Apoptosis

, Volume 14, Issue 12, pp 1484–1495

Diabetes and apoptosis: lipotoxicity

Authors

  • Christine M. Kusminski
    • Department of Internal Medicine, Touchstone Diabetes CenterUniversity of Texas Southwestern Medical Center
  • Shoba Shetty
    • Department of Internal Medicine, Touchstone Diabetes CenterUniversity of Texas Southwestern Medical Center
  • Lelio Orci
    • Faculty of Medicine, Department of Cell Physiology and MetabolismCMU, University of Geneva
  • Roger H. Unger
    • Department of Internal Medicine, Touchstone Diabetes CenterUniversity of Texas Southwestern Medical Center
    • Department of Internal Medicine, Touchstone Diabetes CenterUniversity of Texas Southwestern Medical Center
    • Department of Cell BiologyUniversity of Texas Southwestern Medical Center
Diabetes and Apoptosis

DOI: 10.1007/s10495-009-0352-8

Cite this article as:
Kusminski, C.M., Shetty, S., Orci, L. et al. Apoptosis (2009) 14: 1484. doi:10.1007/s10495-009-0352-8

Abstract

Obesity is an established risk factor in the pathogenesis of insulin resistance, type 2 diabetes mellitus and cardiovascular disease; all components that are part of the metabolic syndrome. Traditionally, insulin resistance has been defined in a glucocentric perspective. However, elevated systemic levels of fatty acids are now considered significant contributors towards the pathophysiological aspects associated with the syndrome. An overaccumulation of unoxidized long-chain fatty acids can saturate the storage capacity of adipose tissue, resulting in a lipid ‘spill over’ to non-adipose tissues, such as the liver, muscle, heart, and pancreatic-islets. Under these circumstances, such ectopic lipid deposition can have deleterious effects. The excess lipids are driven into alternative non-oxidative pathways, which result in the formation of reactive lipid moieties that promote metabolically relevant cellular dysfunction (lipotoxicity) and programmed cell-death (lipoapoptosis). Here, we focus on how both of these processes affect metabolically significant cell-types and highlight how lipotoxicity and sequential lipoapoptosis are as major mediators of insulin resistance, diabetes and cardiovascular disease.

Keywords

DiabetesApoptosisLipotoxicityPancreatic β-cellsLeptinAdiponectin

Abbreviations

T2DM

Type 2 diabetes mellitus

CVD

Cardiovascular disease

TG

Triglyceride/triacylglycerol

FFA

Free fatty acid

GLUT4

Glucose transporter-4

NO

Nitric oxide

AMPK

Adenosine monophosphate-activated protein kinase

CoA

Coenzyme-A

SREBP-1c

Sterol regulatory element binding protein-1c

PPAR-γ2

Peroxisome proliferator-activated receptor

ACC

Acetyl coenzyme A carboxylase

FAS

Fatty acid synthetase

GPAT

Glycerol-3-phosphate acyl transferase

CPT-1

Carnitine palmityl transferase-1

ACO

Fatty acyl-CoA oxidase

PGC-1α

Peroxisome proliferator-activated receptor-γ coactivator-1α

UCP-2

Uncoupling protein-2

VLDLs

Very low-density lipoproteins

ROS

Reactive oxygen species

ER

Endoplasmic reticulum

SPT-1

Serine palmitoyl transferase

Bcl2

B-cell lymphoma 2

Bad

Bcl2-antagonist of cell death

Bax

Bcl2-associated X protein

Bid

BH3 interacting domain death agonist

Bim

Bcl2-like 11

NF-κB

Nuclear factor-κB

iNOS

Inducible nitric oxide synthase

ACS

Acyl CoA synthase

ECM

Extracellular matrix

FAT-ATTAC

Fat apoptosis through targeted activation of caspase-8

FKBP

Peptidyl-prolyl cis–trans isomerase

AICAR

5-Amino 4-imidazolecarboxamide riboside

MCD

Malonyl CoA decarboxylase

Introduction

We appreciate that obesity is a major risk factor in the pathogenesis of type 2 diabetes mellitus (T2DM), cardiovascular disease (CVD) and the associated co-morbitites. Obesity is further characterized as a chronic state of positive energy balance that saturates the normal capacity of adipose tissue to store triglycerides (TG), resulting in ectopic TG deposition outside the realms of adipose tissue compartments. Indeed, obesity leads to the over-accumulation of lipids in non-adipose tissues, notably, skeletal muscle, liver, heart, and pancreas. Such excessive deposition of fat (referred to as steatosis in the liver) trigger organ-specific toxic reactions that may lead to cellular dysregulation and functional tissue impairment. This detrimental phenomenon of ectopic over-accumulation of reactive lipid moieties in non-adipose tissues is now commonly referred to as “lipotoxicity”. However, the term lipotoxicity, encompasses toxicity that results not only from a lipid overload induced by an excess of unoxidized fatty acids (FAs), but also from endogenous lipids synthesized from excess glucose, through the process of de novo lipogenesis.

The lipotoxicity in liver, pancreatic β-cells, myocardium and skeletal muscle, ultimately leads to insulin resistance, cardiomyopathy and T2DM; such a constellation of pathophysiological changes is referred to as the metabolic syndrome, as defined by Reaven [1]. A pertinent causative link between adipose tissue distribution and the metabolic syndrome resides in the accumulation of lipids in non-adipose sites. When lipids over-accumulate ectopically in non-adipose tissues during periods of overnutrition, they are ineffectively oxidized. These FAs can be biosynthetically utilized in other, deleterious non-oxidative pathways; ceramide biosynthesis is one such example, in which excess ceramides trigger apoptosis of lipid-laden pancreatic β-cells and cardiomyocytes, through increased nitric oxide formation. Such a cascade of lipid-induced programmed cell-death, a sequel to lipotoxicity, is termed ‘lipoapoptosis’; however, the underlying metabolic causes that result in lipoapoptosis are yet to be fully unraveled. Here, we therefore summarize the available data on the molecular mechanisms and pathophysiology believed to underlie lipotoxicity and lipoapoptosis.

Lipocentric perspectives

Traditionally, the interpretation of insulin resistance in the pathogenesis of T2DM was sculptured into a somewhat ‘glucocentric’ milieu, occurring primarily through glucotoxicity, such that overt hyperglycemia was inevitably perceived as a first and foremost target for therapeutic intervention. Whilst this glucocentric concept was the general consensus, Randle’s hypothesis introduced us to an amalgamation of both glucose and lipid metabolism. The “glucose-fatty acid cycle” placed emphasis on a free fatty-acid (FFA)-induced impairment of glucose metabolism that results in insulin resistance [2]. Indeed, over the decades, such a glucolipotoxic viewpoint was strengthened with evidence of FFAs inhibiting insulin-mediated glucose uptake through interference of the glucose transporter GLUT4 and, its’ translocation to the plasma membrane from intracellular vesicles, thus effectively preventing glucose uptake into myocytes [3, 4]. Similarly in the liver, FFAs were shown to inhibit insulin-mediated suppression of glycogenolysis and gluconeogenesis [5].

In recent years, however, studies accentuated a dysregulation in lipid homeostasis as the primary metabolic defect in T2DM. Elevated circulating levels of FFAs were observed prior to the onset of glucose intolerance [6, 7]; on the whole, this caused a major paradigm shift in the understanding of the metabolic syndrome. Consequently, FFAs are now recognized as significant contributors to insulin resistance pathophysiology, with lipotoxicity identified as the primary perpetuator of the metabolic syndrome. Such a ‘lipocentric’ theory therefore portrays the hyperglycemia-induced insulin resistance and pancreatic β-cell loss in obesity, as secondary to the metabolic disturbances resulting from ectopic lipid deposition and lipotoxicity [8, 9]. If this is indeed the case, hyperglycemia could theoretically, be rectified through the removal of the ectopic overaccumlation of lipids [10].

The evolution of the adipocyte

Fuel storage

The evolutionary conversion of the adipocyte allowed the cell to acquire adaptive responses to fluctuations in nutrient availability to promote longevity [11], by storing excessive quantities of FA in the form of triacylglycerol (TG). Such an effective homeostatic fuel system thus permits adipocytes to store FAs during periods of overnutrition and timely release them during periods of nutrient deprivation. The adipocyte as such, may not be essential for life when food is abundant, however there must be evolutionary pressure for its existence. Indeed, the lack of adipocytes, even in the presence of only a modest caloric surplus, results in ectopic deposition of TGs in non-adipose tissues (primarily in the liver), with resulting steatosis; such lipoatrophic conditions indeed exist clinically, for instance, in the case of congenital generalized lipodystrophies. However, the rarity of generalized lipodystrophy may reflect an extremely pertinent, if not essential, role for adipocytes in extending the survival period during famines [12].

Anti-steatotic and anti-lipotoxic actions

In addition to excess caloric storage, adipocytes have several key physiological functions to protect other cells that have a more restricted capacity to store TGs under conditions of a positive energy balance. Such protection can be orchestrated through the endocrine functions of the adipocyte; secreting anti-steatotic hormones that determine the partitioning of surplus dietary lipid in times of varying caloric cycles.

The liporegulatory hormone leptin is secreted from adipocytes in proportion to whole body fat mass [13]. During periods of overnutrition, leptin exerts its physiological functions that protect non-adipose tissues from the ectopic deposition of non-oxidative metabolic products of FAs by repressing food-intake and increasing energy expenditure. Hyperleptinemia functions to regulate food-intake via hypothalamic circuits [14, 15], whilst more direct functions of leptin reduce lipogenesis and enhance the compensatory FA oxidation capacity of non-adipocytes during ectopic lipid overload [16, 17]. Collectively, such actions dissipate redundant energy from FA excess as heat and ultimately, minimize FA entry into non-oxidative lipotoxic pathways. However, when the peripheral homeostatic leptin-system is either congenitally deficient or a leptin resistance develops, harmful products of non-oxidative metabolism such as ceramides accumulate in non-adipose tissues. For instance, generalized lipid accumulation will develop in pancreatic-islets [18] and the heart [19], resulting in NO-mediated lipotoxicity; sequential lipoapoptosis will then ensue; such lipid overload can be corrected by restoring leptin action [9].

Evidence for the anti-steatotic action of leptin originates from several rodent models. In normal rats fed a 60% high-fat diet, acute 24 h measurements reveal circulating levels of leptin to increase proportionately with adipocyte fat accumulation [18]. Secondly, rodents on high-fat diets exhibit marginal lipid deposits in the heart, liver, pancreatic-islets and skeletal-muscle [20], irrespective of chronic adipocyte hypertrophy. Conversely, in leptin-deficient rodents, or rodents with non-functional leptin receptors, non-adipose tissues rapidly accumulate high levels of TG, despite normal fat-intake (6%) [18]. Furthermore, adenoviral transfer of the wild-type leptin receptor (OB-Rb) to obese Zucker diabetic fatty (ZDF) rats exhibiting non-functional leptin receptors, reduces plasma and hepatic lipid content, however, sensitizes only the liver to endogenous hyperleptinemia [18]. Taken together, these data suggest that one of the major functions of the dietary-induced hyperleptinemia is to protect non-adipose tissues from lipid over-accumulation.

Similarly, another adipocyte-derived hormone, adiponectin [21], may also mediate protective effects against lipotoxic insult to maintain FA homeostasis [22], thus complementing leptin in anti-steatotic and anti-apoptotic functions. Adiponectin is an endogenous insulin sensitizer with the capacity to normalize dyslipidemia [23] and exert anti-atherogenic actions [24]. Such insulin sensitizing effects have been attributed to a sequence of events that induce AMPK activation, inactivate acetyl coenzyme-A carboxylase (ACC) and decrease malonyl coenzyme-A (CoA) levels, a process that may stimulate FA oxidation in skeletal-muscle [2527]. However, it is more likely that endogenous adiponectin exerts its effects in a skeletal muscle-independent way, mainly through action in organs such as the liver. The liporegulatory properties of adiponectin have further been highlighted by its’ ability to regulate food-intake and cause TG depletion in liver, to ultimately, minimize ectopic lipid deposition in non-adipose tissues [22].

The liporegulatory system

When dietary intake matches energy expenditure, the liporegulatory system remains in equilibrium. However, upon FA dysregulation, the intricate mechanisms of the liporegulatory system are implemented to utilize the anti-lipogenic and pro-oxidative, transcriptional aspects of leptin and potentially, adiponectin. For instance, in pancreatic-islets, leptin exerts two distinct functions on the periphery to minimize plasma FA entry into non-adipose tissues: (1) promote the oxidative machinery in non-adipose cells to cope with an influx of surplus FA and, (2) downregulate lipogenesis, the process of de novo FA synthesis from glucose, in non-lipogenic tissues.

In particular, leptin down-regulates lipogenic transcription factors [28], namely sterol regulatory element binding protein-1c (SREBP-1c) in liver [29] and peroxisome proliferator-activated receptor (PPAR)-γ2 [30] in fat; this reduces the expression of specific lipogenic enzymes, such as ACC, fatty acid synthetase (FAS) and glycerol-3-phosphate acyl transferase (GPAT) [31]. In addition to reducing ACC formation, the protein activity of ACC is diminished through AMPK-mediated phosphorylation [32]; this lowers malonyl-CoA, a potent inhibitor of carnitine palmityl transferase-1 (CPT-1)-mediated FA oxidation; this phenomenon is occasionally referred to as the “McGarry effect” [33, 34]. An illustration of the McGarry effect is provided in Fig. 1. At present, however, how leptin and adiponectin fall into this equation of AMPK-mediated FA homeostasis, is yet to be fully unraveled.
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-009-0352-8/MediaObjects/10495_2009_352_Fig1_HTML.gif
Fig. 1

The McGarry effect (adapted from [95])

Hyperleptinemia can simultaneously induce oxidative enzymes, by permitting FAs to upregulate PPAR-α [35], a transcription factor mainly targeting the mitochondrial and peroxisomal enzymes of FA oxidation; this sequentially induces the mitochondrial enzyme CPT-1 and peroxisomal enzyme fatty acyl-CoA oxidase (ACO) [31]. Leptin can further up-regulate peroxisome proliferator-activated receptor-γ coactivator-1α (PGC-1α) [30], a master regulator of mitochondrial biogenesis [36].

The combination of an increase in FA oxidation and a decrease in FA synthesis may therefore account for reduced lipid content in a cell. An increase in mitochondrial uncoupling protein-2 (UCP-2) expression, induced by hyperleptinemia [35], causes the excess energy generated by FA oxidation to be dissipated as heat. Such an effect occurs through the ability of UCP-2 to uncouple oxidative phosphorylation from ATP synthesis, thus allowing the excess energy to be dispersed as heat. Therefore, ectopic accumulation of TG in non-adipose tissues is minimal as long as the leptin system is operative.

Fatty acid overload and lipotoxicity

When caloric intake exceeds energy expenditure, the adipocyte begins to store surplus calories to expand its TG pool. Circulating levels of leptin increase concomitantly [18, 37], whilst plasma levels of adiponectin, paradoxically decrease [38]. Such circumstances initiate the liporegulatory system to upregulate oxidative metabolism of FA and, reduce de novo FA synthesis in lean non-adipose tissues. Ectopic deposition of surplus lipids is therefore minimal and partitioning of body-fat is maintained correctly. The ability to appropriately expand adipose tissue in the context of an energy surplus is therefore instrumental.

Non-adipose tissues also have the capability to dispose of surplus FAs. The liver can export TG as very low-density lipoproteins (VLDLs) [39], whereas skeletal muscle lowers lipid content by increasing muscular activity to enhance FA oxidation and reduce the risk of lipotoxicity. However, while these tissues have compensatory mechanisms to cope with modest surplus in FA levels, other tissues are somewhat more vulnerable to a lipid spillover. For instance, pancreatic β-cells [8, 40] and the myocardium [19, 41] both have limited compensatory FA oxidation as the sole means by which to dispose of excess FA. This may account for the high β-cell loss and cardiovascular risk associated with diabetes and obesity. However, the fact that such complications occur only after a considerable period of overnutrition suggests that initially, such tissues are effectively protected against steatosis. In light of this, Table 1 highlights the organ-specific effects that result from a lipotoxic event.
Table 1

Organ-specific consequences that result from the metabolic dysregulation associated with lipid accumulation and lipotoxicity

Target organ

Potential lipotoxic consequences

Pancreas

Reduced insulin secretion, pancreatic β-cell failure, islet dysfunction, impaired proliferative capacity, and lipoapoptosis

Heart

Myocardial insulin resistance, cardiac steatosis, cardiomyopathy, and heart failure

Liver

Fatty liver, nonalcoholic steatohepatitis, and cirrhosis

Muscle

Reduced insulin-stimulated glucose-uptake and insulin resistance

When surplus caloric intake greatly exceeds the adipocyte hypertrophy and hyperplasia necessary to accommodate the sudden influx of excess FA, or if leptin, or leptin receptors are deficient or non-functional, the liporegulatory system breaks down [42]. The net result is an overspill of FAs from the adipocyte to non-adipocyte cells and unoxidized FA excess is deposited in non-adipose tissues. Circulating levels of FFAs, of which are higher in obese states, are taken up; the intracellular, unoxidized surplus FA that exceeds the oxidative requirements of a cell therefore is esterified to TG.

While tissues initially respond adequately to facilitate the storage of surplus TGs, which once sequestered are considered passive and harmless to the cell, saturation of this neutral TG store, can however, exceed the cells’ storage capacity and drive hydrolysis back into fatty-acyl CoA, which surpasses the oxidation capacity of a cell and feeds the unoxidized excess into non-oxidative toxic pathways of FA metabolism, such as de novo ceramide formation [43]. Such a plethora of deleterious reactive lipids accumulate in skeletal muscle, liver, heart, and pancreatic β-cells; this contributes to the pathophysiology of insulin resistance, steatotic liver disease, impaired myocardial function and β-cell failure, respectively. A representative set of examples of cellular lipid accumulation is shown in Figs. 2 and 3. Ultimately, the net result of these cellular lipid accumulations is apoptosis [1].
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-009-0352-8/MediaObjects/10495_2009_352_Fig2_HTML.jpg
Fig. 2

Lipid accumulation and cell death in various cell-types. a Lipid accumulation in a 12 week-old liver from an ob/ob mouse, b TUNEL stain of an apoptotic fat-pad, c oil red O stain of a human cardiac muscle from an individual with a BMI of 42, d oil red O stain of a human cardiac muscle from an individual with a BMI of 28, e oil red O stain of the cardiac muscle from an obese ZDF fa/fa rat. f oil red O stain of the skeletal muscle from an obese ZDF fa/fa rat

https://static-content.springer.com/image/art%3A10.1007%2Fs10495-009-0352-8/MediaObjects/10495_2009_352_Fig3_HTML.jpg
Fig. 3

Comparison of lipid accumulation in the exocrine pancreas. a Oil red O stain of the pancreas of a 20 week-old normal lean ZDF (+/+) rat, b of an 8 week-old pre-diabetic obese ZDF (fa/fa) rat, c of a 16 week-old diabetic obese ZDF (fa/fa) rat and, d electron photomicrograph of a 14 week-old diabetic obese ZDF rat pancreas. Despite the increase in lipids, apoptosis is not discerned in these exocrine cells, perhaps because the lipids are in droplets of neutral fat. In β-cells, by contrast, lipid droplets are few and apoptosis is far more common [96]

In mechanistic terms of defective liporegulation, for example in unleptinized pancreatic-islets of obese (fa/fa) ZDF rats, steatosis develops through enhanced lipogenic capacity and depleted FA oxidation [44]. Inevitably, an increase in SREBP-1c [30], PPAR-γ, ACC, FAS and GPAT gene expression [44] is evident; this parallels a downregulation in PPAR-α message and its’ target oxidative enzymes, CPT-1 and ACO. In contrast, the reduction in TG content, as mediated by overexpression of the functional leptin receptor in ZDF rat islets [9] is associated with improved β-cell function and subsequent protection from lipotoxicity.

Additional perspectives on lipotoxicity

The teleological reasons for the manifestations of various downstream events of lipid overload, such as insulin resistance, could also be perceived as a self-protective survival measure for a cell [9], rather than the entirely detrimental sequence of events originally proposed. For example, under hyperglycemic conditions, lipid-laden myocytes or hepatocytes may acquire resistance to insulin-stimulated entry of surplus glucose. Such exclusion of glucose from the cell, limits intracellular substrate availability for de novo lipogenesis and minimizes, or even postpones the onset of severe lipotoxic damage and lipoapoptosis.

Conventionally, insulin binds to its receptor, ultimately triggering phosphorylation of Akt1 and translocation of GLUT4 to the plasma membrane, thus allowing glucose entry into the cell [45]. The excess of intracellular glucose, not stored as glycogen, nor oxidized via glycolytic pathways, enters lipogenic pathways as acetyl-CoA, to undergo carboxylation to malonyl-CoA; an intermediate in the synthesis of FA, triacylglycerol and ceramides. In addition to this, malonyl-CoA can inhibit CPT-1, to obstruct FA oxidation [46].

Alternatively, glucometabolic insulin resistance may occur to maintain the cells’ capacity to oxidize FA, prevent ceramide overaccumulation through de novo lipogenesis from glucose, and ultimately, avert severe lipotoxicity [5, 9]. In particular, FAs or ceramide themselves can block GLUT4 translocation to the plasma membrane at the level of Akt1 activity [47], therefore excluding extracellular glucose from the cell. This reduces malonyl-CoA formation, which lowers intracellular FA and ceramide accumulation; importantly, this eliminates the inhibitory sequestration of CPT-1 and allows an effective rate of FA oxidation to resume. Paradoxically, FA and ceramides in moderation may culminate in glucometabolic insulin resistance, to the detriment of preventing the more fatal lipotoxic consequences.

How does such a scenario extrapolate to β-cells and cardiomyocytes? Considering β-cells are extremely susceptible to hyperglycemia-induced reactive oxygen species (ROS) damage [48, 49], as they exhibit low levels of free-radical quenching enzymes [50]; what protective measures are in place to protect such a vulnerable cell-type from lipotoxicity? Whilst β-cells initially compensate lipotoxic insults through hyperplasia [51], this may only be short-lived; they then lose glucose transporter-2 and glucokinase [52], which attenuates their glucose responsiveness, and overt diabetes sets in. Of note, ceramide has been reported to inhibit glucokinase [53]. Furthermore, moderate levels of FA can cause mild-uncoupling of mitochondrial oxidative phosphorylation in β-cells; a potential defense mechanism to lower hyperglycemia-induced ROS formation [54, 55].

Interestingly, recent evidence suggests that adiponectin may promote survival of pancreatic-islets. Adiponectin has been observed to moderately suppress cytokine- and FA-induced β-cell apoptosis in order to prevent autoimmune and lipotoxicity-induced dysfunction [56]. Furthermore, adiponectin has been reported to stimulate AMPK in cultured MIN6 β-cells and purified rat islets [57, 58], thus demonstrating its’ potential to enhance FA oxidation in β-cells through ACC regulation; this is consistent with the phosphorylation of ACC observed in cultured human islets [59]. These data therefore underline the promising anti-apoptotic functions of adiponectin, which may reduce the formation of toxic lipid metabolites to counteract lipotoxicity-induced β-cell destruction; further encouraging the hypothesis that adiponectin may play a pivotal role in compensatory β-cell growth, as observed in insulin-resistant animals and humans [60].

Leptin and adiponectin in diet-induced obesity and lipotoxicity

The following sequence of events is proposed to occur during diet-induced obesity and lipotoxicity.

Excess dietary intake stimulates hyperinsulinemia, therefore upregulating the lipogenic transcription factor SREBP-1c [61] and increasing the enzymatic machinery for de novo synthesis of FA. As lipogenesis increases, newly synthesized FAs and dietary FAs are transported as VLDLs from the liver to adipocyte compartments for storage as TG, resulting in adipocyte hypertrophy and obesity. Leptin action promotes FA oxidation to utilize surplus fuel by enhancing energy expenditure, whilst simultaneously downregulating SREBP-1c in non-adipose tissues to reduce de novo lipogenesis [9, 62], therefore minimizing ectopic fat deposition. Synergistically, adiponectin promotes storage of surplus dietary fat as TG, by promoting a lipogenic profile in fat; through upregulation of PPAR-γ2 and its targets, adiponectin drives the expansion of adipose tissue, thereby limiting ectopic fat deposition. Our previous observations using ob/ob adiponectin transgenic mice offer primafacie evidence for this scenario [22]. Congenitally obese mice lacking leptin, that were maintaining adiponectin levels comparable to the lean state, displayed massive expansion of subcutaneous adipose tissue and were ironically, protected from diet-induced insulin resistance. Phenotypically, all systemic metabolic parameters were vastly improved, possibly due to an upregulation of PPAR-γ and its’ target genes in adipocytes, enhancing lipogenesis with subsequent adipocyte hyperplasia. The ultimate result included redistribution of lipids from ectopic deposits in the liver, muscle and pancreatic-islets, primarily to the subcutaneous adipose tissue that functions, in this context, as a ‘metabolic sink’; further evident, was an increase in hepatic insulin sensitivity, that paralleled improvements in islet morphology, respectively. As such, adiponectin may be perceived as a starvation signal released by the adipocyte, providing systemic indication that adipocytes can accommodate more lipids to expand stores in adipose tissue [22].

In light of this, adiponectin-null mice have been documented to exhibit a leptin-induced compensatory increase in FA oxidation [63]; furthermore, an adipocyte-specific overexpression of the leptin receptor generates mice incapable of expanding their adipose tissue mass during overnutrition [64]. These studies imply that the lack of one adipokine, intensely exaggerates the functions of the other. However, overexpression of either adipokine, in the context of obesity, will instigate an inhibition of the other.

Regardless of these adipokine-induced protective measures, insulin resistance and glucose intolerance result from the modest increase in unoxidized FA accumulating in skeletal muscle. As obesity progresses, leptin- and adiponectin-mediated protection against lipotoxicity deteriorates, partly due to a saturated leptin system. Leptin no longer has the capacity to inhibit SREBP-1c or maintain the compensatory increase in FA oxidation normally induced by hyperleptinemia. Paradoxically, adiponectin levels decline with increasing adiposity, resulting in failure to preferentially compartmentalize surplus TG to appropriately expand fat-mass; visceral adiposity may therefore prevail.

Lipid-induced insulin resistance may exist secondary to leptin-resistance and adiponectin deficiency. A general systematic imbalance to partition surplus dietary lipids in the correct adipose depot accompanies an overaccumulation of unoxidized FAs and their toxic lipid derivatives in non-adipose tissues. Impaired tissue function and lipoapoptosis inevitably follow.

Mechanisms of lipoapoptosis

Whilst the initial stages of lipid overload are well understood, the precise mechanisms by which the late destructive phases of lipotoxicity leads to apoptosis of non-adipocyte cells remain to be fully elucidated. In pancreatic-islets [40] and cardiac myocytes [19, 65] for instance, ceramide-induced apoptosis appears to be a potent apoptotic stimulus [66]. Additional mechanisms are likely to be in place, however, such as the generation of ROS and lipid peroxidation products, toxic membrane destabilizing effects of FAs, increases in intracellular calcium and, the activation of ER stress pathways [67].

Pancreatic β-cell lipoapoptosis

Lipid-laden pancreatic β-cells of leptin-unresponsive (fa/fa) ZDF rats overexpress the enzyme serine palmitoyl transferase (SPT-1) [68], which catalyzes the condensation of unoxidized palmitoyl-CoA and serine, to form dihydrosphingosine [69]; the first step in de novo ceramide biosynthesis. Apoptosis has been postulated to result from excessive de novo ceramide formation [40] coupled to, and amplified, by the FA-induced downregulation of the anti-apoptotic factor Bcl2 [40, 70]. In light of this, ceramide-induced inhibition of Akt activity [47, 71] has been proposed to impair GLUT4 translocation to the plasma membrane in muscle [47, 72], thereby strengthening a role for ceramides in lipoapoptosis and tissue loss. Notably, inactivation of the pro-apoptotic factors Bcl2 family members, such as Bad, Bax, Bid, and Bim, requires Akt-mediated phosphorylation [73] as a potential cell survival signal. Amongst all the FAs examined, palmitic acid exerts the utmost toxic effects on pancreatic-islets, by hindering proliferation and inducing oxidative stress and β-cell apoptosis [70, 74, 75]. Conversely, unsaturated oleic acid prevents such palmitate-induced apoptosis; potentially via upregulation of Bcl2 [70].

In addition to impairing insulin signaling, other mechanisms may prevail. Ceramide can induce apoptosis by activating NF-κB, which upregulates the expression of inducible nitric oxide synthase (iNOS), thereby enhancing nitric oxide and peroxynitrite formation [76]. Such damaging ROS generation can be blocked by the ceramide inhibitor, fumonisin-B1 [40]. Peroxynitrite may therefore mediate apoptosis, as the iNOS inhibitors, aminoguanidine and nicotinamide, prevent FA-induced apoptosis of β-cells in the islets of ZDF rats.

In the later stages β-cell lipoapoptosis, profound mitochondrial alterations occur, such as ceramide-induced inhibition of the mitochondrial respiratory chain complexes I and III [77, 78], superimposed with enhanced ROS production; this mimics the pharmacological actions of the mitochondrial complex blockers, antimycin A- and rotenone-induced ROS formation in INS-1 β-cells [79]. At this stage, apoptosis now becomes evident. Interestingly, FFA-induced UCP2 expression can govern the apoptosis of β-cells through a decrease in intracellular ATP levels [80]; mitochondria are subsequently uncoupled and glucose-stimulated insulin secretion is modified [81]. In contrast, UCP knockout mice are protected from impaired β-cell function [82]. As long as β-cell proliferation and replenishment is in equilibrium with the lipoapoptotic rate, islet function remains adequate. If however, the apoptotic rate exceeds the rate of compensatory β-cell hyperplasia, then naturally insulin production declines to below minimal levels required to compensate for insulin resistance, islets exist in a functionally compromised state and overt diabetes arises.

Despite these potent effects, the anti-apoptotic properties of leptin can intervene at the level of β-cell lipoapoptosis. The ratio of pro-apoptotic and anti-apoptotic Bcl2 family members is crucial for the susceptibility of a cell to noxious lipid derivatives. When islets are exposed to surplus FAs, Bcl2 levels decline rapidly; however, such suppression can be prevented by leptin [83]. Several lines of evidence further describe a protective function of leptin against β-cell lipoapoptosis, through a reduction in ceramide synthesis and iNOS expression, in addition to prevention of FA-induced reduction of Bcl2, therefore inhibiting cell-death [40, 51, 83].

Cardiomyocyte lipoapoptosis

Lipid overaccumulation in the heart is associated with cardiac dysfunction and heart failure, which may form a platform for diabetic cardiomyopathy [19, 84]. The ceramide pathway may participate in the cardiomyopathy of ZDF (fa/fa) rats, as progressive cardiac lipoapoptosis has resulted in severe loss of myocardial function [85]. Indeed, at an early age, obese ZDF (fa/fa) rats develop echocardiographic evidence of impaired myocardial contractility [19]. However, post 20-weeks of age, severe lipotoxic alterations are evident in the heart, occurring as a consequence of cardiomyocyte lipoapoptosis [19]. In parallel, in vitro studies have demonstrated that cardiomyocytes cultured with palmitate exhibit reduced FA oxidation through elevated malonyl-CoA levels and lowered AMPK activity; concomitantly, TG and ceramide content increases, superimposed with induced caspase-3-like activity and DNA laddering [86].

Further evidence of the existence of lipotoxic heart disease in rodents comes from a study by Chiu and colleagues. These authors demonstrated the apoptotic consequences of FA overload in their cardiomyocyte-specific acyl CoA synthase (ACS)-overexpressing mice, a mouse model of lipotoxic cardiomyopathy [65]. Such an overexpression of ACC was shown to cause an increase in cardiac TG and ceramide, coupled with cytosolic cytochrome-c extrusion from mitochondria, a universal feature of apoptotic cell death. Collectively, symptoms of lipid cardiomyopathy, congestive heart failure and premature death were evident [65]. Human studies have reinforced such findings, as a marked increase in intramyocardial lipid accumulation was observed in subjects with heart failure and obesity-associated diabetes [87]. Taken together, human data further implicates that the cardiac accumulation of TG is associated with FFA exposure, generalized ectopic FA deposition, vascular overload and hypertrophy [41, 88].

The adipocyte and apoptosis

Adipocyte hypertrophy serves to provide anti-steatotic protection to non-adipose tissues from lipoapoptosis; however, how do adipocytes and adipose tissue on the whole, cope with a “lipotoxic insult”? If indeed such an event exists for adipocytes.

Apoptosis of the mature adipocyte may play a fundamental role in cellular turnover and adipose tissue remodeling, to ultimately maintain tissue homeostasis. In fact, it has been postulated that adipocyte number is essentially programmed during childhood and adolescence [89]; whereas during adulthood, a delicate balance between adipocyte generation and adipocyte cell-death remains tightly synchronized and constant [89]. However, an imbalance in such processes may also occur, especially in pathological states, such as diet-induced obesity and diabetes.

Obesity in particular, progresses through adipocyte hypertrophy, which in itself increases the possibility that enlarged fat-cells will encounter necrotic cell-death at an increased frequency; this may further initiate local inflammatory cascades. However, the precise mechanisms that trigger such apoptotic or necrotic pathways are currently undetermined. Our recent studies highlight one plausible scenario, whereby enhanced plasticity and remodeling of the extracellular matrix (ECM) can accommodate adipocyte growth during adipose tissue expansion [90]. Indeed, obesity can upregulate ECM constituents, such as collagens [91], which enhance the connective fiber content of the adipose matrix, thus counteracting its’ tensility and capacity to expand. Such structural pressures may therefore increase the number of necrotic adipocytes [92] and subsequent inflammation. To address this issue, we generated ob/ob mice deficient in collagen-VI. Regardless of the current dogma that hypertrophic adipocytes exacerbate inflammation, these mice paradoxically exhibited vast improvements in metabolic and inflammatory stature [90], despite an increase in adipocyte size. A more flexible adipocyte matrix may therefore be more amenable to adipocyte expansion, reduce the rate of adipocyte cell-death and minimize necrosis-induced inflammation to facilitate the storage of excess lipids, to ultimately prevent ectopic lipid deposition and lipotoxicity.

To further emphasize the significance of the ECM in adipose tissue remodeling, a recent study by Strissel and colleagues provides an elegant example of adipocyte cell-death and macrophage-orchestrated dynamics in states of obesity. The group documented that mice fed a high-fat diet for 20 weeks exhibit a marked increase in adipocyte death, from <0.1% at baseline, to 80% at week 16; concomitant with extensive collagen deposition, infiltrating adipose tissue macrophages, an upregulation in inflammatory genes and, associated insulin resistance [92]. Surprisingly by week 20, adipocyte number was restored with smaller adipocytes, coincident with reduced adipocyte death and improved insulin resistance. Although the incidence of adipocyte death promotes inflammation in adipose, a pro-inflammatory setting at sites of tissue remodeling is merely transient [93]. This clears the way for a remodeling program that ameliorates inflammation and restores the capability of adipose to expand. Thus a phenotypic switch from hypertrophic to hyperplasic obesity may ultimately prevail.

With regards to major technological advances that enable the study of adipocyte cell-death and de novo adipogenesis in vivo, we recently generated the first inducible and reversible model of lipodystrophy, termed the FAT-ATTAC mouse (fat apoptosis through targeted activation of caspase-8) [94]. Administration of a chemical compound that forces dimerization of an adipocyte-specific caspase-8-FKBP fusion protein, results in absolute loss in functional adipose, due to apoptosis of adipocytes. Surprisingly, the lack of adipocytes causes a reduced level of systemic inflammation despite accumulation of a large number of macrophages in the remnant fat-pad, further underlining the importance of the adipocytes as an essential component of the local cellular cross-talk that results in a pro-inflammatory cascade [94]. Notably, functional adipocytes can be recovered upon cessation of treatment in this particular model.

Concluding remarks

Lipotoxic effects prevail in almost every cell-type contributing towards homeostatic control of energy metabolism. In all cases, the cytoplasmic enrichment of unesterified FAs or triglycerides is associated with a deterioration of insulin sensitivity, overall cellular function and an increased rate of cell-death. This highlights the important role of the adipocyte and adipose tissue in maintaining systemic insulin sensitivity. In this somewhat “adipocentric” view of energy homeostasis, the ability of adipose tissue to appropriately respond to excess caloric intake by engaging in hypertrophic and hyperplastic responses with minimal associated inflammation is key to maintaining normal energy homeostasis, thereby avoiding ectopic lipid deposition in non-adipose tissue. The challenge for the future will be to identify the most limiting step(s) in this process and to pin-point the areas open for pharmacological intervention.

Acknowledgments

We kindly thank Kate McCorkle for assistance with the figures. Our research is supported by NIH grants R01-DK55758 and R01-CA112023 (PES) and a JDRF post-doctoral fellowship 3-2008-130 (CMK).

Copyright information

© Springer Science+Business Media, LLC 2009