Role for Stearoyl-CoA Desaturase-1 in Leptin-Mediated Weight Loss

See allHide authors and affiliations

Science  12 Jul 2002:
Vol. 297, Issue 5579, pp. 240-243
DOI: 10.1126/science.1071527


Leptin elicits a metabolic response that cannot be explained by its anorectic effects alone. To examine the mechanism underlying leptin's metabolic actions, we used transcription profiling to identify leptin-regulated genes in ob/ob liver. Leptin was found to specifically repress RNA levels and enzymatic activity of hepatic stearoyl–CoA desaturase-1 (SCD-1), which catalyzes the biosynthesis of monounsaturated fatty acids. Mice lacking SCD-1 were lean and hypermetabolic. ob/ob mice with mutations in SCD-1 were significantly less obese than ob/ob controls and had markedly increased energy expenditure. ob/ob mice with mutations in SCD-1 had histologically normal livers with significantly reduced triglyceride storage and VLDL (very low density lipoprotein) production. These findings suggest that down-regulation of SCD-1 is an important component of leptin's metabolic actions.

Leptin is an adipocyte-derived hormone that regulates energy balance, metabolism, and the neuroendocrine response to altered nutrition (1, 2). The metabolic program that leptin elicits is not explained by its effects on food intake alone (3, 4). Replacing leptin in leptin-deficient (ob/ob) mice and humans leads to the depletion of lipid in adipose tissue, liver, and other tissues (5–8). Leptin treatment also improves insulin sensitivity and reduces fat content in lipodystrophic mice and humans (9,10).

To elucidate the mechanism by which leptin reduces hepatic lipid content, we used microarrays to identify genes in liver that were differentially regulated by leptin or by food restriction (pair-feeding). Leptin-treated ob/ob mice lose significantly more weight than pair-fed mice, indicating that leptin stimulates energy expenditure [fig. S1 (11)]. Liver RNA from these animals was hybridized to microarrays, and the data were analyzed using a K-means clustering algorithm (12). We identified 15 clusters of genes with distinct patterns of expression, six of which correspond to genes specifically regulated by leptin, but not by pair-feeding (fig. S2). To prioritize leptin-regulated genes for functional analysis, we developed an algorithm to identify and rank genes that are specifically repressed by leptin (11). The gene encoding SCD-1 ranked the highest in this analysis (table S1).

The microsomal enzyme SCD-1 is required for the biosynthesis of the monounsaturated fats palmitoleate and oleate from saturated fatty acids (13, 14). SCD-1 RNA levels were highly elevated in untreated ob/ob liver (Fig. 1A). SCD-1 RNA levels in leptin-treatedob/ob mice were normalized at 2 days, and by 4 days, they fell to levels below that of lean controls, a result consistent with previous studies (15, 16). Pair-fed mice showed a smaller and delayed decrease in SCD-1 gene expression.

Figure 1

Leptin-specific down-regulation of SCD-1 RNA levels and enzymatic activity. (A) As shown from an independent time-course experiment, Northern blots of liver RNA samples were hybridized with radioactively labeled cDNA probes specific for SCD-1 and the small mitochondrial RNA pAL-15 (loading control). (B) SCD enzymatic activity was measured as in (11, 19). Activity is expressed as nanomoles minute−1 milligram−1 protein. Error bars indicate the SEM, n = 3 for each group. *P < 0.05 versus saline treated,# P < 0.05 versus pair fed, **P < 0.0005 untreated lean versus untreatedob/ob.

SCD enzymatic activity was elevated 700% in livers of untreatedob/ob mice relative to wild type (P < 0.0005) and remained significantly elevated in saline-treated, freely fed ob/ob controls (Fig. 1B). Leptin treatment normalized SCD enzymatic activity, whereas pair-feeding reduced enzymatic activity to a lesser extent (P < 0.005 leptin versus freely fed, P < 0.05 leptin versus pair-fed at each time point). Levels of hepatic monounsaturated 16:1 and 18:1 fatty acids, the products of SCD-1, were elevated inob/ob mice and normalized by 12 days of leptin treatment, but not by pair-feeding (table S2).

To investigate whether SCD-1 might mediate some of leptin's metabolic effects, we studied asebia mice (abJ/abJ ), which carry mutations in SCD-1 (17, 18). The weight of male abJ/abJ mice was indistinguishable from that of littermates, and femaleabJ/abJ mice weighed significantly more than littermates (Fig. 2A, P < 0.005 at all ages after 6 weeks). However,abJ/abJ mice had significantly reduced fat mass relative to controls (from 12.7 to 8.3% in females, P < 0.005; from 13.5 to 6.9% in males,P < 0.0005, Fig. 2B) and plasma leptin levels (females, 2.3 ± 0.4 ng/ml versus 5.0 ± 0.9 ng/ml; males, 2.7 ± 0.5 ng/ml versus 7.8 ± 1.4 ng/ml, P< 0.01 for both sexes).

Figure 2

Attenuation of the obese phenotype in mice with a mutation in SCD-1 (abJ/abJ ). (A) Weight curves of ob/ob (purple squares), abJ/abJ; ob/ob (green diamonds),abJ/abJ (gold circles), and control (magenta triangles) mice. Error bars indicate the SEM, n ≥ 9 for each group. *P < 0.01abJ/abJ versus control or P < 0.0001abJ/abJ; ob/obversus ob/ob. (B) Carcass analysis of female and male ob/ob (purple),abJ/abJ; ob/ob(green), abJ/abJ (gold), and control (magenta) mice was performed as in (11,26). Error bars indicate the SEM, n ≥ 8 for each group. *P < 0.005 versus lean control,# P < 0.05 versus ob/ob.

To explore the effects of SCD-1 deficiency on the ob/obphenotype, we intercrossed ob/+ andabJ/+ orabJ/abJ mice (11). abJ/abJ; ob/ob mice showed a dramatic reduction in body weight at all ages compared with ob/ob littermate controls (Fig. 2A,P < 0.0001 from 5 weeks of age). At 16 weeks of age, weight was reduced by 29% in females (P < 10−6) and 34% in males (P < 10−4). Fat mass in 16-week-old double-mutant females was 32.1% versus 51.0% in ob/ob (P < 10−4) and 28.1% in double-mutant males versus 49.9% inob/ob (P < 10−5) (Fig. 2B).abJ/abJ; ob/ob mice of both sexes also showed a significant increase in percent lean mass relative to ob/ob littermates (from 15.3 to 18.5% in females, P < 0.005; from 15.3 to 20.1% in males,P < 0.05), indicating that double mutants do not suffer from a generalized growth defect.

To analyze energy balance, we measured food intake and energy expenditure in ob/ob and lean littermates with and without homozygous SCD-1 mutations. ob/ob mice were hyperphagic compared with lean controls (Fig. 3, A and B). Despite being significantly leaner,abJ/abJ; ob/ob mice consumed 35% more food than ob/ob littermates (9.0 ± 0.9 g/day versus 6.6 ± 0.4 g/day, P < 0.05) (Fig. 3, A and B).abJ/abJ mice consumed 53% more than lean controls (6.3 ± 0.2 g/day versus 4.2 ± 0.2 g/day, P < 10−6) with an average food intake indistinguishable from that of ob/obmice (Fig. 3, A and B). Relative to ob/ob controls,abJ/abJ; ob/ob mice showed increases of 96% (female) and 56% (male) in total oxygen consumption (V O2) and increases of 174% (female) and 71% (male) in restingV O2 (Fig. 3, C and D;P < 0.03). Total and resting oxygen consumption in female double mutants were indistinguishable from those of lean controls, and in male double mutants, they were even greater than those of lean controls (P < 0.05 for restingV O2). LeanabJ/abJ mice expended more energy with increases of 30% and 46% in total oxygen consumption and increases of 43% and 62% in resting oxygen consumption for females and males, respectively (Fig. 3, C and D;P < 0.02). Double-mutantabJ/abJ; ob/ob mice had increased plasma levels of ketone bodies (β-hydroxybutyrate) relative to ob/ob littermates, suggesting increased fatty acid oxidation (females: 6.6 ± 1.2 mg/dl versus 3.4 ± 0.6 mg/dl; P < 0.05; males: 4.5 ± 1.4 mg/dl versus 3.5 ± 0.7 mg/dl; NS).

Figure 3

Increased energy expenditure, despite persistent hyperphagia, accounts for reduced adiposity inabJ/abJ; ob/ob andabJ/abJ mice. (A) Average daily food intake by group and (B) scatterplot of average daily food intake for individual mice in each group of ob/ob (purple),abJ/abJ; ob/ob(green), abJ/abJ (gold), and control (magenta) mice. Mice were acclimated in individual cages and 24-hour food consumption was measured for eight consecutive days. Food intake for each mouse was averaged over the 8 days, and these values were averaged for the mice in each group. Mice were 12 to 18 weeks of age. Food intake for males and females was indistinguishable and has been pooled. Error bars indicate SEM,n ≥ 4 for each group. *P < 0.05 versus lean control, # P < 0.05 versusob/ob. (C) Total and (D) resting oxygen consumption of female and male ob/ob(purple),abJ/abJ; ob/ob(green), abJ/abJ (gold), and control (magenta) mice. 14- to 16-week-old mice were placed in an Oxymax indirect calorimeter (Columbus Instruments, Columbus, OH) and allowed 2 hours to acclimate to the new environment. Measurements were taken for 5 hours during the middle of the light cycle. The totalV O2is the average of these readings. The restingV O2 is the average of all readings that are one standard deviation below the totalV O2, as these readings represent periods of inactivity. Error bars indicate SEM,n ≥ 3 for each group. *P < 0.05 versus lean control, *P < 0.05 versusob/ob.

ob/ob mice have massively enlarged livers that are engorged with lipid. Gross inspection revealed that the hepatomegaly and the steatosis of ob/ob mice were normalized inabJ/abJ; ob/ob mice. Histological sections of ob/ob liver showed large lipid-filled vacuoles, whereas those ofabJ/abJ; ob/ob mice showed little or no vacuolation and were indistinguishable from those of wild-type mice (Fig. 4A). The levels of liver triglyceride were substantially increased in ob/obmice (P < 0.005 versus all other groups), whereas triglycerides in abJ/abJ; ob/ob mice were reduced to levels comparable to lean controls (Fig. 4B). As previously shown, triglyceride levels in leanabJ/abJ mice were reduced below those of lean controls (Fig. 4B) (19).

Figure 4

Reduced hepatic lipid storage and VLDL production in abJ/abJ; ob/ob mice. (A) Hematoxylin and eosin (H&E)–stained liver sections from representative ob/oband abJ/abJ; ob/ob mice. Images are 200× magnifications; scale bars denote 100 μm. (B) Liver triglyceride content of ob/ob(purple),abJ/abJ; ob/ob(green), abJ/abJ (gold), and control (magenta) mice. Levels were determined as in (11, 26). Error bars indicate SEM, n ≥ 4 for each group. *P < 0.0005 versus ob/ob. (C) VLDL production in ob/ob (purple),abJ/abJ; ob/ob(green), abJ/abJ (gold), and control (magenta) mice. Mice fasted for 5 hours were injected with 0.5 mg/kg tyloxapol (Triton-1339, Sigma) via the tail vein as described in (11, 20). Tail bleeds were done at 0, 45, and 90 min and plasma triglycerides were assayed using an enzymatic reagent. The slope of the line indicates the rate of VLDL production. Error bars indicate SEM, n ≥ 3 for each group. For abJ/abJ; ob/ob and control, P < 0.05 versusob/ob. ForabJ/abJ ,P = 0.07 versus ob/ob.

Palmitoleate and oleate in triglycerides and cholesteryl esters are major constituents of very low density lipoprotein (VLDL) particles, which transport fatty acids from liver to adipose tissue and other sites. We assayed the rate of hepatic lipid export in mice injected with tyloxapol, an inhibitor of VLDL hydrolysis, permitting the measurement of VLDL production (20). VLDL production was increased 140% in ob/ob mice relative to lean controls (Fig. 4C, P < 0.05). Compared with ob/obcontrols, VLDL production was reduced 72% inabJ/abJ; ob/ob mice (8.2 mg/dl per min versus 2.3 mg/dl per min, P < 0.02) (Fig. 4C).

These data show that SCD-1 is required for the fully developed obese phenotype of ob/ob mice and suggest that a significant proportion of leptin's metabolic effects may result from inhibition of this enzyme. The basis for the metabolic effects of SCD-1 deficiency is not known. One possibility is that reduced activity of SCD-1 may decrease adiposity by decreasing cellular levels of malonyl CoA, thereby reducing fatty acid biosynthesis and de-repressing fatty acid oxidation. In the absence of SCD-1, a reduced rate of triglyceride and VLDL synthesis increases the intracellular pool of saturated fatty acyl CoAs leading to an increase in fatty acid oxidation. Monounsaturated fats are necessary for normal rates of triglyceride and cholesteryl ester synthesis, which are required for hepatic lipid storage and VLDL synthesis (19). Saturated fatty acyl CoAs, but not monounsaturated fatty acyl CoAs, potently allosterically inhibit acetyl-CoA carboxylase (ACC), reducing cellular levels of malonyl CoA (21, 22). Malonyl CoA is required for fatty acid biosynthesis and also inhibits the mitochondrial carnityl palmityl transferase shuttle system, the rate-limiting step in the import and oxidation of fatty acids in mitochondria (23). This putative mechanism is similar to that described in mice lacking acetyl-CoA carboxylase 2, which also have increased fatty acid oxidation in skeletal muscle and a lean phenotype (24). The mechanism by which leptin increases fatty acid oxidation in liver may be similar to that in skeletal muscle in that both may operate by reducing ACC activity. However, in skeletal muscle, leptin—acting directly and indirectly via the CNS—inhibits ACC via α2 AMP-kinase, which phosphorylates and inhibits ACC (25).

Alternative mechanisms could also account for the metabolic effects of SCD-1 deficiency. Changes in SCD-1 activity could alter the levels of ligands for peroxisome proliferator–activated receptors PPARα and PPARγ, or other nuclear hormone receptors. Changes in the ratio of saturated to unsaturated fatty acids in phospholipids can also alter membrane fluidity, which could affect signal transduction. The observed increase in energy expenditure associated with SCD-1 deficiency also suggests that uncoupling activity and/or futile cycles are induced.

The effects of leptin on SCD-1 in liver are likely to require central action, as mice lacking the leptin receptor in brain have enlarged, fatty livers, whereas livers of mice with a liver-specific knockout of the leptin receptor appear normal (26). Leptin also reduces hepatic SCD-1 activity when administered intracerebroventricularly (27). The CNS signals that modulate liver metabolism in response to leptin are unknown. SCD-1 deficiency also appears to modulate CNS pathways that regulate food intake, perhaps secondary to the increased oxygen consumption.

Leptin may also modulate the production of monounsaturated fatty acids in tissues other than liver. Down-regulation of SCD enzyme activity by leptin in other tissues, including brain, could contribute to some of the observed metabolic effects. Although SCD-1 is the only isoform normally expressed in liver, both SCD-1 and SCD-2, a similar enzyme, are expressed in most other tissues. Both enzymes catalyze the same reaction, but their cellular distribution and substrate preferences may be different.

In summary, a deficiency of SCD-1 ameliorates the obesity ofob/ob mice and completely corrects the hypometabolic phenotype of leptin deficiency. These findings suggest that leptin-specific down-regulation of SCD-1 is an important component of the novel metabolic response to leptin and suggests that inhibition of SCD-1 could be of benefit for the treatment of obesity, hepatic steatosis, and other metabolic disorders.

  • * To whom correspondence should be addressed. E-mail: friedj{at}


Stay Connected to Science

Navigate This Article