Introduction
Obesity represents a global health pandemic affecting both developed and, increasingly, developing countries(Reference Bluher1). The burden posed by obesity on people’s health and the health care systems worldwide is a direct consequence of its life-threatening comorbidities. Indeed, obesity is not to be merely considered as an abnormal excess of body fat, instead the excess of body weight and the overexpansion of the adipose tissue trigger a plethora of pathophysiological derangements that disrupt cardiometabolic health(Reference Brown, Fujioka and Wilson2). In support of this, obesity is closely associated with the onset of type 2 diabetes (T2DM)(Reference Zatterale, Longo and Naderi3), cardiovascular diseases(Reference Koliaki, Liatis and Kokkinos4), non-alcoholic fatty liver disease(Reference Spengler and Loomba5), certain types of carcer(Reference Avgerinos, Spyrou and Mantzoros6) and mental health issues(Reference Mazon, de Mello and Ferreira7).
In terms of its pathophysiology, obesity arises from a long-term positive energy balance which is the direct consequence of energy overconsumption, fostered by the ready availability of cheap, highly palatable, nutrient-dense foods and physical inactivity(Reference Cordain, Eaton and Sebastian8–Reference Hall, Ayuketah and Brychta11). These are two pivotal paradigms which have been fuelling obesity and its comorbidities in high-income countries and are now spreading to low- and middle-income countries due to rapid behavioural and cultural changes prompted by the pressure of globalisation(Reference Cecchini, Sassi and Lauer12). Remarkably, obesity develops despite the human subject’s body being equipped with a sophisticated and finely tuned centre for the control of energy balance, located in the hypothalamus, which is tasked to match energy intake and expenditure to one another to maintain body weight within a tight range(Reference Rui13). This is achieved by the ability of the hypothalamus, and particularly the arcuate nucleus, to sense peripheral hormonal, nervous and nutritional cues related to the nutritional status and mount physiological responses to maintain energy homoeostasis(Reference Rui13–Reference Cowley, Smart and Rubinstein16). Nonetheless, this homoeostatic system appears to fail in obesity, thereby predisposing to weight and adiposity gain(Reference Williams17). From a nutritional perspective, long-chain saturated fatty acids and sugar represent the primary nutritional culprits in promoting hypothalamic dysfunction(Reference Sergi and Williams18,Reference Gao, Bielohuby and Fleming19) .
Thus, obesity arises as a consequence of nutrient oversupply, with this energy excess being stored in the adipose tissue in the form of triglycerides. The adipose tissue triglycerides represent an energy reservoir to be used in response to fasting or increased energy demand, as in the case of prolonged physical activity. In particular, the adipose tissue releases free fatty acids as a result of triacylglycerols lipolysis which, in turn, is a finely tuned process under the control of hormonal, nervous and nutritional inputs(Reference Zechner, Zimmermann and Eichmann20). Therefore, under physiological conditions, this allows the correct matching of free fatty acid supply to and demand of extra adipose tissues. However, this tight physiological regulation of lipid metabolism becomes impaired in obesity, with a mismatch between fatty acid supply to metabolically active tissues and their ability to catabolise them(Reference Fucho, Casals and Serra21–Reference Kim, Hickner and Cortright24). In this regard, adipose tissue dysfunction represents a pivotal driver of free fatty acid oversupply to metabolically active tissues such as the skeletal muscle and liver. Indeed, dysfunctional adipocytes are characterised by insulin resistance with promotes the disinhibition of lipolysis leading to an uncontrolled release of free fatty acids from the adipose tissue(Reference McGillicuddy, Reynolds and Finucane25). Uncontrolled fatty acid spill over from the adipose tissue and fatty acids derived from chylomicrons and VLDL, in the face of impaired fatty acid oxidative capacity and reduced energy demand, promote the ectopic accumulation of lipid in tissue not suited for lipid storage, namely skeletal muscle(Reference Aas, Kase and Solberg26), liver(Reference Anstee and Goldin27), pancreas(Reference Assimacopoulos-Jeannet28) and heart(Reference Zhou, Grayburn and Karim29), a phenomenon termed lipotoxicity(Reference Unger30). Lipotoxicity, in turn, has been proposed as one of the key effectors underlaying the link between obesity and its comorbidities, therefore representing a putative target to improve cardiometabolic health(Reference Sezer31). However, not all the lipid species appear to be detrimental when accumulating in non-adipose tissues. In keeping with this, while triacylglycerols intracellular accumulation does not appear to promote insulin resistance(Reference Bosma, Hesselink and Sparks32,Reference Metcalfe, Smith and Turner33) ; the opposite is true for ceramides which have been shown to hamper insulin signalling and impair insulin sensitivity(Reference Chavez, Knotts and Wang34–Reference Chaurasia and Summers37).
Saturated fatty acids have been widely proposed as a pivotal metabolic culprit in fostering obesity(Reference Sergi and Williams18) and insulin resistance(Reference Sears and Perry38), with ceramides being one of the potential mediators driving the metabolically detrimental effects of these dietary fatty acids(Reference Chavez and Summers36,Reference Chaurasia, Tippetts and Mayoral Monibas39) . In light of this, the aim of this review is to provide an overview on the molecular mechanisms underpinning the effects of ceramides on insulin resistance as well as their role as an innovative biomarker for T2DM. Additionally, this manuscript will discuss the role of saturated fatty acid quality in the modulation of ceramide metabolism.
Ceramide biosynthesis and breakdown
Ceramides are the precursor of complex sphingolipids that are integral components of cell membranes. Cellular ceramide content is regulated by hormonal, metabolic and inflammatory cues which include substrate availability, specifically palmitoyl-CoA, the CoA-derivate of the long-chain saturated fatty acid palmitic acid, pro-inflammatory cytokines, adiponectin and fibroblast growth factor 21 signalling(Reference Chaurasia and Summers37,Reference Chaurasia, Talbot and Summers40) . Ceramide de novo synthesis takes place in the endoplasmic reticulum (ER) and begins with the condensation of palmitoyl-CoA and the amino acid serine catalysed by serin-palmitoyl transferase. The resulting product, 3-ketosphinganine, is then converted to sphinganine by 3-ketosphinganine reductase. The next step is represented by a N-acylation reaction mediated by a family of enzymes termed ceramide synthases, which encompasses six isoforms. In light of this, part of this step provides a wide variety of dihydroceramides differing in their acyl chain length are synthetised, ranging from 14 to 34 carbon atoms. The final step of the de novo ceramide synthesis is catalysed by dihydroceramide desaturases which introduces a 4,5-trans-double bond into the dihydroceramide molecules forming ceramides(Reference Chaurasia and Summers37,Reference Chaurasia and Summers41) . These enzymatic reactions, particularly those mediated by ceramide synthases, give raise to a variety of ceramide pool which potentially confer different pathophysiological roles to these sphingolipids and is most likely influenced by the availability of specific fatty acids as well as the ceramide synthase substrate preferences(Reference Turpin-Nolan and Bruning42).
Ceramides can also be synthetised following alternative pathways, namely the catabolism of sphingomyelin and the salvage pathway. The former pathway generates ceramides via the sphingomyelinase-mediated catabolism of sphingomyelins(Reference Gault, Obeid and Hannun43), whereas the latter is based on the recycling of sphinganine derived from ceramidase-induced ceramide catabolism(Reference Kitatani, Idkowiak-Baldys and Hannun44). A further pathway responsible for ceramide synthesis, albeit less frequent, relies on the hydrolysis of glycosphingolipids by glycosidases(Reference Kitatani, Sheldon and Rajagopalan45). Once synthetised, ceramide can be exported to the Golgi apparatus where head groups are added to the 1-hydroxyl position to produce ceramide-1-phosphate, sphingomyelin or glycoceramides as extensively reviewed elsewhere(Reference Hannun and Obeid46).
Beside their de novo synthesis and regeneration, the metabolism of ceramides also includes a catabolic branch (Figure 1). Indeed, ceramides can be breakdown by ceramidase-mediated diacylation leading to the formation of sphingosine, which, in turn, is converted to sphingosine-1-phosphate by specific cytosolic sphingosine kinases. Sphingosine-1-phosphate is finally broken down to hexadecenal and phosphoethanolamine by sphingosine-1-phosphate lyase localised in the ER (Figure 1)(Reference Gault, Obeid and Hannun43). Ceramide degradation is mediated by ceramidases which in mammals include five proteins classified according to their pH optimum in acid, neutral and alkaline ceramidases(Reference Chaurasia and Summers41). Beside these classical ceramides, the adiponectin receptor has also been reported to possess intrinsic ceramidase activity which is induced by adiponectin binding(Reference Holland, Miller and Wang47,Reference Vasiliauskaite-Brooks, Sounier and Rochaix48) .
The impact of ceramides on insulin resistance and the onset of type 2 diabetes mellitus
Ceramides are at the forefront of metabolic dysfunction integrating nutrient overload, particularly in the form of long-chain saturated fatty acids, obesity and insulin resistance. Indeed, these lipotoxic lipid species have been reported to disrupt insulin signal transduction pathway and promote insulin resistance(Reference Chavez and Summers36,Reference Chaurasia, Tippetts and Mayoral Monibas39) , trigger metabolic inflammation(Reference Chaurasia, Talbot and Summers40,Reference Sergi, Morris and Kahn49,Reference Boon, Hoy and Stark50) and disrupt energy and glucose homeostasis by inducing hypothalamic dysfunction(Reference Sergi and Williams18,Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51) , which not surprisingly are all pivotal features of obesity.
From a mechanistic perspective, ceramides have been shown to hamper insulin signalling via different mechanisms. The first relies on ceramide-induced activation of protein kinase Cζ which, by phosphorylating AKT/protein kinase B on a regulatory site of Pleckstrin homology domain, inhibits its translocation to the plasma membrane(Reference Powell, Hajduch and Kular52,Reference Bourbon, Yun and Kester53) (Figure 2). On the contrary a further mechanism of ceramide-induced insulin resistance involves the activation of protein phosphatase 2A and the subsequent dephosphorylation and inactivation of AKT/protein kinase B(Reference Salinas, Lopez-Valdaliso and Martin54) (Figure 2). Additionally, ceramides are able to impair insulin signalling by targeting the insulin receptor substrate. This effect is mediated by ceramide-induced activation of the double-stranded RNA dependent protein kinase – c-jun N-terminal kinase (JNK) axis which culminates with the phosphorylation of insulin receptor substrate on serine 307 and its inhibition(Reference Hage Hassan, Pacheco de Sousa and Mahfouz55). Finally, ceramides are also able to disrupt insulin signal transduction pathway by upregulating the transcription factor Pbx-regulating protein 1 and its cofactor p160 and favouring their interaction. In turn, Pbx-regulating protein 1–p160 interaction hampers insulin signalling leading to a decrease in glycogen synthesis and glucose uptake by L6 myotubes(Reference Cimmino, Lorenzo and Fiory56). These mechanistic data are extensively supported by in vivo studies both in human subjects as well as animal models. Indeed, myriocin-mediated inhibition of serin palmitoyl transferase, the first and rate-limiting enzyme in the ceramide de novo synthesis pathway, improved systemic saturated fatty acid and dexamethasone-induced insulin resistance(Reference Holland, Brozinick and Wang57), as well as enhanced adipose tissue insulin sensitivity in rats fed a high-fat diet(Reference Blachnio-Zabielska, Hady and Markowski58). Furthermore, specifically targeting of ceramides 4,5-trans double bond in their sphingoid backbone by whole body or liver and/or adipose tissue dihydroceramide desaturase 1 knock-out countered insulin resistance and hepatic steatosis induced by either a high-fat diet or leptin deficiency, indicating that the deleterious effects ascribed to ceramides are not dependent of other lipid intermediary in the ceramide biosynthetic pathway(Reference Chaurasia, Tippetts and Mayoral Monibas39). Remarkably, the effects of the inhibition of ceramide production are not limited to insulin resistance, with myriocin preventing the onset of fasting hyperglycaemia, and the decline in β-cell mass in Zucker diabetic fatty rats(Reference Holland, Brozinick and Wang57). Thus, apart from improving insulin sensitivity, the inhibition of ceramide synthesis also exerts a protective effect on pancreatic β cells, a further tissue target of lipotoxicity(Reference Ye, Onodera and Scherer59), thereby preventing the progression from prediabetes to full blown T2DM(Reference Holland, Brozinick and Wang57). The protective effects of the inhibition of ceramide accumulation upon high-fat diet feeding does not appear to be tissue specific as similar improvements in insulin sensitivity were also reported upon specific inhibition of ceramide synthesis in skeletal muscle(Reference Turpin-Nolan, Hammerschmidt and Chen60) or brown adipose tissue(Reference Turpin, Nicholls and Willmes61).
The evidence relative to the metabolic effects of ceramide in humans are limited by the impossibility to perform pharmacological or genetic manipulation to decrease ceramide synthesis as reported in animal models. However, this does not negate the metabolically detrimental effects of ceramide in humans. Indeed, circulating ceramides, particularly circulating C16:0 ceramide, correlated with insulin resistance assessed using the homoeostasis model assessment: insulin resistance(Reference Lemaitre, Yu and Hoofnagle62). This also held true relative to the accumulation of ceramide in metabolically active tissues. In agreement with this, an increase in ceramide levels in the adipose tissue correlated with hepatic lipid content and insulin resistance independently of obesity(Reference Kolak, Westerbacka and Velagapudi63). Furthermore, an increase in C16:0 ceramide and the up-regulation of ceramide synthase 6 was reported in the adipose tissue of obese individuals and positively correlated with adiposity, hyperglycaemia as well as insulin resistance assessed by euglycemic hyperinsulinemia clamps(Reference Turpin, Nicholls and Willmes61). The buildup of ceramides has also been reported in the skeletal muscle and, not surprisingly, also in this case it correlated with insulin resistance(Reference Coen, Dube and Amati64–Reference Dube, Amati and Toledo66). Most importantly, lifestyle, such as physical activity, or pharmacological interventions (metformin and pioglitazone), as well as bariatric surgery apart from improving glycaemic control also resulted in a decrease in intramyocellular and circulating ceramide levels(Reference Dube, Amati and Toledo66–Reference Warshauer, Lopez and Gordillo69) supporting the notion that a decrease in muscular and circulating ceramide levels goes hand in hand with improved insulin resistance. However, the cause–effect relationship between ceramides and insulin resistance in humans remains to be fully elucidated. Indeed, it cannot be excluded that improved insulin sensitivity in response to the aforementioned interventions may be exquisitely dependent on improved lipid metabolism of which a decrease in ceramide levels may only be a read out. Furthermore, albeit unlikely, the improvement in insulin sensitivity may be a cause, rather than a consequence, of the decrease in tissue and circulating ceramide levels. Nonetheless, the animal studies described that fat provides sufficient support to the conundrum that ceramides represent a pivotal driver of insulin resistance.
Nevertheless, the fact that not all ceramides are created equal must not be overlooked. In line with this notion, the metabolically detrimental effects described above appear to be ceramide specific. Indeed, evidence from animal models suggest that lowering C18:0 ceramide by knocking out ceramide synthase 1 both within the skeletal muscle and globally improved glucose homoeostasis(Reference Turpin-Nolan, Hammerschmidt and Chen60). Nevertheless, lowering C16:0 ceramide by targeting ceramide synthase 5 and 6 in the skeletal muscle did not improve insulin sensitivity in high-fat-fed animals(Reference Turpin-Nolan, Hammerschmidt and Chen60). On the contrary, other evidence, always collected as part of animal studies, supported the role of C16:0 ceramide as being instrumental for the development of insulin resistance(Reference Turpin, Nicholls and Willmes61,Reference Raichur, Wang and Chan70) . This was also corroborated in human subject studies reporting that in the adipose tissue, the de novo synthesis of C16:0 ceramide was a major driver of insulin resistance(Reference Chathoth, Ismail and Alghamdi71). While the involvement of specific ceramide species in promoting insulin resistance remains a matter of debate, indistinctly lowering ceramide levels by dihydroceramide desaturase 1 knock out has metabolically beneficial effects of insulin resistance and hepatic steatosis(Reference Chaurasia, Tippetts and Mayoral Monibas39).
Ceramide accumulation within the hypothalamus contributes to impaired peripheral metabolic health and obesity
The metabolically detrimental effects of ceramide accumulation are not limited to peripheral tissues, as described so far. Indeed, ceramide buildup has also been reported in the hypothalamus of rodents fed a high-fat diet leading to a deterioration of metabolic health(Reference Borg, Omran and Weir72). Hypothalamic lipotoxicity, and particularly ceramide accumulation, is not without metabolic consequences. In support of this, direct hypothalamic infusion of C6 ceramide, a cell permeable ceramide, induced hypothalamic ER stress and inflammation(Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51), two pivotal mechanisms responsible for promoting insulin and leptin resistance, posing the pathogenetic basis for obesity and impaired glycaemic control(Reference Posey, Clegg and Printz73–Reference Ozcan, Ergin and Lu76). As such, animals receiving intracerebroventricular C6 ceramide administration displayed a decrease in insulin sensitivity and developed obesity secondary to decreased energy expenditure and brown adipose tissue thermogenesis(Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51). Remarkably, these effects were countered by inhibition of hypothalamic ER stress which, not only normalised body weight, but also reversed central ceramide-induced peripheral insulin resistance(Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51), supporting the notion that ER stress represents an effector of lipotoxicity. These effects were recapitulated in obese Zucker rats which are generally characterised by hypothalamic ER stress and ceramide accumulation. In these animals, inhibition of ER stress was sufficient to improve their metabolic phenotype marked by an improvement in insulin as well as leptin resistance, reduced body weight gain and hepatic steatosis(Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51). In consideration of this, besides their role in insulin resistance, ceramides may also promote obesity. This was confirmed by the protective effect of ceramide synthase 6 knock out against high-fat diet-induced obesity(Reference Turpin, Nicholls and Willmes61). Always in support of the metabolically detrimental role of ceramide accumulation in the hypothalamus, inhibition of its synthesis within the hypothalamus resulted in an improvement in insulin sensitivity, an effect that was also confirmed on cultured hypothalamic neurons(Reference Campana, Bellini and Rouch77). Additionally, the inhibition of ceramide synthesis within the hypothalamus improved glycaemic control by enhancing glucose-stimulated insulin secretion and pancreatic β-cell mass(Reference Campana, Bellini and Rouch77). Thus, ceramides appear to impair metabolic health and particularly insulin sensitivity by accumulating both in peripheral tissues and in the hypothalamus, with these effects being presumably dependent on the activation of protein kinase Cζ, protein phosphatase 2A, metabolic inflammation and ER stress(Reference Chavez and Summers36,Reference Chaurasia, Talbot and Summers40,Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51,Reference Han and Kaufman78) . Particularly, ER stress may be at the interface between lipotoxicity and metabolic dysfunction. However, ER stress may mediate the deleterious effect of ceramide in concert with inflammation given the cross-talk between these pathophysiological responses(Reference Ajoolabady, Lebeaupin and Wu79) and the ability of ceramides to trigger both(Reference Sergi, Morris and Kahn49,Reference Contreras, Gonzalez-Garcia and Martinez-Sanchez51) . Nevertheless, ER stress may not be responsible for the detrimental effects of ceramide on insulin resistance in in vitro skeletal muscle models. This possibility is supported by the fact that the inhibition of ER stress is not sufficient to rescue palmitic acid-induced impairment in insulin signalling in C2C12 and human myotubes(Reference Rieusset, Chauvin and Durand80,Reference Hage Hassan, Hainault and Vilquin81) .
As described thus far, ceramides represent a particularly strong candidate in linking obesity, dysfunctional lipid metabolism and insulin resistance. In addition, the inhibition of ceramide synthesis has been reported to exert beneficial effects on a full spectrum of obesity and insulin resistance-associated cardiometabolic aberrations, including metabolic dysfunction-associated steatotic liver disease, cardiomyopathy, atherosclerosis and hypertension(Reference Chaurasia, Tippetts and Mayoral Monibas39,Reference Holland, Brozinick and Wang57,Reference Glaros, Kim and Quinn82–Reference Kasumov, Li and Li84) . In light of this, modulating lipid metabolism to counter ceramide synthesis represents a promising therapeutic strategy to improve cardiometabolic health.
Ceramides as a biomarker of insulin resistance and progression to type 2 diabetes mellitus
The deleterious effects of ceramides on skeletal muscle insulin signalling stem beyond their intracellular synthesis and may be dictated by plasma ceramides which are primarily derived from the liver in both rodents and humans(Reference Lightle, Tosheva and Lee85) (Figure 2). These hepatic-synthetised ceramides travel in the blood stream packed in LDL and VLDL(Reference Wiesner, Leidl and Boettcher86) (Figure 2) and have been reported to target the skeletal muscle to elicit insulin resistance(Reference Boon, Hoy and Stark50). Indeed, infusion of LDL containing C24:0 ceramide in lean mice induced an impairment in insulin signalling via AKT, thereby hampering insulin-mediated glucose uptake in skeletal muscle(Reference Boon, Hoy and Stark50). The ability of LDL-derived ceramides to elicit insulin resistance may also hold true in humans as the transport of ceramide in LDL is increased in obese and type 2 diabetic individuals and correlated with insulin resistance(Reference Boon, Hoy and Stark50). This notion is also supported by the fact that insulin resistance, assessed by homoeostasis model assessment: insulin resistance, is positively associated with saturated ceramides(Reference Neeland, Singh and McGuire87).Thus, an intracellular mismatch between fatty acid supply and oxidation may not be the only driver of ceramide accumulation, but these lipotoxic lipid species may also act in an endocrine fashion to target extra hepatic tissues as demonstrated for skeletal muscle(Reference Boon, Hoy and Stark50). In consideration of the ability of circulating ceramides to trigger insulin resistance and the positive correlation between circulating ceramides and the onset of insulin resistance(Reference Boon, Hoy and Stark50,Reference Haus, Kashyap and Kasumov88) , these sphingolipids are emerging as a predictive marker of insulin resistance and progression to T2DM. Additionally, total dihydroceramides were reported to be elevated up to 9 years prior to the diagnosis T2DM(Reference Wigger, Cruciani-Guglielmacci and Nicolas89) as well as in obese and type 2 diabetic individuals(Reference Zarini, Brozinick and Zemski Berry90) and are associated with the severity of metabolic dysfunction-associated steatotic liver disease(Reference Carlier, Phan and Szpigel91), further supporting ceramide and the metabolites in its biosynthetic pathways as potential novel prognostic biomarkers of T2DM. Furthermore, not only serum dihydroceramides correlate with insulin resistance, but they are also able to impair insulin signalling in primary myotubes(Reference Zarini, Brozinick and Zemski Berry90). However, the ability of dihydroceramides to directly promote insulin resistance remains controversial(Reference Chaurasia, Tippetts and Mayoral Monibas39). This further confirms the implication of ceramide metabolism, particularly its anabolism, as a pivotal driver of insulin resistance and impaired glycaemic control, suggesting that defects in ceramide homeostasis may prevent the onset of overt T2DM.
Ceramides link fatty acid oversupply to insulin resistance
Energy overload is a sine qua non condition to shift ceramide synthesis beyond cellular physiological requirements as a precursor of complex sphingolipids(Reference Gault, Obeid and Hannun43) and promotes its metabolically detrimental accumulation(Reference Summers, Chaurasia and Holland92). Particularly, excess fatty acid supply represents a pivotal trigger for ceramide synthesis which takes place as a compensatory mechanism to buffer the excess fatty acid supply and prevent their accumulation and disruption of cellular the bilayer structure(Reference Summers, Chaurasia and Holland92). Fatty acids, upon entry into the cells, are activated via the esterification with a CoA molecule to generate acyl-CoAs whose metabolic fate is dictated by the cellular energy status. In the condition of negative energy balance, acyl-CoAs are shuttled into the mitochondria via the carnitine system to be β-oxidised to generate energy and heat. When energy needs are met, fatty acids are diverged towards the synthesis of triacylglycerols, the main form of energy storage in the body along with glycogen, and glycerophospholipids which are key components of cellular lipid bilayers. However, when fatty acid supply overcomes intracellular storage capacity, fatty acid excess is funnelled towards the synthesis of ceramides(Reference Summers, Chaurasia and Holland92,Reference Summers93) .
The effect of ceramides on lipid metabolism
Ceramide accumulation intracellularly is able to modulate lipid metabolism(Reference Summers, Chaurasia and Holland92,Reference Summers93) . In support of this, ceramides can rewire lipid metabolism to increase fatty acid uptake and promote their storage in the form of triglycerides. In line with this, lowering ceramide levels in the liver or adipose tissue by overexpressing ceramidases in animal models resulted in a down-regulation in CD36(Reference Xia, Holland and Kusminski94), suggesting ceramide being able to increase fatty acid uptake. Similarly, knocking out dihydroceramide desaturase 1, thereby lowering ceramide content in the liver, was sufficient to dampen hepatic lipid intake(Reference Chaurasia, Tippetts and Mayoral Monibas39). This effect appears to be dependent upon ceramide-induced protein kinase Cζ activation as the overexpression of a dominant-negative form of this enzyme was sufficient to abrogate the effects of ceramides on fatty acid uptake(Reference Xia, Holland and Kusminski94). Apart from increasing CD36-mediated fatty acid uptake, ceramides also increase fatty acid esterification via the up-regulation of sterol response element binding proteins and the consequent indication of genes involved in triacylglycerols synthesis(Reference Jiang, Xie and Li95). Additionally, ceramides decrease the uptake of glucose(Reference Wang, O’Brien and Brindley96) and amino acids(Reference Guenther, Peralta and Rosales97,Reference Hyde, Hajduch and Powell98) , possibly to favour the catabolism of fatty acid as the main energy substrate, even though it is also true that C16:0 ceramide impairs mitochondrial function(Reference Raichur, Wang and Chan70) and may therefore negatively impact fatty acid catabolism. In line with this, ceramides, and particularly C16:0 ceramide accumulation, have been shown to impair mitochondrial respiration and the activity of electron transport chains II(Reference Raichur, Wang and Chan70) and IV(Reference Zigdon, Kogot-Levin and Park99) in vitro, leading to a decrease in β-oxidation as reported in cultured hepatocytes(Reference Raichur, Wang and Chan70). Always in agreement with their ability to affect lipid metabolism, ceramides inhibit isoproterenol-induced phosphorylation of adipocyte hormone sensitive lipase(Reference Chaurasia, Tippetts and Mayoral Monibas39) with this effect being a potential mechanism by which these sphingolipids may counter further oversupply of NEFA to metabolically active tissues. Nevertheless, despite this appearing a protective mechanism to prevent fatty acid oversupply to liver and skeletal muscle for example, ceramide accumulation within adipocytes may contribute to increase adipocyte size and impair mitochondrial respiration(Reference Chaurasia, Kaddai and Lancaster100). In agreement with this, genetic approaches aimed at lowering adipocyte ceramide levels led to a decrease in adipocyte size, increased mitochondrial function and enhanced insulin sensitivity(Reference Chaurasia, Tippetts and Mayoral Monibas39,Reference Chaurasia, Kaddai and Lancaster100) . Thus, on one hand ceramides increase fatty acid uptake and intracellular storage and influence adipocyte lipid metabolism possibly to buffer fatty acid excess and the potential perturbation of cell membranes by the detergent-like effects of NEFA; however, on the other hand these mechanisms contribute to increasing intracellular lipid storage and adipocyte hypometabolism, thereby paving the way for the development of insulin resistance(Reference Summers93).
Long- versus medium-chain saturated fatty acids as the dietary drivers of ceramide synthesis
As already discussed, fatty acid overload is the main driver of ceramide synthesis. However, not all fatty acids promote ceramide synthesis, a paradigm which is in line with the fact that not all fatty acids are metabolically detrimental(Reference Luukkonen, Sadevirta and Zhou101,Reference Wanders, Blom and Zock102) . Indeed, long-chain saturated fatty acids, as opposed to mono- and poly-unsaturated fatty acids, increase circulating ceramide levels(Reference Luukkonen, Sadevirta and Zhou101,Reference Rosqvist, Kullberg and Stahlman103,Reference Tuccinardi, Di Mauro and Lattanzi104) in human subject feeding trials, an effect that was accompanied by insulin resistance(Reference Luukkonen, Sadevirta and Zhou101,Reference Tuccinardi, Di Mauro and Lattanzi104) (Figure 3a). The accumulation of ceramides was also demonstrated in the skeletal muscle and liver of rats upon the infusion of lard oil, mainly composed of long-chain saturated fatty acids, and was paralleled by insulin resistance(Reference Holland, Brozinick and Wang57). The ability of long-chain saturated fatty acids to promote ceramide buildup may be dependent on an increase in serine palmitoyltransferase and a concomitant decrease in ceramidase activity. The induction of serine palmitoyltransferase activity by long-chain saturated fatty acids is not surprising given that palmitoyl-CoA, the CoA derivate of palmitic acid, along with serine, is a substrate for this enzyme(Reference Blachnio-Zabielska, Baranowski and Zabielski105). Long-chain saturated fatty acids are also able to trigger inflammatory response in a variety of tissues(Reference Sergi, Luscombe-Marsh and Heilbronn106–Reference Williams, Campbell and Drew111) and inflammation itself has been proposed as one of the triggers of ceramide synthesis, thus representing a further mechanism linking saturated fatty acids and ceramide buildup(Reference Chaurasia, Talbot and Summers40) (Figure 3a). Indeed, the chronic low-grade inflammation typical of obesity not only represents a key pathophysiological mechanism underpinning the cardio-metabolic complications linked with obesity, but also promotes the synthesis of ceramide(Reference Chaurasia, Talbot and Summers40). This may be at the basis of a vicious cycle in which metabolic inflammation promotes the synthesis of ceramide and ceramide itself fuels inflammation(Reference Chaurasia, Talbot and Summers40), with both stimuli hampering insulin signal transduction pathway(Reference Chavez and Summers36,Reference Hotamisligil112,Reference Hotamisligil113) . In support of the nexus between inflammation and ceramide synthesis, in cell cultures cytokines stimulate the synthesis of ceramide by promoting the expression of genes involved in de novo ceramide synthesis and sphingomyelin breakdown(Reference Xu, Yeh and Chen114–Reference Kolesnick, Haimovitz-Friedman and Fuks116). This effect has been further confirmed in vivo as demonstrated by the ability of TNFα to increase ceramide levels in epithelial cells via the activation of neutral sphingomyelinase(Reference Mallampalli, Peterson and Carter117). Always in line with the ability of inflammation to modulate ceramide metabolism, toll-like receptor 4 activation resulted in increased ceramide synthesis(Reference Schilling, Machkovech and He118) in primary macrophages, while its pharmacological inhibition led to a decrease in skeletal muscle ceramide levels(Reference McKenzie, Reidy and Nelson119). Particularly, the toll-like receptor 4 appears to mediate the synergistic effects of its cognate ligand, lipopolysaccharide and palmitic acid in mediating ceramide buildup(94). However, the mechanisms underpinning this synergy between lipopolysaccharide and palmitic acid remain to be fully elucidated, also in consideration of the fact that the ability of palmitic acid to act as a toll-like receptor 4 agonist remains controversial(Reference Lancaster, Langley and Berglund120–Reference Erridge and Samani122).
As already anticipated, ceramides also impact metabolic health by accumulating in the hypothalamus. In light of this, it is not surprising that diet may also impact hypothalamic ceramide accumulation. Indeed, a high-fat diet rich in long-chain saturated fatty acids promoted ceramide accumulation within the hypothalamus(Reference Borg, Omran and Weir72) with this effect appearing to be sex specific and more marked in male rodents(Reference Morselli, Fuente-Martin and Finan123). These in vivo data are also supported by in vitro studies in hypothalamic neuronal cell cultures confirming the ability of long-chain saturated fatty acids to promote ceramide accumulation not only in peripheral tissues, but also within the hypothalamus. Saturated fatty acids, and particularly palmitic acid, promoted C16:0 ceramide accumulation within neuronal hypothalamic cell lines with this effect being accompanied by inflammatory responses(Reference Sergi, Morris and Kahn121,Reference McFadden, Aja and Li124) and impaired insulin signalling(Reference Campana, Bellini and Rouch77). Remarkably, increasing fatty acid oxidation in hypothalamic neurons not only lowered ceramide levels with cultured hypothalamic neurons, but also mitigated inflammation(Reference McFadden, Aja and Li124). To a similar extend, specifically inhibiting ceramide synthesis using pharmacological or molecular tools was sufficient to rescue hypothalamic insulin signalling(Reference Campana, Bellini and Rouch77), confirming the role of ceramides as pivotal mediators of the metabolically detrimental role of long-chain saturated fatty acids also with the hypothalamus.
Nevertheless, the effects of saturated fatty acids on ceramide accumulation appear to be chain-length specific as long- but not medium-chain saturated fatty acids are able to increase ceramide synthesis. Indeed, animals fed a high-fat diet, when switched to a high-fat diet supplemented with medium-chain triacylglycerols (C8 and C10), not only experienced an improvement in metabolic health independently of changes in adiposity but also displayed a decrease in hepatic ceramide content(Reference Mourad, Abdualkader and Li125) (Figure 3b). This effect may be dependent upon the ability of medium-chain triglycerides to down-regulate the expression of key enzymes involved in ceramide biosynthesis and sphingomyelin hydrolysis, namely ceramide synthase 6 and sphingomyelin phosphodiesterase 3, respectively(Reference Mourad, Abdualkader and Li125). Another potential mechanism underpinning these effects may be dependent on the ability of medium-chain saturated fatty acids to modulate mitochondrial oxidative metabolism and the fact that, compared with long-chain saturated fatty acids, they are more effectively β-oxidised(Reference DeLany, Windhauser and Champagne126). With regard to their impact on mitochondrial oxidative metabolism, lauric acid, a medium-chain saturated fatty acids did not impair mitochondrial membrane potential or promote mitochondrial fission in the primary myotubes of human subjects as opposed to the long-chain saturated fatty acid palmitic acid(Reference Sergi, Luscombe-Marsh and Naumovski107). Furthermore, in rodents, a high-fat diet enriched in medium, compared with long-chain fatty acids, induced a greater increase in skeletal muscle markers of mitochondrial metabolism with this effect being associated with an improvement in insulin sensitivity in skeletal muscle and adipose tissue(Reference Turner, Hariharan and TidAng127). Thus, it is tempting to hypothesise that the preserved mitochondrial function in response to medium-chain saturated fatty acids would favour mitochondrial oxidative metabolism, thereby preventing intracellular ceramide buildup. However, this effect was tissue-specific as medium-chain fatty acid supplementation led to hepatic steatosis and liver-selective insulin resistance in rodents while preserving skeletal muscle and adipose tissue insulin sensitivity(Reference Turner, Hariharan and TidAng127). These effects may also be related to the divergent effect of palmitic and lauric acid on metabolic inflammation, with the former, but not the latter, being able to trigger the activation of the pro-inflammatory NFκB signalling in primary myotubes of human subjects(Reference Sergi, Luscombe-Marsh and Naumovski107) (Figure 3), which, in turn, has been reported to play a key role in disrupting mitochondrial function in myotubes(Reference Nisr, Shah and Ganley128). Always in line with the effect of medium-chain fatty acid on inflammation, they have been shown to inhibit the pro-inflammatory effect of lipotoxicity by activating the G-protein-coupled receptor 84, as demonstrated on hepatic macrophages(Reference Ohue-Kitano, Nonaka and Nishida129). Finally, medium chain fatty acids are also able to act as signalling molecules given by their ability to increase intracellular AMP(Reference Takikawa, Kumagai and Hirata130) while lowering ATP levels(Reference Debeer, Mannaerts and De Schepper131) (Figure 3b). The drop in ATP/AMP ratio, in turn, activates AMP-activated protein kinase which is pivotal in switching off anabolic pathways, including ceramide biosynthesis(Reference Erickson, Smith and Anthonymuthu132), while stimulating ATP-generating pathways such as β-oxidation(Reference Herms, Bosch and Reddy133). Thus, the activation of intracellular signalling pathways that boost oxidative metabolism while inhibiting anabolic pathways represents a further mechanisms by which medium-chain fatty acids counter ceramide accumulation.
Conclusions
Ceramides are at the nexus between the derangements in lipid metabolism underlaying metabolically unhealthy obesity and insulin resistance. Indeed, the intracellular accumulation of ceramides is a pivotal driver of insulin resistance and arises as a direct consequence of the mismatch between fatty acids supply, both from the adipose tissue and the diet, and their catabolism via mitochondrial β-oxidation. Remarkably, ceramides not only represent a pathogenetic factor able to impair insulin signal transduction, but the raise in circulating levels of total and specific dihydroceramide species such as Cer(d18:0/22:0). Cer(d18:0/24:0) may also predict the onset of overt T2DM. In keeping with this, the levels of these dihydroceramide species were reported to be up-regulate up to 9 years before the diagnosis of T2DM, suggesting that the defect in lipid metabolism that derange ceramide homoeostasis precede the onset of the disease. The overconsumption of dietary saturated fatty acids has been widely reported to hamper insulin sensitivity with this effect being potentially mediated, at least in part, by their ability to foster ceramide synthesis. However, this effect is chain length-specific as long-chain saturated fatty acids, and particularly palmitic acid, foster ceramide accumulation by providing the building blocks for its synthesis and by triggering inflammatory responses. On the contrary, medium-chain saturated fatty acids not only have not been associated with an increase in ceramide synthesis, but some reports also suggest their ability to counter lipotoxicity.
Financial support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors declare none.
Authorship
D.S. conceived and wrote the first draft of the review. E.Z., C.C., M.P. and G.Z. contributed to writing, editing and critically reviewed the manuscript. All authors have read and approved the final version of the manuscript.