A compendium of G-protein–coupled receptors and cyclic nucleotide regulation of adipose tissue metabolism and energy expenditure

Abstract

With the ever-increasing burden of obesity and Type 2 diabetes, it is generally acknowledged that there remains a need for developing new therapeutics. One potential mechanism to combat obesity is to raise energy expenditure via increasing the amount of uncoupled respiration from the mitochondria-rich brown and beige adipocytes. With the recent appreciation of thermogenic adipocytes in humans, much effort is being made to elucidate the signaling pathways that regulate the browning of adipose tissue. In this review, we focus on the ligand–receptor signaling pathways that influence the cyclic nucleotides, cAMP and cGMP, in adipocytes. We chose to focus on G-protein–coupled receptor (GPCR), guanylyl cyclase and phosphodiesterase regulation of adipocytes because they are the targets of a large proportion of all currently available therapeutics. Furthermore, there is a large overlap in their signaling pathways, as signaling events that raise cAMP or cGMP generally increase adipocyte lipolysis and cause changes that are commonly referred to as browning: increasing mitochondrial biogenesis, uncoupling protein 1 (UCP1) expression and respiration.

Introduction

Obesity is a disorder that leads to complications such as insulin resistance and diabetes. Changes in diet and exercise are the most commonly recommended methods to treat obesity, but rarely succeed over the long term [1]. More recently, invasive procedures such as bariatric surgery have also been prescribed [2]. In addition, pharmacological approaches to reduce obesity, including blockade of fat absorption and central nervous system stimulants, have been developed [2], but include unacceptable side effects or eventual lack of efficacy. The lack of effective therapeutics for obesity is a serious deficiency in the currently available pharmacological compendium, as recent studies have reported that approximately 30% of Americans are clinically obese [3] and the prevalence of obesity worldwide is rising [4]. Obesity is co-morbid with Type 2 diabetes and related diseases such as heart disease, nephropathy and sleep apnea. Thus, finding ways to treat or reduce obesity are important for decreasing the amounts of diabetes and related diseases.

A novel strategy for achieving weight loss that was attempted early in the 20th century targeted ‘uncoupled respiration’ by using the drug 2,4-dinitrophenol (DNP), which globally and indiscriminately uncouples oxidative phosphorylation from ATP production in mitochondria of all cells. This therapy was employed extensively between 1933 and 1938, but its use was largely discontinued after a number of deaths, and the US FDA labeled this drug as “extremely dangerous and not fit for human consumption” [5]. While DNP is not an acceptable anti-obesity agent, it did shed light on the fact that wasteful metabolic processes may be an attractive means of achieving weight loss. Moreover, recently a modified form of DNP that targets the liver has been shown in rodents as well as non-human primates to reverse dyslipidemia and hepatic steatosis by increasing fat oxidation in the liver [6,7].

One naturally occurring process that involves regulated energy-expenditure is non-shivering thermogenesis. This is the production of heat in warm-blooded organisms without skeletal muscle shivering. An important site of thermogenesis in mice and humans is brown adipose tissue (BAT), which is a type of adipose tissue that has large numbers of mitochondria, contributing to its brown color [8]. Activation of the BAT signature protein, uncoupling protein 1 (UCP1), allows protons to cross the inner mitochondrial membrane without ATP production. This uncoupled electron transport causes heat to be the main product of metabolism [9]. It has recently been appreciated that adult humans possess significant amounts of UCP1 and mitochondria-rich brown adipocytes [10–12], and the activation of UCP1 is now seen as a potential therapeutic target for obesity and metabolic disease as a mechanism to consume energy and reduce or prevent obesity. Even in the absence of appreciable weight loss, reducing plasma glucose and fatty acids could be very beneficial for risk of diabetes and cardiovascular disease. In addition to BAT, organisms possess white adipose tissue (WAT) that primarily functions to store caloric energy in the form of triglycerides. However, cells in the WAT depot can increase a population of cells, called “beige” or “brite” adipocytes, that have elevated levels of mitochondria, express UCP1, and are more thermogenic. This process is called “beiging” or “browning” [8].

The therapeutic implications of a pharmacological agent(s) which could safely increase a subject’s uncoupled-energy expenditure via their endogenous UCP1 proteins are attractive. Indeed, recent studies have identified a wide range of neural, circulating, and dietary factors that purportedly increase adipose browning. As there are too many agents that have alleged browning effects to be included in this review, we have decided to only focus on the pathways that enhance/inhibit adipocyte browning and BAT thermogenesis via their regulation of the cyclic nucleotides, cAMP and cGMP (cyclic Adenosine MonoPhosphate and cyclic Guanosine MonoPhosphate). As such, much of this review will focus on G-Protein Coupled Receptors (GPCRs) whose downstream signaling events regulate the production of cAMP. In the event that the reader is unfamiliar with this family of receptors and their downstream signaling, we have suggested a few recent reviews [13–15]. We will also review browning via the much smaller receptor–enzyme families of guanylyl cyclases (GC), which produce cGMP, because there is large overlap in the downstream signaling pathways of cAMP and cGMP. Finally, we also review the emerging concept of modulating adipocyte thermogenesis by targeting the phosphodiesterase (PDE) enzymes, which degrade cAMP and cGMP. We chose to focus on these areas because we estimate that they represent some of the most fertile ground for developing potential pharmaceutical agents that augment adipocyte browning as an estimated 35% of all approved drugs target GPCRs [16] and large portions of the much smaller GC and PDE families are the targets of many currently approved therapeutics.

In this review, we discuss in alphabetical order the GPCR, GC, and PDE families that have been implicated in the context of regulating adipocyte cyclic nucleotide concentrations, lipolysis, oxygen consumption, mitochondrial biogenesis, UCP1 expression, and uncoupled respiration starting with the adrenergic receptors. In Table 2 at the end of this chapter we have listed many of these receptor/ligand signaling pathways in order to provide a compendium of the currently available information about these signaling pathways and how they affect adipocyte lipolysis and thermogenesis. This is not meant to be a comprehensive list as not all receptor/ligand signaling pathways were able to be included. Hopefully, this will allow us to better understand the subtle differences in these signaling pathways that will help us choose the best pathway(s) to target for the development of therapeutics that augment adipocyte thermogenesis.

Table 2.

Summary of ligand–receptor signaling pathways that regulate adipocyte lipolysis and browning

Ligand	Receptor	G-protein	Cyclic nucleotide	Lipolysis	Browning/Thermogenesis
Adenosine	A1R [466,467]	Gαi	Adenosine opposes the actions of other agents that raise cAMP [468–472]. A1 R knock-out blocks the 2-chloroadenosine-evoked reduction in cAMP [473].	Adenosine opposes the actions of other agents which raise lipolysis both in vitro/ex vivo [468,469,471,472,474–480] and in vivo [481]. A1 R knock-out blocks the 2-chloroadenosine-evoked reduction in lipolysis [473,482].	Adenosine inhibits adipocyte respiration [479,483,484].
Adenosine	A2AR, A2BR [466]	Gαs	Adenosine increases cAMP in some cell lines [485]. A2AR and A2BR agonists increase cAMP [485].	Adenosine increases lipolysis in some cell lines [485]. A2AR and A2BR agonists increase lipolysis [485].	Activation of A2AR increases thermogenic gene expression in human and mouse brown and white adipocytes [485].
Adrenergic: (norepinephrine, epinephrine)	β1AR, β2AR, β3AR	Gαs	Increase cAMP [486]	Increase lipolysis [486].	Increase browning [486].
Adrenergic: (norepinephrine, epinephrine)	α1AR	Gαq/11			Augment the βAR response in brown adipocytes to increase thermogenesis [38,39].
Adrenergic: (norepinephrine, epinephrine)	α2AR	Gαi	Decrease cAMP by inhibiting adenylyl cyclase [486].	Decrease lipolysis [25,40].	Central nervous system activation of α2AR can suppress peripheral SNS activation of BAT [487]. Peripheral actions variable in literature.
Bile acids (cholic acid, chenodeoxycholic acid)	GPBAR1 (aka TGR5) [466,488–493]	Gαs	Bile acids (including cholic acid) increase cAMP in BAT cells [488].		Cholic acid and chenodeoxycholic acid both have been reported to increase adipocyte thermogenesis [488,490,491,494–496]. GPBAR1 selective agonists have similar browning effects [489,492,493,497,498].
Gastric inhibitory polypeptide (GIP)	GIPR [147,167–177,466]	Gαs	GIP raises cAMP concentrations and PKA-evoked signaling [95,167,171,174,177,188–191].	The effects of GIP on lipolysis have observed it to be increased [168,181,185,189], unaffected [193] and decreased [116,192]. GIP increased both glycerol release and fatty acid re-esterification, which resulted in an overall decrease in free fatty acid release despite the increased lipolysis [185].	Both GIP treatment and Gipr knockdown were reported to increase UCP1 in immortalized BAT cells along with an increase in their maximal oxygen consumption [204]. Similar findings were observed in Gipr−/− mice where Ucp1 mRNAwas increased in brown adipose tissue of mice fed a high fat diet [151].
Glucagon	GCGR [75,85–96,466]	Gαs	Glucagon increases intracellular cAMP in adipocytes [85,104,107,110,111,113,115,117, 118,120,124–128,284,285,469,499,500]	Glucagon increases circulating glycerol and free fatty acids in vivo [60,65,66,77–84]. Glucagon increases adipocyte lipolysis in vitro and ex vivo [75,87,89,97–103,105–110,112–114,116,118–123,126–130,501–504].	Glucagon increases body temperature, oxygen consumption, metabolic rate and thermogenic markers in vivo, but some studies indicate that this can be blocked by inhibiting β-adrenergic signaling [56–76,135]. Glucagon increases oxygen consumption and thermogenesis in vitro and ex vivo in adipocytes and adipose tissue [65,71,75,122,129–133].
Glucagon-like peptide-1 (GLP-1)	GLP1R [85,86,94,95,142–147,466]	Gαs	GLP-1 has been reported to increase [85,113,126,128] or not alter [117] intracellular cAMP in adipocytes.	GLP-1 has been reported to increase [85,113,126,128] or not alter [148,149] lipolysis.	Central administration of synthetic GLP1R agonists increase adipose tissue browning [134,152,153]. Peripheral administration of synthetic GLP1R agonists increase UCP1 in adipocytes [155–159].
Melanocyte stimulating hormones (MSH): α-MSH, β-MSH, γ-MSH etc. Adrenocorticotropic hormone (ACTH)	MC2R [228–233,236], MC5R [228,230–233,236,466] Human adipocytes lack MC2R but may express other receptors [229,239,240]	Gαs	ACTH, which binds to all melanocortin receptors, increases cAMP in adipocytes [104,107,110,115,120,219,222–224,228,230,233,234,236,284,285,292,469,499,500,505–507]. α-MSH, which does not bind to MC2R, similarly increases cAMP in adipocytes [228,230,233].	ACTH stimulates lipolysis [97–99,101,103,106,107,109,110,120, 217,218,220–223,226,227,229,231,233,234,501,502] α-MSH and β-MSH, which do not bind MC2R, stimulate lipolysis [103,226,231,233,235,241]. Lipolysis is not affected by ACTH in primates or canines [229,237,238,240], though a-MSH may have a small lipolytic effect [240].	ACTH increases oxygen consumption and thermogenic gene expression in adipocytes [59,130,133,225,232,234,236]. α-MSH, which does not bind to MC2R, increases thermogenic gene expression in adipocytes [235,236,242].
Neuropeptide Y (NPY), peptide YY (PYY), pancreatic polypeptide (PP)	YR1, YR2, YR5 [147,466,508–513]	Gαi		Both NPY and PYY inhibit lipolysis in vitro [238,241,514–516]. NPY and PYY inhibit lipolysis in vivo in sympathectomized rats [517].	NPY administration reduces [518] while global NPY knockout increases adipocyte thermogenic markers [519,520]. NPY suppress thermogenic gene expression and oxygen consumption in C3H10T1/2 adipocytes [521].
Parathyroid hormone (PTH), Parathyroid hormone-related protein (PTHrP)	PTH1R, PTH2R [236,266,268]	Gαs,Gαq/11	PTH stimulates cAMP in adipocytes [260–264].	PTH stimulates lipolysis in adipocytes [238,258–261,265,266].	Both PTH and PTHrP increase respiration and promote the expression of thermogenic genes in adipocytes [236,267,268].
Pituitary adenylate cyclase-activating peptide (PACAP), Vasoactive intestinal peptide (VIP)	PAC1, VPAC1, VPAC2 [236,466,511,522–528]	Gαs	VIP increases cAMP in adipocytes [85,111,115,264,529].	PACAP- and VIP-stimulated adipocyte lipolysis [112,116,123,149,181,226,264,524]. This can only be inhibited by a VPAC2 antagonist [524].	PACAP has been reported to stimulate Ucp1 gene expression and has a small stimulatory effect on respiration in brown adipocytes [236]. Central administration of either PACAP or VIP has been shown to promote brown adipose tissue thermogenesis [530,531].
Prostaglandin E2 (PGE2)	EP2, EP4 [236,466,532–538]	Gαs	In the absence of elevated cAMP, high concentrations of PGE2 by itself have been reported to increase cAMP in adipocytes [288,293]. Under similar circumstances, PGE1, a similar PGE analog, has also been reported to increase cAMP in adipocytes [284,539].		PGE2 has been shown to increase UCP1 [236,303,304,540]. 16,16-dimethyl-PGE2, an EP2/EP3/EP4 agonist, has been reported to also increase the expression of thermogenic gene expression in BAT in vivo [297].
Prostaglandin E2 (PGE2)	EP3 [296,466,532,533,535–538,541]	Gαi	PGE2 opposes the actions of other agents which raise cAMP [286–290,292,295,542–544]. PGE1 also lowers cAMP after it has been raised by other agents. [284,285,506,542,545–547]. In the absence of raising cAMP with other agents, PGE2 by itself lowers cAMP, but only at low concentrations [288,293]. Sulprostone, an EP1 and EP3 selective agonist, has also been shown to oppose actions of other agents that raise cAMP in adipocytes [293,544]. The EP3 antagonist, L-798,106, blocks the PGE2-evoked reduction in cAMP [295].	PGE2 opposes actions of other agents that raise lipolysis both in vitro/ex vivo [276,286,289,290,292,294,296,303,482,502,505,541,543,544,548–552] and in vivo/in situ [553,554]. PGE1 also reduces lipolysis after it has been raised by other agents both in vitro/ex vivo [434,435,500–502,505,515,542,547, 555–560] and in vivo/in situ [103,502,553,561]. Sulprostone, an EP1 and EP3 selective agonist, has also been shown to oppose actions of other agents that raise lipolysis in adipocytes [294,544]. EP3 knock-out or the antagonist L-826,266 blocks the PGE2-evoked reduction in lipolysis [296,541].
Prostacyclin (PGI2)	IP [466,532,562]	Gαs	PGI2 raises cAMP in adipocytes [279–283,563].	PGI2 increases lipolysis [276].	PGI2 and its synthetic analog carbaprostacyclin have been observed to increase Ucp1 and other thermogenic gene expression in adipocytes [298–301].
Secretin	SCTR [85,236,466,564–566]	Gαs	Secretin increases cAMP in adipocytes [85,107,111,124,125,545,564].	Secretin stimulates lipolysis both in vivo [79] and in vitro [102,106,107,112,114,116,545,564, 565,567,568].	Secretin increases oxygen consumption and thermogenic gene expression in brown adipocytes [71,236,566].
Thyroid-stimulating hormone (TSH, aka thyrotropin)	TSHR [236,466,569–582]	Gαs	TSH increases intracellular cAMP levels in adipocytes [104,285,499,570,573,574,579,580].	TSH induces lipolysis [103,501,502,577,583–589].	TSH increases Ucp1 and oxygen consumption in adipocytes in vitro [124,579,582].
Atrial natriuretic peptide (ANP), B-type natriuretic peptide (BNP)	NPRA (aka GC-A) [315,383,390–396,399,402,404,406,412,418,590]	–	ANP signaling increases intracellular cGMP concentrations in adipocytes [393,404,410,418]. BNP signaling also increases intracellular cGMP concentrations in adipocytes [404].	ANP stimulates lipolysis in vitro [395,406,411,413,414,418,421,591] and in vivo [395,402,414,417,419,591]. BNP stimulates lipolysis in vitro [419] and in vivo [395,419,420].	ANP increases oxygen consumption and UCP1 expression in adipocytes in vitro [315,[404,[406,421,[422]. BNP increases UCP1 expression in vitro [404] and in vivo [315,400,423].
C-type natriuretic peptide (CNP)	NPRB (aka GC-B) [392,394,396,404]	–	CNP signaling increases intracellular cGMP concentrations in adipocytes [393,409,410].		Mice overexpressing CNP in adipose tissue have increased oxygen consumption and UCP1 in inguinal WAT [409].

G-protein–coupled receptors

Adrenergic receptors

The adrenergic receptors are in many respects the prototypical GPCR. Owing to the ability to purify significant quantities of the β₂AR protein and G-protein complexes in some laboratories, many of the fundamental characteristics of this family of receptors were described in studies of the β₂AR [17]. The gene and cDNA for the β₂AR was also the first of this large family of seven transmembrane domain-containing proteins to be cloned [18], followed soon thereafter by the β₁AR [19]. There are nine adrenergic receptors (ARs): three alpha₁- (α₁ARs), three α₂ARs and three beta-(β)ARs (Figure 1). Each of these receptors is the products of separate genes. The βARs are by far the most studied and discussed regarding adipocyte catabolism and increasing energy expenditure through non-shivering thermogenesis, and there are so many review articles on this topic that the reader can easily find one (Figure 2). Therefore, we will not spend a great deal of time on this topic here, except to make some general statements and points of interest to us. In general, β₁AR and β₂AR are broadly expressed throughout organs and cell types, and both can be found expressed together in some tissues such as the heart, adipose tissues and vasculature. The β₃AR, which was the last of the ARs to be cloned and characterized [20–22], tends to be found primarily in adipocytes; although there are reports of functionally relevant β₃AR in organs such as urinary bladder. In rodent adipose tissue, β₃AR is the most highly expressed relative to β₁AR and β₂AR [22,23], while in human adipose tissues β₃AR represents a much smaller proportion compared with β₁AR and β₂AR [22,24–26]. Further, β₃AR is found in all rodent adipose depots – white and brown – but its lower level in primate adipose tissue at this point also appears to be more restricted to ‘brown’ or ‘beige’ adipocytes that are able to express UCP1 [27–29]. This difference in the level of expression of the β₃AR between rodents and primates is important to remember. In humans and non-human primates, the β₁AR is preferentially activated by norepinephrine, and β₂AR can also be activated in human adipocytes through circulating epinephrine [26]. A recent report suggests that β₁AR, and not β₃AR, may be more important for activating human brown adipocytes [30]. There are also subtle and not-so-subtle differences in the signaling properties of the three βARs [31,32]. While β₁AR and β₂AR utilize β-arrestins in their signaling scenarios, β₃AR is unique in this regard in that it does not recruit β-arrestin [33], although all three βARs are nevertheless able to activate ERK pathways by different means [31]. Activation of p38 MAPK downstream of PKA is also an important feature of βAR signaling in adipocytes and is required for orchestrating the signaling components to increase transcription of the Ucp1 and Pgc1α genes, among others [31,34]. Finally, a more recent discovery in the complexity of βAR signaling in the adipocyte for driving ‘browning’ and brown fat development is the unexpected activation of the rapamycin-sensitive mTOR complex 1 (mTORC1) by PKA [35]. These signaling pathways through the βARs in adipocytes also afford remarkable plasticity between the energy-storing ‘white’ adipocyte and the energy-consuming, thermogenic ‘brown’ or ‘beige’ adipocyte [36,37]. The α₁AR and α₂AR are also expressed in adipocytes, the former having been reported to also contribute to brown adipocyte function [38,39] and the latter largely involved in blocking the actions of the βARs through their coupling to the heterotrimeric Gα_i protein and inhibition of adenylyl cyclase [25,40]. Moreover, increases in α₂ARs in human obesity are reported to contribute to the refractoriness of the tissue for catabolism.

Figure 1. Catecholamine biosynthetic pathway and signaling.

Epinephrine and norepinephrine (also called adrenaline and noradrenaline, respectively) are made from the amino acid tyrosine via the catecholamine biosynthetic pathway. Epinephrine and norepinephrine signal through nine GPCRs called the adrenoceptors. The α₁ receptors are Gα_q-coupled, α₂ receptors are Gα_i-coupled, and the β-receptors are Gα_s-coupled. The α-receptors are activated by epinephrine and norepinephrine, with epinepherine having a lower potency than norepinephrine on the α₂ receptors. For the β-receptors, in β₁ norepinephrine is more potent than epinephrine, in β₂ epinephrine is more potent than norepinephrine, and in β₃ epinephrine and norepinephrine have similar potencies (adrenoceptors in the IUPHAR/BPS Guide to Pharmacology Database.

Available from: https://doi.org/10.2218/gtopdb/F4/2019.4).

Figure 2. Sympathetic activation of adipocyte lipolysis and thermogenesis.

Canonical activation of adipocyte lipolysis and thermogenesis occurs via the sympathetic nervous system. Sympathetic neurons release norepinephrine which activate β-adrenergic receptors on the adipocytes. Activation of these receptors turn on their associated G-proteins by causing the Gβγ (not shown) and Gα_s subunits to dissociate. The Gα_s subunit in turn augments adenylyl cyclase activity. The cAMP produced by adenylyl cyclase causes the regulatory subunits of PKA to dissociate from the catalytic subunits thereby liberating its activity. PKA phosphorylates and activates Hormone Sensitive Lipase (HSL) and other components of the lipolytic pathway. PKA also phosphorylates and activates p38 MAPK that leads to increased Ucp1 transcription and the expression of other pro-thermogenic genes.

Glucagon/PACAP/secretin family

The B1 subfamily of GPCRs consists of 15 members, all of which bind polypeptide hormones and couple to Gα_s, although some can also couple to other G-proteins [41]. Nine of the ligands for this family (Glucagon, Glucagon-Like Peptides [GLP-1, GLP-2], Gastric Inhibitory Polypeptide [GIP], Growth-Hormone Releasing Hormone [GHRH], Pituitary Adenylate Cyclase-Activating Peptide [PACAP], peptide histidine methionine [PHM], Secretin, Vasoactive Intestinal Peptide [VIP]) are structurally related, many of which can bind to and activate each other’s receptors at high concentrations [41,42]. Furthermore, glucagon, GLP-1 and GLP-2 are all derived from the same proglucagon gene [43,44]. This family of receptors is also referred to as the Secretin Family, because the first hormone discovered in this (or any) family was secretin [45]; the PACAP Family, because secretin is absent in fish and PACAP appears to be the most ancient gene in the family [46]; or the Glucagon Family, though this generally refers only to the receptors for glucagon, GLP-1, GLP-2 and GIP, while other closely related receptors such as GHRH and secretin are also sometimes included [42]. Another similarity that is shared between GLP-1, GLP-2, GIP, secretin and VIP is that they are secreted from the intestines in response to food. These hormones are collectively referred to as “Incretins” due to their ability to enhance glucose stimulated insulin secretion in response to food intake [47]. Several of these family members, including glucagon, GLP-1 and GIP, are degraded by dipeptidyl peptidase-4 (DPP-4) [48]. DPP-4 inhibitors, also called gliptins, are currently used as therapeutic treatments for Type 2 diabetes due to their ability to prolong the half-life of these incretins [49]. Here, we will review yet another commonality among several of these closely related ligands and receptors: their ability to induce adipocyte browning and thermogenesis. Readers should be advised that there is considerable variation in perspective among studies in that it is still not clear what effects are ‘direct’ effects of the peptides on adipose tissue per se versus ‘indirect’ effects that may stem from inter-organ cross-talk or central effects. Therefore, as they will become apparent in the following sections, the literature is still somewhat conflicted.

Glucagon

Glucagon is best known as the hormone that opposes insulin, acting as a counter-regulatory hormone produced by α-cells in the pancreatic islets of Langerhans [50]. Glucagon was discovered while studying methods of insulin purification; it was found that a certain fraction was able to cause a rise in blood glucose in pancreatectomized dogs [51]. After pancreatic glucagon was discovered, it was found that a similar substance was also produced in the intestinal mucosa [52]. The gene that encodes glucagon also encodes GLP-1 and GLP-2, which are produced by the L-cells in the intestines [43,44,53,54]. The canonical function of glucagon is to promote glycogenolysis and gluconeogenesis in the liver thereby raising blood glucose levels [55].

In addition to its gluconeogenic effects, glucagon has long been known to exhibit thermogenic properties with numerous reports of glucagon increasing body temperature, oxygen consumption, metabolic rate and thermogenic markers in vivo [56–76]. The concomitant lipolytic effects, such as increasing circulating glycerol and free fatty acids, are also observed in response to glucagon administration in vivo [60,65,66,77–84]. Glucagon has apparent direct effects on adipose tissue as the glucagon receptor (Gcgr) is expressed in adipocytes [75,85–96]. There are numerous reports of glucagon increasing intracellular cAMP and adipocyte lipolysis in vitro and ex vivo [75,85,87,97–128]. Furthermore, several studies have shown that glucagon increases oxygen consumption and thermogenesis in isolated adipose tissue or cells [65,71,75,129–133]. Nevertheless, in vivo studies present a much more complicated picture of how glucagon regulates body temperature and adipose thermogenesis. Glucagon appears to partially exert its pro-thermogenic effects through an indirect, centrally mediated mechanism, as central administration of glucagon increases brown adipose tissue thermogenesis [134]. Additionally, these thermogenic-effects of glucagon may be indirect, as they are oftentimes blocked by propranolol or adrenalectomy [56,57,71]. Somewhat contradicting these findings are studies showing that denervation is ineffective at altering glucagon-evoked changes in thermogenic mitochondrial markers or lipolysis [67,81]. In one instance the effects of glucagon on energy expenditure were partially blunted in mice lacking Fgf21, further suggesting that much of its effects are indirect [75]. Further complications regarding glucagon’s thermogenic role in vivo arise from the observations that the glucagon-evoked rise in body temperature is intact in mice that have a Ucp1 gene knockout [75] and in swine, which have an ancestral disruption of the Ucp1 gene [135,136], indicating that the effects of glucagon may not be entirely dependent upon uncoupled respiration.

Despite the uncertainty behind the mechanism, the glucagon family of peptides appear to be an important component of the thermogenic response, as mice with a global gene knockout of the pro-glucagon gene have impaired responses to cold or β₃-agonist treatment [137]. Additionally, mice with a global gene knockout of Gcgr, though they remain cold tolerant and have unaltered sympathetic tone, exhibit impaired cold-induced browning of white adipose tissue [76]. From these studies, it appears that glucagon augments energy expenditure through both direct and indirect effects. While studies have demonstrated that glucagon directly binds its own receptor on adipocytes and promotes fatty acid liberation and thermogenesis, glucagon also acts through secondary pathways, such as catecholamines and FGF21, to promote adipocyte lipolysis and browning. Though the extent at which glucagon acts through direct versus indirect mechanisms to augment energy expenditure remains a bit unclear, this remains an important component of its activities and has been a therapeutic goal for agonists that target the glucagon receptor.

Glucagon like peptide-1

GLP-1 and GLP-2 are derived from the same gene as glucagon and are produced in the L-cells of the intestines and released in response to food intake [43,44]. As such, one of their main functions as incretins is to augment glucose stimulated insulin secretion from the islet β-cell [138]. While GLP-2 has been less well studied, the effects of GLP-1 on the β-cell have been exploited for therapeutic benefit, as several GLP-1 receptor agonists have been developed to treat Type 2 diabetes. These include exenatide (synthetic exendin-4, Byetta), liraglutide (Victoza), dulaglutide (Trulicity), albiglutide (Tanzeum), lixisenatide (Adlyxin) and semaglutide (Ozempic). In addition to their effect on the β cell, these GLP-1 agonists also have roles in other tissues through which they can regulate food intake and weight gain [139]. This ‘side effect’ of GLP-1 agonists first became clinically applicable in December 2014, when liraglutide was also approved for weight loss under the product label Saxenda.

Although GLP-1 has a well-established role in the β cell, its function in adipose tissue is a bit more controversial. Even the question of whether or not the GLP-1 receptor is expressed in adipose tissue has produced conflicting results. For example, a few studies utilizing Northern Blots failed to observe the GLP-1 receptor (Glp1r) mRNA [140,141], most studies using more sensitive methods, such as PCR or radioligand binding, did find the mRNA or protein in adipose tissue [85,86,94,95,142–147]. Similarly, GLP-1 has been observed to both increase [85,113,126,128] or not alter [117,148,149] intracellular cAMP and lipolysis in adipocytes. Globally, GLP-1 was suggested to play some role in thermogenesis in a study of mice with a whole body knockout of Glp1r: Glp1r^−/− mice had relatively lower energy expenditure at room temperature, but this difference was ablated at thermoneutrality [150]. Furthermore, although chow fed Glp1r^−/− mice showed a very modest increase in oxygen consumption in response to norepinephrine treatment, high fat diet fed Glp1r^−/− mice had a significant reduction in response to norepinephrine [150]. By contrast, another group found increased oxygen consumption in Glp1r^−/− mice on both regular chow and high fat diet, which was attributed to increased locomotor activity [151]. Despite these conflicting results for direct effects of GLP-1 on adipose tissue, central administration of GLP1R agonists: GLP-1, liraglutide and exendin-4; all found increased adipose tissue browning and energy expenditure [134,152,153]. Nevertheless, it has been reported that signaling through the GLP1R in vagal afferent neurons decreases energy expenditure and brown adipose tissue thermogenesis [154]. These findings make interpreting results from peripheral administration of these agonists difficult, as the contribution of the indirect stimulation of adipose tissue remains possible. In studies where humans were administered GLP-1, no effect on energy expenditure has been observed [72,73]. However, when exendin-4, liraglutide, GLP-1 [32–36] amide (LVKGR amide) and supaglutide (a tool compound that is not in the clinic) were administered to rodents, oxygen consumption was increased and adipose tissue browning was observed [155–159]. This effect was lost in mice housed at thermoneutrality, as 6 weeks of liraglutide treatment failed to affect thermogenic gene expression in the adipose tissue of mice and any changes in oxygen consumption in response to norepinephrine were comparable with that of untreated mice that were pair-fed with the liraglutide treated mice [150]. This indicates that these apparent changes in energy expenditure may be secondary to changes in body weight. While the central effects of GLP1R agonists on weight gain and energy expenditure are well established, their peripheral effects that directly impact the adipocyte are much less understood and it is questionable how much of an impact adipocyte GLP1R have on thermogenesis.

Gastric inhibitory polypeptide

Entero-gastrone was the name first given to what was then a putative hormone that was reported to be released in response to dietary fat and inhibits gastric secretion [160]. Later efforts to purify Entero-gastrone yielded the hormone that we now call Gastric Inhibitory Polypeptide or Glucose-dependent Insulinotropic Polypeptide (GIP) [161–164]. GIP is expressed in the K cells of the intestines [165] and is well known for its role augmenting glucose stimulated insulin secretion as an incretin [166]. The GIP receptor (GIPR) is also highly expressed in mature rodent and human adipocytes [147,167–177]. It appears that GIP–GIPR signaling on adipocytes has surprisingly similar effects to that of insulin: they both promote glucose and fatty acid uptake and lipid accretion in adipocytes [173,178–186]. In vivo, this may not be surprising given the close association between GIP and insulin as elevation of GIP levels will result in a concomitant rise in insulin in the fed state that may itself be mediating these effects [187]. However, in vitro, where there is no accompanying insulin release, this association is somewhat unexpected in adipocytes as GIPR is Gα_s coupled and raises cAMP concentrations and PKA-evoked signaling [95,167,171,174,177,188–191]. Though most studies find that GIP raises cAMP in adipocytes, its effects on lipolysis are somewhat controversial with some studies finding an increase [168,181,185,189], and others a decrease [116,192] in lipolysis, and yet others with no effects [193]. One study somewhat reconciled the conflicting observation of GIP promoting both fatty acid uptake and lipolysis: they found that GIP increased both glycerol release and fatty acid re-esterification, which resulted in an overall decrease in free fatty acid release despite the increased lipolysis [185].

GIP–GIPR signaling also has an important role in weight gain in response to high fat diet. Several maneuvers including GIPR knockout mice, GIPR antagonists, Gip-expressing K cell ablation and neutralization of GIP or GIPR with an antibody all exhibited resistance to diet-induced obesity [151,177,184,194–203]. GIPR signaling in adipose tissue appears to contribute significantly to this effect as mice with adipose tissue specific GIPR knockout are resistant to high fat diet-mediated weight gain [201], and re-expression of GIPR in the adipose tissue of whole body GIPR knockout restores the weight gain phenotype in response to high fat diets [198]. Brown adipose tissue was not considered to be mediating the effect, since mice with a Gipr knockout driven by Myf5-Cre had similar body weights upon high fat diet feeding [204]. However, the specificity of Myf5-Cre is not confined to the brown adipocyte, but also includes skeletal muscle; and there have been unexpected phenotypes associated with this Cre driver [205]. The effect of GIPR on lipid metabolism may mediate much of its effects on weight gain as GIPR ablation reduces the respiratory quotient, indicative of greater lipid oxidation. However, that said, most, but not all, studies observe no effects on energy expenditure [177,184,199,204,206]. Paradoxically, administration of a DPP-4-resistant GIP analog or GIP overexpression also appears to reduce weight gain when fed a high fat diet [207,208]. Along with this, both GIP treatment and Gipr knockdown were reported to increase UCP1 in immortalized BAT cells along with an increase in their maximal oxygen consumption [204]. Similar findings were observed in Gipr^−/− mice where Ucp1 mRNA was increased in brown adipose tissue of mice fed a high fat diet [151]. These studies indicate that GIP–GIPR signaling in adipocytes exists and may regulate metabolic phenotype, but its role is unclear at this point and more work is needed to clarify it.

The engineered polyagonists

An innovative concept has emerged in this area that takes advantage of the similarities between these incretin peptides and their receptors in order to marshal the potential therapeutic benefits of stimulating each of these receptors, by developing single molecules that can activate multiple receptors in this family, to unique benefit [209]. The first ‘designer’ dual-agonists contain features and activities at the glucagon and GLP-1 receptors [210,211]. The idea behind this was to combine the thermogenic and lipolytic properties of glucagon with the anti-hyperglycemic properties of GLP-1 in order to counteract the glucose raising glycogenolytic and gluconeogenic properties of glucagon. Moreover, being a single molecular entity has clear benefits for patenting purposes. These peptides reduce body weight in mice with diet induced obesity [210,211] as well as reducing steatohepatitis [212]. While there is an effect to decrease food intake, despite this, energy expenditure is also increased, respiratory quotient is decreased, and some versions of this dual agonist increase UCP1 in brown adipose tissue. Another dual-agonist that was developed is a peptide that activates the GLP-1 and GIP receptors [213]. This peptide also reduces food intake, body weight and blood glucose, but it had no effect on energy expenditure or respiratory quotient. A tri-agonist that combines the activities of glucagon, GLP-1, and GIP has also been made [191]. This tri-agonist was found to reduce body weight and fat mass, reduce food intake, increase energy expenditure, decrease respiratory quotient, and improve glucose homeostasis in mice with diet induced obesity. Some of the weight-lowering effects of the tri-agonist were clearly due to increased energy expenditure as the weight lost in tri-agonist-treated mice was greater than that of their pair-fed controls. The glucagon-like actions of the tri-agonist were largely responsible for the reduction in body weight and food intake as these effects were lost in Gcgr^−/− mice. The GLP-1-like actions were important for glycemic control as the tri-agonist increased blood glucose in Glp1r^−/− mice. Also, the loss of body weight and food intake were more modest in these Glp1r^−/− mice. The GIP-like functions were also important for the glucose-lowering effect as this was lost in Gipr^−/−mice. Apparently, the activation of Glp1r prevented the tri-agonist from raising blood glucose in these Gipr^−/− mice. The body weight, food intake and fat mass lowering effects of the tri-agonist were unaltered in the Gipr^−/− mice, indicating that the GIP-like effects were not responsible for these phenotypes. This tri-agonist probably has a role in adipose tissue as it stimulates cAMP in 3T3-L1 adipocytes [191], and increased circulating free fatty acid levels in vivo [214]. Although less work has been done to understand the adipose browning potential of these polyagonists, it remains possible that some of their beneficial metabolic effects stem from adipocyte thermogenesis. Some of these polyagonists are now in clinical trials, and it is an exciting new area of basic and translational research.

In addition to these peptide polyagonists, chemical hybrid molecules have also been designed that have potential browning effects. One such molecule is a glucagon and T3 (triiodothyronine) hybrid that consists of the glucagon peptide linked to a T3 molecule [96]. This glucagon/T3 was designed because it would deliver the T3 primarily to the liver and adipose tissue and thereby avoid the adverse effects of T3 in other organs. Glucagon/T3 potently lowered body weight due to increased energy expenditure as it did not alter food intake and was significantly blunted in Ucp1^−/− mice. This increased the thermogenic gene profile of inguinal white adipose tissue but had minimal effect on the brown adipose tissue. It is unclear if the glucagon- or T3-activities are the primary drivers of this browning phenotype as the glucagon receptor is necessary for the targeting of glucagon/T3 to the adipose tissue, which is lost in Gcgr^−/− mice.

Melanocortins

The endogeneous melanocortins are all products of a single gene, produced via cleavage of the proopiomelanocortin (POMC) polypeptide by prohormone convertases. These melanocortins are agonists for five different Gα_s-coupled receptors that are termed MC1R–MC5R. MC1R is expressed in melanocytes and white blood cells, consequently being involved in pigmentation and immune responses. MC2R is expressed in adrenal glands and is an important regulator of glucocorticoid production. MC3R and MC4R are both expressed in the central nervous system and are important regulators of food intake and energy homeostasis. MC5R is expressed in skeletal muscle and adipocytes where it regulates fatty acid oxidation and lipolysis, in addition to exocrine glands where it is associated with sebum production. There are many biologically active peptides that are derived from POMC; these include α-melanocyte stimulating hormone (MSH), β-MSH, γ-MSH and adrenocorticotropic hormone (ACTH; also known as corticotropin), in addition to several others. In addition, there are two endogeneous antagonists/inverse-agonists of the melanocortin receptors that are products of their own genes: agouti and agouti-related protein (AGRP). α-MSH, β-MSH, γ-MSH, ACTH are all agonists for MC1R, MC3R, MC4R and MC5R, though their affinities for the receptors vary. MC2R is unique in that ACTH is its only endogenous ligand; hence, ACTH is the only melanocortin that can bind to and activate all five receptors. The expression patterns of melanocortins vary, with α-MSH being expressed in the central nervous system and hypothalamus, β-MSH in the hypothalamus (except in rodents where POMC has an amino acid substitution that disrupts a cleavage site necessary to produce β-MSH), γ-MSH in the pituitary and hypothalamic arcuate, and ACTH in the anterior pituitary corticotrophs. The ‘Agouti’ peptide serves as an antagonist for MC1R and MC4R and is expressed in skin, testis, ovary and adipose tissues. AGRP is an antagonist for MC3R and MC4R and is produced in the adrenal cortex and certain regions of the central nervous system (reviewed in [215,216]).

ACTH is well known for its ability to stimulate the release of glucocorticoids (cortisol in humans and corticosterone in rodents) from the adrenal cortex and thereby regulate one of the most important aspects of glucose metabolism. In addition to its effects on glucocorticoids, in many species ACTH acts directly upon MC2R in adipocytes increasing intracellular cAMP–PKA signaling and thereby stimulating lipolysis, oxygen consumption and uncoupled thermogenesis [59,97–99,101,103,104,106,107,109,110,115,120,130,133,217–236]. However, MC2R is not expressed in the adipocytes of some species, including primates; as a result, ACTH has no effect on intracellular cAMP, lipolysis or uncoupled thermogenesis in human or canine adipocytes [229,237–240]. In vivo administration of ACTH increases adipocyte browning in rodents; however, this effect is significantly blunted by the concomitant rise in corticosterone, which is a potent inhibitor of adipocyte browning [234]. In addition to MC2R, MC5R is also expressed on rodent adipocytes [228,230–233,236]. α-MSH, which does not bind to MC2R, similarly increases cAMP, lipolysis and uncoupled thermogenesis in these adipocytes; however, the relevance of this pathway in vivo is questionable as the endogenous levels of ACTH and α-MSH in adipose tissue are not likely to be high enough to fully activate MC5R [103,226,228,230,231,233,236,241,242]. Central administration of melanocortins also induces adipocyte thermogenesis, but this effect is dependent upon the central nervous system as effects of centrally administered melanocortins are lost in denervated mice [243–245]. Though agonists activating the melanocortin receptors are potent browning agents in many species, the clinical utility of this pathway in humans is negligible due to the low receptor expression in human adipose tissue.

Parathyroid hormones

Parathyroid hormone (PTH) is an 84-amino acid peptide that is produced by the ‘chief’ cells in the parathyroid glands for which it is named [246,247]. PTH is best known for its effects on the kidney and bone. In the kidney, PTH promotes calcium reabsorption, inhibits phosphate reabsorption, and promotes the synthesis of the vitamin D metabolite, 1,25 dihydroxy-vitamin D3, which itself also promotes calcium reabsorption [248]. In bone, intermittent PTH augments bone growth but continuous PTH signaling favors bone resorption [248]. Certain cancers had long been known to produce a factor that behaved like PTH, but when it was demonstrated that PTH was in fact not the factor, Parathyroid Hormone-related Protein (PTHrP) was identified [249–251]. Mature human PTHrP is a 141-, 139-, or 173-amino acid peptide, depending on the splicing pattern of the transcript [252,253]. PTHrP is produced in a wide variety of tissues in which it acts in a paracrine manner [254]. PTHrP, like PTH, is crucial for bone development, but unlike PTH^−/− mice that have bone abnormalities, PTHrP^−/− mice die immediately after birth via respiratory failure due to defective rib cage development [255,256]. Aside from its role in bone, PTHrP also plays a role in other diverse processes such as placental calcium transport, lactation and smooth muscle relaxation [254].

There are two receptors for PTH and PTHrP: Parathyroid Hormone 1 Receptor (PTH1R, also known as the PTH/PTHrP Receptor) and Parathyroid Hormone 2 Receptor (PTH2R). Both PTH and PTHrP are agonists for PTH1R, but PTH2R is not activated by PTHrP [257]. Both receptors are primarily coupled to Gα_s, but can also be coupled with Gα_q/11 [248].

The ability of PTH to stimulate cAMP production and, through the cAMP-PKA signaling is well established, and in adipocytes can stimulate lipolysis [238,258–266]. However, it has only been recently been observed that PTH and PTHrP can have ‘browning’ properties as tumor-derived PTHrP was found to promote cachexia via adipose browning [267]. Shortly thereafter, the same authors found that PTH also promotes browning and cachexia in a 5/6 nephrectomy model of kidney failure [268]. Both PTH and PTHrP promote browning and cachexia via PTH1R signaling through the Gα_s–PKA pathway [267,268]. PTH and PTHrP both increase respiration and promote the expression of thermogenic genes in adipocytes [236,267,268].

Prostaglandins

The ‘E-series’ prostaglandins are known to raise the body temperature by inducing fever. In fact, during the first study examining the effect of prostaglandin E₂ (PGE₂) infusion in humans, a ‘feeling of warmth’ was among the first symptoms reported [269]. This pyretic response of PGE₂ is mediated through the EP3 receptor that is located in hypothalamic neurons of the preoptic nucleus [270,271]. Downstream neural signaling events lead to classical SNS activation of BAT and thereby raise body temperature [272,273]. It appears that PGE₂ primarily regulates fever-induced thermogenesis and has minimal effects on normal body temperature regulation (reviewed in [274]; the role of other prostaglandins in hyperthermia are reviewed in [275]). Despite this, prostaglandins have very important roles in adipocytes, including adipogenesis, lipolysis and the regulation of thermogenic adipocytes.

Prostaglandins are bioactive lipid-signaling molecules that are oxidative metabolites of arachidonic acid; they are produced by cyclooxygenase enzyme (COX1 and COX2) metabolism of arachidonic acid into an intermediary metabolite (PGH₂) followed by its rapid conversion into one of the five bioactive prostanoids (PGE₂, PGD₂, PGF_2α, PGI₂, TXA₂) by their respective synthases (Figure 3). Prostaglandins act locally in an autocrine or paracrine manner that is dependent on the activation of a family of specific GPCRs. The two predominant prostaglandins in adipose tissue appear to be PGE₂ and PGI₂. These generally have an opposing effect in adipose tissue [276] with prostacyclin (PGI₂) signaling through its lone Gα_s-coupled receptor, IP, and PGE₂ signaling through its predominantly Gα_i-coupled receptor, EP3 [277,278]. As such, many studies have shown that in adipocytes, signaling through IP raises cAMP [279–283] while PGE₂ and its synthetic analogs lower cAMP [284–295]. It should be noted that three other receptors for PGE₂ exist, with EP2 and EP4 being Gα_s-coupled and EP1 being Gα_q-coupled [277,278]; hence the actions of PGE₂ may raise cAMP depending on the circumstances. EP3 antagonists or gene knockout blocks the cAMP- and lipolysis-lowering effects of PGE₂ in adipocytes [295,296], indicating that these effects of PGE₂ occur via EP3.

Figure 3. Prostaglandin biosynthetic pathway and signaling.

Prostaglandins are oxidative metabolites of arachidonic acid that are produced by cyclooxygenase (COX) enzymes. Arachidonic acid is liberated from the plasma membrane by phospholipase A₂ (PLA₂). COX converts arachidonic acid into the short lived PGH₂ (not shown) which is converted into five primary bioactive prostanoids: PGE₂, PGD₂, PGF_2α, PGI₂ (prostacyclin) and TXA₂ (thromboxane). These prostanoids act locally in an autocrine or paracrine manner. The local action of prostaglandins depends on activation of a family of GPCRs designated EP (for E prostanoid), DP, FP, IP and TP receptors, for the other prostaglandins, respectively.

COX inhibition with the nonselective inhibitor indomethacin, the COX2 selective inhibitor celecoxib, or COX2 gene deletion have been shown to inhibit beiging of inguinal WAT [297,298]. Consistent with its ability to stimulate Gα_s-coupled receptors, PGI₂ and its synthetic analog carbaprostacyclin have been observed to increase Ucp1 and other thermogenic gene expression in adipocytes [298–301]. Additionally, the PGI₂ analog beraprost increases skin temperature in mice [302], though it is unclear if this is due to a direct effect on the adipose tissue or via neural pathways. The roles of PGE₂ and other prostaglandins are less clear. PGE₂ has been shown to increase UCP1 mRNA expression in human omental white adipose tissue explants [303], mouse primary brown adipocytes [236], and in mice inguinal white adipose tissue using an adipose-targeted nanoparticle delivery system [304]. 16,16-dimethyl-PGE₂, an EP2/EP3/EP4 agonist, has been reported to also increase the expression of thermogenic gene expression in BAT in vivo [297]. However, 16,16-dimethyl-PGE₂ was also shown to elicit a dose dependent decrease in UCP1 expression in rosiglitazone-treated hMADS [305]. The EP4 antagonist, AH-23848, and to a lesser extent the EP2 antagonist, AH-6809, attenuated isoproterenol-evoked Ucp1 mRNA and thermogenic gene expression in Retinoblastoma-null adipocytes [297]. These studies indicate that in adipocytes PGE₂ and PGI₂ may have a pro-thermogenic role (Figure 4). PGD₂ may also have a role in BAT as the expression of one of its synthases, lipocalin prostaglandin D synthase, is induced in BAT by cold exposure and global knockout is associated with a change in fuel utilization during cold-exposure [306]. Additionally, the subcutaneous WAT of these knockout mice showed a 4-fold increase in Ucp1 mRNA expression over the wild-type [307].

Figure 4. Prostaglandin thermogenic signaling.

PGE₂−EP3 signaling is well known to be an important regulator of the increased thermogenesis that occurs during fever. In addition, recent studies have shown that prostaglandins are important regulators of adipocyte browning.

Guanylyl cyclases

In addition to cAMP, another cyclic nucleoside, cGMP, is an important second messenger. Unlike the hundreds of GPCRs that influence cAMP levels, cGMP is regulated by a much more finite group of upstream signaling mediators. cGMP is made by guanylyl cyclases that convert GTP into cGMP in much the same way that adenylyl cyclase converts ATP into cAMP. Guanylyl cyclases can be divided into two categories: soluble and membrane bound (also called ‘particulate’). Soluble guanylyl cyclases are heterodimeric being made up of α and β subunits, both of which have two isoforms, and are activated by nitric oxide (NO) [308,309]. Nitric oxide synthases (NOS), of which there are three isoforms termed endothelial (eNOS), inducible (iNOS), and neuronal (nNOS) [310], mediate the endogenous production of NO by converting l-arginine into citrulline and NO [311]. In addition, nitrate (NO₃⁻) and nitrite (NO₂⁻), which had been considered inert end products of NO metabolism, have been recognized for their contribution to NO production and signaling via the nitrate–nitrite–NO pathway [312]. There are seven membrane bound guanylyl cyclases (though not all are found in humans) termed Guanylyl Cyclase A-G (GC-A and -B are also commonly referred to as Natriuretic Peptide Receptor A and B, respectively); they are homodimers consisting of an extracellular ligand binding domain, a transmembrane domain, followed by the intracellular kinase homology domain, hinge region, and guanylyl cyclase catalytic domains [308,313]. cGMP activates two serine/threonine kinase termed protein kinase G-I and -II (PKG), which are similar to PKA in that they phosphorylate many of the same targets and are both activated by cyclic nucleoside binding to their regulatory subunits. PKG differs from PKA in that its regulatory subunit is part of the same polypeptide as its catalytic subunit. Like activation of PKA, activation of PKG increases the browning of adipose tissue, increasing mitochondrial biogenesis and UCP1 expression, and these effects are ablated with PKG inhibitors or PKG-I gene knockout [314–316]. Though there have been numerous studies that have demonstrated increased adipocyte browning by targeting GPCRs and thereby cAMP, targeting of cGMP-regulated pathways remains less studied (Figure 5).

Figure 5. Guanylyl cyclase dependent activation of adipocyte lipolysis and thermogenesis.

Activation of both soluble and membrane bound guanylyl cyclases in adipocytes increases intracellular cGMP concentrations. This leads to increased PKG activity, which phosphorylates and activates signaling pathways that cause increased lipolysis and thermogenic gene expression.

Soluble guanylyl cyclases

Drugs that activate soluble guanylyl cyclases by acting as NO donors, called nitrovasodilators, have long been utilized to relax the vascular endothelium and reduce blood pressure. Nitroglycerine is one of the earliest nitrovasodilators; it was suspected of being a vasodilator a few years after its initial synthesis [317,318]. Within three decades, it was already being utilized therapeutically to treat angina pectoris and reduce blood pressure [319]. A century later, the mechanism behind nitroglycerine’s vasodilation was elucidated when it was found to be a NO donor around the same time that NO produced by the endothelium was recognized as an important signaling molecule for the relaxation of smooth muscle [311,320–323]. Nitroglycerin and other traditional vasodilators aremain important pharmacological tools and are still used to treat acute heart failure [324].

In addition to its effects on vasodilation, NO has important effects on energy expenditure and adipose tissue. Several studies have shown that NO stimulates mitochondrial biogenesis in adipocytes [325–327]. In vitro, it appears that NO-cGMP signaling disrupts pre-adipocyte proliferation and pushes them into differentiating into a more brown-like adipocyte [314,316,328–332]. This mitochondria-generating effect is not limited only to adipose, as this phenomenon has been reported in several other tissues [333]. In addition to promoting mitochondrial biogenesis, NO itself directly affects mitochondrial respiration. NO can directly bind to the O₂ binding site in cytochrome oxidase inhibiting its turnover and thereby mitochondrial respiration [334–337]. As with other tissues, this reaction also occurs in brown adipose tissue [338].

NO–cGMP evoked pathways appear to be an important component of the cold-induced browning phenotype as cold-norepinephrine-cAMP-evoked signaling induces iNOS and eNOS expression and activity in adipocytes [339–343]. This appears to be of functional importance as inhibition of NOS activity with N^ω-nitro-l-arginine methyl ester (l-NAME) impairs this catecholamine-induced browning [339,344–347]. In vivo treatment with l-NAME actually decreases body weight and adipose tissue mass, but this appears to be primarily the result of decreased food intake [345,348–351]. Conflicting results were obtained for l-NAME’s effect on UCP1 in BAT [351,352], but l-NAME administration to rat pups reduced their oxygen consumption during cold stress [353].

Targeting of the individual NOS’s produces somewhat different results. Though Enos^−/− mice maintain similar body weights and food consumption; oxygen consumption, Ucp1 expression, mitochondrial β-oxidation, and cold-induced mitochondrial biogenesis are all reduced [325,327,354,355]. Overexpression of eNOS is beneficial as it increases mitochondrial biogenesis in adipose tissue conferring a brown color, increases oxygen consumption, and reduces weight gain conferring resistance to diet induced obesity [356].

On the other hand, iNOS appears to be maladaptive as iNOS^−/− mice are protected against proinflammatory macrophage infiltration of adipose tissue and the associated extracellular matrix deposition, fibrosis and insulin resistance [357–359]. On a normal standard chow diet, iNOS^−/− mice maintain body weights similar to that of wild-type, thought they have been reported to be slightly heavier or to have reduced fat mass [360,361]. One study reported that these mice have reduced food intake, heat production, CO₂ output and respiratory exchange ratio [361]. However, in Lep^ob/ob iNOS^−/− mice, body weight was reduced, core body temperature was increased, and the brown adipose tissue had increased UCP1 and fluorodeoxyglucose uptake [362,363]. Similarly, the decreased expression of genes associated with mitochondrial biogenesis upon high fat diet feeding was rescued in iNOS^−/− mice [357].

Though iNOS produced NO appears to be detrimental to a healthy metabolic phenotype, activation of the soluble guanylyl cyclases in adipose tissue still appears to be beneficial. Mice lacking the β₁ subunit of soluble guanylyl cyclase have impaired thermogenesis due to reduced mitochondria and UCP1 in the brown adipose tissue [332]. In addition, activation of soluble guanylyl cyclases with the agonist, BAY41–8543, increases whole-body energy expenditure, induces browning of WAT, and protects against diet-induced weight gain [332]. Similarly, the addition of sodium nitrate to the diet, which leads to the production of biological NO independent of NOS via the nitrate–nitrite–NO pathway, has beneficial metabolic effects. When sodium nitrate is added to the drinking water, body weight is reduced in both wild-type and eNOS^−/− mice [364,365], and significant morphological changes to the white adipose tissue, inducing the formation of multilocular adipocytes and the expression of Ucp1 and other thermogenic genes.

As with most other agents that induce adipocyte browning, NO–cGMP signaling also plays a role in adipocyte lipolysis. NO donors stimulate lipolysis, increasing glycerol release from adipocytes [366–369]. However, NO donors also reduce and NOS inhibitors augment isoproterenol-stimulated lipolysis [366,370–372]. This occurs because NO can directly interact with isoproterenol altering its chemical composition converting it into a molecule that is not a β-agonist [373]. NO-signaling may be an important component of PKA-mediated lipolysis, as NOS inhibitors have been reported to reduce the lipolytic response to β₃- and cAMP-stimulated lipolysis [369,374]. It has been posited that the antioxidant properties of NO mediate this effect as antioxidants rescue the lipolysis lost to NOS inhibitors [374]. Additionally, cAMP-signaling increases nitrite levels and this may be an important source of NO for lipolysis [368,369]. Contradicting these observations of NO-stimulated lipolysis, the NOS inhibitor NG-monomethyl l-arginine (l-NMMA), but not its inactive isomer d-NMMA, increase lipolysis [370,375]. A similarly enhanced lipolytic response was observed it the NOS inhibitor l-NAME, but its inactive isomer d-NAME also enhanced lipolysis indicating that it was caused by an off-target effect [374].

Membrane bound/particulate guanylyl cyclases – natriuretic peptide receptors

Natriuretic peptides (NPs) are peptides that are well known for their abilities to lower blood pressure via diuresis and the eponymous natriuresis [376]. There are three types of NPs: Atrial Natriuretic Peptide (ANP), B-type Natriuretic Peptide (also called Brain Natriuretic Peptide, BNP), and C-Type Natriuretic Peptide (CNP). Mature ANP is 28-, CNP is 22-, and BNP can vary between 26- and 45-amino acids depending on species, with each containing a 17-amino acid loop bound by a disulfide bond [377]. ANP and BNP are secreted primarily from atrial cardiomyocytes in response to mechanical stretch, whereas CNP is produce by a variety of tissues including vasculature, brain and pituitary gland [378].

NPs act on three receptors termed Natriuretic Peptide Receptor A (NPRA, also called Guanylyl Cyclase A, GC-A), Natriuretic Peptide Receptor B (NPRB, also called Guanylyl Cyclase B, GC-B), and Natriuretic Peptide Receptor C (NPRC, also called the Natriuretic Peptide Clearance Receptor; note NPRC is not the same as GC-C which is the guanylyl cyclase-coupled receptor for guanylin). NPRA binds to both ANP and BNP with pM affinities but does not interact with CNP at physiological concentrations. Conversely, NPRB only interacts with CNP, which it binds in the pM range, and does not appreciably bind ANP and BNP whose affinities are in the nM range. All NPs interact with NPRC with pM affinities [377,379].

NPRA and NPRB have a similar structure, containing an extracellular ligand-binding domain, a single trans-membrane region, and the kinase homology domain (which represses the guanylyl cyclase activity in the inactive state) and guanylyl cyclase catalytic domain in the intracellular portion [313,377,380]. Due to their guanylyl cyclase activity, NPRA and NPRB are also referred to as Particulate Guanylyl Cyclases or Membrane-Bound Guanylyl Cyclases [313]. Circulating NPs bind to inactive homodimers of these NPRs, and this is thought to produce a conformational change that leads to the activation of their intrinsic guanylyl cyclase activity. NPRC also contains an extracellular ligand-binding domain and single trans-membrane region, but only has 37-amino acids in the intracellular portion and hence does not synthesize cGMP [313,377,380]. NPRC was first described as a ‘clearance’ receptor because early studies found that though it has no effect on cGMP levels, the NPRC selective ligand C-ANP_4–23 increases circulating NPs in vivo (i.e. it is thought that NPRC clears NPs from the circulation because the NPRC selective ligand competes with the other NPs to be cleared) [381,382]. NPRC also reduces the efficacy of ANP in vitro, as we have shown that adding NPRC to cells causes a rightward shift in the ANP dose–response curve [383]. Despite its minimal intracellular domain, NPRC has been reported to promote clathrin-mediated endocytosis [384] and some studies have indicated that NRPC is capable of activating Gα_i signaling [385–389].

Both white and brown adipose tissues are important sites of NP action, demonstrating high levels of NP binding and high expression of all three receptors [315,390–396]. Numerous studies have shown that in adipose tissue the expression levels of the NP receptors are dynamic and that the ratio of the guanylyl cyclase-coupled receptors, NPRA and NPRB, to the clearance receptor, NPRC, is an important component of metabolic function (i.e. with a higher NPRA/NPRC or NPRB/NPRC ratio, NPs will have a stronger signaling effect and will be more able to raise intracellular cGMP concentrations). In obese patients and animal models of obesity, the adipose tissue NPRA&B/NPRC ratio is lower indicating impaired NP signaling [396–405]. Additionally, in adipocyte NPRA/NPRC ratio can be lowered by insulin, which is elevated during obesity associated Type 2 diabetes [396,403,406]. Conversely, dieting/fasting and cold exposure raises the NPRA&B/NPRC ratio in adipose tissue indicating improved NP signaling [315,398,407,408]. Consistent with this, increased NP signaling has been demonstrated to improve metabolic dysfunction [383,398–400,409]. In adipose tissue, NP signaling increases intracellular cGMP concentrations [393,410], which have effects that are largely similar to increasing intracellular cAMP concentrations. For instance, NPs promote glucose uptake in adipocytes [383,411,412] and are potent lipolytic agents [383,395,402,406,409,411,413–421]. As expected, in obese subjects with a reduced NPRA/NPRC ratio the lipolysis response is reduced, but is increased in patients on a low-calorie diet [402,419].

Perhaps not surprisingly, given that PKA and PKG phosphorylate highly similar peptide motifs (RRXS*/T*), cardiac NPs have also been shown to stimulate brown fat thermogenesis as well as promote ‘browning’ in white adipose tissue [315,398,400,404,421–423]. Treatment of adipocytes with ANP increases mitochondrial gene expression and increases uncoupled respiration [315,421]. Like the βARs and PKA, the NPs and PKG similarly activate p38 MAPK [315,424] as well as the mTORC1 machinery [423]. It has also been shown that mice overexpressing BNP (driven by the serum amyloid P promoter which expresses BNP in the liver) also have increased oxygen consumption, although increased mitochondria in other tissues such as skeletal muscle may also be playing a role in that model [398]. The metabolically beneficial effects of the cardiac NPs can also be augmented by either eliminating NPRC or by compounds that selectively bind NPRC and effectively block the binding of ANP/BNP [315,383,425]. Increased adipose thermogenesis in mice fed a high fat diet was observed not only in Nprc^−/− mice with a global gene knockout, but also in mice with an adipose tissue specific Nprc gene deletion [383]. In both mouse models, there was improved insulin sensitivity, and a generally improved metabolic phenotype [383]. Muscle-specific Nprc knock-out had a minimal impact on metabolic phenotype with regard to high-fat diet feeding, indicating that the beneficial metabolic effects of NPRC ablation are primarily due to changes in the adipose tissue [383]. However, we do not rule out a role for NPRC in skeletal muscle for affecting fatty acid oxidation and outcomes such as muscle performance or endurance.

Phosphodiesterases

While many studies have demonstrated that stimulating the production of cyclic nucleotides in adipose tissue leads to increased browning, thermogenesis and decreased obesity, the fact remains that levels of cAMP and cGMP can also be elevated by decreasing their degradation. The phosphodiesterase (PDE) enzymes inactivate cAMP and cGMP by breaking the phosphodiester bond in these cyclic nucleotides. There are 11 families of PDEs and each family has between one and four members for a total of 24 genes (Table 1) (there is a PDE12, but it breaks the phosphodiester bond in the poly(A) tails of mitochondrial RNA [426]). In addition, complexity increases as each gene product can have multiple splice variants, each of which can then play a role in their subcellular localization and action. Despite this class of enzymes being considerably smaller than GPCRs, they are nevertheless important and well-known therapeutic targets. Of these 11 families, currently 3 families are targets of FDA approved drugs. These PDE inhibitors serve as therapeutics for many diverse conditions such as chronic obstructive pulmonary disease (nonselective: theophylline [427]), acute heart failure (PDE3: milrinone, aka Primacor), intermittent claudication (PDE3: cilostazol, aka Pletaal [428]), chronic obstructive pulmonary disease (PDE4: roflumilast, aka Daliresp), psoriatic arthritis (PDE4: apremilast, aka Otezla [429]), atopic dermatitis (PDE4B: crisaborole, Eucrisa [430]), erectile dysfunction (PDE5: sildenafil, aka Viagra; vardenafil, aka Levitra; tadalafil, aka Cialis; avanafil, aka Stendra [431]), and pulmonary arterial hypertension (PDE5, sildenafil, aka Revatio; tadalafil, aka Adcirca [431]).

Table 1.

Phosphodiesterase family members

Family	Gene(s)	Affinity (Km) (μM)	Miscellaneous
1	Pde1a Pde1b Pde1c	cAMP 0.3–120 cGMP 0.6–6.0	Ca2+/calmodulin stimulated
2	Pde2a	cAMP 30–50 cGMP 10–30	cGMP stimulated
3	Pde3a Pde3b	cAMP 0.02–0.4 cGMP 0.02–0.2	cGMP inhibited
4	Pde4a Pde4b Pde4c Pde4d	cAMP 1.5–10 cGMP NA	cAMP specific
5	Pde5a	cAMP NA cGMP 1–6.2	cGMP specific
6	Pde6a Pde6b Pde6c	cAMP NA cGMP 10–17	cGMP specific
7	Pde7a Pde7b	cAMP 0.1–0.2 cGMP NA	cAMP specific
8	Pde8a Pde8b	cAMP 0.06–0.6 cGMP NA	cAMP specific IBMX-insensitive
9	Pde9a	cAMP NA cGMP 0.17–0.7	cGMP specific IBMX-insensitive
10	Pde10a	cAMP 0.17–1 cGMP 0.3–14	cAMP inhibited
11	Pde11a	cAMP 2.0–3.2 cGMP 0.4–2.1

In adipocytes, PDEs have been studied extensively and have a well-established role. Canonical insulin inhibition of lipolysis occurs via activation of PDE3B which in turn degrades cAMP thereby decreasing lipolysis. Non-selective PDE inhibitors, such as IBMX, theophylline and caffeine, have long been known to increase adipocyte lipolysis and thermogenesis [432–437]. Resveratrol, which is found in red wine and other foods [438], is also a non-selective PDE inhibitor, inhibiting PDEs 1, 3 and 4 [439], and has been shown to increase whole body oxygen consumption and Ucp1 expression in mouse WAT and BAT [440]. Despite the common knowledge that PDE inhibition increases adipocyte thermogenesis and the fact that highly selective PDE inhibitors are readily available, much less work has been done to elucidate the role of the individual PDEs.

PDE1 has the unique distinction of being the only family of PDEs that are regulated by calcium-calmodulin signaling [441]. PDE1 was suggested to influence cAMP accumulation in brown adipocytes by an early study that showed that the calcium/calmodulin-mediated decrease in cAMP could be blocked by a non- or PDE1-selective inhibitor [442]. PDE1 activity in brown adipocytes was independently verified in a subsequent study [443]. PDE1B was more conclusively identified as having a role in browning because it is a target of microRNA-378 (miR-378) [444]. miR-378 comes from an intron of PGC-1β, an important regulator of mitochondrial biogenesis and brown fat-specific gene expression; consequently, it has a similar expression pattern. miR-378 overexpression in mice induces brown adipocyte hyperplasia that improves the metabolic phenotype. This miR-378 overexpression also increased cAMP levels and lipolysis in BAT.

PDE3 is perhaps the PDE most commonly associated with adipocytes. This is due, in no small part, to its canonical role in the insulin-mediated inhibition of lipolysis [445]. As expected for an enzyme that degrades cAMP in adipocytes, inhibition or gene knock-out of PDE3 increases browning [446,447]. The PDE3 inhibitor, cilostamide, augments isoproterenol-evoked cAMP and Ucp1 mRNA accumulation in brown adipocytes in vitro [443]. In the absence of β-adrenergic stimulation, cilostamide only increased cAMP and Ucp1 mRNA in the presence of the PDE4 inhibitor, rolipram [443]. In vivo studies by Vincent Manganiello’s group found that Pde3b^−/− mice are resistant to high fat diet feeding [447]. This appears to be due to increased mitochondrial activity and thermogenic gene expression, which was observed in epididymal WAT in response to Pde3b knockout and cilostamide [446,447]. Additionally, Pde3b knockout and cilostamide enhanced the effects of β₃-adrenergic stimulation on mitochondrial activity, thermogenic gene expression, oxygen consumption, respiration and weight loss [446,447]. This was attributed to a PKA–AMPK mediated signaling pathway [447].

PDE4 is one of the three cAMP specific PDEs. PDE4 inhibitors have recently garnered some attention as a possible therapeutic target for obesity [448,449]. Potential weight loss benefits of PDE4 inhibition has already been observed in chronic obstructive pulmonary disease patients treated with roflumilast, whom display reductions in body weight, though the decline in body weight was greater in the patients that experienced the side effects of diarrhea, nausea, vomiting or headache [450]. Another PDE4 inhibitor, rolipram, also showed a beneficial metabolic response, but in mice [439]. In addition to reduced weight gain when fed a high fat diet, these mice had increased oxygen consumption, body temperature and Ucp1; although increased skeletal mitochondria may be, at least partially, mediating some of these effects [439]. Genetic evidence for a role of PDE4 comes from Pde4b knockout mice, which have a slightly reduced body weight and improved white adipose tissue when fed a high fat diet [451]. However, these mice also displayed increased locomotor activity which, again, may be mediating some of these effects [451]. The protein stability of PDE4D5 is positively regulated by Rheb; consequently, Rheb knockout mice have increased adipocyte cAMP and browning [452]. Though there is significant evidence implicating PDE4 in the augmentation of adipocyte thermogenesis, inhibition of both PDE3 and PDE4 in brown adipocytes results in greater UCP1 expression than inhibiting either alone [443].

PDE5, which is cGMP specific, is perhaps best known as the primary target of sildenafil, the active ingredient in ‘the little blue pill’ Viagra. There is some evidence that PDE5 inhibitors might improve the metabolic phenotype and increase adipocyte thermogenesis. In one study, sildenafil has been shown to improve insulin sensitivity during prediabetes in a randomized, controlled clinical trial [453]. Tadalafil, another PDE5 inhibitor, has been reported to improve forearm glucose uptake in people following a meal [454]. Additionally, mice treated with sildenafil had reduced body weight and fat mass due to increased energy expenditure [455]. This was accompanied by improved insulin sensitivity with increased glucose uptake into the skeletal muscle [455]. Similarly, Wistar rats made insulin resistant with an isocaloric fructose-enriched diet have reduced plasma triglycerides and improved oral glucose tolerance when treated with sildenafil [456]. Given that PDE5 inhibition improves capillary permeability and blood flow, and that blood flow to insulin sensitive tissues is a critical factor regulating insulin action [457], it is possible that some of these beneficial effects of PDE5 inhibition are through its vascular effects. However, direct effects of PDE5 inhibition on adipocytes have been observed. Sildenafil treatment of mice induced browning of inguinal WAT [316], but did not affect UCP1 expression in BAT [455]. Similar results were found in humans, where sildenafil increased UCP1 and mitochondrial respiration in subcutaneous adipose tissue, but not BAT [458] and vardenafil increased PPARGC1A and mitochondrial DNA in omental adipose tissue [326]. Additionally, the PDE5 inhibitor, udenafil, has been found to increase the oxygen consumption rate in 3T3-L1 adipocytes [459]. PDE6 is structurally similar to PDE5, and many of these inhibitors also block PDE6 activity, but PDE6 is primarily expressed in the rod and cone cells of the retina and is therefore unlikely to mediate these effects [460,461].

Pde8a, one of two genes in the IBMX-insensitive and cAMP specific PDE8 family, may also play a role in brown adipose tissue [443].

In addition to the aforementioned PDEs, PDE7B, PDE9A and PDE10A, have been detected in human adipose tissue via Western Blot [462]. Though PDE10A is highly expressed in the striatum with minimal expression in peripheral tissues, PDE10A expression in human supraclavicular BAT and mouse BAT can be seen using the PDE10A radioligand, [¹⁸F]-AQ28A [463]. Early studies noted that Pde10a^−/− mice weighed less than their Pde10a^+/+ counterparts, though no explanation was given [464]. It was only some years later that THPP-6, a PDE10 inhibitor, was found to increase energy expenditure and suppress food intake explaining the reduced body weight [465]. Another PDE10 inhibitor, MP-10, had no effect on food intake, but similarly increased energy expenditure, which was due to increased thermogenesis from brown and visceral white adipose tissue [463]. Similar to other means of elevating cyclic nucleotides, MP-10 also elevated lipolysis in cultured brown adipocytes [463]. These PDE10 inhibitors both caused a net improvement in the metabolic phenotype, improving glucose tolerance and insulin sensitivity [463,465].

Summary/Conclusion

In this review, we have covered studies of second messenger signaling pathways implicated in brown and beige adipocytes from studies performed in cell culture models, rodents and humans (Figure 6 and Table 2). Each of these approaches has its limitations. For example, one procedural – almost philosophical – aspect of studies in rodents is the environmental housing temperature that is used. It is currently a topic of discussion in the field as to whether rodent studies should be conducted at their thermoneutral temperature (~28–30°C; 82°F) instead of our ambient indoor temperature that is thermoneutral for humans. This is because at the latter temperature (~23°C), there is a slight thermal stress to rodents and could impact energy expenditure. We increasingly see publications that include experiments at both temperatures, but the jury is still out. Nevertheless, it is accepted in the field that measurements of energy expenditure should be corrected to body mass (lean mass) and not total body mass.

Figure 6. G-protein coupled receptors and cyclic nucleotide regulation of adipose tissue lipolysis and energy expenditure.

Many hormones utilize GPCR-cAMP- or GC-cGMP-mediated pathways to regulate adipocyte lipolysis and energy expenditure. These pathways are inherently intertwined as the liberation of fatty acids by lipolysis supplies the fuel for the thermogenesis. It is interesting to note that so many of these hormones utilize the same downstream signaling mechanism. For this reason, treatment of adipocytes with these hormones produces (at least superficially) similar results.

The vast majority of studies on this topic of brown and beige adipocytes, energy expenditure and fuel utilization, originally stem from the catecholamines, and particularly the βARs, because of the importance of the sympathetic innervation of adipose depots and to a lesser extent circulating catecholamines. These receptors and cAMP are by far and away the most heavily studied for understanding adipocyte metabolism, and especially the expansion and activation of energy-consuming brown/beige adipocytes. The current optimism in the field that these thermogenic cells may provide some therapeutic benefit for improving insulin sensitivity, reducing inflammation and promoting overall metabolic health has led to exploration of other potential pathways. In this review, we have drawn attention to these many other receptors and mediators of cyclic nucleotide signaling and summarize their effects on adipocytes and adipose tissue. While some of these receptors and ligands are already well studied, others might benefit from deeper investigation into their mechanisms and physiology. In our view, there is never a shortage of surprises to be uncovered when it comes to dissecting signal transduction mechanisms and the web of interactions that they mediate. Perhaps new insight and discoveries from these lesser studied receptors and pathways will yield important new research directions. It is important to remember that “You can’t make the puzzle if you don’t have all the pieces”.

Abbreviations

GPCR

G-protein–coupled receptor

DNP

2,4-dinitrophenol

BAT

brown adipose tissue

UCP1

uncoupling protein 1

WAT

white adipose tissue

cAMP

cyclic adenosine monophosphate

cGMP

cyclic guanosine monophosphate

GC

guanylyl cyclase

PDE

phosphodiesterase

AR

adrenergic receptor

PKA

protein kinase A

PPARγ

peroxisome proliferator-activated receptor gamma

PGC-1α

PPARγ coactivator-1α

mTORC1

mTOR complex 1

GLP

glucagon-like peptide

GIP

gastric inhibitory peptide

GHRH

growth-hormone releasing hormone

PACAP

pituitary adenylate cyclase-activating peptide

PHM

peptide histidine methionine

DPP-4

dipeptidyl peptidase-4

HSL

hormone sensitive lipase

GCGR

glucagon receptor

T3

triiodothyronine

POMC

proopiomelanocortin

MSH

melanocyte stimulating hormone

ACTH

adrenocorticotropic hormone

AGRP

agouti-related peptide

PTH

parathyroid hormone

PTHrP

PTH-related protein

PG

prostaglandin

COX

cyclooxygenase

PLA₂

phospholipase A₂

NO

nitric oxide

PKG

protein kinase G

l-NAME

N^ω-nitro-l-arginine methyl ester

l-NMMA

NG-monomethyl l-arginine

NP

natriuretic peptide