Insulin action in adipocytes, adipose remodeling, and systemic effects

Abstract

On this 100^th anniversary of the discovery of insulin, we recognize the critical role that adipocytes, which are exquisitely responsive to insulin, have played in determining the mechanisms for insulin action at the cellular level. Our understanding of adipose tissue biology has evolved greatly and it is now clear that adipocytes are far more complicated than simple storage depots for fat. A growing body of evidence documents how adipocytes, in response to insulin, contribute to the control of whole-body nutrient homeostasis. These advances highlight adipocyte plasticity, heterogeneity and endocrine function, unique features that connect adipocyte metabolism to the regulation of other tissues important for metabolism (e.g., liver, muscle, pancreas).

This year the medical and scientific communities as well as society “at large” celebrate the 100^th anniversary of the discovery that a pancreatic extract could lower blood glucose in diabetic dogs. In 1921, insulin was identified as the active component of the pancreatic extract. Soon thereafter insulin was shown to be effective in lowering blood glucose in humans with type 1 diabetes, leading to what has been referred to as “one of the miracles of modern medicine” (Bliss, 1982). Before this discovery, type 1 diabetes was a deadly disease, usually in childhood, and starvation or severe restriction of carbohydrates was the only therapy that could delay demise. Insulin (with diet and exercise) is still the sole treatment for Type 1 diabetes. Insulin’s effect to lower blood glucose is dependent on inhibition of hepatic glucose output as well as stimulation of glucose uptake into adipocytes and muscle, activities that during the past 100 years have been extensively studied at all levels of biology (i.e., molecular, cellular and whole-body). The topic of this article is the role of adipocytes and adipose tissue in insulin’s control of metabolism, highlighting some unexpected findings and identifying major unanswered questions.

Insulin-stimulated glucose uptake.

In studies of isolated adipocytes and muscle cells, insulin was shown to rapidly (within minutes) and reversibly increase glucose uptake by 2 to 10 fold. In both these cells types, enhanced glucose uptake is achieved by increasing the amount of the GLUT4 glucose transporters in the plasma membrane. GLUT4 is a member of the GLUT family of facilitative hexose transporters; GLUT4 is most highly expressed in adipocytes, and skeletal and cardiac muscle cells (Mueckler and Thorens, 2013). The increased amount of GLUT4 in the plasma membrane in response to insulin is achieved by a redistribution of GLUT4 from intracellular storage pools to the cell surface (Cushman and Wardzala, 1980; Klip et al., 1987; Suzuki and Kono, 1980; Wardzala and Jeanrenaud, 1981) (Figure 1). The stimulation of glucose uptake is accounted for by the increased amount of GLUT4 in plasma membrane without any change in intrinsic transporter characteristics (Km for glucose or flux). Elevated GLUT4 plasma membrane levels are dynamically maintained in the presence of insulin by the continued rapid recycling of internalized GLUT4 back to the plasma membrane for as long as the insulin receptor is activated (Govers et al., 2004; Karylowski et al., 2004; Zeigerer et al., 2002). When blood insulin levels drop, recycling of GLUT4 back to the plasma membrane returns to the pre-stimulation rate, and GLUT4 is dynamically re-sequestered intracellularly, poised to be translocated to the plasma membrane by another round of insulin stimulation. Consequently, the acute regulation of glucose uptake by adipocytes and muscle cells is controlled post-translationally with each molecule of GLUT4 participating in multiple rounds of recruitment to the plasma membrane followed by intracellular sequestration as determined by blood insulin levels.

Figure 1.

1A. Insulin-regulated pathways in adipocytes. See text for detailed discussion. The dotted arrow from insulin receptor signifies that signal transduction cascade from the receptor to Rab10 and control of lipolysis involves multiple signaling intermediaries. In control of GLUT4 we highlight Rab10 because that is the only protein along the pathway discussed in the text. GLUT4= Glucose transporter type 4; Rab10= Ras-related protein 10; FFAs= free fatty acids; TG=triglycerides; DNL=de novo lipogenesis; ChREBP= Carbohydrate-responsive element-binding protein; Acetyl-CoA= acetyl coenzyme A.

1B. Insulin action in white adipose tissue (WAT) alters fuel metabolism in liver and muscle via inter-tissue communication. Insulin signaling in adipocytes suppresses WAT lipolysis, decreasing the flux of free fatty acids (FFAs) and glycerol to the liver, and thereby contributing to the suppression of hepatic glucose production (HGP). FFAs, once converted to acetyl CoA, act as positive allosteric modulators of pyruvate carboxylase (PC), the enzyme catalyzing the first step of hepatic gluconeogenesis. Glycerol is a gluconeogenic substrate. The reduced flux of these metabolites to the liver contributes to the reduction in liver glucose output in the fed state. HGP is also regulated by direct action of insulin to promote hepatic glycogen synthesis and by the effect of glucagon and other counter-regulatory hormones on glycogenolysis. Insulin exerts an anti-lipolytic effect in WAT which maintains insulin sensitivity in muscle by reducing the availability of fatty acids and their derivatives. In addition, insulin promotes de novo lipogenesis in WAT, indirectly by stimulating glucose transport through GLUT4 which activates the lipogenic transcription factor, ChREBP. This results in synthesis and secretion of metabolically beneficial signaling lipids which increase muscle insulin sensitivity. Adipocytes also synthesize and release adipo-cytokines which alter metabolism in liver and muscle. The role of insulin in the regulation of adipo-cytokines needs further investigation. Black arrows indicate insulin direct effects, green arrows indicate insulin indirect effects. GLUT4= Glucose transporter type 4; ChREBP= Carbohydrate-responsive element-binding protein; Acetyl-CoA= acetyl coenzyme A; PC= pyruvate carboxylase.

Insulin control of GLUT4 trafficking, “translocation model”, as a mechanism to regulate glucose uptake was the first example of a transporter protein’s net activity being regulated by the control of cellular compartmentalization. This mode of regulation is distinct from regulation by gating the transport activity. In addition, because the insulin receptor is a tyrosine kinase, this opened a new field of study of the regulation of membrane trafficking downstream of a tyrosine kinase receptor. GLUT4 trafficking pathways and the intricate cell biology underlying insulin regulation of this trafficking have been extensively studied in both adipocytes and muscle cells. Readers are referred to recent detailed review on this topic (Klip et al., 2019).

Metabolic fate of glucose in adipocytes.

Although insulin stimulates glucose uptake by both adipocytes and muscle cells by eliciting GLUT4 translocation to the plasma membrane, the fate of the glucose differs between these cell types. In muscle cells the glucose in excess of what is needed to fuel anabolic activities is stored primarily as glycogen, to be used during fasting or when a sustained burst of exercise is required, thus providing a cell-intrinsic energy source (DeFronzo et al., 1981; Shulman et al., 1990). Key metabolic fates of glucose in adipocytes are schematized in Figure 1. Very little of the glucose taken up by adipocytes is stored as glycogen (<5% of glucose taken up; (Flatt and Ball, 1964)), although the dynamics of glycogen in adipocytes has a role in coordinating glucose and lipid metabolism (Markan et al., 2010). The increased glucose uptake stimulated by insulin contributes to triacylglyceride (TG) synthesis by providing glycerol-phosphate as the backbone for free fatty acid esterification as well as through the de novo synthesis of free fatty acids (de novo lipogenesis, DNL), with at most 50% of glucose stored as TG (Flatt and Ball, 1964). A considerable amount of the glucose taken up by adipocytes is metabolized to lactate and secreted from cells (Lagarde et al., 2020). The lactate can be used by other cells as a fuel source or perhaps as a signaling metabolite by activating the lactate receptor. The biology of lactate production by adipocytes is an active area of research. Another metabolic fate of the glucose is oxidation to CO₂ which leaves the adipocytes, creating a net loss of carbon from adipocytes.

It is perplexing that only about half of the glucose taken up by adipocytes ends up stored as TG or glycogen, the two major energy stores in cells. This is particularly surprising in light of the pronounced effect insulin has on increasing adipocyte uptake of glucose. What are the other physiologic impacts of insulin-stimulated glucose uptake by adipocytes in addition to energy storage? A number of studies support the hypothesis that adipocytes sense whole-body glucose homeostasis based on insulin control of glucose flux. Deletion of GLUT4 from adipocytes, thereby markedly reducing insulin-stimulated glucose uptake, causes hyperglycemia and peripheral insulin-resistance without altering the number or size of adipocytes (Abel et al., 2001). Whereas over-expression of GLUT4 by ~6-9-fold selectively in adipocytes in mice results in improved glucose tolerance despite increased adiposity due to adipocyte hyperplasia with no change in adipocyte size (Shepherd et al., 1993). Because less than 20% of total postprandial glucose is disposed of by adipocytes (Marin et al., 1992), these data reveal that insulin-control of glucose transport into adipocytes has an important impact on whole body glucose homeostasis that extends beyond the amount of glucose that is disposed of by adipocytes. Consistent with this hypothesis, deletion of Rab10 from adipocytes, a protein key for insulin-stimulated GLUT4 translocation to the plasma membrane, phenocopies adipose GLUT4 knockout, inducing liver insulin resistance and glucose intolerance (Eguez et al., 2005; Vazirani et al., 2016). Interestingly, Rab10 knockout only reduces insulin-stimulated glucose uptake by 50%, demonstrating how finely tuned whole-body glucose homeostasis is to insulin control of glucose transport and metabolism in adipocytes. Furthermore, the 2 fold change caused by Rab10 knockout is within the normal physiologic range of changes into glucose flux into adipocytes suggesting that physiological changes in insulin-stimulated glucose transport in adipocytes are likely to regulate whole body glucose homeostasis.

Consequently, although adipocytes were originally considered to be inert storage depots designed to supply energy when food was unavailable, it is now clear that adipocytes are master regulators of both systemic energy balance and glucose/insulin homeostasis. In the following sections, we review how adipocytes exert these effects on peripheral metabolism.

Insulin regulation of systemic metabolism

De Novo Lipogenesis (DNL)

An important fate of glucose in adipocytes is conversion to fatty acids via DNL. This has a major impact on whole-body metabolism not because of the amount of glucose that is metabolized in this pathway. But rather because some of the lipids that are synthesized from glucose serve as signals that regulate metabolism and inflammation in other tissues. Several well-studied transcription factors act as metabolic switches to regulate DNL including sterol regulatory element-binding protein (SREBP)1c, Carbohydrate Response Element Binding Protein (ChREBP), and liver X receptor (LXR). The dominant lipogenic factor in adipose tissue is ChREBP (Horton et al., 2003), which is activated by simple carbohydrates such as glucose and fructose. Insulin does not directly activate ChREBP but rather has a permissive role in its regulation by controlling glucose uptake. In addition to DNL, ChREBP regulates glycolysis and the pentose phosphate pathway which provide substrates for DNL (Baraille et al., 2015; Iizuka et al., 2004).

In rodents and humans, ChREBP expression in adipose tissue is tightly linked with insulin sensitivity (Herman et al., 2012), underscoring the importance of adipocyte glucose metabolism for system metabolic regulation. Downregulation of ChREBP selectively in adipocytes results in systemic insulin resistance with impaired insulin action in the liver, muscle, and WAT (Vijayakumar et al., 2017). In humans, WAT ChREBP and lipogenic enzyme expression strongly correlate with insulin sensitivity even among obese people (Eissing et al., 2013; Kursawe et al., 2013; Roberts et al., 2009). Improving insulin sensitivity in insulin-resistant people restores WAT ChREBP expression (Eissing et al., 2013; Kursawe et al., 2013). In humans and mouse models, the expression of GLUT4 and ChREBP in WAT are highly correlated (Hammarstedt et al., 2018; Herman et al., 2012). Furthermore, overexpression of GLUT4 selectively in adipocytes in mice is sufficient to induce ChREBP expression in WAT in association with upregulation of lipogenic enzymes and DNL. ChREBP expression is even linked to the ability of insulin to stimulate GLUT4 translocation since adipose-specific knockout of Rab10 results in reduced ChREBP expression. Surprisingly, there seems to be a reciprocal relationship between GLUT4 expression/translocation and ChREBP since knockout of ChREBP in adipocytes in mice results in a defect in insulin-stimulated glucose transport due to impaired GLUT4 translocation to the plasma membrane (Vijayakumar et al., 2017). This is in spite of normal insulin-stimulated Akt phosphorylation. The mechanism for this is an important area of investigation.

ChREBP drives the synthesis of many types of fatty acids, and its activation can alter the ratio of monounsaturated to saturated fatty acids (Denechaud et al., 2008). However, one of the most consequential effects of ChREBP activation is increased synthesis of lipids that function as signaling molecules, often exerting their effects through cell surface receptors. One such family is the structurally new class of endogenous lipids, the branched Fatty Acid Hydroxy Fatty Acids (FAHFAs) which consist of a fatty acid and a hydroxy fatty acid bound by an ester bond (Yore et al., 2014). In 2014, a subfamily of FAHFAs, the Palmitic Acid esters of Hydroxy Stearic Acids (PAHSAs), was first demonstrated to have anti-diabetic and anti-inflammatory effects (Syed et al., 2018a; Yore et al., 2014). In 2018, Pflimlin et al. failed to find beneficial metabolic effects with acute or subchronic treatment of HFD-fed mice with PAHSAs (Pflimlin et al., 2018). Numerous methodological issues contributed to these negative results, as discussed in subsequent publications (Kuda, 2018; Syed et al., 2018b). Since then, other labs have shown that PAHSAs and other FAHFAs have beneficial metabolic effects (Bandak et al., 2018; Benlebna et al., 2020a; Benlebna et al., 2020b; Schultz Moreira et al., 2020; Wang et al., 2018; Wen et al., 2020; Zhou et al., 2019) although some FAHFAs may not be beneficial.

While PAHSAs are present in all tissues studied, their levels are highest in WAT and BAT. PAHSA levels are reduced in serum and subcutaneous (SQ) adipose tissue of insulin-resistant humans and mice, and these levels correlate highly with insulin sensitivity (Yore et al., 2014). PAHSAs improve glucose homeostasis in insulin-resistant obese mice by enhancing insulin action to reduce hepatic glucose production which results at least in part, from improved suppression of lipolysis by insulin in adipocytes (Zhou et al., 2019).

PAHSAs also lower adipose tissue inflammation in obese mice (Yore et al., 2014). Their anti-inflammatory effects are further evident by the fact that they markedly reduce Type 1 diabetes incidence (Syed et al., 2019) and colitis severity in mouse models (Lee et al., 2016). In human islets, PAHSAs augment glucose-stimulated insulin secretion and protect against metabolic stresses (Yore et al., 2014). Undoubtedly, some of the beneficial effects of PAHSAs result from their local production in tissues. However, adipose tissue contributes significantly to circulating PAHSA levels (personal communication). Multiple additional FAHFAs are also biologically active have anti-diabetic (Benlebna et al., 2020a) and anti-inflammatory effects (Kolar et al., 2019; Paluchova et al., 2020) as well as other beneficial effects such as improving cognition in obese mice (Wen et al., 2020). The abundance of PAHSA isomers is reduced in WAT and serum of AdChREBP KO mice, and 9-PAHSA supplementation completely reverses their insulin resistance (Vijayakumar et al., 2017). The mechanism appears to involve decreased WAT inflammation and increased insulin-stimulated adipocyte glucose transport. This illustrates how ChREBP activation, in response to increased glucose flux into adipocytes, has an important role in modulating systemic insulin sensitivity.

In addition to FAHFAs, other signaling lipids are synthesized and released from adipocytes such as diacylglycerols (DAGs), acylcarnitines, ceramides, prostaglandins, lysophosphatidic acid, palmitoleate, oxylipins and N-acyl amino acids as extensively reviewed in (Li et al., 2020b)(Chaurasia et al., 2020; Gancheva et al., 2018; Petersen and Shulman, 2018; Yang et al., 2018). Their roles include antagonizing or augmenting insulin action in other tissues, and stimulation or inhibition of fatty acid uptake, cellular respiration and mitochondrial uncoupling. Here, we briefly discuss aspects of the regulation of insulin sensitivity by some of these lipids, while acknowledging there is still much to be discovered about the functions of these lipids in metabolic regulation.

Some human and rodent studies have shown a correlation between increased DAGs and hepatic insulin resistance (Kumashiro et al., 2011; Luukkonen et al., 2016; Ruby et al., 2017). A number of studies seeking to establish a causal relationship have provided varying results (Petersen and Shulman, 2018). DAGs may promote insulin resistance by activating novel PKC isoforms, which alter the phosphorylation of insulin signaling molecules thereby impairing signal transduction (Erion and Shulman, 2010; Samuel et al., 2010). Additional studies have aimed to establish a causal relationship between novel PKCs and lipid-induced insulin resistance (Petersen and Shulman, 2018). The structural complexity of DAGs based on differences in fatty acid chains supports the possibility that structurally distinct DAGs might have different effects on insulin action. In addition, the effects of PKC on insulin signaling might be influenced by targeting of PKC to specific subcellular membrane compartments. Consequently, additional studies are needed to identify the mechanism by which specific DAGs might cause insulin resistance.

Similarly, circulating ceramides are elevated in subjects with insulin resistance, type 2 diabetes and non-alcoholic fatty liver disease (Haus et al., 2009; Jensen et al., 2019; Wigger et al., 2017). Also, ceramide levels are elevated in adipose tissue biopsies from insulin-resistant subjects (Kolak et al., 2007). However, the association between ceramides and insulin resistance has not been observed in all studies and their mechanism of action is still ill-defined (Chaurasia et al., 2020; Petersen and Shulman, 2018) . As is the case for DAGs, structural variations among ceramide species (acyl chain length, saturation of fatty acids, and subcellular localization of ceramides) affect their biological functions (Park et al., 2014), complicating studies of the link between ceramides and insulin resistance. Recent studies using genetic mouse models indicate that adipose tissue and liver ceramides play a role in the pathogenesis of hepatic steatosis and insulin resistance caused by leptin deficiency or obesogenic diets. On the contrary, dihydroceramides do not show the same detrimental effects. These lipids are precursors of the toxic ceramides and other sphingolipids and the ratios of ceramides/dihydroceramides are inversely associated with insulin sensitivity (Chaurasia et al., 2019). In some models of insulin resistance, ceramide changes were paralleled by DAG changes (Bruce et al., 2012; Bruce et al., 2013). This observation supports a debate on the importance of ceramides vs DAGs in insulin resistance in liver, muscle and adipose tissue. More studies are needed to further elucidate the role of specific ceramide stereoisomers in the regulation of insulin sensitivity and the mechanism underlying their pro-inflammatory actions in adipose tissue (Chaurasia et al., 2020).

In addition to DAGs and ceramides, acylcarnitines might also play a role in muscle insulin resistance. Acylcarnitines accumulate in muscle mitochondria from insulin resistant rats because the persistent lipid load causes high rates of β-oxidation outpacing the metabolic capacity of the TCA cycle (Koves et al., 2008; Muoio and Koves, 2007; Yang et al., 2018). In vitro studies and human studies have sought to establish a causal role for acylcarnitines in muscle insulin resistance (Aguer et al., 2015; Bruls et al., 2019). Nonetheless, it remains unclear whether acylcarnitines simply reflect defective lipid metabolism in insulin resistant muscle or cause muscle insulin resistance (Schooneman et al., 2013).

Insulin regulation of lipolysis and its systemic consequences.

Insulin suppression of lipolysis in adipocytes is a major action by which it regulates systemic metabolism. The precise mechanisms by which insulin suppresses lipolysis are still not fully understood but appear to involve inhibition of the adrenergic cAMP/PKA signaling pathway that stimulates lipolysis (Choi et al., 2006; Duncan et al., 2007; Jaworski et al., 2007; Vaughan, 1964). This anti-lipolytic effect of insulin links adipose tissue metabolism to hepatic gluconeogenesis, by regulating the flux of fatty acids and glycerol to the liver (Bergman, 2000; Perry et al., 2015; Sindelar et al., 1997) (Figure 1B). Glycerol is a gluconeogenic substrate, while fatty acids, once converted to acetyl CoA, are positive allosteric modulators of pyruvate carboxylase, which catalyzes the carboxylation of pyruvate to oxaloacetate, which is the first step of gluconeogenesis (Figure 1B). The relative importance between cell-autonomous and indirect effects of insulin on the regulation of hepatic glucose production is unresolved. A confounding factor is that the effects of insulin on the hepatocyte metabolism are influenced by the nutrient state of the liver. For example, data from several groups support the hypothesis that hepatic glucose production (HGP) is controlled by insulin’s direct effect to promote glycogenolysis when glycogen levels are higher (Edgerton et al., 2017; Perry et al., 2015; Titchenell et al., 2016). In contrast, in the fasted state when insulin levels fall, HGP is regulated more by the indirect effect of insulin on adipose tissue lipolysis. Thus, increased gluconeogenesis occurs by a substrate (fatty acid and glycerol) push mechanism resulting from the fact that insulin no longer inhibits lipolysis.

As is the case for the liver, insulin action in adipocytes affects muscle insulin sensitivity (Figure 1B). Knockout of GLUT4 or ChREBP selectively in adipocytes in mice results in decreased glucose transport in skeletal and cardiac muscle (Abel et al., 2001; Vijayakumar et al., 2017). The effects of insulin on adipocyte lipid metabolism as well as changes in adipo-cytokines are important for adipose-muscle communication. However, it is not clear whether insulin signaling in adipocytes directly affects adipokine release. Adipocytes also regulate whole body insulin sensitivity, inflammation and vasoactivity through the synthesis and release of hormones, cytokines and other proteins as previously reviewed (Chawla et al., 2011; Li et al., 2020b; Makki et al., 2013). The discovery of leptin which regulates energy balance and neuro-endocrine function was seminal in understanding the importance of the adipocyte as an endocrine cell (Friedman, 2016). Some of these worsen insulin resistance and increase the risk for Type 2 diabetes while others enhance insulin sensitivity. Whether and how insulin regulates each of these molecules deserves further investigation. Data suggest that the synthesis and secretion of leptin and adiponectin, two well-studied adipokines, are not regulated by insulin. However, the fatty acid-binding protein (FABP) 4 is transcriptionally regulated by insulin (Trojnar et al., 2019), illustrating that insulin has some important effects on the adipose tissue secretome.

Adipose Tissue Remodeling in Adaptive Energy Homeostasis: the role of insulin.

Adipocytes can rapidly expand or contract in response to changes in the nutritional status. This dynamic phenomenon is possible in part due to changes in various stromal vascular cells, including immune cells, and extra cellular matrix that constitute adipose tissue mass. These events are known as “adipose tissue remodeling” and they are responsible for the adaptation to a spectrum of metabolic states (Choe et al., 2016; Vegiopoulos et al., 2017) (Figure 2). In addition to alterations in adipocyte metabolism (fatty acid/glycerol fluxes), remodeling is accompanied by altered secretion of adipo-/lipokines mentioned above.

Figure 2.

A. Adipose tissue remodeling in normal physiology and pathophysiology. Adipocytes undergo rapid and reversible morphologic changes in the transition between fasting and fed state. Upon food ingestion, adipocytes expand to store excess fuel including circulating lipids and glucose, as triglycerides (TG) in the lipid droplet. This is accompanied by altered secretion of lipids, cytokines and other adipokines. Adipocytes can expand by two different mechanisms: hypertrophy (increase in size) and hyperplasia (increase in number). Once the expansion limit is reached, adipocytes become dysfunctional. Chronic overnutrition is associated with hypertrophic adipocytes presenting crown-like structures which contain necrotic cells and macrophages. Aberrant adipose tissue remodeling with chronic overnutrition is accompanied by immune cell recruitment and activation, hypoxia, fibrosis.

Figure 2B. Insulin effects on glucose and lipid metabolism in functional and dysfunctional adipocytes (Left panel) Healthy adipocytes are sensitive to the metabolic actions of insulin. (Right panel) Overloaded or dysfunctional adipocytes are resistant to many actions of insulin. This has consequences for ectopic lipid accumulation in other tissues. Dysfunctional adipocytes also have reduced secretion of lipo-cytokines which have beneficial metabolic effects and increased secretion of detrimental proteins and lipids which exert adverse effects in other organs.

Fasted and fed states.

The ratio of insulin to glucagon controls the short-term systemic availability of fuels, including glucose and fatty acids, in the physiological transition between fasting and food ingestion, one of the major situations in which adipocyte size changes (Figure 2). As noted above, in the postprandial state, which is accompanied by high blood insulin levels, insulin promotes triglyceride (TG) storage in lipid droplets and inhibits TG release by lipolysis as fatty acids and glycerol. Insulin promotes the formation of TGs by at least four different mechanisms: (i) insulin-stimulated glucose uptake provides glycerol-3-phosphate to which fatty acids can be esterified to synthetize TGs; (ii) insulin activates lipoprotein lipase, an endothelial enzyme degrading circulating TGs that are embedded in VLDLs and chylomicrons, liberating fatty acids and glycerol (Goldberg et al., 2009); (iii) insulin also promotes fatty acid uptake in adipocytes by increasing the translocation of fatty acid transporters to the plasma membrane (Stahl et al., 2002), and; (iv) DNL, as described above. As also noted above, insulin is the most potent anti-lipolytic hormone, rapidly suppressing plasma levels of non-esterified fatty acids through the inhibition of TG lipolysis in adipocytes(Griffin et al., 1999; Lewis et al., 1997).

With fasting there is an increase in the mobilization of lipids from adipose tissue because the reduced circulating insulin levels are not sufficient to inhibit lipolysis. The shift from lipogenesis to lipolysis in fasting results in a reduction of adipocyte size. There are important adipose depot-specific differences in the response to fasting re-feeding. Specifically, visceral adipose tissue is more rapidly “consumed” during fasting in mice while subcutaneous adipose tissue is preferentially remodeled during re-feeding (Tang et al., 2017). Understanding the molecular mechanisms for such differences among adipose depots could provide novel therapeutic targets for metabolic diseases. Adipose tissue remodels with fast fed transitions as part of normal tissue homeostasis. The prolonged fasting (24 hr for mice) lies outside that normal range and therefore physical changes in adipose tissue associated with more prolonged fasting might involve additional mechanisms.

Pathological remodeling of adipose tissue and insulin resistance.

In addition to the significant expansion and contraction of adipose tissue that occurs with the daily fed to fasted transitions, more extensive adipose remodeling accompanies states of over nutrition and involves changes in macrophages, other components of the stromovasculature and extracellular matrix (Choe et al., 2016; Rosen and Spiegelman, 2014)(Figure 2). Adipocytes first expand (hypertrophy) to accumulate more lipids which prevents ectopic lipid accumulation in cells of other tissues (e.g., liver, muscle and pancreas). Once the limit of expansion for an adipocyte is reached, hyperplasia takes place through proliferation and/or differentiation of pre-adipocytes (Wang et al., 2013). In humans, the order of these responses is adipose depot-specific (Tchoukalova et al., 2010). In rodents, adipocyte hypertrophy might be followed by adipocyte death and eventually by the appearance of new adipocytes (MacKellar et al., 2010; Strissel et al., 2007). These changes provide safe storage of fat in adipocytes, preventing the negative metabolic consequences of ectopic fat accumulation in other tissues. When adipose mass cannot expand and store more triglycerides, ectopic lipid accumulation occurs. This can result in insulin resistance in liver and muscle and impaired insulin secretion from pancreatic beta cells, thereby increasing the risk for Type 2 diabetes. Severe forms of ectopic lipid deposition seen in lipodystrophy syndromes are associated with some of the most extreme forms of insulin resistance as reviewed in (Vegiopoulos et al., 2017). Prolonged adipocyte hypertrophy is accompanied by sustained release of pro-inflammatory mediators and other lipids that reduce insulin sensitivity (for example, IL-6, TNF-α, DAGs, ceramides) (Martinez-Santibanez and Lumeng, 2014; Sun et al., 2011). “Adipose dysfunction”, characteristic of the state of adipocyte hypertrophy, encompasses the failure of adipose tissue to expand and accommodate increasing fat stores as well as to remodel without resulting in chronic proinflammatory stimulation.

Insulin resistance and Type 2 diabetes.

Many unanswered questions remain regarding the causes of adipose tissue insulin resistance and its consequences on other organs. Resistance to the anti-lipolytic effect of insulin increases hepatic gluconeogenesis in mice on HFD (Perry et al., 2015), but this might be limited to specific experimental paradigms (i.e. hyperinsulinemic-euglycemic clamp). Indeed, FFA levels are not necessarily elevated in humans with T2D (Fraze et al., 1985; Groop et al., 1989; Reaven et al., 1988; Swislocki et al., 1987). FFA fluxes rather than levels have been proposed to be important in determining glucose output by the liver (Petersen and Shulman, 2018). However, other mechanisms might also link adipose tissue dysfunction and whole-body insulin resistance in obesity including the altered release of adipo-cytokines.

To define the contribution of adipose tissue dysfunction to whole-body insulin resistance, significant attention has focused on the phenomenon known as “adipose tissue inflammation”. This is defined as immune cell infiltration and activation of macrophages to a proinflammatory state resulting in proinflammatory cytokine secretion in adipose tissue. As pointed out by others, this term might be inaccurate because classical signs of inflammation (e.g. pain) are missing in the case of adipose tissue chronically exposed to over nutrition (Rosen and Spiegelman, 2014). Also, the role of some immune cells in the initial remodeling of adipocytes in response to over nutrition is protective. In particular, macrophages are known to buffer FFAs liberated by excessive lipolysis (Prieur et al., 2011; Shapiro et al., 2013), facilitate angiogenesis and extracellular matrix deposition (Bourlier et al., 2008; Cho et al., 2007; Murdoch et al., 2005; Yancopoulos et al., 2000). These are all actions necessary for proper adipose tissue remodeling in response to nutrient overload and they actually prevent ectopic lipid accumulation in other organs. Only when over nutrition is prolonged do the immune cells resident in AT or actively recruited there, contribute to insulin resistance. The changes in the secretome of immune cells, the switch of macrophages to a proinflammatory phenotype in obesity, and the plethora of immune cells recruited to AT during over nutrition are very well documented (Chawla et al., 2011; Li et al., 2020a; Makki et al., 2013). However, new cell types continue to be discovered which may be involved in the process. For example, the coupled use of flow cytometry and electron microscopy has enabled observation of the recruitment of dendritic cells into adipose tissue of mice on High Fat Diet (Stefanovic-Racic et al., 2012).

It is not clear whether adipose tissue inflammation has a causal role in the onset of insulin resistance, because some studies have shown that adipose tissue insulin resistance develops independently of adipose tissue macrophage infiltration (Lyu et al., 2021; Nishimura et al., 2009). However, (Perry et al., 2015) demonstrated that adipose tissue inflammation (e.g. IL-6) can have a causal role in systemic insulin resistance in obesity using elegant flux studies. The latter evidence is consistent with a model in which adipose tissue inflammation is linked to lipotoxicity and whole-body insulin resistance. Many other pro- and anti-inflammatory adipo-/cyto-kines also regulate systemic insulin sensitivity (see above).

Future directions

Adipocytes have been a powerhouse for understanding insulin action. While major advances have been made since the discovery of insulin 100 years ago, in defining the molecular events underlying insulin action in adipocytes and the systemic consequences, there are still many unanswered questions. We pointed out how the mechanisms for some of insulin’s effects are still unclear. This is partly because the dominant regulatory pathways for metabolic regulation in adipose tissue, liver and muscle differ depending on the metabolic condition. In addition, we are just beginning to understand how insulin action in adipocytes integrates with other inputs such as neural innervation and activation (Guilherme et al., 2019). The role of brown and beige adipose tissue in the regulation of insulin sensitivity is a very productive research area and how insulin influences the “browning” or “beiging” of adipocytes is of high interest. Single cell RNAseq has been extremely informative for many research questions and is particularly useful for the study of adipose tissue. Going forward, a better understanding of the integrated network of cell types and pathways upstream and downstream of insulin in adipose tissue could lead to approaches to enhance systemic insulin sensitivity to prevent and improve the treatment of Type 2 diabetes.