissue-dependent, with AT, kidneys, and skeletal muscles requiring less pegvisomant toreduce GH actions compared to the liver, where more pegvisomant is necessary to reduce IGF-Iproduction [83]. The differences in GH sensitivity in various tissues might be explained by thedifferent densities of GH receptors in these organs.

The Impact on Body FatTreatment for acromegaly promotes changes of body composition. Indeed, a few studies haveshown that, despite a similar body weight 1 year after transsphenoidal adenomectomy (surgeryto remove the pituitary adenoma), total fat mass significantly increased whereas muscle, skin,and visceral organs decreased (i.e., the organs themselves decreased, not their fat content)[4,49,51,52,84]. Fat accumulation tended to be more pronounced in men than in women.Moreover, SAT and VAT depots increased, while fat depots in limbs, head, and neck regionsdecreased [84].

The very first study assessing changes of body composition after SA treatment lasted only 4weeks and did not demonstrate any changes in fat mass, thus suggesting that a longer time isneeded to observe an impact on AT [85]. In fact, a later study demonstrated that fat massincreased after 12 weeks of SA treatment [86]. Recently, it has been shown that pegvisomantincreased intra-abdominal fat without significant alterations in SAT after 6 months of therapy [87].Co-treatment with SAs and pegvisomant for 24 weeks versus SA monotherapy did not impactadversely on the classical fat disposition, but increased liver fat and decreased intramyocellularfat [56]. This effect is most probably caused by the antagonizing effect of pegvisomant on GHsignaling in liver and muscle.

Concluding RemarksEarlier studies have demonstrated that insulin resistance in acromegaly is located at both hepaticand extra-hepatic levels. Among the extra-hepatic tissues, AT is an important candidate.Although decreased, AT should be considered a contributor to the systemic insulin resistancein active acromegaly. During chronic GH stimulation, as in acromegaly, AT alters glucosehomeostasis by releasing inflammatory adipokines that can recruit inflammatory cells andpromote low-grade chronic inflammation in a paracrine/autocrine or endocrine matter. Theincreased lipolysis in acromegaly results in high circulating FFA levels that negatively impact oninsulin action in muscle and liver. Lastly, at least in mouse and cell culture models, excess GHmight also impair adipocyte differentiation, thus inhibiting AT expansion that would then lead toectopic fat deposition and consequent insulin resistance. Although this is an attractive hypothe-sis, this mechanism remains to be proven in humans. Treatment modalities impact differently on

Outstanding QuestionsHow can we elucidate the role of FFAsand ectopic fat deposition on insulinresistance in acromegaly? A studydesign involving quantification of intra-myocellular and intrahepatic lipid con-tent, coupled with clamp studies toassess insulin resistance, and the mea-surement of circulating FFAs in patientswith acromegaly and age-, gender-,body composition-matched controlswill help to resolve this issue.

What are the molecular characteristicsof AT in acromegaly? We can addressthis by using gene profiling analysesand histological studies of AT biopsies,from SAT and VAT depots, taken frompatients with active disease before andafter successful treatment. Assessingthe adipocyte senescence, necrosis,and death, as well as extracellularmatrix deposition in these biopsies, willgive important information.

Does GH affect the ability of humanMSCs to differentiate towards theadipogenic lineage? If GH decreasesadipogenesis at the expense of osteo-blastogenesis in human MSCs, it wouldcontribute to our understanding of thedecreased fat and increased bone massin acromegaly.

How relevant are the results describedin mice, with varying levels of GH sig-naling and depot-specific differences inAT, for human acromegaly?

Is brown and beige/brite AT affected,and could it be a target for medicaltreatment in active acromegaly?

Would it be useful to apply an individu-alized approach for the management ofdiabetes in patients with acromegaly –

based on the identified mechanisms ofAT-driven insulin resistance?

234 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4

glucose metabolism, but not on AT deposition that is predominantly in the trunk and visceraldepots. After successful treatment the storage of fat in the central depots seems appropriatebecause these depots were the most affected in active disease. However, the question remainsas to whether the properties of the newly acquired AT are similar to those of healthy fat.

Possible directions to further deepen our understanding of the impact of AT on glucosemetabolism in acromegaly are summarized in the Outstanding Questions. It is uncertain towhat extent the medical therapies used in the treatment of obesity-associated diabetes can beused for the successful treatment of diabetes in acromegaly. Identifying the specific mechanismsof AT-driven insulin resistance in acromegaly will also set the stage for personalized diabetestreatment. Decreasing the insulin resistance in acromegaly will not only improve the cardiovas-cular risk and morbidity of the patients but also potentially help to achieve better IGF-I control viaeffects on the liver.

AcknowledgmentsWe would like to thank Thor Ueland for valuable discussions and Emil Păun for assistance with the Figures. This work was

supported by a studentship from the Faculty of Medicine, University of Oslo. Work in our department was supported by

unrestricted research grants from Pfizer and Novartis.

References1. Vega, G.L. et al. (2006) Influence of body fat content and distribu-

tion on variation in metabolic risk. J. Clin. Endocrinol. Metab. 91,4459–4466

2. Fieffe, S. et al. (2011) Diabetes in acromegaly, prevalence, riskfactors, and evolution: data from the French Acromegaly Registry.Eur. J. Endocrinol. 164, 877–884

3. Katznelson, L. et al. (2014) Acromegaly: an endocrine societyclinical practice guideline. J. Clin. Endocrinol. Metab. 99, 3933–3951

4. Olarescu, N.C. et al. (2012) Adipocytes as a source of increasedcirculating levels of nicotinamide phosphoribosyltransferase/visfa-tin in active acromegaly. J. Clin. Endocrinol. Metab. 97, 1355–1362

5. Moller, N. and Jorgensen, J.O. (2009) Effects of growth hormoneon glucose, lipid, and protein metabolism in human subjects.Endocr. Rev. 30, 152–177

6. Saltiel, A.R. and Kahn, C.R. (2001) Insulin signalling and theregulation of glucose and lipid metabolism. Nature 414, 799–806

7. Rosen, E.D. and Spiegelman, B.M. (2014) What we talk aboutwhen we talk about fat. Cell 156, 20–44

8. Lafontan, M. (2014) Adipose tissue and adipocyte dysregulation.Diabetes Metab. 40, 16–28

9. Defronzo, R.A. (2009) Banting Lecture. From the triumvirate to theominous octet: a new paradigm for the treatment of type 2diabetes mellitus. Diabetes 58, 773–795

10. Gustafson, B. et al. (2015) Insulin resistance and impaired adipo-genesis. Trends Endocrinol. Metab. 26, 193–200

11. Fisker, S. et al. (2004) Gene expression of the GH receptor insubcutaneous and intraabdominal fat in healthy females:relationship to GH-binding protein. Eur. J. Endocrinol. 150,773–777

12. Nielsen, C. et al. (2008) Growth hormone signaling in vivo in humanmuscle and adipose tissue: impact of insulin, substrate back-ground, and growth hormone receptor blockade. J. Clin. Endo-crinol. Metab. 93, 2842–2850

13. Vendelbo, M.H. et al. (2015) GH signaling in human adipose andmuscle tissue during ‘feast and famine’: amplification of exercisestimulation following fasting compared to glucose administration.Eur. J. Endocrinol. 173, 283–290

14. Richard, A.J. and Stephens, J.M. (2011) Emerging roles of JAK-STAT signaling pathways in adipocytes. Trends Endocrinol.Metab. 22, 325–332

15. Samuel, V.T. and Shulman, G.I. (2016) The pathogenesis of insulinresistance: integrating signaling pathways and substrate flux. J.Clin. Investig. 126, 12–22

16. Kahn, S.E. et al. (2014) Pathophysiology and treatment of type 2diabetes: perspectives on the past, present, and future. Lancet383, 1068–1083

17. Gregor, M.F. and Hotamisligil, G.S. (2011) Inflammatory mecha-nisms in obesity. Ann. Rev. Immunol. 29, 415–445

18. Palmer, A.K. et al. (2015) Cellular senescence in type 2 diabetes: atherapeutic opportunity. Diabetes 64, 2289–2298

19. Odegaard, J.I. and Chawla, A. (2013) Pleiotropic actions of insulinresistance and inflammation in metabolic homeostasis. Science339, 172–177

20. Nguyen, M.T. et al. (2007) A subpopulation of macrophagesinfiltrates hypertrophic adipose tissue and is activated by free fattyacids via Toll-like receptors 2 and 4 and JNK-dependent path-ways. J. Biol. Chem. 282, 35279–35292

21. Boni-Schnetzler, M. et al. (2009) Free fatty acids induce a proin-flammatory response in islets via the abundantly expressed inter-leukin-1 receptor I. Endocrinology 150, 5218–5229

22. Yeop Han, C. et al. (2010) Differential effect of saturated andunsaturated free fatty acids on the generation of monocyte adhe-sion and chemotactic factors by adipocytes: dissociation of adi-pocyte hypertrophy from inflammation. Diabetes 59, 386–396

23. Pal, D.et al. (2012) Fetuin-A acts as an endogenous ligand of TLR4 topromote lipid-induced insulin resistance. Nat. Med. 18, 1279–1285

24. Donath, M.Y. (2014) Targeting inflammation in the treatment oftype 2 diabetes: time to start. Nat. Rev. Drug Discov. 13, 465–476

25. Ohashi, K. et al. (2014) Role of anti-inflammatory adipokines inobesity-related diseases. Trends Endocrinol. Metab. 25, 348–355

26. Samuel, V.T. and Shulman, G.I. (2012) Mechanisms for insulinresistance: common threads and missing links. Cell 148, 852–871

27. Samuel, V.T. et al. (2010) Lipid-induced insulin resistance: unrav-elling the mechanism. Lancet 375, 2267–2277

28. Hocking, S. et al. (2013) Adiposity and insulin resistance inhumans: the role of the different tissue and cellular lipid depots.Endocr. Rev. 34, 463–500

29. Neeland, I.J. et al. (2012) Dysfunctional adiposity and the risk ofprediabetes and type 2 diabetes in obese adults. JAMA 308,1150–1159

30. Gesta, S. et al. (2007) Developmental origin of fat: tracking obesityto its source. Cell 131, 242–256

31. Tchkonia, T. et al. (2013) Mechanisms and metabolic implicationsof regional differences among fat depots. Cell Metab. 17, 644–656

32. Bjorntorp, P. (1990) ‘Portal’ adipose tissue as a generator of riskfactors for cardiovascular disease and diabetes. Arteriosclerosis10, 493–496

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 235

33. Lafontan, M. and Girard, J. (2008) Impact of visceral adiposetissue on liver metabolism. Part I: heterogeneity of adipose tissueand functional properties of visceral adipose tissue. DiabetesMetab. 34, 317–327

34. Hansen, I. et al. (1986) Insulin resistance in acromegaly: defects inboth hepatic and extrahepatic insulin action. Am. J. Physiol. 250,E269–E273

35. Karlander, S. et al. (1986) Increased glucose turnover and glucosecycling in acromegalic patients with normal glucose tolerance.Diabetologia 29, 778–783

36. Foss, M.C. et al. (1991) Peripheral glucose metabolism in acro-megaly. J. Clin. Endocrinol. Metab. 72, 1048–1053

37. Bolinder, J. et al. (1986) Insulin action in human adipose tissue inacromegaly. J. Clin. Investig. 77, 1201–1206

38. Nielsen, S. et al. (2001) Pharmacological antilipolysis restoresinsulin sensitivity during growth hormone exposure. Diabetes50, 2301–2308

39. Nielsen, T.S. et al. (2014) Dissecting adipose tissue lipolysis:molecular regulation and implications for metabolic disease. J.Mol. Endocrinol. 52, R199–R222

40. Yin, D. et al. (1998) Somatotropin-dependent decrease in fattyacid synthase mRNA abundance in 3T3-F442A adipocytes is theresult of a decrease in both gene transcription and mRNA stability.Biochem. J. 331, 815–820

41. Ottosson, M. et al. (1995) Growth hormone inhibits lipoproteinlipase activity in human adipose tissue. J. Clin. Endocrinol. Metab.80, 936–941

42. Clasen, B.F. et al. (2013) Gene expression in skeletal muscle afteran acute intravenous GH bolus in human subjects: identification ofa mechanism regulating ANGPTL4. J. Lipid Res. 54, 1988–1997

43. Boero, L. et al. (2012) GH levels and insulin sensitivity are differentlyassociated with biomarkers of cardiovascular disease in activeacromegaly. Clin. Endocrinol. 77, 579–585

44. Moller, N. et al. (1992) Basal- and insulin-stimulated substratemetabolism in patients with active acromegaly before and afteradenomectomy. J. Clin. Endocrinol. Metab. 74, 1012–1019

45. Higham, C.E. et al. (2009) Pegvisomant improves insulin sensitivityand reduces overnight free fatty acid concentrations in patientswith acromegaly. J. Clin. Endocrinol. Metab. 94, 2459–2463

46. Nellemann, B. et al. (2014) Growth hormone-induced insulin resis-tance in human subjects involves reduced pyruvate dehydroge-nase activity. Acta Physiol. 210, 392–402

47. Ueland, T. et al. (2010) Associations between body composition,circulating interleukin-1 receptor antagonist, osteocalcin, andinsulin metabolism in active acromegaly. J. Clin. Endocrinol.Metab. 95, 361–368

48. Olarescu, N.C. et al. (2014) Inflammatory adipokines contribute toinsulin resistance in active acromegaly and respond differently todifferent treatment modalities. Eur. J. Endocrinol. 170, 39–48

49. Olarescu, N.C. et al. (2015) The metabolic risk in newly diagnosedpatients with acromegaly is related to fat distribution and circulat-ing adipokines and improves after treatment. NeuroendocrinologyPublished online January 9, 2015. http://dx.doi.org/10.1159/000371818

50. Freda, P.U. et al. (2008) Lower visceral and subcutaneous buthigher intermuscular adipose tissue depots in patients with growthhormone and insulin-like growth factor I excess due to acromeg-aly. J. Clin. Endocrinol. Metab. 93, 2334–2343

51. Reyes-Vidal, C. et al. (2014) Prospective study of surgical treat-ment of acromegaly: effects on ghrelin, weight, adiposity, andmarkers of CV risk. J. Clin. Endocrinol. Metab. 99, 4124–4132

52. Reyes-Vidal, C.M. et al. (2015) Adipose tissue redistribution andectopic lipid deposition in active acromegaly and effects of surgicaltreatment. J. Clin. Endocrinol. Metab. 100, 2946–2955

53. Barclay, J.L. et al. (2011) GH-dependent STAT5 signaling plays animportant role in hepatic lipid metabolism. Endocrinology 152,181–192

54. Cordoba-Chacon, J. et al. (2015) Growth hormone inhibits hepaticde novo lipogenesis in adult mice. Diabetes 64, 3093–3103

55. Winhofer, Y. et al. (2014) No evidence of ectopic lipid accumulationin the pathophysiology of the acromegalic cardiomyopathy. J. Clin.Endocrinol. Metab. 99, 4299–4306

56. Madsen, M. et al. (2012) Fat content in liver and skeletal musclechanges in a reciprocal manner in patients with acromegaly duringcombination therapy with a somatostatin analog and a GH recep-tor antagonist: a randomized clinical trial. J. Clin. Endocrinol.Metab. 97, 1227–1235

57. Olarescu, N.C. et al. (2015) GH action influences adipogenesis ofmouse adipose tissue-derived mesenchymal stem cells. J. Endo-crinol. 226, 13–23

58. Wabitsch, M. et al. (1996) Mitogenic and antiadipogenic propertiesof human growth hormone in differentiating human adipocyteprecursor cells in primary culture. Pediatr. Res. 40, 450–456

59. Lu, C. et al. (2010) A novel effect of growth hormone on macro-phage modulates macrophage-dependent adipocyte differentia-tion. Endocrinology 151, 2189–2199

60. Berryman, D.E. et al. (2011) Growth hormone and adipose tissue:beyond the adipocyte. Growth Horm. IGF Res. 21, 113–123

61. Benencia, F. et al. (2015) Male bovine GH transgenic mice havedecreased adiposity with an adipose depot-specific increase inimmune cell populations. Endocrinology 156, 1794–1803

62. Ciresi, A. et al. (2012) Visceral adiposity index is associated withinsulin sensitivity and adipocytokine levels in newly diagnosedacromegalic patients. J. Clin. Endocrinol. Metab. 97, 2907–2915

63. Lam, K.S. et al. (2004) Serum adiponectin is reduced in acromeg-aly and normalized after correction of growth hormone excess. J.Clin. Endocrinol. Metab. 89, 5448–5453

64. Silha, J.V. et al. (2003) Perturbations in adiponectin, leptin andresistin levels in acromegaly: lack of correlation with insulin resis-tance. Clin. Endocrinol. 58, 736–742

65. White, U.A. et al. (2016) The modulation of adiponectin by STAT5-activating hormones. Am. J. Physiol. Endocrinol. Metab. 310,E129–E136

66. Ciresi, A. et al. (2015) Serum visfatin levels in acromegaly: Corre-lation with disease activity and metabolic alterations. GrowthHorm. IGF Res. 25, 240–246

67. Liu, T.F. et al. (2012) NAD+-dependent sirtuin 1 and 6 proteinscoordinate a switch from glucose to fatty acid oxidation during theacute inflammatory response. J. Biol. Chem. 287, 25758–25769

68. Froy, O. (2010) Metabolism and circadian rhythms–implications forobesity. Endocr. Rev. 31, 1–24

69. Brooks, A.J. et al. (2014) Mechanism of activation of protein kinaseJAK2 by the growth hormone receptor. Science 344, 1249783

70. Dominici, F.P. et al. (2005) Influence of the crosstalk betweengrowth hormone and insulin signalling on the modulation of insulinsensitivity. Growth Horm. IGF Res. 15, 324–336

71. Brooks, A.J. and Waters, M.J. (2010) The growth hormone recep-tor: mechanism of activation and clinical implications. Nat. Rev.Endocrinol. 6, 515–525

72. Feng, X. et al. (2014) Suppressors of cytokine signaling (SOCS)and type 2 diabetes. Mol. Biol. Rep. 41, 2265–2274

73. del Rincon, J.P. et al. (2007) Growth hormone regulation ofp85alpha expression and phosphoinositide 3-kinase activity inadipose tissue: mechanism for growth hormone-mediated insulinresistance. Diabetes 56, 1638–1646

74. Takano, A. et al. (2001) Growth hormone induces cellular insulinresistance by uncoupling phosphatidylinositol 3-kinase and itsdownstream signals in 3T3-L1 adipocytes. Diabetes 50, 1891–1900

75. Jessen, N. et al. (2005) Evidence against a role for insulin-signalingproteins PI 3-kinase and Akt in insulin resistance in human skeletalmuscle induced by short-term GH infusion. Am. J. Physiol. Endo-crinol. Metab. 288, E194–E199

76. Henry, R.R. et al. (2013) Hyperglycemia associated with pasireo-tide: results from a mechanistic study in healthy volunteers. J. Clin.Endocrinol. Metab. 98, 3446–3453

77. Tzanela, M. et al. (2011) Glucose homeostasis in patients withacromegaly treated with surgery or somatostatin analogues. Clin.Endocrinol. 75, 96–102

78. Steffin, B. et al. (2006) Effects of the long-acting somatostatinanalogue Lanreotide Autogel on glucose tolerance and insulinresistance in acromegaly. Eur. J. Endocrinol. 155, 73–78

79. Shishko, P.I. et al. (1994) Insulin-like growth factors and bindingproteins in patients with recent-onset type 1 (insulin-dependent)

236 Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4

diabetes mellitus: influence of diabetes control and intraportalinsulin infusion. Diabetes Res. Clin. Pract. 25, 1–12

80. Leung, K.C. et al. (2000) Insulin regulation of human hepatic growthhormone receptors: divergent effects on biosynthesis and surfacetranslocation. J. Clin. Endocrinol. Metab. 85, 4712–4720

81. Samson, S.L. (2015) Pasireotide in acromegaly: an overview ofcurrent mechanistic and clinical data. Neuroendocrinology 102,8–17

82. Lindberg-Larsen, R. et al. (2007) The impact of pegvisomant treat-ment on substrate metabolism and insulin sensitivity in patients withacromegaly. J. Clin. Endocrinol. Metab. 92, 1724–1728

83. Neggers, S.J. et al. (2011) Hypothesis: extra-hepatic acromegaly:a new paradigm? Eur. J. Endocrinol. 164, 11–16

84. Brummer, R.J. et al. (1993) Adipose tissue and muscle volumedetermination by computed tomography in acromegaly,before and 1 year after adenomectomy. Eur. J. Clin. Investig.23, 199–205

85. Hansen, T.B. et al. (1994) Body composition in active acromegalyduring treatment with octreotide: a double-blind, placebo-con-trolled cross-over study. Clin. Endocrinol. 41, 323–329

86. O'Sullivan, A.J. et al. (1994) Body composition and energyexpenditure in acromegaly. J. Clin. Endocrinol. Metab. 78,381–386

87. Plockinger, U. and Reuter, T. (2008) Pegvisomant increases intra-abdominal fat in patients with acromegaly: a pilot study. Eur. J.Endocrinol. 158, 467–471

Trends in Endocrinology & Metabolism, April 2016, Vol. 27, No. 4 237