Systemic Effects of White Adipose Tissue Dysregulation and Obesity-Related Inflammation
Amir J. Guri,
Nutritional Immunology and Molecular Medicine Laboratory, CyberInfrastructure Division, Virginia Bioinformatics Institute, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA
Nutritional Immunology and Molecular Medicine Laboratory, CyberInfrastructure Division, Virginia Bioinformatics Institute, Virginia Polytechnic Institute and State University, Blacksburg, Virginia, USA
Over the past 20 years, there has been an alarming increase in the prevalence of overweight and obesity (1). Current estimates indicate that ∼66% of the US population is overweight with >30% obese, a state of affairs that raises numerous health and fiscal concerns. With regard to the former, recent research has established that obesity significantly increases the risk of chronic diseases such as type 2 diabetes (T2D), cardiovascular disease, nonalcoholic fatty liver disease (NAFLD), colon cancer, and obstructive sleep apnea (2). Fiscally, obesity-related illness encompassed 10% of all health-care costs in the United States in 2008 (1,3), a rate that is anticipated to grow if trends in childhood obesity continue on their current trajectory. If the epidemic in obesity is not effectively dealt with, this reality forewarns of very serious health and financial problems for the United States.
The problems associated with obesity develop initially as a problem related to energy storage, thereby placing WAT at the front lines. Although WAT's principal function deals with energy storage, it serves other important homeostatic functions as well. WAT acts to buffer the amount of circulating fatty acids (FAs) that occurs during the postprandial period, protecting peripheral tissues from excessive lipid accumulation (4), and most prominent energy-storing cells in WAT, adipocytes, secrete hormones called adipokines that affect a diverse array of local and systemic functions. WAT also consists of a stromal vascular fraction comprised of preadipocytes, endothelial cells, fibroblasts, macrophages, monocytes, and lymphocytes, which also holds important metabolic and immunological roles. To gain greater perspective on the importance of WAT for normal metabolic function, researchers developed the A-ZIP/F-1 mouse, a model almost entirely deficient in WAT (5,6). Mice without fat experience a severe form of insulin resistance (IR), similar to humans born with genetic fat deficiency. The IR is ameliorated if WAT is transplanted into the body (5,6).
The A-ZIP/F-1 mouse exemplifies the importance of WAT in the maintenance of systemic insulin sensitivity, though it is still unclear how precisely WAT regulates systemic insulin response and glucose homeostasis. Differentiating the relative importance of WAT's energy storage vs. endocrine functions has become a complex task, and both of these functions appear to be involved WAT's systemic effects. The objective of this review is to describe how obesity-induced imbalances in WAT can lead to significant systemic consequences. A familiar outcome of WAT dysfunction is the initiation of a local inflammatory response in the various peripheral tissues, a response which can feedback on itself and perhaps potentiates a number of local metabolic disorders (Figure 1).
One explanation tying the energy-storing and endocrine functions together involves the deleterious effects of adipocyte hypertrophy. Adipocytes are well-known for their pliable nature and have the ability to expand to very large sizes. Human adipocytes can grow to ∼20-fold in diameter and several thousand-fold in volume (7). There appears to be a breaking point, however, when adipocytes become too large to function effectively mainly due to variety of stress signals released by the expanding cell. One source of these signals is thought to be endoplasmic reticulum (ER) stress (8). Nutrient excess is a well-known activator of ER stress, and when a certain threshold is passed a signaling cascade known as the unfolded protein response is initiated. The unfolded protein response triggers the activation of numerous stress and inflammatory pathways and the production of cytokines and chemokines, all of which have been shown to antagonize the insulin pathway (9).
There is also evidence that adipocyte stress signals may result from a hardening of its extracellular matrix. Extracellular matrix components such as collagen VI are upregulated in the obesity and act to increase the mechanical rigidity of adipocytes (10), thereby making it more difficult to expand. In mice deficient in collagen VI, adipocytes are able to enlarge “stress-free” during periods of positive energy balance, and resultantly these mice have a significantly improved inflammatory profile (10).
Because of the stress hypertrophic adipocytes are under, the cells are largely dysfunctional. The cells show a poor ability to take up and release FFA appropriately, and overall WAT from obese individuals demonstrates a net efflux of FAs (11). Thus, peripheral tissues, including SM, liver, pancreas, heart, kidney, and blood vessels, must increase their lipid intake to compensate, exposing them to the potentially harmful affects of lipotoxic compounds.
The adipokine profile of WAT of obese individuals is also remarkably different in comparison to lean individuals. The proportion of proinflammatory adipokines and cytokines, such as leptin, interleukin-6 (IL-6), IL-8, monocyte chemoattractant protein-1 (MCP-1), resistin, retinol-binding protein-4, and interferon γ inducible protein-10 increases significantly while adiponectin, the main anti-inflammatory compound secreted by WAT, decreases (12). Adipokines play central roles in the regulation of food intake, insulin action, lipid and glucose metabolism, angiogenesis and vascular remodeling, regulation of blood pressure, and coagulation (13), and therefore alterations can have significant consequences.
Immune cell infiltration into the stromal vascular fraction
Under obese conditions, stressed adipocytes begin recruiting immune cells into the stromal vascular fraction. In 2003, Weisberg et al. (14) and Xu et al. (15) demonstrated that macrophages are increased in WAT of obese mice and humans, and recently Nishimura et al. discovered that CD8+ T cells precede and contribute to the macrophage influx (16). The physiological role of infiltrating adipose tissue macrophages is thought to be debris clearing in nature. Once extravasated into the tissue, macrophages localize around dead adipocytes and form crown-like structures that can persist for long lengths of time (17). The F4/80+CCR2+CD11c+ macrophages recruited to WAT are significantly more inflammatory than resident adipose tissue macrophages from lean individuals (18,19,20) and are responsible for the majority of WAT proinflammatory cytokine and chemokine secretion. Eventually, a positive feedback cycle is formed whereby infiltrating macrophages recruit and sustain the presence of new macrophages, thus resulting in further propagation of inflammation (21,22).
The remainder of this article will review the systemic direct and indirect effects of obesity and adipocyte hypertrophy-related manifestations on important regions in the body.
Proper maintenance of blood glucose levels relies on concerted regulation of a number of metabolically important tissues. As a tissue responsible for ∼80% of insulin-stimulated glucose disposal, SM is a region that, when compromised, can create numerous complications for the remaining tissues in the body. Much of investigation on SM IR has focused on lipid-induced effects, but SM inflammation is beginning to enter the dynamic.
FA transport and entry
Numerous studies have linked to the development of SM IR with nonesterified FA (NEFA) overexposure. Insulin signaling in SM and cultured myocytes is inhibited by NEFA overexposure (23,24), and increasing FA levels in rodents and humans to high levels abolishes insulin sensitization of the insulin receptor substrate-1/phosphoinositide kinase-3 insulin signaling pathway (23,24,25).
The inhibition of SM insulin response is preceded in vivo by an upregulation in SM FA transporters and increase in intramyocellular fat content (26,27). FA transporters are required for long-chain FAs (LCFAs) such as palmitate to enter the myocyte. FA translocase (FAT/CD36), plasma membrane-associated FA-binding protein (FABPpm), and FABP1–6 are all enhanced in the plasma membrane under obese conditions (27,28).
FAs entering the myocyte are immediately acted upon by acyl-CoA synthetase, which attaches an acyl-coenzyme A (acyl-CoA) moiety to form an FA-acyl-CoA. The FA-acyl-CoA can migrate toward the mitochondria to be oxidized or is utilized in the production of intramyocellular lipid (29). The entry of LCFAs into the mitochondria requires further modification from the protein, carnitine palmitoyltransferase-1, making it an essential point in allosteric regulation of β-oxidation (29).
The mechanisms of lipodystrophy
The increased presence of FAs promotes FA oxidation, in part because FAs directly activate genes necessary for β-oxidative processes. FAs and their lipid derivatives function as ligands for the peroxisome proliferator-activated receptor (PPAR) family of nuclear receptors, which include PPARs α, γ, and δ (30,31). PPAR-γ activation in the myocyte promotes FA storage and glucose utilization, whereas activation of PPARs α and δ, the two more abundant PPARs in SM, increase the production of pro-oxidative proteins (32).
The increase in FA abundance and oxidation in the myocyte ultimately affects insulin-stimulated glucose transport into the cell. As confirmed by magnetic resonance spectroscopy, glucose transport is the rate-controlling step for insulin-stimulated glucose disposal into SM (33). One theory underlying the connection between FA overabundance and IR purports that mitochondria in obese and T2D patients become dysfunctional under such chronic conditions. The dysfunctional mitochondria are less able to perform their oxidative capacities, resulting in the eventual build-up of intramyocellular fat and LCFA-CoAs. LCFA-CoAs can be converted into ceramide and diacylglycerol, noted in vitro activators of stress kinase pathways protein kinase C and IκB kinase-β (IKKβ) that inhibit insulin by phosphorylating serine residues of insulin receptor substrates 1 and 2 (33). Patients with T2D have ∼30% fewer mitochondria and significantly reduced mitochondrial density and mass than their healthy counterparts, key pieces of evidence supporting mitochondrial dysfunction (34,35). Insulin-resistant offspring of T2D patients also have decreased mitochondrial density (36), a finding which suggests that reduced mitochondrial density may be a primary cause of pathology in some individuals.
Despite these associations, evidence supporting a direct association between mitochondrial dysfunction and IR has been conflicting. In support, overweight/obese individuals undergoing a 4-month intervention of physical activity and food restriction have increased mitochondrial size and electron transport chain activity alongside improvements in insulin sensitivity (37). Treatment with the antidiabetic drug rosiglitazone for 8 weeks, however, enhances SM insulin sensitivity without significantly altering mitochondrial function (38). Also, dietary restriction in the absence of physical activity improves SM IR without significantly altering mitochondrial metabolism (39). Rather, it may be the case that a deficiency in insulin signaling adversely affects mitochondrial metabolism. In support of this latter possibility, studies have shown that administering insulin increases mitochondrial ATP production in healthy individuals, but not in those with T2D (40). Muscle mitochondrial protein content is also enhanced by insulin treatment in healthy, but not insulin-resistant, individuals (40,41).
A separate theory put forward by Koves et al. and Muoio contends that SM IR is more closely related to lipid-induced mitochondrial stress rather than mitochondrial dysfunction (42). This stress occurs when the increase in β-oxidation is not matched by a concomitant increase in tricarboxylic acid cycle genes in the mitochondria and results in the release of incompletely oxidized intermediates, such as acylcarnitines, which activate the stress kinase pathways (42). In support of this theory, mice lacking the enzyme malonyl-CoA decarboxylase, an enzyme which increases the rate of β-oxidation by relieving malonyl-CoA-mediated inhibition of carnitine palmitoyltransferase-1, are protected from high-fat-diet-induced glucose intolerance despite a high presence of LCFA-CoAs in the cytosol (42). Studies investigating the effect of exercise intervention in T2D patients and obese Zucker rats also support the assertion that high levels of intramuscular fat and LCFA-CoAs are not directly responsible for IR in SM (43,44).
The theory put forward by Koves and Muoio also suggest an explanation for the so-called “athlete's paradox,” a term used to describe the observation that exercise-trained athletes have enhanced insulin sensitivity despite having intramuscular fat levels similar to that of obese and T2D patients (45). The latter groups, of course, tend to be more prone to IR. The authors have suggested greater levels of the mitochondrial flux protein PPAR-γ coactivator-1α brought about by training may place athletes in a better position to respond to an increased presence in intramuscular fat than untrained individuals by tightening the link between oxidation and the tricarboxylic acid cycle (42). It was recently determined, however, that overexpression of PPAR-γ coactivator-1α does not correct obesity-induced imbalances in FA oxidation or mitochondrial DNA content (46), a finding which suggests that other than PPAR-γ coactivator-1α are involved.
An alternative explanation is that muscles from exercise-trained athletes are better able to divert potentially toxic lipid substrates into more innocuous, less oxidizing forms than untrained individuals. Endurance athletes have been shown to express greater levels of the triglyceride (TG)-synthesis protein diacylglycerol acyltransferase-1, an enzyme which catalyzes the final step in TG synthesis (47). Liu et al. have shown that transgenic overexpression of diacylglycerol acyltransferase-1 in mice mimics the insulin-sensitizing effects of exercise in muscle (47), lending significant credence to this possibility.
These findings do show a definitive role for NEFA overexposure and their transport into SM in the development of SM IR, though the mechanisms by which they lead to dysfunction still require greater elucidation.
Adipokines and SM
Adipokines have also been shown to mediate important aspects of SM metabolism function, and therefore the significant change in adipokine secretion that occurs during obesity may play a role in the tissue's dysfunction. SM also develops resistance to both adiponectin and leptin during the course of obesity (26,48). The reductions in adiponectin activity following high-fat feeding occur in as little as 3 days and precede the increase in FA transport (2 weeks) and decrease in insulin activity (4 weeks) (49). Some of this resistance may be attributable to decreases in the adiponectin receptors AdipoR1/R2 but findings that SM (50), but the findings that AdipoR1 protein content is similar among lean and obese humans detracts from this possibility (51,52). For leptin, the protein suppressor of cytokine signaling-3 has been linked with the diminished activity of its receptor (53). Suppressor of cytokine signaling-3 prevents leptin-induced activation of adenosine monophosphate-activated kinase (AMPK) by inhibiting STAT3, a main signaling component (53,54).
Both leptin and adiponectin protect SM from lipotoxicity and increase the activity of AMPK, a simulator of FA oxidation (55). The pro-oxidative effects of leptin have been shown to be effective in muscle of lean, but not obese, humans (56), an observation that has been attributed to leptin resistance. Fuentes et al. recently demonstrated that the 170-kDa (OB-R170) receptor, the main receptor linked to leptin's intracellular signaling (57), is significantly decreased in muscles of obese individuals, as too was leptin signaling, as reflected by lower levels of phosphorylated Tyr705-STAT3, Thr172-AMPKα, and Ser221-ACCβ (48).
Adiponectin has been shown to enhance FA oxidation by inducing mitochondrial biogenesis and PPAR-α-responsive gene expression in SM (58,59). Mice lacking adiponectin are born with pre-existing mitochondrial dysfunctions (59), adding support to the notion that its decline is related to the mitochondrial stress observed in obese patients. Both leptin and adiponectin are also implicated in decreasing FA transport abundance in the plasma membrane (26). In lean rats, plasma membrane concentrations of FAT/CD36 and FABPpm decline significantly following treatment with exogenous leptin (60).
Resistin is another adipose-derived factor that has been shown to influence SM substrate metabolism and insulin sensitivity. Like leptin, resistin is increased in genetic and diet-induced models of obesity (61), though its effect on FA oxidation are opposed to that of leptin. Treatment of L6 myotubes with resistin significantly attenuates AMPK activation and FA oxidation and causes an increase in intramuscular lipid content within 24 h (62). Resistin also significantly impairs insulin signaling in cultured myotubes and in isolated rat soleus muscles (63,64), though the presence of palmitate appears requisite for its effects on oxidation and insulin sensitivity (64). Interestingly, resistin's insulin-antagonizing effects in the latter study did not involve an increase in ceramide levels, though reducing ceramide levels still mitigated its effect on glucose transport and insulin signaling. Today, resistin's role in pathogenesis of SM IR in obesity remains complex and somewhat undefined.
Evidence of a local immune response in SM
One possibility why researchers have been unable to explain SM IR solely in terms of lipids, lipid derivatives, mitochondrial irregularities, or adipokines is that these factors are only part of the overall picture. Integrating knowledge about immune cells, cytokines, and adipokines is essential for a comprehensive understanding of SM dysfunction. For instance, in regard to proinflammatory cytokine involvement, it has been shown that tumor necrosis factor-α (TNF-α) levels are increased in SM of obese humans (65), that acute TNF-α infusion decreases SM insulin sensitivity (66), and that in vitro exposure of proinflammatory cytokines to myotubes impairs glucose transport and insulin action (67). Acute treatment with the anti-inflammatory cytokine IL-10 protects SM from IL-6- and lipid-induced IR (68). Interestingly, although exogenous treatment of IL-6 has been shown to have insulin-antagonizing effects, it is also produced in high amounts by exercising muscle and may also induce anti-inflammatory effects under certain conditions. IL-6 has been shown to suppress TNF-α in exercise conditions (69), and transient increases in plasma IL-6 have been shown to increase the production of anti-inflammatory cytokines IL-1ra and IL-10 (70).
However, the conditions and environment in obesity differ from that of exercise. In the former, inflammation has been shown to be predominantly immune-cell derived, rather than emanating from distressed myocytes. Deletion of IKKβ in myocytes does not affect FA-induced IR (71), whereas deletion of IKKβ in myeloid cells protects the mouse from high-fat-diet-induced IR (72). Although it is possible that adipose tissue macrophages exert a distal affect on SM insulin response, there is now strong evidence of a local immune response in SM of obese mice and humans. Macrophages infiltrate SM in obese mice and humans and are associated with increased TNF-α, IL-6, and chemokine receptor-2 (73,74). We have found similar results in our lab as well. C57BL6/J mice fed high-fat diets for 70 and 140 days and obese db/db mice each showed significantly greater immune cell infiltration in the hindlimb muscles than mice fed low-fat diets (Table 1). The percent of F4/80+CD11b+ macrophages were also significantly enhanced in mice fed high-fat diets after 140 days (Figure 2).
Table 1. Effect of diet-induced and transgenic obesity on skeletal muscle immune cell infiltration and area of intramuscular adipocytes
The mechanisms underlying SM macrophage infiltration are unclear, though myocyte stress signals, initiated perhaps through incompletely oxidized FA intermediates, oxidative stress, are thought to be involved. We suspect that the local immune response in SM, as well as in other tissues, stem from the combination of factors prevalent in obesity, including NEFA overexposure, adipokine imbalance, and systemic low-grade inflammation.
The liver is responsible for regulating ∼80% of endogenous glucose production (75) and holds other key roles in regulating lipid and protein metabolism, lipoprotein metabolism, and detoxification (76). Hepatic lipid content and IR are associated with increased size and abundance of TG-rich very-low-density lipoproteins, which contribute heavily to cardiovascular disease and atherosclerosis (77). Both lipid accumulation and inflammation have been proposed to play significant roles in the dysfunction of this tissue, which because of its diverse array of metabolic activities can hold significant systemic consequences.
Liver TG accumulation and hepatic inflammation
Similar to SM, the liver increases its uptake and synthesis of fat during the onset of obesity (78), and obese patients overexpress genes involved in lipid storage and FA metabolism, including PPAR-γ2, acyl-CoA synthetase long-chain family member-4, FABP4, FABP5, and lipoprotein lipase (79). NAFLD, which encompasses obesity-induced liver steatosis, nonalcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma, is also directly and independently associated with IR (80). Yet, even with this correlation and evidence revealing that mice lacking liver FABPs are protected from obesity and obesity-induced IR (81), it is still unclear whether or not liver steatosis damages insulin sensitivity. Some studies showing that TG pools in the liver may be protective against FA-induced lipotoxicity. Those pools created by unsaturated FAs have been shown to buffer the apoptotic effects of palmitate in various cell lines (82). Moreover, treatment of palmitate to HepG2 liver cells impairs IR without increasing TG accumulation, whereas the unsaturated FA oleate acts to increase intracellular TG accumulation without impairing insulin signaling (83).
One possibility is that chronic lipid infiltration induces stress kinase pathways and nuclear factor κB (NF-κB)–induced inflammation in hepatocytes which, over time, can worsen and significantly impair hepatic function. Mitochondrial dysfunction in hepatocytes is a potential source of such stress as its development precedes that of hepatic IR in Otsuka Long-Evans Tokushima Fatty rats (84). Eventual involvement of NF-κB is also well documented. Obesity-induced increases in hepatic expressions of TNF-α, MCP-1, MIP-1α, Rantes, and IP-10 do not manifest in mice lacking hepatic IKK∍ (85), a key component of the NF-κB signaling pathway. Likewise, mice deficient in hepatocyte IKKβ are immune from hepatic, though not systemic, IR (72). To gauge the importance of hepatocyte NF-κB in the hepatic dysfunction, Cai et al. transgenically expressed IKKβ in lean mice to values similar to that of obese mice. The transgenic mice developed both hepatic and SM IR without steatosis and had increased IL-1β plasma concentrations and TNF-α and IL-6 liver production (86). The activation of Kupffer macrophages, the resident macrophage population in the liver, appeared to play a critical role, particularly in promoting SM IR distally. In line with these findings, disrupting Kupffer cell phagocytic and secretory capacities with gadolinium chloride significantly improves glucose tolerance and hepatic insulin signaling (87). Also, the development of hepatic steatosis and IR following a high-fat or high-sucrose diet is prevented in mice with gadolinium chloride-induced Kupffer cell depletion (88).
The effect of adipokines on hepatic function
The effects of leptin and adiponectin have been well-characterized in the liver. Both compounds stimulate hepatic FA oxidation through AMPK (89). Leptin has also been shown to be profibrogenic in the liver and activates Kupffer cells (90,91). Its profibrogenic effects appear to be dependent on sympathetic nervous system activation of norepinephrine (92), a noted promoter of hepatic fibrosis (93,94).
Studies examining the effect of adiponectin in various models of hepatic injury have shown it to be hepatoprotective. Chronic adiponectin treatment significantly improves hepatomegaly and steatosis in ob/ob and ethanol-fed mice (95). Various other forms of liver damage, including those induced by carbon tetrachloride, lipopolysaccharide/d-galactosamine, pharmacological compounds, bile duct ligations, and a methionine-deficient diet, are ameliorated with adiponectin treatment or worsened with its deficiency (96,97,98,99,100). Adiponectin treatment improves steatosis and necroinflammation in an experimental model of nonalcoholic steatohepatitis, in part through inhibition of TNF-α (95), and unlike leptin, adiponectin blocks NF-κB activation, inhibits IL-6 and chemokine production, and protects hepatocytes from Fas-mediated cell death (101,102), thereby eliciting an anti-inflammatory effect.
Other adipose-derived factors have also been associated with the development of hepatic steatosis and IR, though for number of these compounds there is limited mechanistic knowledge. One such example is the recently identified adipokine chemerin, which has been independently associated with obesity and NAFLD (103,104). It remains unclear, however, whether chemerin exacerbates NAFLD, whether its increase is caused by NAFLD, or whether its increase is coincidental to NAFLD. The full-length chemerin has also been shown to have a potent proinflammatory effect on activated macrophages expressing the chemerin receptor chemokine-like receptor-1 (105), though interestingly its short form, chemerin-15, has been shown to have anti-inflammatory effects (106).
Retinol-binding protein-4 is another compound that has been reported to be implicated in obesity-induced IR and NAFLD (107,108), though there are conflicting reports on its effects as well (109,110,111). Retinol-binding protein-4 is a specific transport protein for vitamin A that is secreted by hepatocytes as well as adipocytes. Its potential effect on hepatic steatosis is thought to be linked to its upregulation of gluconeogenic enzyme phosphoenolpyruvate carboxykinase in mice (112). Still, although some human data show positive correlations of serum retinol-binding protein-4 levels and NAFLD in obese or T2D patients (107,113), others have found no relationship at all (114).
Thus, although the cause of hepatic inflammation remains unclear, there are clear indications that activation of resident Kupffer macrophages and alterations in adipokines secreted from WAT may be key contributors to the imbalance that occurs in this tissue.
As the sole producer of insulin, the pancreas is a critically important tissue in the maintenance of glucose homeostasis. Insulin is produced by the β-cells of the islets of Langerhans, and these cells become progressively dysfunctional and apoptotic in the obese state (115). In fact, T2D does not occur in the absence of β-cell failure (115), making the understanding of how pancreatic cells become apoptotic a very important area of investigation.
During the early stages of diet-induced obesity the macronutrient environment is changing. Postprandial glucose levels are elevated for longer periods of time due to the progression of systemic IR and levels of NEFAs are also increased. In the healthy state, transient glucose and NEFAs elevations are important to regular pancreatic functioning, and both stimulate insulin action (116). NEFAs also serve as endogenous ligands for PPAR-α and γ, which have both been shown to protect against β-cell failure (117,118).
At the onset, β-cells adapt well to the changing macronutrient environment, in what has been termed β-cell compensation (119). Their mass and proliferation increase to compensate for the peripheral IR and insulin secretion is increased to maintain consistent blood glucose concentrations (119,120). It has been suggested that glucose, a stimulator of β-cell proliferation and mass both in vivo and in vitro, may contribute to adaptation, though human studies have failed to confirm this theory (121).
Eventually a decline in β-cell mass and function is observed as obese conditions persist. The term glucolipotoxicity has been coined to describe the combined effects of glucose and lipid overexposure on pancreatic dysfunction, and these two events are said to act synergistically in pancreatic degeneration. Studies have shown that the pancreas can adapt to elevated glucose or lipid levels alone by increasing its oxidative capacity (122,123,124), but because increased oxidation of one excludes the other, it cannot respond adequately when both occur concomitantly. Under such conditions, lipid esterification and fat deposition in the pancreas are favored due to glucose-induced elevation of malonyl-CoA, and eventually the pancreas can accumulate proapoptotic lipid derivatives such as ceramides and diacylglycerol (119,124). High glucose levels also inhibit PPAR-α and UCP-2 expression, another mechanism in which the tissue could respond to elevated lipids (125). In support of the glucolipotoxicity theory it has been shown that the damaging effects of lipid overexposure are only apparent in the context of elevated glucose (126,127,128), findings which have found in vivo support (129,130). In the next section, we will discuss how elevated glucose and lipid levels are purported to induce β-cell failure.
The mechanisms of glucose- and lipid-induced β-cell failure
The mechanisms linking the toxic effects of glucose and lipid overexposure to β-cell failure include mitochondrial dysfunction and reactive oxygen species (ROS) production, ER stress, and intracellular calcium increase (116). When occurring alone, chronic glucose exposure shows inconsistent effects with regard to inducing apoptosis (131). Exposing β-cells to glucotoxic conditions results in a decrease in glucose-stimulated insulin secretion and an abolishment of glucose-stimulated insulin secretion under regular physiological concentrations (132). Glucose also increases intracellular concentrations of calcium to high levels that persist even after glucose levels are normalized (133). Calcium is an integral component in the mechanism of glucose-stimulated insulin secretion, and the persistently elevated levels may contribute to apoptotic signaling in the β-cell (133). There is also evidence that the increase in glucose metabolism resulting from hyperglycemia would act to increase oxidative phosphorylation and reactive oxygen species (ROS) production (134), thereby enhancing the inflammatory response. The pancreas has very low levels of ROS-detoxifying enzymes and is, therefore, susceptible to the damaging effects of oxidative stress. In the pancreas, oxidative stress inhibits transcription of the insulin gene and activates stress kinase pathways, which phosphorylate insulin receptor substrate-1 serine residues (135,136,137,138), and exposure to antioxidants has shown effectiveness in reducing these outcomes. Diabetic rats treated with antioxidants N-acetyl-l-cysteine or aminoguanidine are protected from glucose toxicity in the pancreas (134), and there have been similar improvements in β-cell function observed in antioxidant-treated islets taken from T2D patients (139).
In the context of high glucose exposure, chronically elevated lipid levels also inhibit glucose-stimulated insulin secretion and appear to be an important component in obesity-induced β-cell failure (140,141,142). Long-term exposure of the pancreas to NEFAs does increase fat accumulation in both the exocrine and endocrine regions of the pancreas (143), though there remains some ambiguity as to whether TG accumulation itself is damaging to pancreatic function. A number of reports indicate that lipotoxicity in pancreatic β-cells is specific to saturated FAs, whereas pools created by unsaturated FAs are innocuous and protective (82,144,145). Conversely, increasing TG synthesis though adenoviral overexpression of DGAT induces β-cell failure in the presence of high glucose, suggesting that process FA esterification itself may be involved in lipotoxicity (146).
Although there is some in vitro and in vivo evidence linking certain lipid derivatives such as ceramide, the specific metabolites and mechanisms underlying these effects remain largely elusive. Ceramide produced de novo by the pancreas can induce apoptosis through the mitochondrial death pathway and caspase elevation (142) and inhibits insulin expression (141). The unfolded protein response and ER stress have also been implicated in lipid-induced apoptosis (147,148,149). Luciani et al. have shown that the unloading of ER calcium stores by palmitate can induce mitochondrial dysfunction and suggest that ER stress may be the initiating event leading to several other apoptotic pathways (150).
Obesity-related inflammation and pancreatic dysfunction
Although simultaneous glucose and lipid overexposure appear to be the most likely sources of pancreatic dysfunction in obesity, more recent studies also implicate inflammation, mediated largely by cytokines IL-1β and IL-18, in β-cell failure (151). A number of recent studies demonstrate that macrophages are increased in islets of obese db/db mice and T2D patients (152). Macrophages from islets of T2D patients were mainly located around endocrine cells, as opposed to those from healthy patients. which were on the periphery, and were also more inflamed, as determined by having greater CD68+ cell colocalization (153). The method of macrophage recruitment to the pancreas is still unknown, and the significance of these findings has yet to be determined.
The rate of chronic kidney disease (CKD) has paralleled that of obesity, yet there is still ambiguity as to whether obesity itself classifies as an independent risk factor for CKD (154). Recent evidence does point to a causal link, particularly with regard to the development of diabetic nephropathy (155). The kidney undergoes many significant structural and metabolic changes with obesity, including hyperfiltration of the glomeruli, thickening of the glomerular basement membrane, proliferation of mesangial cells, thickening of the mesangial matrix, and expansion of Bowman's capsule (156). The causes for such clinical manifestation vary, and include IR, oxidative stress, alterations in the inflammatory cytokine production, activation of the aldosterone-renin-angiotensin system, the release of profibrotic factors, and altered lipid metabolism (157,158,159,160).
In humans, microalbuminuria is the earliest clinical manifestation of obesity-related kidney damage and diabetic nephropathy, which is characterized by tubulointerstitial fibrosis and glomerulosclerosis (161,162). The passage of albumin into the urine is typically prevented by the selective sieving by the glomerular capillary wall, in which the visceral glomerular epithelial cell, or podocyte, has been shown to play a very critical role (163,164). In the ZDF rat, the degeneration of podocytes has been reported to be responsible for the development of glomeruleroscerlosis (163). Podocytes are insulin-sensitive cells, which utilize GLUT1 and GLUT4 transporters (165), and the insulin response in podocytes is dependent on the protein nephrin for GLUT-1 translocation (166). It has been suggested that IR in these cells may be a key factor in the development of nephropathy (165). Mutations in the gene NPHSI, which encodes nephrin, result in congenital nephrotic syndrome (167).
Ectopic fat accumulation in the kidney
The association between fat accumulation and renal disease was described in 1858 by Virchow, who described successive stages of fatty metamorphosis and fatty detritus in the renal epithelium in Bright's disease (168). In 1936, Kimmelstiel and Wilson also posited lipid accumulation in the kidney as being a contributor to renal disease (169). High-fat-fed C57BL/6J mice have increased lipid accumulation in the glomerular and tubulointerstitial cells of kidney and that mRNA of SREBP1, SREPB2, and lipogenic genes acetyl-CoA carboxylase and FA synthase are all enhanced (170). The high-fat-fed mice also exhibited signs of glomerulosclerosis and proteinuria, including mesangial expansion with increased extracellular matrix protein content and a significant increase of urinary albumin excretion, that were not present in low-fat-fed mice (170). The effects of high-fat feeding on glomerulosclerosis and proteinuria were inhibited in SREPB1c knockout mice (170).
One effect of chronic NEFA exposure in obesity may also be to inhibit insulin sensitivity in podocytes, thereby inhibiting their function. Podocyte insulin sensitivity is impaired in vitro by exogenous palmitate and reversed by rosiglitazone (165).
Renal inflammation and immune cell infiltration
Inflammation in the kidney may also contribute to dysfunction resulting from obesity, particularly with regard to tissue fibrosis. Tubulerofibrosis is involved in nephropathy, and key to its development are the release of profibrotic factors such as transforming growth factor-β1 and plasminogen activator inhibitor-1 (PAI-1) (171,172). Chronic leptin infusion has been shown to induce proteinuria and glomerulosclerosis with accompanied increases in transforming growth factor-β1 and type 2 collagen, perhaps an indication that its elevated levels may be involved in obesity-related kidney dysfunction (173).
Recent clinical evidence also suggests that obesity-induced reductions in adiponectin may be involved in renal dysfunction by affecting podocyte biology (174). Studies by Sharma et al. revealed that mice deficient in adiponectin have increased albuminuria and fusion of podocyte foot processes, perturbations that were normalized with adiponectin treatment (174). In cultured podocytes, adiponectin reduced albumin permeability and alleviated podocyte dysfunction (174).
Macrophage infiltration and inflammation into the kidney has also been reported to be involved in the pathogenesis of obesity-related renal dysfunction (175,176). The infiltration of macrophages into the kidney correlates strongly with tissue injury and is also associated with MCP-1 production (177). Macrophages have been shown to play a pathogenic role in diabetic nephropathy and act as effector cells in initiating tubulointerstitial damage (162,175). Deficiency of MCP-1 in db/db mice significantly ameliorates renal inflammation, macrophage infiltration, and kidney damage (177). The production of MCP-1 by renal cells is thought to be induced by elevated glucose levels, tubular reabsorbed protein, advanced glycosylated end products, and also the activation of the renin-angiotension system (178), and the reduction in renal MCP-1 by angiotensin-– inhibitors was associated with improvements in renal function (162). MCP-1 may also exert direct effects on renal function aside from promoting macrophage infiltration. Binding of MCP-1 to its receptor chemokine receptor-2 in human podocytes in vitro has been shown to decrease nephrin mRNA through a Rho-dependent mechanism (179).
These findings show that lipid infiltration and inflammation are important factors in obesity-related kidney damage, though mechanistically there are few studies describing how these two occurrences relate to each other. Recently Kelly et al. proposed that lipid oxidation and increased expression of the oxidized low-density lipoprotein transporter LOX-1 may be the impetus behind renal inflammation (180). LOX-1 would increased uptake of oxidized lipids and change the epithelial phenotype to a proinflammatory, dysfunctional one (180). There is also evidence that PPAR-γ may play a protective role in the kidney by decreasing renal inflammation and fibrosis (181), though concerns over the safety of the synthetic PPAR-γ agonists, thiazolidediones, have limited their appeal.
Obesity-related comorbidities result from a combination of different interrelated factors, and chronic inflammation, alterations in adipokine secretion, elevated glucose levels, and lipid overexposure are just a sample of the many changes occurring in an obese individual. Because of its prominent role in energy storage, WAT is among the initial tissues compromised in obesity. The result of which leads directly to many of the factors listed above, and thus WAT dysfunction may play an etiopathogenic role in mediating other systemic complications (Figure 3). Thus, to disassociate obesity from its comorbidities, preventing or treating dysfunction in WAT and low-grade chronic inflammation should be a primary goal of pharmacological intervention. Treatments aimed at increasing WAT storage capacity (i.e. increasing adipocyte differentiation), reducing the total amount of fat by enhancing energy expenditure, or decreasing low-grade chronic inflammation address this outcome.