Absence of hormone-sensitive lipase inhibits obesity and adipogenesis in Lep ob/ob mice.

Hormone-sensitive lipase (HSL) plays a crucial role in the hydrolysis of triacylglycerol and cholesteryl ester in various tissues including adipose tissues. To explore the role of HSL in the metabolism of fat and carbohydrate, we have generated mice lacking both leptin and HSL (Lep(ob/ob)/HSL(-/-)) by cross-breeding HSL(-/-) mice with genetically obese Lep(ob/ob) mice. Unexpectedly, Lep(ob/ob)/HSL(-/-) mice ate less food, gained less weight, and had lower adiposity than Lep(ob/ob)/HSL(+/+) mice. Lep(ob/ob)/HSL(-/-) mice had massive accumulation of preadipocytes in white adipose tissues with increased expression of preadipocyte-specific genes (CAAT/enhancer-binding protein beta and adipose differentiation-related protein) and decreased expression of genes characteristic of mature adipocytes (CCAAT/enhancer-binding protein alpha, peroxisome proliferator activator receptor gamma, and adipocyte determination and differentiation factor 1/sterol regulatory element-binding protein-1). Consistent with the reduced food intake, hypothalamic expression of neuropeptide Y and agouti-related peptide was decreased. Since HSL is expressed in hypothalamus, we speculate that defective generation of free fatty acids in the hypothalamus due to the absence of HSL mediates the altered expression of these orexigenic neuropeptides. Thus, deficiency of both leptin and HSL has unmasked novel roles of HSL in adipogenesis as well as in feeding behavior.

Tokyo, Japan), and FFA (NEFA C; Wako Chemicals) were measured enzymatically. Plasma glucose was measured by ANTSENSE II (Bayer Medical, Tokyo, Japan); plasma insulin was measured by the mouse insulin ELISA kit (Morinaga, Tokyo, Japan); and plasma corticosterone was measured by the EIA kit (Diagnostic Systems Laboratories, Inc., Webster, TX).
Immunoblot Analysis-Hypothalamic regions were excised from 8-month-old male mice, homogenized in 0.25 M sucrose buffer (20 mM Tris-HCl, pH 7.4, 0.25 M sucrose, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 2.8 g/ml aprotinin), and centrifuged at 15,000 rpm for 15 min. Fifty g of the supernatant fraction was subjected to SDS/PAGE analysis. S100 protein (10 g) from BAT of wild type mouse was used as a positive control. Immunoblot analysis was performed using the ECL kit (Amersham Biosciences) as described previously (4).
Northern Blot Analysis-Total RNA was isolated by Trizol reagent (Invitrogen). Hypothalamic RNA was isolated as described previously (10). Five g of hypothalamic RNA from each mouse or 7.5-10 g of white adipose tissue (WAT) RNA equally pooled from 3-5 mice of each genotype were electrophoresed in a 1% agarose gel containing formaldehyde and transferred to a nylon membrane. The membranes were hybridized with probes, which were labeled with [␣-32 P]dCTP using the Megaprime DNA labeling system kit (Amersham Biosciences) in Rapidhyb Buffer (Amersham Biosciences) or ULTRAhyb hybridization buffer (Ambion, Austin, TX), and analyzed by a BAS2000 bioimaging analyzer (Fuji Photo Film, Tokyo, Japan). Loading was normalized by the expression of 36B4.
Histology-Sixteen-week-old mice were sacrificed by decapitation. Tissues were excised, fixed in 10% neutral buffered formalin, embedded in paraffin, and stained with hematoxylin-eosin. Cell size was estimated by NIH Image. After blocking with 0.5% goat serum, sections of epididymal fat pads were incubated with anti-mouse S-100 antibody (Sigma) overnight and were incubated with an anti-rabbit secondary antibody followed by staining with the avidin-biotin complex-alkaline phosphatase method and counterstaining with Harris's hematoxylin. Apoptotic cells were detected by terminal deoxynucleotidyltransferasemediated dUTP nick end-labeling using an in situ apoptosis detection kit (Takara Biomedicals Otsu, Japan) according to the manufacturer's instructions, with counterstaining with methyl green (Wako Chemicals). Proliferating cells were detected by bromodeoxyuridine immunostaining. Two h before sacrifice, bromodeoxyuridine (Sigma) was administered intraperitoneally at 50 mg/kg body weight in phosphatebuffered saline. After fixation with 10% neutral buffered formalin, mouse epididymal fat was embedded in paraffin. Immunostaining was performed by a bromodeoxyuridine In-Situ Detection Kit (PharMingen, San Diego, CA) and counterstained with Mayer's hematoxylin (Wako Chemicals).
Detection of DNA Ladder-Adipose tissue DNA was extracted as described previously (4). Five g of DNA end-labeled with [␣-32 P]dCTP by Klenow was subjected to electrophoresis in a 1.5% agarose gel and then transferred to nylon membranes, as described previously (11). As a control, mouse thymocyte apoptotic DNA was loaded. In brief, mouse thymus was ground over a mesh, and the isolated thymocytes were induced to undergo apoptosis by culturing in RPMI containing 10% calf serum and 1 M dexamethasone at 37°C for 6 h.
Food Intake and Body Temperature-Food intake and body weight were measured for 3 days after the mice were individually housed and adapted for 1 week. Feeding efficiency was defined as an increase in body weight divided by food intake during 3 days (n ϭ 8 -10). Body temperature was measured by an NK-YSI precision N550 thermometer and its probe (Nikkiso-YSI, Japan). These experiments were performed at the end of light cycle at the age of 14 weeks.
power, which is proportional to body surface area.
Statistics-Statistical differences between groups were analyzed by one-way analysis of variance and a post hoc Tukey-Kramer test, unless otherwise stated.

HSL Deficiency Ameliorates Obesity in Lep ob/ob
Background-In the Lep ϩ/ϩ background, there was no difference in body weight among HSL ϩ/ϩ , HSL ϩ/Ϫ , and HSL Ϫ/Ϫ mice on a normal chow diet (Fig. 1a). In the Lep ob/ob background, however, body weight of doubly homozygous mice was reduced by 26% compared with the Lep ob/ob /HSL Ϫ/Ϫ mice at the age of 16 weeks (Fig. 1a). The difference was discernible at the age of 11 weeks in males and 9 weeks in females. Furthermore, the age-dependent weight gain was negligible in Lep ob/ob /HSL Ϫ/Ϫ mice after 10 weeks of age.
HSL Ϫ/Ϫ Mice Have an Increased Number of Preadipocytelike Cells in WAT-In the Lep ϩ/ϩ background, WAT from HSL Ϫ/Ϫ mice contained clusters of small cells devoid of lipids in addition to lipid-filled mature adipocytes (Fig. 2a), as we reported previously (4). In the Lep ob/ob background, the small cells, which appeared slightly increased in Lep ob/ob /HSL ϩ/ϩ mice, were robustly increased in Lep ob/ob /HSL Ϫ/Ϫ mice (Fig. 2,  a and b).
To determine the identity of these small cells, we performed immunohistochemistry for S-100 protein, which is used as a marker of adipocytes (12), particularly of preadipocytes, and neural cells (Fig. 2c). The small cells were strongly positive for S-100, indicating that they were preadipocytes. Upon Northern blot analyses of WAT (Fig. 3a), the expression of adipose differentiation-related protein (13) and CAAT/enhancer-binding protein ␤, which is dominant during the early stage of differentiation of adipocytes (14), was increased in HSL Ϫ/Ϫ WAT compared with HSL ϩ/ϩ WAT in both Lep ϩ/ϩ and Lep ob/ob backgrounds. On the other hand, the expression of CAAT/enhancerbinding protein ␣, which is dominant during the late stage of differentiation (14), adipocyte determination and differentiation factor 1/sterol regulatory element-binding protein-1 (15), fatty acid synthase (16), and peroxisome proliferator activator receptor ␥ (17) was decreased in HSL Ϫ/Ϫ WAT compared with HSL ϩ/ϩ WAT in both Lep ϩ/ϩ and Lep ob/ob backgrounds. These results strongly support the results of immunohistochemistry and the conclusion that preadipocytes were increased in HSL Ϫ/Ϫ WAT. No significant differences were observed in the expression of lipoprotein lipase and uncoupling protein-2 between Lep ob/ob /HSL ϩ/ϩ and Lep ob/ob /HSL Ϫ/Ϫ mice.
To rule out the possibility that the small cells are apoptotic due to overaccumulation of intracellular TG, we performed terminal deoxynucleotidyltransferase-mediated dUTP nick end-labeling staining and DNA ladder detection. Although Lep ob/ob WAT contained an increased number of terminal deoxynucleotidyltransferase-mediated dUTP nick end labelingpositive apoptotic cells and increased DNA ladder formation compared with Lep ϩ/ϩ WAT, there was no difference between HSL ϩ/ϩ and HSL Ϫ/Ϫ WAT (Fig. 3b). Furthermore, we performed in vivo bromodeoxyuridine labeling to determine Food Intake Is Reduced in Obese HSL Ϫ/Ϫ Mice-To determine the causes of the lower adiposity in Lep ob/ob /HSL Ϫ/Ϫ mice, we examined food intake. In the Lep ϩ/ϩ background, there was no difference in food intake between HSL ϩ/ϩ and HSL Ϫ/Ϫ mice. In the Lep ob/ob background, food intake, which was increased compared with the Lep ϩ/ϩ background, was reduced by 19% in HSL Ϫ/Ϫ mice compared with HSL ϩ/ϩ mice at the age of 14 weeks (Fig. 4a). The reduced food intake was apparent at least after 8 weeks of age (data not shown). Lep ob/ob /HSL Ϫ/Ϫ mice had significantly lower feeding efficiency than Lep ob/ob /HSL ϩ/ϩ mice (Fig. 4b), indicating that the reduced food intake was not the only mechanism contributing to the leanness of Lep ob/ob /HSL Ϫ/Ϫ mice.
Adrenalectomy reduces food intake (18,19) in Lep ob/ob mice, which are in hyper-ACTH status, and HSL deficiency reduces the corticosterone response to ACTH (20,21). Therefore, we measured basal plasma corticosterone levels to examine whether adrenal insufficiency underlies the reduced food intake. There was no significant decrease in the plasma corticos-  4. Food intake and plasma corticosterone levels. a, food intake was measured at the age of 14 weeks (n ϭ 10). b, feeding efficiency was determined at the age of 14 weeks (n ϭ 8 -10). c, basal plasma corticosterone levels at the age of 9 weeks (n ϭ 10).
Hypothalamus expressed HSL protein in wild-type mice but not in HSL Ϫ/Ϫ mice (Fig. 5a). The suppressed food intake in Lep ob/ob mice led us to examine the expression of various neuropeptides that govern appetite and satiety in the hypothalamus (Fig. 5, b and c). The expression levels of NPY and AgRP, both of which were induced in Lep ob/ob background, were de-creased nearly to the level of lean Lep ϩ/ϩ mice (Fig. 5, b and c). The expression levels of pro-opiomelanocortin, which were suppressed in the Lep ob/ob background, were not different between HSL ϩ/ϩ and HSL Ϫ/Ϫ mice (Fig. 5, b and c).
HSL Deficiency Decreases Plasma Levels of FFA and Glycerol-Lep ob/ob /HSL Ϫ/Ϫ mice showed significant reduction in plasma levels of FFA and glycerol compared with Lep ob/ob / HSL ϩ/ϩ mice (FFA, 663 Ϯ 67 versus 1,174 Ϯ 17 M, p Ͻ 0.01; glycerol, 10.0 Ϯ 2.8 versus 43.4 Ϯ 6.0 mg/dl, p Ͻ 0.01). In the lean Lep ϩ/ϩ /HSL Ϫ/Ϫ mice, plasma cholesterol levels were elevated, whereas plasma TG levels were decreased compared with those in Lep ϩ/ϩ /HSL ϩ/ϩ mice, which was due to an increase in high density lipoproteins and a decrease in very low density lipoproteins, respectively (data not shown), as reported previously (22). In the Lep ob/ob background, no significant difference was observed between HSL ϩ/ϩ and HSL Ϫ/Ϫ mice.
Leanness and low plasma FFA levels are commonly associ-  6. Body temperature and oxygen consumption. a, core body temperature was measured at the age of 14 weeks (n ϭ 10). b, oxygen consumption was measured under a 12-h light/dark cycle and free access to food and water at the age of 12 weeks (n ϭ 8 -9). Oxygen consumption was normalized by body surface area. E, Lep ϩ/ϩ /HSL ϩ/ϩ ; Ⅺ, Lep ϩ/ϩ /HSL Ϫ/Ϫ ; q, Lep ob/ob /HSL ϩ/ϩ ; f, Lep ob/ob /HSL Ϫ/Ϫ . Each value represents the mean Ϯ S.E. ated with increased insulin sensitivity, whereas impaired adipogenesis is associated with insulin resistance (23). To estimate insulin sensitivity, we measured blood levels of glucose and insulin. There were no differences in the blood glucose and plasma insulin levels between Lep ob/ob /HSL ϩ/ϩ and Lep ob/ob / HSL Ϫ/Ϫ mice when fasted (data not shown), indicating that insulin sensitivity was not significantly different between them. DISCUSSION In the present studies, we use Lep ob/ob /HSL Ϫ/Ϫ mice to show that the absence of HSL inhibits adipogenesis and feeding behavior, thereby decreasing adiposity in leptin-deficient obesity. Resistance to obesity was also observed in a high fat diet-induced obesity model (24).
Adipocytes in WAT of Lep ob/ob /HSL Ϫ/Ϫ mice displayed exaggerated size heterogeneity (Fig. 2, a and b). In particular, small cells were remarkably increased. Although adipocytes from Lep ob/ob WAT are prone to apoptosis upon insulin depletion (25) and indeed Lep ob/ob WAT contained a substantial number of apoptotic cells, HSL deficiency did not further increase the number of apoptotic cells in either the Lep ϩ/ϩ or Lep ob/ob settings (Fig. 3c). Since the majority of the small cells were positive for S-100 protein (Fig. 2c), we consider the small cells to be preadipocytes. This view is consistent with the gene expression profiles of WAT; adipose differentiation-related protein and CAAT/enhancer-binding protein ␤, which characterize the early phase of adipocyte differentiation, were up-regulated, and CAAT/enhancer-binding protein ␣, peroxisome proliferator activator receptor ␥, adipocyte determination and differentiation factor 1/sterol regulatory element-binding protein-1, and fatty acid synthase, which characterize the mature stage of adipocyte differentiation, were down-regulated (Fig. 3a). Based on these results, we speculate that HSL deficiency inhibits adipogenesis in a certain population of preadipocytes, thereby stimulating their accumulation. Substances released from adipocytes of Lep ob/ob /HSL Ϫ/Ϫ mice may mediate the inhibition of adipocyte differentiation. In this context, it is of note that FFA, whose production was reduced in HSL Ϫ/Ϫ mice, induces differentiation of preadipocytes (26 -29). Moreover, Janke et al. (30) have recently reported that angiotensin II secreted by mature adipocytes inhibits differentiation of preadipocytes in culture.
It is also possible that intracellular lipids such as DG, which are accumulated in the absence of HSL (6), directly inhibit adipogenesis. It is tempting to speculate that protein kinase C activation by DG inhibits glycogen synthase kinase-3 (31), thereby enhancing signaling of Wnt, a negative regulator of adipogenesis (32). The impaired adipogenesis observed in Lep ob/ob /HSL Ϫ/Ϫ mice appears to be inconsistent with our previous findings that HSL Ϫ/Ϫ embryonic fibroblasts are indistinguishable from wild-type embryonic fibroblasts in adipogenic ability (8). In the experiments in vitro, a pharmacological dose of insulin, methyl-isobutylxanthine, or dexamethasone in the differentiation induction medium may have overcome the reduced adipogenic potential of HSL Ϫ/Ϫ embryonic fibroblasts. Indeed, it was reported that exposure to methyl-isobutylxanthine or cAMP is sufficient to suppress expression of Wnt10b in 3T3-L1 cells (33).
Hyperphagia of leptin-deficient Lep ob/ob is mediated by the up-regulation of the orexigenic neuropeptides, NPY (34) and AgRP (35), with concomitant suppression of the pro-opiomelanocortin gene, which encodes the anorexigenic ␣-melanocytestimulating hormone (36). Since disruption of NPY (37) or its cognate receptor (Y2 receptor) (38) attenuates the obesity of Lep ob/ob mice, the reduced expression of NPY and possibly AgRP may account for the decreased food intake in Lep ob/ob / HSL Ϫ/Ϫ mice (Fig. 5, b and c). Together with the hypothalamic expression of HSL (Fig. 5a), these results strongly indicate the crucial role of hypothalamic metabolism of fatty acids in the control of feeding behavior. In this context, it is noteworthy that inhibition of FFA synthesis by fatty acid synthase inhibitors shows similar suppressing effects on food intake (39) as well as on fasting-induced increase in the hypothalamic expression of NPY and AgRP (40). Likewise, inhibition of hypothalamic carnitine palmitoyltransferase-1 decreases food intake (41). Neither body temperature nor oxygen consumption was different between Lep ob/ob /HSL Ϫ/Ϫ and Lep ob/ob /HSL ϩ/ϩ mice (Fig. 6). Thus, the reduced food intake as well as the defective differentiation of preadipocytes may account for the lower adipocity in Lep ob/ob /HSL Ϫ/Ϫ mice at least in mice aged older than 8 weeks (Fig. 1). It is interesting to note that the lower adiposity was not observed in mice younger than 8 weeks. We speculate that non-HSL TG lipase compensates for the deficiency of HSL until about 8 weeks of age in terms of growth and adiposity.
Leanness or resistance to obesity has been reported in other murine models with mutated genes that are instrumental in fatty acid metabolism. For example, mice lacking the RII␤ subunit of protein kinase A (42) or perilipin (43) have increased lipolysis; mice lacking acyl-CoA:diacylglycerol transferase (44) or mitochondrial glycerol-3-phosphate acyltransferase (45) have reduced synthesis of TG. In particular, mice with a mutation in the peroxisome proliferator activator receptor ␥-binding site in the gene for the cytosolic form of phosphoenolpyruvate carboxykinase, which lack expression of phosphoenolpyruvate carboxykinase activity in WAT, are similar to Lep ob/ob /HSL Ϫ/Ϫ mice in that their WAT also has accumulation of lipodystrophy-like small cells dispersed among normal large fat cells (46). The significance of this similarity is currently unknown.
In conclusion, Lep ob/ob /HSL Ϫ/Ϫ mice have impaired adipogenesis and resistance to obesity at least partially due to the reduced food intake. These findings should provide the basis for understanding the pathophysiology of obesity and can be exploited to develop novel therapy for the endemic disease.