Lipotoxicity in the Pancreatic Beta Cell: Not Just Survival and Function, but Proliferation as Well?

Abstract

Free fatty acids (FFAs) exert both positive and negative effects on beta cell survival and insulin secretory function, depending on concentration, duration, and glucose abundance. Lipid signals are mediated not only through metabolic pathways, but also through cell surface and nuclear receptors. Toxicity is modulated by positive signals arising from circulating factors such as hormones, growth factors and incretins, as well as negative signals such as inflammatory mediators and cytokines. Intracellular mechanisms of lipotoxicity include metabolic interference and cellular stress responses such as oxidative stress, endoplasmic reticulum (ER) stress, and possibly autophagy. New findings strengthen an old hypothesis that lipids may also impair compensatory beta cell proliferation. Clinical observations continue to support a role for lipid biology in the risk and progression of both type 1 (T1D) and type 2 diabetes (T2D). This review summarizes recent work in this important, rapidly evolving field.

Introduction

The subject of toxic effects of lipid species on pancreatic beta cells is broad and growing. Early biochemical work illustrated how the intracellular metabolism of lipids can either promote or inhibit the insulin secretory response to glucose, depending on the context. Lipids are now known to act not only through biochemical nutrient pathways, but also through signaling via cell surface and nuclear receptors. Newer findings link lipotoxicity to inflammation, oxidative, nitrosative and endoplasmic reticulum (ER) stress pathways and autophagy. Many outstanding minds and teams have touched upon this field. Here we will cover recent advances specific to toxic effects of lipids on beta cell survival, insulin secretory function, and beta cell mass (Figure 1). As such, we do not discuss the well-established beneficial effects of lipids on insulin secretion via FFAR1 and other pathways. The review is organized from an outside-in perspective, beginning with extracellular factors, and from a physiological rather than a biochemical perspective. A section is devoted to the evolving new concept that lipotoxicity may impact beta cell mass by reducing beta cell proliferation, with an emphasis on how our own work integrates with the field. Given the broad nature of this topic, coverage of each concept is brief; the reader is encouraged to read the original sources. By design, the review is restricted to work from the past year or two.

Figure 1.

Free fatty acids exert both positive (green) and negative (red) effects on beta cell mass and function. FFAs signal through receptors such as FFAR1 and PPARs, or through metabolic pathways as comprehensively reviewed in [1, 2]. Positive effects are mediated predominantly through FFAR1 and PPARs. Negative effects are mediated through inflammation, cellular stress mechanisms, and possibly inhibition of the cell cycle. Negative effects of FFAs are modulated by growth factors and incretins. Whether autophagy plays a net positive or net negative role is controversial.

Extracellular signals influencing lipotoxicity

Lipid effects on the beta cell are modulated by extracellular factors. Signals that promote beta cell mass and function, such as lactogens, estrogens, and incretins generally protect beta cells against lipotoxicity. As reviewed in detail [1, 2], toxic effects of lipids are usually manifested only when high glucose is also present. Extracellular beta cell toxins such as inflammatory cytokines synergize with lipotoxicity to further impair beta cell function and survival, as summarized below.

Growth factors and hormones

Circulating growth factors impact lipotoxicity in the beta cell. While lactogens protect beta cells against lipotoxic cell death via activation of Jak-Stat signaling [3], hepatocyte growth factor actually promotes lipotoxicity [4]. The insulin signaling nuclear factor FoxO1 mediates some aspects of lipotoxicity; a recent study mapped out which genes are regulated by FoxO1 in a beta cell line [5]. The female sex steroid hormone estradiol was found to regulate islet lipid synthesis; deletion of the ER-alpha receptor predisposed mice to lipotoxic beta cell dysfunction [6].

Significant recent effort has been directed towards understanding how incretin hormones, in particular Glp-1, interact with lipotoxicity. Increasing Glp-1 signaling is a new T2D therapeutic approach that has generated excitement because improved insulin secretion is accompanied by weight loss and, possibly, beta cell regeneration. Lipid exposure negatively impacts incretin signaling, both by downregulation of the Glp-1 receptor [7] and by interfering with downstream cAMP signaling [7, 8]. Treating diabetic mice with a combination of lipid-lowering therapy and Glp-1 agonist improved beta cell mass and function better than either alone [7]. Incretins were found to promote the interconnected network of beta cells in human islets, and exposure to lipids disrupted this connectivity, and impaired insulin secretion [9]. A number of teams have found that incretin signaling promotes insulin secretion and beta cell survival to counteract glucolipotoxicity in vitro, through effects on mitochondria [10], insulin signaling intermediates such as Akt and mTor [11–13], oxidative and endoplasmic reticulum stress [14], and the nuclear factor SREBP1 [13]. The protective effects of Glp-1 signaling against lipotoxicity have been extended to human islets [15].

Inflammatory mediators

Inflammation negatively impacts beta cell survival and function through multiple mechanisms [16]. Fatty acids can directly activate inflammatory pathways themselves, to potentiate inflammatory toxicity. In some cases inflammatory signals are propagated via intracellular lipid species as described below. Lipid exposure in vivo increases islet inflammation, and inflammatory cytokines and immune cells are present in the pancreas in T2D (recent examples: [17, 18]). Treating beta cells with palmitate increased chemokine production and recruitment of pro-inflammatory macrophages, via TLR4-Myd88 [19]. High fat feeding, and palmitate treatment in vitro, increased islet production of macrophage migration inhibitory factor (MIF); deletion of MIF protected beta cells against lipotoxic cell death [20]. Lipopolysaccharides (LPS) treatment reduced insulin secretion and expression of beta cell differentiation markers Pdx1 and MafA, though TLR4 and NF-kβ [21]. Lipids potentiated IL-1beta-induced endoplasmic reticulum stress through IRE1/Xbp activation [22]. Cytokines impaired insulin secretion and increased cell death via a pathway involving 12-lipoxygenase and its downstream lipid product, 12-hydroxyeicosatetraenoic acid (12-HETE) and Nox-1 and reactive oxygen species [23]. Additionally, blocking palmitoylation protected INS-1 cells against cytokine-induced oxidative and nitrosative stress [24]. Thus, lipotoxicity promotes and potentiates islet inflammation, supporting the concept that modulating inflammation might be a therapeutic approach for diabetes [18].

Membrane receptors influencing lipotoxicity

Initially, all lipid effects on the beta cell were thought to arise from effects on metabolic pathways. The finding that FFAs activate G-protein coupled receptors (GPCRs) has generated enormous excitement because of the biological and therapeutic implications [25]. FFAs activate GPR40 (now known as FFAR1), which is expressed on the surface of human and rodent beta cells, and is, remarkably, thought to mediate many of the positive effects of fatty acids on insulin secretion without negatively impacting beta cell function or survival [26–28]. Glucolipotoxicity may exert some negative effects by interfering with normal FFAR1 function. In one study hyperglycemia decreased, but hyperlipidemia increased, FFAR1 expression [29]. In INS-1 cells, saturated fatty acids decreased FFAR1 expression, whereas unsaturated fatty acids increased FFAR1 expression and protected against lipotoxicity [30]. Supporting a role for endogenous FFAR1 in insulin secretion, a single nucleotide polymorphism at the FFAR1 locus correlated with insulin secretory function in people; genotyping this locus may allow clinical prediction of which patients will respond to FFAR1 agonists [31].

FFAR1 agonists are under development as therapies for T2D. Preclinical studies suggest they potentiate insulin secretion, decrease beta cell apoptosis and maintain beta cell mass [25]. Like other GPCRs, careful ligand design may allow selective activation of positive functions without others that are less desirable [26, 32]. FFAR1 agonists potentiate glucose-dependent insulin secretion in vivo in diabetic rodents [33–36]. Although the majority of analyses suggest that FFAR1 activation doesn’t lead to lipotoxicity, in fact protects against lipotoxicity, a few studies have found that some lipotoxic effects may be mediated by FFAR1 signaling. Extended exposure of human islets to palmitate decreased insulin content and secretion, which was preventable by a FFAR1 antagonist [37]. In a mouse model of beta cell overload with failure, blocking FFAR1 reduced insulin secretion, circulating proinsulin, and beta cell apoptosis, suggesting the possibility that in the setting of beta cell failure increasing insulin secretion by FFAR1 agonism might increase stress by further overloading the beta cell [38, 39].

Intracellular mechanisms of lipotoxicity

Classically, lipid toxicity to beta cells was thought to arise from chronic alteration of biochemical substrate flux patterns, rendering the beta cell less responsive to glucose, and through protein kinase C signaling. More recently, lipid exposure has also been shown to activate cell stress responses including oxidative stress, endoplasmic reticulum stress, and autophagy. Much work remains to be done to identify how each of these mechanisms relates to human diabetes.

Metabolic pathways

Decades of research have illuminated many details of the biochemistry of lipid metabolism in beta cells. The roles of metabolic pathways in the positive and negative effects of fatty acids on insulin signaling, interaction between glucose and lipid metabolism, and lipotoxicity in general have been extensively and carefully reviewed [1, 2]. Upon entering the beta cell, free fatty acids are activated by acyl-CoA synthase, and then either oxidized or re-esterified for storage or glycerolipid cycling [2]. Glucose, fatty acids and amino acids interact at a biochemical level to influence many important cellular processes through the Krebs cycle, pyruvate cycling, and the glycerolipid-free fatty acid cycle [2, 40]. Over the past 1–2 years some new work has contributed to understanding of the role of FFA metabolism in lipotoxicity.

At the level of lipid entry into beta cells, one histological analysis has found that lipoprotein lipase (LPL), important for cleaving triglyceride to allow fatty acid entry into cells, is surprisingly not intravascular in mouse islets, but instead appears to be intracellular, where it would not have access to circulating triglyceride [41]. Islet expression of LPL was not altered by fasting-fed state, but was regulated by leptin [41]. Islet lipid uptake may be regulated by signals that change LPL location to extracellular; disruption of these signals may contribute to reduction of the lipid component of glucose stimulated insulin secretion (GSIS). A study seeking to determine which acyl-CoA synthase participates in GSIS determined that Acsl4 is required for fatty acid potentiation of GSIS in INS1 cells [42]. However, Acls4 was not required for metabolism of long chain FA, but instead appeared to be protective by sequestering a toxic fatty acid species, epoxyeicosatrienoic acids (EETs). Once inside the cell, fatty acid oxidation is thought to not be pathogenic, but instead protective against lipotoxicity as measured by ER stress [43, 44]. The insulin signaling intermediate mTor may impact rates of fatty acid storage versus utilization in human beta cells [45].

New findings regarding mechanisms of lipotoxicity include a pathway linking lipid exposure to altered glycosylation patterns in beta cells, resulting in impaired glucose transport [46]. Palmitate interferes with glucose uptake, calcium signaling, mitochondrial respiration and insulin content [47, 48]. Palmitate treatment altered calcium handling and insulin secretion through neprilysin, a secreted protease not previously known to impact beta cell function [49]. The metabolic lipotoxicity pathways underlying impaired insulin secretion and increased cell death were found to be mechanistically distinct in INS1 cells [48]. Also in INS1 cells, a comprehensive analysis of metabolite levels in the context of mRNA expression and histone modification cataloging is an important resource for the community [50]. Another mechanism of palmitate toxicity may be aberrant palmitoylation of proteins; blocking palmitoylation prevented lipotoxic beta cell death [51]. In a technical breakthrough, the interaction of glucose and fatty acids at the level of the electron transport chain was directly visualized using a sophisticated confocal-based imaging technique [52].

Other lipid-related species may also play a role in beta cell function and dysfunction. Ceramide and sphingolipids impact these processes [53]. Mice fed an isocaloric diet in which long chain FAs were replaced by medium chain FAs showed glucose intolerance and impaired insulin secretion [54]. The role of cholesterol in beta cell function and mass is unclear. The cholesterol transport protein ABCG1 was increased in insulinoma tissue, and correlated with insulin secretion [55]. The intracellular cholesterol transport protein Npc1, genetically associated with risk of obesity in humans, may also play a role in insulin secretion [56]. LXR alpha, a receptor for cholesterol-related compounds, is important for insulin secretion in vitro and in vivo by altering glucose metabolism, ATP production and calcium channel flux; modulating its activity deregulated lipid metabolism via SREBP [57].

Stress pathways: oxidative stress, ER stress and autophagy

Oxidative stress is caused by generation of reactive oxygen species (ROS) that exceeds reducing capacity. Beta cells have limited anti-oxidative defense mechanisms and are particularly susceptible to oxidative damage. Consequences of redox imbalance include lipid peroxidation, oxidation of proteins, DNA damage and interference of reactive species with signal transduction. Excesses of lipids and glucose induce oxidative and nitrosative stress in beta cells; short-term activation of ROS increases GSIS, but excessive ROS impairs insulin secretion [24, 58]. Human islets from both diabetic and nondiabetic individuals have detectable lipid peroxide protein adducts, suggesting oxidative damage [59]. In mice, lipid infusion increased islet ROS, and beta cell dysfunction was prevented by treatment with a reducing agent or inhibition of NADPH oxidase [60, 61]. Nicotinamide protected INS1 cells against lipotoxic cell death through sirtuins [62]. Reducing agents or antioxidants may improve beta cell function by protecting against oxidative stress.

ER stress refers to failure to maintain homeostasis of the ER, the site of protein folding for all secretory peptides. The beta cell is a workhorse for insulin synthesis and secretion, and is sensitive to ER stress [16, 63]. Oxidative stress increases ER stress because redox state is important for proper ER function; however, agents that stress the beta cell ER may not alter its redox state [64]. Under conditions of moderate ER load the unfolded protein response (UPR) compensatory mechanism engages, but if the stress cannot be resolved cell death ensues. Fatty acids have long been known to increase ER load, by affecting protein processing, trafficking, Ca2+ regulation and oxidative stress [16, 63]. Saturated fatty acids such as palmitate induce ER stress, whereas unsaturated fatty acids exert protective effects; in INS1 cells, palmitate activated UPR via the PERK and IRE1 pathways, and the effect was prevented by co-treatment with oleate [65]. In bTC3 cells, palmitate activated all three UPR pathways, through store-activated calcium entry [66]. Palmitate also induced ER stress in a human beta cell line [67]. Protection by unsaturated fatty acids may be mediated by cellular inhibitor of apoptosis-1 (cIAP1), an E3 ligase for the cell death director CHOP [68]. A screen for effectors of palmitate-induced apoptosis discovered ER communication with the intrinsic mitochondrial apoptosis pathway, through BH3-only proteins DH5 and PUMA [69]. Another screen found ubiquitin C-terminal hydrolase L1 to be required for ER function, beta cell survival and insulin secretion when exposed to lipotoxicity [70]. ER morphological and functional distress caused by fatty acid treatment was traced to AMPK acting through a GTPase called dynamin related protein 1 (DRP1), known to regulate mitochondrial fission [71]. Another study found that palmitate influenced the makeup of ER lipid rafts, which determined the cellular response to ER stress [72]. The immediate early gene Npas4, rapidly induced by palmitate, promoted beta cell survival during ER stress [73]. Treatment with incretins protected beta cells against palmitate-induced ER stress [74]. Combined with the large body of clinical literature suggesting beta cell exhaustion contributes to T2D, and possibly T1D, management of ER stress may assist in preservation of beta cell mass and function.

Autophagy is used to recycle unnecessary or dysfunctional cellular components, and to ensure survival during starvation by maintaining cellular energy levels. Autophagy increases during nutrient stress; controversy exists as to whether autophagy is predominantly detrimental or protective to beta cells. Palmitate increases autophagy in rat and human beta cells and is associated with ER distension [75]. Atg7 overexpression sensitized cells to palmitate-induced autophagy, which was found to increase inflammatory mediators via cathepsin B and the NLRP3 inflammasome, linking cellular nutrient stress to inflammation [76]. On the other hand, mice deficient in beta cell autophagy due to deletion of Atg7 showed impaired ER adaptation and heightened sensitivity to ER stress, resulting in frank diabetes when mated onto a leptin deficient background [77]. Although fatty acids increase autophagosomes, suggesting increased autophagy, a dynamic study found that in fact autophagic flux, a measure of activity, was reduced in beta cells treated with oleate or palmitate [78]. Consistent with the concept that autophagy is activated by starvation, and fatty acid treatment represents nutrient excess, simulating starvation by rapamycin treatment partially restored autophagic flux [78]. The role of autophagy in beta cell failure in T2D requires further investigation.

Lipotoxicity effects in the nucleus

Some negative effects of lipids occur at the level of the nucleus. PPARs, classic nuclear lipid receptors, exert mostly positive effects on beta cell mass and function. Our own work has supported a new role for lipotoxicity in preventing beta cell proliferation, described in some detail in the cell cycle section below.

Lipid receptors

The peroxisome proliferator-activated receptor (PPAR) nuclear receptors are transcription factors regulating many genes involved in differentiation, development and metabolism. They play an important role in T2D because of their effects on circulating glucose and lipids, and both PPARα and PPARγ are targets of medications currently in use. PPARs are important for beta cell function. Overexpression of PPARα in obese mice preserved insulin secretion without affecting beta cell mass in one study [79]; in another study, however, treatment of obese mice with pan-PPAR agonist bezafibrate also maintained glucose homeostasis, but in this case by preventing weight gain [80]. Expansion of beta cell mass in leptin deficient mice is dependent on PPARγ [81]. PPARγ activation improved insulin secretion by increasing expression of FFAR1, as well as beta cell differentiation genes, in a pathway dependent on glucose transport, FFAR1, and phospholipase C [29, 82]. Thus, recent literature suggests that most of the cellular effects of lipids via PPARs are positive.

Cell cycle regulation: lipid effects on beta cell proliferation

A new, less well characterized form of lipotoxicity has been postulated: that in addition to impairment of insulin secretion and induction of beta cell death, lipid exposure may prevent beta cell mass expansion by inhibiting beta cell proliferation. Data supporting this hypothesis include cell culture studies showing that long chain FFAs reversibly blocked glucose induced beta cell proliferation in INS-1 cells; intriguingly, fatty acid oxidation was not required, and the dose of fatty acids used did not interfere with glucose metabolism [83, 84]. Palmitate has been shown to reduce proliferation in cultured human beta cells, an effect that was mitigated by co-incubation with oleate [85]. FFAR2, a receptor for short chain fatty acids, was upregulated in islets at a stage of pregnancy when beta cells proliferate [86]. In vivo, raising circulating levels of FFA by direct infusion of triglyceride with heparin prevented glucose-induced mouse beta cell proliferation [87]. FFA also reduced proliferation when directly applied to primary mouse beta cells in vitro, and the mechanism was traced to induction of cell cycle inhibitor proteins p16 and p18 [87].

Other studies contradict these findings, however, and FFAs have even been postulated to promote compensatory beta cell proliferation [88]. In seeming contradiction, beta cell proliferation increases in mice overfed with a high fat diet, but FFAs don’t increase until after beta cell proliferation begins, and proliferation may be driven by other changes associated with overnutrition [89]. In cultured rat islets, palmitic and oleic acids increased beta cell proliferation and insulin secretion synergistically with prolactin treatment [90]. A 1:1 mixture of oleic:palmitic acid stimulated tritiated thymidine incorporation in rat islets [45]. In Zucker fatty rats subjected to partial pancreatectomy, beta cell regeneration exceeded that of non-hyperlipidemic controls, with robust beta cell proliferation [91]. In seeming direct contradiction to the infusion study in mice, intravenous infusion of triglyceride and glucose into 6-month old rats resulted in increased beta cell mass and proliferation [92]. Many experimental differences may explain the seemingly contradictory findings, including the species, age, degree of hyperglycemia, insulin resistance and infusion procedure. A prior study of lipid infusion in rats also concluded that lipids increased beta cell mass and proliferation [93]. Thus, whether in vivo exposure to lipids promotes or prevents beta cell proliferation remains an open question, and how this relates to human biology is uncertain.

Evidence lipotoxicity is relevant to living human beings

Circulating lipids influence risk of developing T2D [94]. New data confirm prior findings that triglyceride levels positively correlate with risk of T2D, and specifically with beta cell dysfunction [95, 96]. In adolescents, acute elevation of triglyceride by overnight infusion reduced insulin secretion in response to hyperglycemic clamp; whether this effect is race-dependent remains unclear [97, 98]. Elevation of free fatty acids was strongly associated with reduced beta cell function in both children and adults, with, intriguingly, a more consistent effect seen on insulin secretory capacity than on insulin sensitivity [99]. However, the potentiation of acute glucose-stimulated insulin secretion by prior infusion of insulin was found to be independent of free fatty acids [100].

The link between cholesterol metabolism and diabetes is an active area of investigation. HMG-CoA reductase inhibitors (statin-class medications) are now known to slightly increase the risk of new onset T2D, but the mechanism remains unknown [101]. Seemingly contradictory to this, atorvastatin preserves beta cell function in some patients with early T1D; the reason behind this observation is equally unclear [102]. The cardioprotective high-density lipoprotein (HDL) cholesterol particle, which transports cholesterol out of tissues and back to the liver for clearance, appears to be protective against development of T2D [103]. Protection was correlated with larger particle size, possibly implicating flux of cholesterol transport in diabetes prevention. Blood taken from subjects treated with a CETP inhibitor, which elevates HDL levels, increased insulin secretion from a beta cell line, an effect that may have been related to efflux of cholesterol from the cultured cells [104]. In diabetic mice, chronic infusion of HDL improved blood glucose and pancreatic islet architecture [105]. HDL protected beta cells against ER stress-mediated cell death [106].

A novel concept based on observations of diabetes remission after bariatric surgery links beta cell dysfunction in T2D with fat accumulation in the pancreas itself [107]. Consistent with this hypothesis, pancreatic steatosis in people with genetic ATGL deficiency was associated with impaired insulin secretory function without impairment in insulin sensitivity [108]. On the other hand, an imaging-based study found that in nondiabetic individuals with mild obesity, pancreatic lipid content varied by ethnicity, and in both African American and Caucasian subjects pancreatic triglyceride was positively correlated with first phase insulin secretion after IV glucose challenge [109]. Whether pancreatic lipid content is a marker of a global process, such as insulin resistance or failure of the adipocyte storage system, or has direct effects on beta cell function, remains unknown.

Conclusion

Beta cell lipotoxicity takes many forms, with respect to lipid species, cellular location of action, pathways involved, and end effects on beta cell survival, mass, and function. Since lipid biology clearly interacts with diabetes risk and complications in people, and lipid signaling is amenable to therapeutic intervention, this is an important field of study that directly impacts human health. Much work needs to be done to identify rational basis for new therapies that target lipotoxicity to prevent and treat beta cell failure in diabetes.