Role for Stearoyl-CoA Desaturase-1 in Leptin-Mediated Weight Loss Paul Cohen, Makoto Miyazaki, Nicholas D. Socci, Aaron Hagge-Greenberg, Wolfgang Liedtke, Alexander A. Soukas, Ratnendra Sharma, Lisa C. Hudgins, James M. Ntambi, and Jeffrey M. Friedman

Time course. The time-course experiment was performed as in (S1). Liver RNA was isolated from groups of ob/ob mice treated with subcutaneous leptin (infusion rate of 4.8 g/24 hours via an Alzet miniosmotic pump, Alza Co., Palo Alto, CA) for 2, 4, and 12 days, and from untreated ob/ob mice and ob/ob mice that received saline. As an additional control, RNA was isolated from ob/ob mice that were pair-fed to the leptin-treated group also for 2, 4, and 12 days of treatment. The body weight and food intake of these groups of mice is shown in Fig. S1. As previously shown, leptin-treated ob/ob mice lose significantly more weight than pair-fed mice, indicating that leptin has metabolic actions that are not purely due to its effects on food intake (S2).

Microarrays. RNA from the livers of these animals was hybridized to Affymetrix murine 6500 gene chips according to established methods. Data were analyzed by using a K-means clustering algorithm. This computational method groups genes with similar patterns of expression over the time-course experiment. For inclusion in the cluster analysis, a gene had to meet preestablished selection criteria as in (S1) with some minor modifications.

Identifying and ranking leptin-regulated genes. In order to prioritize leptin-regulated genes for detailed functional analysis, we developed an algorithm to identify and rank genes that are specifically repressed by leptin. This algorithm ranked genes based on the extent to which their expression was (i) increased in ob/ob liver compared with wild type, (ii) repressed by leptin treatment, and (iii) different between leptin treatment and pair-feeding. Such genes are candidates contributing to the obese phenotype. Specifically, we selected genes whose expression was increased in ob/ob relative to wild-type and corrected by leptin administration. The selected genes were then scored by using the following two criteria: (i) the magnitude of the difference in expression between the untreated ob/ob sample and the day 12 leptin-treated sample and (ii) the magnitude of the distance between the leptin expression profile and the pair-fed expression profile. The above distance was defined to be the Euclidean distance between the normalized 3-dimensional vectors composed of the leptin-treated and pair-fed profiles on days 2, 4, and 12. To combine the two scores, the genes were ranked according to both parameters and then the pair of ranks was averaged for each gene. This final average rank served as the score for each gene. Genes with the lowest scores are those most strongly leptin-regulated. The results of this analysis are shown in Table S1.

SCD activity assays. These assays measure the conversion of [1-¹⁴C]stearoyl-CoA to [1-¹⁴C]oleate as in (S3). Tissues were homogenized in 10 volumes of buffer A (0.25 M sucrose, 1 mM EDTA, 10 mM Tris-HCl, 1 mM PMSF, pH 7.4). The microsomal fractions (100,00g) were isolated by sequential centrifugation. Reactions were performed at 37°C for 5 min with 100 g protein homogenate and 6 M of [1-¹⁴C]stearoyl-CoA (60,000 rpm), 2 mM NADH, 0.1 M Tris-HCl, pH 7.2. After the reaction, fatty acids were extracted and methylated with 10% acetic chloride/methanol. Saturated and monounsaturated fatty acid methyl esters were separated by 10% AgNO₃-impregnated thin-layer chromatography (TLC) using hexane:diethyl ether (9:1) as the developing solution. The plates were sprayed with 0.2% 2’,7’-dichloroflourescein in 95% ethanol, and the lipids were identified under UV light. Fractions were scraped off of the plate, and radioactivity was measured by using a liquid scintillation counter.

Lipid composition. We extracted total lipids from liver with Folch solvent (chloroform/methanol 2:1). At the start of the liver extraction, internal standards of C17:0 triglyceride, cholesteryl ester, and phosphatidylcholine were added to quantitfy each lipid class (Nu Check Prep, Elysian, MN and Sigma). The lipid classes in the liver extract were separated by TLC using silica gel G and a developing solvent of hexane/diethyl ether/acetic acid (60:40:1). The fatty acids in each lipid class were methylated with fresh 5% methanolic HCl and analyzed with a gas chromatograph (model 5890, Hewlett-Packard, Palo Alto, CA) equipped with a 100 m by 0.25 mm SP2560 fused silica capillary column (Supelco, Bellefonte, PA).

Mouse crosses. Asebia (ab^J/ab^J) mice were obtained from the Jackson Laboratory (Bar Harbor, ME). These mice have a spontaneous deletion of the first four exons of the SCD-1 gene and express no SCD-1 RNA or protein (S4, S5). SCD-1 is highly expressed in the sebaceous glands of the skin and the harderian and meibomian glands of the eye. Consequently, ab^J/ab^J mice have cutaneous and ocular defects. The phenotype of these animals is identical to that of mice with an induced mutation in SCD-1 (S6). Double-mutant ab^J/ab^J; ob/ob mice were generated by two successive crosses. First, ab^J/ab^J mice (background strain information is available at http://www.informatics.jax.org/external/festing/mouse/docs/ABJ.shtml) were crossed to ob/+ mice (C57BL/6 background). Then, progeny were intercrossed to generate the following groups of mice: (i) double-mutant mice (ab^J/ab^J; ob/ob), (ii) ob/ob controls (ab^J/+; ob/ob and +/+; ob/ob), (iii) asebia mice (ab^J/ab^J; ob/+ and ab^J/ab^J; +/+), and (iv) controls (ab^J/+; ob/+, ab^J/+; +/+, +/+; ob/+, and +/+; +/+). Genotypes were determined by genomic PCR and Southern blotting.

Carcass analysis. Carcasses were weighed and oven dried until weight was constant. Total body water was calculated as the difference between the weights before and after drying. The carcass was then homogenized in a blender and duplicate aliquots were extracted with a Soxhlet apparatus by using a 3:1 mixture of chloroform/methanol. The extracted homogenate was dried and weighed to calculate fat mass and lean mass.

Liver triglyceride quantitation. From each mouse, 40 to 100 mg of liver was homogenized in 4 ml of chloroform-methanol (2:1). A total of 0.8 ml of 50 mM NaCl was added to each sample. Samples were then centrifuged and a drop of Triton X-100 was added to duplicate 50-l aliquots from the lower phase. These aliquots were then evaporated under N₂ gas. Triglycerides were assayed by using an enzymatic reagent (Sigma).

VLDL production assay. Mice were fasted for 5 hours before injection with tyloxapol (Triton-1339, Sigma) as described in (S7). Mice were injected via the tail vein with 0.5 mg/kg tyloxapol dissolved in saline at a concentration of 0.15 g/ml. Tail bleeds were done just before injection (t₀) and 45 and 90 min following injection. Plasma triglycerides were assayed by using an enzymatic reagent (Sigma). The slope of the line denotes the rate of VLDL production.

Supplemental Figure 1. Leptin has novel metabolic effects. (A) Percent change in body weight and (B) daily food intake in ob/ob mice either freely fed and treated with saline (purple squares), pair fed and treated with saline (gold triangles), or leptin treated (green circles). Groups of animals were killed, and liver RNA was prepared at days 0, 2, 4, and 12. Error bars indicate the SEM, n = 4 for each group at each time point. *P < 0.05 pair fed versus leptin; for both pair fed and leptin, P < 0.05 versus freely fed at all time points.

Medium version | Full size version

Supplemental Figure 2. Leptin elicits a specific program of gene expression in the liver. K-means cluster analysis of genes from the time course. In the colorgram on the left, each row represents a gene and each column represents a group. The color denotes the expression level of each gene relative to wild type. The graphs on the right are another means of representing the data. In these curves, the expression profile of the average gene in each cluster is depicted relative to untreated lean littermate controls for which expression is given a value of zero. The expression levels in leptin-treated and pair-fed mice are shown in green and purple, respectively. We identified 15 clusters of genes with distinct patterns of expression, six of which correspond to genes that are specifically regulated by leptin relative to pair-feeding (Fig. S2). A complete list of the genes in each cluster is available at http://hal.rockefeller.edu/arrays/leptin/liver. These data confirm that leptin elicits a novel program of gene expression in liver. As the genes in each of these clusters are specifically regulated by leptin (as compared with pair-feeding), they could, in some cases, mediate some of the effects of the hormone.

Medium version | Full size version

Table 1. Leptin-dependent gene expression in ob/ob liver.

Medium version | Full size version

Table 2. Liver fatty acid composition.

Medium version | Full size version

References

S1. A. Soukas, P. Cohen, N. D. Socci, J. M. Friedman, Genes Dev. 14, 963 (2000).

S2. N. Levin, C. Nelson, A. Gurney, R. Vandlen, F. de Sauvage, Proc. Natl. Acad. Sci. U.S.A. 93, 1726 (1996).

S3. M. Miyazaki, Y.-C. Kim, M. P. Gray-Keller, A. D. Attie, J. M. Ntambi, J. Biol. Chem. 275, 30132 (2000).

S4. A. H. Gates, M. Karasek, Science 148, 1471 (1965).

S5. Y. Zheng et al., Nature Genet. 23, 268 (1999).

S6. M. Miyazaki, W. C. Man, J. M. Ntambi, J. Nutr. 131, 2260 (2001).

S7. M. Merkel et al., J. Clin. Invest. 102, 893 (1998).