Role of stearoyl-coenzyme A desaturase in regulating lipid metabolism

Multiple isoforms of mouse and human SCD

Four SCD isoforms (Scd1–Scd4) have been identified in the mouse, which reside within a 200 kb region of mouse chromosome 19 [6–10]. Scd1 and Scd2 expression is observed in a variety of lipogenic tissues, including liver and adipose tissue, Scd3 is primarily expressed in the skin, Harderian gland and preputial gland, and Scd4 is mainly expressed in the heart [5,6,9]. SCD3 is a palmitoyl-CoA desaturase with minimal desaturase activity towards stearoyl-CoA, while SCD1, SCD2, and SCD4 desaturate both palmitoyl-CoA and stearoyl-CoA to palmitoleoyl-CoA and oleoyl-CoA, respectively [11]. The physiological roles of Scd1 and Scd2 have been explored in mice with a targeted deletion of these genes [2,12]. Scd2 is critical for MUFA synthesis in liver and skin of neonatal mice, during which time Scd1 expression is very low [12]. Scd1, however, is the predominantisoforminmostlipogenictissuesinadultmice [2,5,12].One exception isthebrain,forwhich SCD2 contributes most δ-9 desaturase activity, and Scd1 deletion has minimal effect on brain MUFA content [2,12]. Interestingly, SCD2, but not SCD1, was recently shown to be required for adipogenesis in 3T3-L1 cells through maintaining adequate levels of peroxisome proliferator-activated receptor (PPAR)-γ [13^•]. Two SCD isoforms have been foundinhumans,SCD1andSCD5.WhereashumanSCD1 shares 85% amino acid identity with mouse Scd1–Scd4, human SCD5 shares limited identity to the mouse Scd genes and is unique to primates [14,15]. While both SCD1 and SCD5 are expressed in a variety of tissues in adult humans, SCD1 is highest in adipose and liver, but SCD5 is most abundant in brain and pancreas [14–16].

SCD1: a highly regulated gene but an unstable protein

Scd1 gene expression is altered by a remarkable number of nutrients (i.e. cholesterol, glucose, fructose, fatty acids, retinol), hormones (i.e. insulin, leptin, thyroid hormone) and environmental factors, and this regulation has been reviewed elsewhere [5,17,18]. Recent studies have provided additional insights into the molecular components controlling the transcription of Scd1. Scd1 expression is positively regulated by liver X receptor (LXR) binding to an LXR response element in the Scd1 promoter, as well as LXR-mediated activation of sterol regulatory-element binding protein-1c (SREBP-1c) transcription [19–21]. The effect of dietary carbohydrates, such as glucose, fructose and sucrose, to robustly increase hepatic Scd1 is due to both SREBP-1c-dependent and independent mechanisms [4,22]. These SREBP-1c-independent mechanisms could potentially be due to carbohydrate-induced activation of LXR or carbohydrate response-element binding protein (ChREBP) [23^•,24]. Intracerebroventricular glucose administration strongly decreases hepatic Scd1, suggesting that central nutrient sensing also regulates hepatic lipid metabolism [25^••]. This brain–liver connection is further exemplified by the ability of central administration of the melanocortin receptor agonist MTII or leptin to suppress hepatic Scd1 expression, with the effect of leptin established to be independent of insulin and SREBP-1c [26–28]. A novel link between the hepatic xenobiotic response and lipogenesis was discovered by the observation that activation of the pregnane-X-receptor (PXR) increases Scd1 expression via a SREBP-1c-independent mechanism [29^•,30]. Activation of the cellular immune response via toll-like receptor 2 also increases the transcription of Scd1, potentially via the numerous nuclear factor kB elements in the Scd1 promoter [31]. The binding of SREBP-1 to the SREBP response element of the Scd1 promoter is decreased by dietary n-3 and n-6 polyunsaturated fatty acids, partly due to repression of SREBP-1c maturation [21]. In contrast, dietary saturated fat strongly induces Scd1 expression, more so than dietary MUFA, in a mechanism that may involve fatty acid upregulation of PPARγ coactivator 1β and subsequent coactivation of SREBP-1c and LXR [32^•,33].

Cellular control of SCD activity also occurs at the level of the protein. The normal half life of SCD1 ranges between 2 and 4 h [34,35,36^•]. Scd1 transcription and the half life of the protein are increased in the presence of the PPARα-agonist clofibric acid [36^•,37]. Mouse SCD1 contains four transmembrane domains with the N and C-termini oriented towards the cytosol [38]. The N-terminus contains a sequence responsible for the rapid degradation of SCD1 in a mechanism that involves the ubiquitinproteasome-dependent system [34,35]. Additionally, the microsomal endopeptidase plasminogen-like protein selectively degrades SCD1 [39]. Human SCD1 and SCD5 have been shown to form dimers and oligomers, which may influence the function or stability of these enzymes [15]. We have reported that SCD1 interacts with DGAT2, potentially channeling newly synthesized MUFA into triglyceride synthesis [40]. The significance and regulation of these protein–protein interactions has yet to be determined.

The effect of SCD1 deficiency on lipid synthesis and oxidation

Mice with a targeted deletion of Scd1 or with spontaneous asebia mutations in Scd1 have provided several novel insights into the metabolic role of SCD1 [2,41]. scd1-deficient mice have reduced fatty acid and triglyceride synthesis in response to high-carbohydrate feeding [1,4,22]. This is partly due to decreased maturation of SREBP-1 protein and expression of SREBP-1 target genes such as fatty acid synthase (Fasn), acetyl-CoA carboxylase α (Acaca) and the fatty acid elongase Elovl6 [3,22,32^•]. Processing of SREBP-1c and lipogenic gene expression is restored by feeding scd1-deficient mice with high levels of dietary MUFA, but not saturated fat [22,32^•]. Thus, normal SREBP-1c activation requires adequate levels of MUFA derived from the diet or from endogenous synthesis. Interestingly, these effects are specific to SREBP-1 and do not affect SREBP-2 [22]. We have recently shown that liver-specific deletion of Scd1 using the Cre-loxP system is sufficient to block this carbohydrate-induced lipogenic program [42^••]. The mechanism for how cellular oleate homeostasis regulates the activation of SREBP-1c is currently unknown.

The lean phenotype of scd1-deficient mice also results from increased energy expenditure and oxygen consumption due to enhanced fatty acid oxidation and thermo-genesis in liver, muscle, and brown adipose tissue, through both transcriptional and posttranscriptional mechanisms [3,43–46]. We hypothesized that the elevated oxidative and thermogenic gene expression was due to increased transactivation of PPARα. PPARα target genes, however, remained increased by SCD1 deficiency in the absence of PPARα [43]. scd1-deficient mice also have elevated AMP-activated protein kinase (AMPK) activity in muscle and liver, which results in inhibitory phosphorylation of acetyl-CoA carboxylase, reduced malonyl-CoA levels and enhanced import of fatty acids into the mitochondria for oxidation [32^•,44,45,47]. The mechanisms responsible for increased AMPK activity and oxidative gene expression, however, remain elusive.

The role of SCD1 in lipoprotein metabolism

Elevated SCD1 activity is correlated with increased plasma triglyceride levels [48], but this is highly dependent upon both genetic and dietary variables. On a chow diet, asebia mice with a naturally occurring Scd1 mutation have severe depletion of plasma triglycerides [1,48]. This effect, however, is more modest in mice with a targeted deletion of Scd1 on a SV129 genetic background [19,22, 43,48] and absent in mice on a C57BL/6 [42^•,49^•] or BTBR [50^•] background. This may be due to genetic differences between C57BL/6 and SV129 mice, which have previously been shown to influence both hepatic Scd1 expression and development of the metabolic syndrome [51]. Scd1 deficiency more consistently leads to lower plasma triglyceride levels under conditions that promote increased hepatic triglyceride synthesis and accumulation, such as high-carbohydrate feeding [22,42^•,49^•], high-fat feeding [42^•,52^•] and LXR activation (Fig. 1) [19,52^•]. Whereas loss of SCD1 function in asebia leptin-deficient ob/ob mice decreased hepatic secretion and fasting levels of plasma triglycerides [53], targeted Scd1 deletion in BTBR ob/ob mice exacerbated the hypertriglyceridemia, potentially due to insufficient insulin secretion [50^••].

Scd1 deficiency also affects plasma levels of HDL and LDL cholesterol. A large increase in HDL cholesterol was observed in chow-fed asebia mice [1], but Scd1 deficiency in gene-targeted mice fed a chow diet has a minimal effect or lowers plasma HDL cholesterol levels [19,49^•,50^•]. scd1-deficient mice treated with the LXR agonist T0901317 have a large increase in HDL cholesterol [19,52^•]. ATP-binding cassette transporter A1 (ABCA1) is a key membrane protein responsible for HDL biogenesis and cellular lipid efflux to HDL [54^•]. Interestingly, LXR-treated scd1-deficient mice have increased expression of ABCA1 protein, but not Abca1 mRNA abundance [52^•]. This effect may be related to the observation that unsaturated fatty acids generated by SCD1 destabilize the ABCA1 protein and decrease cholesterol efflux to HDL (Fig. 1) [55,56].

In contrast to normal LDL cholesterol levels in chow-fed scd1-deficient mice, a very low-fat, high-sucrose diet elevated LDL cholesterol levels in C57BL/6 mice with a global Scd1 deletion due to a decreased rate of LDL clearance [49^••]. This diet also raised LDL cholesterol in mice with a liver-specific deletion of Scd1 [42^••]. Additionally, in leptin-deficient mice, global Scd1 deficiency [50^••] and short-term inhibition of liver Scd1 by RNA interference [57^•] increased plasma LDL cholesterol levels. Unlike humans, mice and rats are able to tolerate a large amount of dietary cholesterol without alterations in plasma LDL cholesterol levels due in part to upregulation of Cyp7a by LXR in rodents [58]. scd1-deficient mice fed a chow diet supplemented with 1% by weight cholesterol, however, develop increased LDL cholesterol levels, suggesting a critical role for SCD1 in LDL cholesterol homeostasis (M.T. Flowers, K. Chu, A.D. Attie and J.M. Ntambi, unpublished data).

In humans and several animal models LDL cholesterol levels have been found to be increased by dietary saturated fatty acids, predominantly through modulation of LDL production and LDL clearance [59–61]. While the mechanisms for this regulation are still unclear, it has been suggested to involve alterations of a regulatory pool of unesterified cholesterol [59] and abundance of the LDL receptor [62]. The relation of SCD1 to LDL cholesterol balance may involve similar mechanisms by decreasing the ratio of MUFAs to saturated fatty acids (Fig. 1). Studies in African green monkeys [63] and Ldlr-deficient, human apolipoprotein B100-overexpressing mice [64,65^•] suggest that LDL composition is also a critical factor in the development of cardiovascular diseases. Both dietary saturated fat and MUFA promote similar coronary artery atherosclerosis and accumulation of cholesteryl oleate in these models, whereas dietary polyunsaturated fat decreases these endpoints concomitant with repression of hepatic Scd1 [63,64,65^•]. It has yet to be determined whether pharmacological or genetic inhibition of Scd1 will reduce atherosclerosis, or whether the antiatherogenic properties of dietary polyunsaturated fats are independent of Scd1.

The conditionally essential requirement of SCD1 for survival

Most diets contain ample unsaturated fat in the form of MUFAs, oleate, or from various n-3 and n-6 fatty acids. Despite the presence of Scd2–Scd4, scd1-deficient mice have a severely blunted capacity for endogenous MUFA synthesis. When fed a very low-fat, high-sucrose diet, scd1-deficient mice develop a progressively severe phenotype characterized by loss of body weight, hypoglycemia, hypercholesterolemia and hepatic dysfunction [49^••]. These phenotypes were corrected by the addition of dietary unsaturated fat, but not saturated fat [49^••]. We also analyzed the hepatic gene expression profile in mice fed the very low-fat diet and found that scd1-deficient mice have a robust increase in the expression of several genes involved in the endoplasmic reticulum stress response and inflammation (M.T. Flowers, A.D. Attie and J.M. Ntambi, unpublished data). Thus, during dietary unsaturated fat insufficiency, SCD1 serves a protective role to allow for continued MUFA synthesis and to maintain cellular lipid homeostasis. Interestingly, hepatic Scd1 and MUFA synthesis are repressed by dietary polyunsaturated fatty acid intake, which may have evolved as a mechanism to maintain cellular un-saturated fatty acid balance. The conditionally essential nature of SCD1 may also explain why SCD1 inhibition reduces the survival of human tumor cells [66^•]. In further support of this notion, mouse and rat strains with a high genetic susceptibility to hepatocarcinogenesis also have high Scd1 expression, highlighting the importance of oleate for continued cell growth and survival [67].

Importance of SCD for skin biology and development

The skin expresses three isoforms of SCD, Scd1–Scd3 [6,12,41]. In adult mice, Scd1 is found in presebocytes and Scd3 in mature sebocytes [6]. Mice with a targeted deletion of Scd1 or naturally occurring asebia mutation have hypoplasia of the sebaceous and meibomian glands resulting in alopecia and closed eye fissure [2,41,68]. The hypoplasia of sebaceous glands in adult scd1-deficient mice results in loss of Scd3 expression in the skin [6]. Thus, it seems that SCD1 is most important for sebocyte maintenance, but the relative importance of the palmitoyl-CoA desaturase activity of SCD3 to sebocyte biology is currently not known.

Both Scd1 and Scd2-deficient mice are born at the expected Mendelian frequency, but unlike Scd1-deficient mice, most Scd2-deficient mice die within 24 h postbirth due to an epidermal permeability barrier defect resulting in severe dehydration [12]. Interestingly, the Scd2-deficient mice that survive into adulthood have normal sebaceous glands and hair growth, indicating that Scd1 and Scd2 have nonredundant roles in sebaceous gland function. The neonatal dependence on Scd2 for survival may be explained by the temporal expression pattern of Scd1 and Scd2. Scd2 was found to be the predominant hepatic SCD isoform in neonates, but hepatic Scd1 expression progressively increases in the 21 days postbirth to become the major hepatic SCD isoform [12].

Scd1-deficient mice have impaired thermoregulation and cold tolerance [46,69^••]. Both gene-targeted Scd1-deficient [69^•] and asebia 2J [68] mice, but not asebia J1 mice [68,70], have a dysfunctional epidermal lipid barrier resulting in increased transepidermal water loss. Interestingly, topical application of an artificial lipid barrier partially rescued the abnormal metabolic rate, water loss and heat loss observed in Scd1-deficient mice [69^•]. Both asebia J1 and asebia 2J mice also have reduced stratum corneum hydration due to reduced glycerol content [70]. The mechanism for the skin barrier defects in adult Scd1-deficient and neonatal Scd2-deficient mice may be related to reduced stratum corneal glycerol content, reduced sebaceous gland lipids, alterations in acylceramide fatty acid composition, or elevated levels of unesterified cholesterol [2,12,69^••,70]. Additionally, antimicrobial properties of skin MUFA may elicit protective effects against infection and inflammation [31].

SCD1 deficiency: friend or foe for diabetes

Insulin resistance and the progression to diabetes are intimately linked with lipid metabolism. Obesity and the accumulation of excess lipids in nonadipose tissues contribute to insulin resistance. Therefore, the decreased lipid synthesis and increased lipid oxidation observed in scd1-deficient mice predicts enhanced insulin sensitivity in these mice. Indeed, scd1-deficient mice have been shown to be protected from both high-fat diet and leptin deficiency-induced obesity and hepatic steatosis [3,50^••,53]. Hyperinsulinemic–euglycemic clamp studies found that Scd1 deficiency elevates whole-body glucose uptake and lowers hepatic glucose output [50^••]. Furthermore, scd1-deficient mice have enhanced insulin sensitivity and basal phosphorylation of insulin signaling components in various tissues such as muscle [50^••,71], white adipose tissue [50^••], brown adipose tissue [72], liver [50^••], and heart [50^••,73^•]. The increased glucose uptake observed in the hearts of scd1-deficient mice [50^•,73^•] occurs concomitant with reduced cardiac fatty acid uptake and oxidation [73^•]. The enhanced insulin signaling in scd1-deficient mice involves decreased expression of the protein-tyrosine phosphatase 1B (PTP1B), but the mechanism for altered PTP1B expression is not known (Fig. 2a) [71,72]. In addition to findings in mice, a recent study of a Swedish human population found SCD1 polymorphisms associated with reduced SCD activity, BMI and waist circumference, as well as improved insulin sensitivity [74^•]. Another report also correlated SCD activity within human adipose tissue with the development of insulin resistance [75^•].

In mice expressing leptin, Scd1 deficiency protects against diet-induced adiposity, hepatic steatosis and insulin resistance, despite an increase in food intake [3,42^••]. In contrast, Scd1 deficiency in leptin-deficient ob/ob mice decreases body weight and hepatic steatosis, but leads to worsened diabetes [50^••,53]. Furthermore, hepatic stea-tosis, but not hyperglycemia, was corrected by Scd1 deficiency in lipodystrophic aP2-nSREBP-1c mice, which are also leptin-deficient [28]. In both leptindeficient models, Scd1 deficiency reduced plasma insulin levels and elevated plasma glucose [28,50^•]. This suggests that some of the beneficial metabolic effects of Scd1 deficiency are dependent upon the presence of leptin. In ob/ob mice, Scd1 deficiency caused the appearance of an abnormal class of pancreatic islets with reduced insulin content, suggesting that β-cell failure is responsible for the diabetic phenotype [50^••]. Consistent with this model, an in-vitro study using the MIN6 β-cell line found that increased expression of Scd1 protects against lipoapoptosis due to the enhanced desaturation of cytotoxic palmitate [76].

The influence of muscle SCD1 expression on metabolism

Many recent studies have documented metabolic changes occurring in muscle as a result of modulation of muscle SCD1 expression. An in-vitro investigation of transiently transfected rat L6 myotubes found increased SCD1 activity to protect cells from fatty acid-induced insulin resistance as measured by 2-deoxyglucose uptake and Akt phosphorylation [77]. It is important to note that another study found L6 myotubes stably transfected with SCD1 to have lower 2-deoxyglucose uptake than cells expressing vector alone [78]. This difference, however, may be due to clonal differences or an effect of SCD1 overexpression on myotube differentiation, as evidenced by differences in GLUT4 (Slc2a4) expression [78]. Despite this effect on basal glucose uptake, cells stably overexpressing SCD1 were resistant to the palmitate-induced decrease in glucose uptake observed in control cells [78]. Acute and chronic endurance exercise in humans elevates the level of SCD1 in muscle [79^•] and, similar to in-vitro overexpression of SCD1 in fatty acid-treated myotubes, enhances myocellular triglyceride synthesis while decreasing ceramide and diacylglycerol accumulation (Fig. 2b) [77,79^•]. In contrast to this model, scd1-deficient mice also have decreased muscle ceramide synthesis due to decreased expression of serine palmitoyltransferase (Fig. 2a) [45]. The explanation for this paradoxical relationship between SCD1 and ceramide is currently not understood, but may involve additional hormonal influences elicited by whole-body SCD1 deficiency. Interestingly, increased muscle SCD1 expression and SCD activity are also observed in two seemingly dissimilar physiological states: obese rodents [77,78] and morbidly obese humans [80]; and semi-starved rats (Fig. 2b) [81]. In both cases, it has been suggested that elevated SCD1 partitions fatty acids towards storage instead of oxidation [80,81], a model that is supported by the observation of increased muscle β-oxidation in scd1-deficient mice (Fig. 2a) [45].

The impact of liver SCD1 deficiency on obesity and insulin resistance

Several strategies have recently been applied to help elucidate the tissue-specific role of SCD1 in mediating metabolic changes. Intraperitoneal injection of mice with scd1-targeted antisense oligonucleotides (ASO) inhibited SCD1 in liver and adipose without affecting the skin, resulting in improved insulin sensitivity and prevention of high-fat diet-induced obesity and hepatic steatosis [82]. These mice also showed increased metabolic rate, increased physical activity, and decreased hepatic fatty acid synthesis [82]. Liver-specific scd1-deficient mice (LKO mice) generated using Cre-lox technology, however, were not protected from high-fat diet-induced obesity and insulin resistance [42^•]. This suggests that resistance to high-fat diet-induced obesity requires SCD1 deficiency in extrahepatic tissues. LKO mice were, however, protected from high-carbohydrate-induced adiposity and hepatic steatosis, and failed to upregulate lipogenic gene expression in response to dietary carbohydrate [42^••]. This blunted lipogenic response was restored by dietary oleate, but not stearate, highlighting a critical role of hepatic SCD1 and its product MUFA in promoting the conversion of dietary carbohydrate into fatty acids [42^••]. Interestingly, LKO mice fed a very lowfat, high-sucrose diet developed hypoglycemia due to impaired gluconeogenesis and glycogen accumulation, and these phenotypes are restored by the addition of dietary oleate [42^••]. The mechanism for how oleate promotes lipogenesis and gluconeogenesis is currently being investigated.

The acute effect of SCD1 inhibition has also been investigated using ASO and RNA interference methods. Using the hyperinsulinemic–euglycemic clamp method, ASO-mediated inhibition of SCD1 was shown to prevent short-term (5-day) high-fat diet-induced hepatic insulin resistance in mice and rats, without affecting whole-body glucose uptake [83]. These effects were observed using both intraperitoneal ASO injection, which affects both liver and adipose, as well as intraportal ASO infusion, which is liver specific [83]. Short-term ASO-mediated SCD1 inhibition also caused enhanced hepatic insulin signaling, suppression of hepatic gluconeogenic gene expression (Pepck and G6pase) and accumulation of hepatic triglycerides [83]. These effects, however, were not observed in liver-specific LKO mice generated by the Cre-lox system [42^••]. Short-term (2-week) inhibition of SCD1 was also investigated in ob/ob mice treated with adenoviral shRNA targeting Scd1 [57^•]. Unlike ASO treatment [83], adenoviral shRNA treatment lowered hepatic neutral lipid content and elevated the expression of G6pase [57^•]. Adenoviral shRNA knockdown of Scd1 increased hepatic phosphatidylcholine levels [57^•], which has also been observed in mice with a whole-body deletion of Scd1 and attributed to higher CTP : choline cytidylyltransferase translocation into the membrane [84]. Dietary-induced suppression of hepatic Scd1 by feeding a lipogenic methionine-choline-deficient diet caused dramatic weight loss and decreased plasma triglyceride levels despite hepatic steatosis [85]. Interestingly, treatment of rats with scd1-targeted ASO or with intracerebroventricular glucose lowered hepatic triglyceride secretion coincident with suppression of Scd1 [25^••], but hepatic triglyceride secretion rates were normal in chowfed LKO mice [42^••]. These disparate outcomes indicate that the metabolic changes due to hepatic Scd1 inhibition may depend on species differences or the duration and method of Scd1 inhibition.