Combined deletion of SCD1 from adipose tissue and liver does not protect mice from obesity.

Stearoyl-CoA desaturase 1 (SCD1) catalyzes the synthesis of monounsaturated fatty acids (MUFA) from saturated FA. Mice with whole-body or skin-specific deletion of SCD1 are resistant to obesity. Here, we show that mice lacking SCD1 in adipose and/or liver are not protected from either genetic- (agouti; Ay/a) or diet-induced obesity (DIO) despite a robust reduction in SCD1 MUFA products in both subcutaneous and epididymal white adipose tissue. Adipose SCD1 deletion had no effect on glucose or insulin tolerance or on hepatic triglyceride (TG) accumulation. Interestingly, lack of SCD1 from liver lowered the MUFA levels of adipose tissue and vice versa, as reflected by the changes in FA composition. Simultaneous deletion of SCD1 from liver and adipose resulted in a synergistic lowering of tissue MUFA levels, especially in the Ay/a model in which glucose tolerance was also improved. Lastly, we found that liver and plasma TG show nearly identical genotype-dependent differences in FA composition, indicating that FA composition of plasma TG is predictive for hepatic SCD1 activity and TG FA composition. The current study suggests that SCD1 deletion from adipose and/or liver is insufficient to elicit protection from obesity, but it supports the existence of extensive lipid cross-talk between liver and adipose tissue.—Flowers, M. T., L. Ade, M. S. Strable, and J. M. Ntambi.

Stearoyl-CoA desaturase 1 (SCD1) catalyzes the desaturation of saturated fatty acids (FA), primarily palmitate (16:0) and stearate (18:0), into the cis -monounsaturated fatty acid (MUFA) products palmitoleate (16:1n7) and oleate (18:1n9), respectively. The metabolic role for SCD1 has been extensively explored in mice with naturally were fasted for 4 h and euthanized by an overdose of isofl urane anesthesia. The genotyping protocol for distinguishing the Scd1 fl ox and wild-type Scd1 alleles has been described previously ( 4,9 ). For breeding strategies involving only one Cre transgene, we used a generic Cre-recombinase genotyping strategy available at the Jackson Laboratories website (jaxmice.jax.org). For LAKO breeding schemes, we designed genotyping primers specifi c to either albumin-Cre or aP2-Cre. Albumin-Cre was amplifi ed using primer Alb-F (5 ′ GCA TGC AGG CAT TCA TCA 3 ′ ) and Cre-R (5 ′ GTG AAA CAG CAT TGC TGT CAC TT 3 ′ ) with a 54°C annealing temperature. AP2-Cre was amplifi ed using aP2-F (5 ′ ATG ATC TGG CCC CCA TTG G 3 ′ ) and Cre-R with a 51°C annealing temperature. All in vivo experimental procedures were performed in accordance with the National Institutes of Health Guidelines for the Care and Use of Animals and were approved by the Institutional Animal Care and Use Committee at the University of Wisconsin-Madison.

Immunoblot analysis of SCD1
Liver and white adipose microsomes were prepared by sequential centrifugation. Tissues were fi rst homogenized at 100 mg tissue/ml buffer in 0.1 M potassium phosphate buffer (pH 7.2) supplemented with 10 g/ml leupeptin and 1 mM PMSF. The homogenate was centrifuged at 10,000 g for 15 min at 4°C. The supernatant was subsequently centrifuged at 100,000 g for 1 h at 4°C. The pellet was rinsed once and resuspended in protease inhibitor free 0.1 M phosphate buffer. Brown adipose tissue was homogenized at 100 mg tissue/ml in lysis buffer containing 1 mM PMSF, 10 g/ml leupeptin, 5 g/ml pepstatin A, and 2 g/ml aprotinin. SCD1 immunoblotting was performed on 10 g of protein using a polyclonal SCD1 antibody (Santa Cruz Biotechnology; sc-14719).

Glucose and insulin tolerance tests
For glucose tolerance tests, mice were fasted for 4 h and subsequently injected intraperitoneally with 10% dextrose at a dose of 1 g/kg body weight. Blood was sampled by retroorbital puncture at 0, 20, 40, 90, and 180 min postinjection. For insulin tolerance tests, nonfasted mice were injected intraperitoneally with 0.75 U/kg body weight of human insulin (Novo Nordisk). Blood was collected by retroorbital puncture at 0, 15, 30, 45, and 60 min postinjection. Plasma glucose was analyzed using a colorimetric glucose oxidase method.

Tissue and plasma lipid analysis
Lipids were extracted from approximately 30 mg of tissue or 100 l of plasma lipids using a modifi cation of the Folch procedure ( 10 ). Samples were homogenized in 3 ml of CHCl 3 :MeOH (2:1) containing 10 mg/100 ml butylated hydroxytoluene. Heptadecanoic acid was added as an internal standard to correct for sample loss. Next, 1 ml of acidifi ed saline (0.01 N HCl, 0.9% NaCl) was added. After vigorous vortexing and centrifugation, the organic layer was isolated, dried under nitrogen gas, and subjected to TLC on silica gel-60 plates (EMD Chemicals) in heptane/isopropyl ether/glacial acetic acid (60/40/3, v/v/v). The bands corresponding to standards were scraped and transferred to screw cap glass tubes containing pentadecanoic acid as an internal standard to control for transmethylation effi ciency. FA were transmethylated in the presence of 14% boron trifl uoride in methanol. The resulting methyl esters were extracted with hexane and analyzed by gas liquid chromatography as previously described ( 11 ). Total lipid contents were calculated from individual FA content in each fraction.

Real-time quantitative PCR
Total RNA was extracted with TRI reagent (Molecular Research), treated with Turbo DNase (Ambion), and reverse transcribed resistance ( 6,7 ). This ASO therapy potently reduced SCD1 expression and activity in the liver and adipose tissue without affecting SCD1 expression in the muscle or skin. This suggests that inhibition of SCD1 in the liver and adipose tissue elicits obesity resistance via a mechanism that is physiologically distinct from that caused by skin SCD1 deletion. Because LKO mice are not protected from high-fat DIO ( 4 ), we hypothesized that deletion of adipose SCD1 (AKO mice) or combined deletion of liver and adipose SCD1 (LAKO mice) would elicit obesity resistance and other improved metabolic parameters previously observed in SCD1 GKO mice and those treated with ASO.
Adipose SCD1 was deleted from both white and brown adipose tissue by introducing the aP2-Cre allele into our SCD1 fl ox/fl ox mice. Surprisingly, AKO and LAKO mice were not protected from obesity despite robust reductions in adipose MUFA content. These results indicate that deletion or inhibition of SCD1 in the white and brown adipose tissue, in the presence or absence of liver SCD1 activity, has minimal effect on whole-body energy expenditure. Our study also reveals that adipose-derived FA affect the FA composition of liver and vice versa. Lastly, we report that the MUFA composition of plasma TG is strikingly similar to that of liver TG and can potentially be used as a predictor of hepatic SCD activity.
To generate conditional Scd1 -defi cient mice also containing the yellow agouti (A y /a) allele, we bred male LKO or AKO mice with female C57BL/6J A y /a mice (Jackson Labs stock #000021 Mice were maintained on a 12 h light/dark cycle with free access to food and water. Breeders were fed Purina 5015, and pups were weaned at 3 weeks of age and subsequently fed a cereal grain-based, low-fat standard diet (SD; Purina 5008). For DIO studies, male mice were individually caged at 8 weeks of age and fed either Purina 5008 or a high-fat diet (Research Diets RD12492; 60% kcal fat) for an additional 18 weeks. For A y /a-induced obesity studies, male mice were individually caged at 4 weeks of age and maintained on Purina 5008 until 26 weeks of age. Animals however, liver weight was reduced by loss of liver Scd1 in A y /a LKO and A y /a LAKO compared with Lox counterparts ( Fig. 2B ). Fasting glucose as well as glucose and insulin tolerance was not signifi cantly affected by adipose Scd1 deletion in a/a mice fed a standard or high-fat diet ( Fig.  3A -C ). However, liver SCD1 deletion in both A y /a LAKO and A y /a LKO improved glucose tolerance ( Fig. 3D ).

Hepatic SCD1 activity infl uences adipose FA composition
Liver and adipose tissue TG levels and their FA composition can be infl uenced directly by several sources. First, dietary FA, primarily in the form of TG-rich chylomicrons, can be taken up into tissues via lipoprotein lipase and lipoprotein receptor-mediated mechanisms. Second, cells can de novo synthesize FA from acetyl-CoA, and the rate of de novo synthesis in liver is especially responsive to dietary nutrients and hormones, such as glucose and insulin, respectively ( 13 ). And fi nally, there is signifi cant lipid crosstalk between the adipose and liver via the reciprocal exchange of adipose-derived free FA and hepatic-derived TG-rich VLDL. Importantly, intracellular saturated FA (palmitate and stearate), regardless of their origin, can be converted to MUFA via SCD1.
Liver SCD1 infl uenced adipose stores of oleate and the 18:1n9/18:0 ratio, but not stores of palmitoleate or 16:1n7/16:0 ratio, in A y /a mice but not in a/a mice ( Fig. 1 and supplementary Fig. I). This suggests that under conditions of high hepatic lipogenesis, liver-derived TG contribute a greater amount to adipose TG stores than during conditions of low hepatic lipogenesis. Additionally, the percentage composition of palmitate, palmitoleate, stearate, and oleate in adipose TG was signifi cantly affected by loss of adipose SCD1 in the subcutaneous fat, but not epididymal fat, of a/a mice. Upon introgression of the A y /a allele, both fat depots were signifi cantly infl uenced by loss of SCD1 from liver or adipose. These fat depotspecifi c changes suggest that the FA composition of fat depots spatially oriented closer to the liver, such as visceral or epididymal fat, is infl uenced to a greater extent by liverderived FA, especially under basal lipogenic conditions. Aside from 16-and 18-carbon saturated FA and MUFA, the other predominant FA found in adipose is linoleic acid (18:2n6), which comprises approximately 25% of adipose FA. In a/a mice, loss of adipose and/or liver SCD1 had no signifi cant effect on the percentage composition of 18:2n6 in epididymal or subcutaneous fat (data not shown). Loss of adipose SCD1 in A y /a AKO mice also had no effect on the abundance of 18:2n6 (supplementary Fig. II). Interestingly, loss of liver SCD1 in both A y /a LKO and A y /a LAKO mice caused elevated levels of 18:2n6 in the epididymal and subcutaneous adipose stores (supplementary Fig. II). To further explore the relationship between liver SCD1 and adipose 18:2n6 levels, we looked at 18:2n6 levels in liver TG and plasma TG. In a/a mice, 18:2n6 levels in liver were nonsignifi cantly elevated by approximately 12% in both LKO and LAKO mice compared with Lox and AKO ( ‫ف‬ 28.7% in LKO and LAKO compared with 25.5% in Lox and AKO), but plasma 18:2n6 levels were unchanged. However, both liver and plasma 18:2n6 using Multiscribe reverse transcriptase (Applied Biosystems). Real-time quantitative PCR was performed on an ABI StepOne Plus instrument using gene-specifi c primers and Power SYBR Green master mix (Applied Biosystems). Quantifi cation of given genes is expressed as mRNA level normalized to a housekeeping gene ( Arbp ) using the ⌬ ⌬ Ct method. Primer sequences are available upon request.

Data and statistical analysis
Data are expressed as mean ± SEM and were compared by oneor two-way ANOVA followed by Tukey's posthoc test in GraphPad Prism. Comparisons with a P < 0.05 were considered signifi cant.

RESULTS
To generate mice defi cient in adipose Scd1 (AKO), we introduced the aP2-Cre transgene into our Scd1 fl ox/fl ox (Lox) colony ( 4,5 ). We also generated mice with liverspecifi c deletion of Scd1 (LKO) that expressed the albumin-Cre transgene ( 4 ), and crossed these with the AKO mice to obtain mice with simultaneous deletion of Scd1 from both the liver and adipose (LAKO). AKO and LAKO mice were born at the expected frequency and were indistinguishable from Lox mice. Importantly, AKO and LAKO mice do not display the closed eye fi ssures, dry skin, and alopecia previously observed in global (GKO) and skinspecifi c (SKO) Scd1 -defi cient mice ( 5 ). Immunoblot analysis confi rmed the deletion of Scd1 from white and brown adipose tissue of AKO and LAKO mice, and from liver of LKO and LAKO mice ( Fig. 1A ). Furthermore, FA composition analysis of TG from both subcutaneous adipose tissue ( Fig. 1B-E ) and epididymal adipose tissue (supplementary Fig. I) refl ects a marked decrease in the abundance of the SCD products palmitoleate (16:1n7) and oleate (18:1n9) and an increase in the SCD substrates palmitate (16:0) and stearate (18:0) in AKO and LAKO mice.

Effect of adipose Scd1 deletion on obesity and glucose tolerance
To test whether adipose and/or liver deletion of Scd1 infl uences the development of obesity, we used both dietary (high-fat DIO) and genetically induced (A y /a allele) obesity models compared with a/a mice fed a standard diet. The high-fat DIO model utilizes a hypercaloric, semipurifi ed diet, which is highly abundant in the SCD product oleate. Despite the ample MUFA content of this high-fat diet, we have previously shown that both GKO and SKO mice, but not LKO mice, are remarkably resistant to obesity on this diet due to a hypermetabolic phenotype ( 4,5 ). A y /a mice have ubiquitous expression of the agouti protein, causing leptin resistance, increased food intake, and decreased energy expenditure ( 2,12 ). Thus, the A y /a model allows for obesity induction by feeding a standard diet. We found no signifi cant difference in body weight or in epididymal or subcutaneous white adipose weights due to adipose and/or liver SCD1 deletion in either the DIO or A y /a models ( Fig. 2A , C, D ). Additionally, food intake was unaffected ( Fig. 2E, F ). Liver weight was not significantly affected in a/a mice fed a standard or high-fat diet; signifi cantly affected by loss of hepatic SCD1 in both the A y /a LKO and A y /a LAKO mice. ( Fig. 4 and Fig. 5 ). Combined loss of liver and adipose SCD1 lowered hepatic TG, the 16:1n7/16:0 ratio, and the 18:1n9/18:0 ratio in high-fat fed LAKO mice as well (supplementary Fig. III). Deletion of liver and/or adipose SCD1 did not infl uence hepatic TG levels in the a/a mice and had a more modest effect on the FA composition. Thus, the hepatic rate of FA synthesis is an important determinant for the accumulation of SCD substrates in the absence of SCD1.
In A y /a mice, the hepatic 16:1n7/16:0 and 18:1n9/18:0 ratios were predominantly determined by liver SCD1. Analysis of the desaturation indices of hepatic lipids also revealed a partial role of adipose SCD1 in determining the FA composition of the liver. In both A y /a and a/a mice, the liver TG 16:1n7/16:0 ratio was independently and additively infl uenced by loss of liver and adipose SCD1 ( Fig. 4 ). Adipose deletion of SCD1 in A y /a AKO mice also reduced the hepatic 18:1n9/18:0 ratio. However, adiposespecifi c deletion of SCD1 did not protect mice from high-fat levels were increased in A y /a LKO and A y /a LAKO compared with A y /a Lox and A y /a AKO mice (supplementary Fig. II). Thus, lack of hepatic SCD1, when combined with the A y /a allele, leads to a dramatic decrease in 18:1n9 availability and a concomitant increase in 18:2n6 enrichment of these liver and plasma TG, which subsequently infl uence the FA composition of the adipose stores.

Effect of liver and adipose SCD1 on hepatic and TG composition
Although A y /a LKO and A y /a LAKO mice were not protected from obesity, their liver weights were signifi cantly lower than the A y /a Lox and A y /a AKO mice ( Fig. 2 ). We measured hepatic TG levels in both A y /a and a/a mice. Compared with their a/a counterparts, A y /a mice had signifi cantly elevated hepatic TG levels ( Fig. 4 ). However, this A y /a effect was severely blunted by loss of liver SCD1 in both A y /a LKO and A y /a LAKO mice. Consistent with this observation, the percentage composition of palmitate, palmitoleate, stearate, and oleate in hepatic TG were all of these particles in hepatocytes. Therefore, we compared the liver and plasma TG masses, desaturation indices, and FA compositions in both A y /a and a/a mice with liver and/or adipose deletion of SCD1 ( Figs. 4 and 5 ). Similar to the pattern observed for liver TG mass, plasma TG mass in a/a mice was not affected by loss of adipose and/or liver SCD1, but loss of liver SCD1 in A y /a LKO and A y /a LAKO led to a reduction of plasma TG levels. Strikingly, the DIO or A y /a-induced hepatic TG accumulation ( Fig. 4  and supplementary Fig. III).

Plasma TG FA composition parallels hepatic TG FA composition
In the fasted state, plasma TG are predominantly transported by liver-derived VLDL, which obtain their neutral lipid core during the presecretory intracellular assembly  expression and concomitant deletion of Scd1 in our studies are likely not restricted to adipocytes. The aP2-Cre transgene has also been detected during embryonic development ( 20 ) in macrophages and macrophage-rich tissues ( 21,22 ), bone marrow ( 21 ), cardiac and skeletal muscle, intestine, stomach, pancreas, and in the central and peripheral nervous system ( 23 ). However, the endogenous aP2 gene is predominantly expressed in adipocytes, and the aP2-Cre-mediated recombination in these other tissues is likely incomplete due to the lower expression level of Cre. Nonetheless, aP2-Cre-mediated deletion of Scd1 from adipocytes and potentially other cell types failed to recapitulate the obesity resistance observed in GKO mice.
We recently reported that skin-specifi c deletion of SCD1 elicited a hypermetabolic phenotype comparable to that observed in GKO mice ( 5 ). These mice resisted adipocyte hypertrophy and maintained insulin sensitivity when challenged with high-fat feeding, which was associated with increased uncoupling protein expression in a variety of peripheral tissues, such as white and brown adipose tissue, muscle, and liver. Thus, altered skin lipid metabolism can indirectly infl uence metabolic fl ux and cell signaling in peripheral tissues. However, these data do not necessarily exclude a role for SCD1 in other tissues, such as brown or white adipose tissue, because the severity of the cutaneous phenotypes in SKO mice may mask the contributions of SCD1 in these tissues to global energy homeostasis. Therefore, it is possible that an extracutaneous mechanism for obesity resistance also exists in the GKO mouse. This is supported by studies in mice treated with Scd1 -targeted ASO, which are reported to not affect the skin desaturation indices ( Fig. 4 ) and FA compositions ( Fig. 5 ) of liver and plasma TG showed the same genotype-dependent changes, suggesting that the FA composition of plasma TG is predictive of both liver TG FA composition and SCD1 activity.

DISCUSSION
Although mice with a global deletion of Scd1 (GKO) display remarkable obesity resistance ( 2-4 ) and insulin sensitivity ( 14 ), the importance of adipose SCD1 for whole-body energy metabolism is unclear. SCD1 is highly expressed in both white and brown adipose tissue. Previous studies in mice with a whole-body deletion of SCD1 have documented increased insulin signaling and uncoupling protein expression in the white and brown adipose tissue (15)(16)(17). Due to the simultaneous deletion of Scd1 from all tissues in GKO mice, it is unclear whether metabolic changes in a particular tissue are a direct effect of local loss of SCD1 function or an indirect effect stemming from loss of SCD1 in distal tissues.
Tissue-specifi c deletion using the Cre-lox system relies upon the use of a tissue-specifi c promoter to drive the expression of the Cre-recombinase. The FABP4(aP2)-Cre transgene is highly expressed in both white and brown adipose tissue ( 8,18 ). The differentiation of preadipocytes into mature adipocytes leads to the activation of several adipose-specifi c genes, including Fabp4 and Scd1 ( 19 ). Therefore, it is likely that the FABP4-Cre-mediated deletion of Scd1 occurs sometime during or immediately after adipocyte differentiation. It is noteworthy to acknowledge that aP2-Cre Different letters indicate signifi cant differences ( P < 0.05) between SCD1 genotypes within the a/a or A y /a group. * P < 0.05 for A y /a versus a/a, n = 5-7 per group.
of published Scd1 -targeted ASO ( 6, 7 ) occur due to the simultaneous inhibition of both Scd1 and Scd2 . This would predict that there are two SCD-mediated mechanisms to obesity resistance in mice: i ) deletion of Scd1 in skin and ii ) inhibition of Scd1 and Scd2 from the liver and adipose.
The failure of Scd1 deletion from the liver to protect against the A y /a-induced obesity is somewhat surprising in light of our previous observations of reduced adiposity in LKO mice fed a semipurifi ed, high-carbohydrate diet ( 4 ). Since the A y /a mice in our study were fed a cereal grainbased, high-carbohydrate, low-fat diet, we hypothesized that lack of liver SCD1 would reduce hepatic conversion of these dietary carbohydrates into TG, reduce liver VLDL TG secretion, and reduce overall body adiposity. Although we did observe reduced hepatic MUFA and TG content, this is presumably insuffi cient to affect overall energy homeostasis. The leptin-resistant phenotype of the A y /a mice results in both increased food intake and decreased energy expenditure. Thus, a hypermetabolic response that increases whole-body energy expenditure, such as that observed in GKO and SKO mice ( 5 ), may be necessary to combat the positive energy balance that exists in the A y /a and DIO obesity challenges.
Palmitoleate (16:1n7) released into circulation from the adipose tissue has been suggested to act as a benefi cial lipokine that improves metabolic derangements, including insulin resistance and hepatic steatosis ( 25 ). The combined adipose defi ciency of the aP2 (FABP4) and mal1 (FABP5) FA binding proteins caused elevated levels of adipose 16:1n7 ( 6,7 ). However, the failure of adipose and/or liver Scd1 deletion to prevent obesity in either the A y /a or DIO model strongly suggests that loss of SCD1 function in these tissues is not the mechanism for obesity resistance in the GKO mice or ASO-treated mice. Furthermore, we did not observe increased expression of thermogenic genes ( Ucp1, Ucp2, Adrb2, Dio2, Ppargc1a ) in the brown adipose tissue of AKO or LAKO mice (supplementary Fig. IV).
Intraperitoneally injected ASO resulted in the accumulation of ASO in several tissues. In addition to the reported effects of SCD1 ASOs on liver and white adipose and brown adipose, we speculate that the ASO-mediated obesity resistance is due to additional inhibition of Scd1 in cells or tissues not targeted by the albumin-Cre and aP2-Cre. In the case of the adipose tissue, the ASO treatment may be eliciting effects on both preadipocytes and mature adipocytes, unlike the aP2-Cre-mediated recombination that requires activation of the aP2 promoter during adipocyte differentiation. Inhibition of SCD1 during the preadipocyte stage may cause distinct metabolic changes compared with deletion of Scd1 in the late stages of or after differentiation. Alternatively, the ASO may be eliciting off-target effects on other genes contributing to obesity resistance, unlike the Cre-lox system, which is specifi c to Scd1 . Unlike the adult mouse liver, which expresses primarily Scd1 , both Scd1 and Scd2 are highly expressed in the mouse adipose tissue ( 24 ). We found that the expression of Scd2 in white and brown adipose tissue was unaffected by deletion of Scd1 via aP2-Cre (supplementary Fig.  IV). Therefore, it is also possible that the anti-obesity effects concomitant with improved metabolic phenotype ( 25 ); this predicts that reducing normal levels of adipose 16:1n7 by deleting SCD1 would worsen metabolism. However, deletion of adipose SCD1 in our study had no effect on body weight, glucose, or insulin tolerance in standard diet-fed mice or on two models of obesity (DIO and A y /a), despite a signifi cant lowering of adipose levels of 16:1n7. Furthermore, simultaneous deletion of liver and adipose SCD1 actually had a benefi cial effect of lowering liver TG levels and improving glucose tolerance compared with control mice. The FA ratios 16:1n7/16:0 and 18:1n9/18:0 in liver TG have been previously shown in human studies to correlate very well with the FA ratios found in plasma TG ( 26 ). In our current study, we found a remarkable similarity between the liver and plasma TG FA compositions that refl ected both A y /a-and SCD1-genotype effects. Although the liver SCD1 genotype comprised a majority of the effect on the liver and plasma 18:1n9/18:0 ratio, we found the 16:1n7/16:0 to be infl uenced by both the liver and adipose SCD1 genotype. Additionally, we found the adipose 18:1n9/18:0 ratio to be infl uenced by hepatic SCD1 in A y /a mice. Therefore, liver and adipose FA stores can reciprocally infl uence the composition of one another, presumably due to the exchange of liver-derived VLDL TG and adipose-derived free FA. We also observed that the magnitude of the hepatic SCD1-genotype effect on these hepatic FA ratios was dramatically infl uenced by the presence or absence of the A y /a allele. Thus, diet and genetic factors that predispose an individual to insulin resistance and increased hepatic lipogenesis can affect the apparent correlation of hepatic FA ratios with hepatic SCD1 levels.