A role for hepatic leptin signaling in lipid metabolism via altered very low density lipoprotein composition and liver lipase activity in mice

Authors

  • Frank K. Huynh,

    1. Departments of Cellular and Physiological SciencesUniversity of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • Ursula H. Neumann,

    1. Departments of Cellular and Physiological SciencesUniversity of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • Ying Wang,

    1. Departments of Molecular and Cellular Pharmacology Group, Faculty of Pharmaceutical SciencesUniversity of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • Brian Rodrigues,

    1. Departments of Molecular and Cellular Pharmacology Group, Faculty of Pharmaceutical SciencesUniversity of British Columbia, Vancouver, British Columbia, Canada
    Search for more papers by this author
  • Timothy J. Kieffer,

    Corresponding author
    1. Departments of Cellular and Physiological SciencesUniversity of British Columbia, Vancouver, British Columbia, Canada
    2. Department of Surgery, University of British Columbia, Vancouver, British Columbia, Canada
    • Department of Cellular and Physiological Sciences, University of British Columbia, 2350 Health Sciences Mall, Vancouver, British Columbia, Canada V6T 1Z3
    Search for more papers by this author
    • fax: 604-822-2316;

  • Scott D. Covey

    Corresponding author
    1. Departments of Biochemistry and Molecular Biology, Life Sciences InstituteUniversity of British Columbia, Vancouver, British Columbia, Canada
    • Department of Biochemistry and Molecular Biology, Life Sciences Institute, University of British Columbia, 2350 Health Sciences Mall, Vancouver, British Columbia, Canada V6T 1Z3
    Search for more papers by this author
    • fax: 604-822-5227.


  • Potential conflict of interest: Nothing to report.

  • Supported by a grant from the Canadian Institutes of Health Research. T. J. K. received a Senior Scholarship from the Michael Smith Foundation for Health Research. F. K. H. received scholarships from the Natural Sciences and Engineering Research Council of Canada and the Canadian Diabetes Association. U. H. N. received the University of British Columbia Summer Student Research Program award sponsored by the Kinsley Brotherton McLeod Endowment and the Florence & George Heighway Endowment Fund.

Abstract

Obesity is highly associated with dyslipidemia and cardiovascular disease. However, the mechanism behind this association is not completely understood. The hormone leptin may be a molecular link between obesity and dysregulation of lipid metabolism. Leptin can affect lipid metabolism independent of its well-known effects on food intake and energy expenditure, but exactly how this occurs is ill-defined. We hypothesized that since leptin receptors are found on the liver and the liver plays an integral role in regulating lipid metabolism, leptin may affect lipid metabolism by acting directly on the liver. To test this hypothesis, we generated mice with a hepatocyte-specific loss of leptin signaling. We previously showed that these mice have increased insulin sensitivity and elevated levels of liver triglycerides compared with controls. Here, we show that mice lacking hepatic leptin signaling have decreased levels of plasma apolipoprotein B yet increased levels of very low density lipoprotein (VLDL) triglycerides, suggesting alterations in triglyceride incorporation into VLDL or abnormal lipoprotein remodeling in the plasma. Indeed, lipoprotein profiles revealed larger apolipoprotein B-containing lipoprotein particles in mice with ablated liver leptin signaling. Loss of leptin signaling in the liver was also associated with a substantial increase in lipoprotein lipase activity in the liver, which may have contributed to increased lipid droplets in the liver. Conclusion: Lack of hepatic leptin signaling results in increased lipid accumulation in the liver and larger, more triglyceride-rich VLDL particles. Collectively, these data reveal an interesting role for hepatic leptin signaling in modulating triglyceride metabolism. (HEPATOLOGY 2013)

Despite the well-accepted link between obesity, diabetes, and dyslipidemia, the molecular mechanisms that drive this association are not understood. The hormone leptin is a potential link between obesity and abnormal lipid metabolism. Leptin is secreted from adipose tissue and acts on the hypothalamus to reduce food intake and increase energy expenditure.1, 2 Thus, leptin-deficient ob/ob mice and leptin receptor-deficient db/db mice are hyperphagic and obese. However, these mice also display hypertriglyceridemia,3 hypercholesterolemia,3 hepatic steatosis,4 and impaired lipid tolerance.5 Several studies suggest that these effects on lipid metabolism are independent of leptin's effects on food intake and obesity. For example, restricting food intake in ob/ob mice cannot improve lipid metabolism as effectively as leptin treatment.6, 7 In addition, lipodystrophic mice and humans, which have little to no adipose tissue and are hypoleptinemic, also display hyperlipidemia and hepatic steatosis, and these symptoms are ameliorated by leptin.8, 9 Clearly, leptin has effects on lipid metabolism independent of its effects on body weight.

The manner by which leptin directly affects lipid metabolism is not well understood. We hypothesized that because the liver plays a role in lipid metabolism, leptin acts directly on the liver to exert some of its metabolic effects. Indeed, leptin receptors are found in the liver,10, 11 and leptin administration to ob/ob mice elicits many changes in the expression of genes involved with lipid metabolism in the liver.6, 8, 12 Furthermore, leptin treatment in ob/ob mice can reverse hepatic steatosis,7 potentially due to direct effects of leptin on the liver.13, 14

To address the direct effects of leptin on the liver, Cohen et al.15 knocked out leptin receptors specifically in hepatocytes. Surprisingly, they found no accumulation of hepatic lipids, but other aspects of lipid metabolism were not explored. We also generated mice with a loss of hepatic leptin signaling wherein the leptin signaling domain is removed specifically from hepatocytes.13 These mice were protected from age- and diet-related glucose intolerance and had increased hepatic insulin sensitivity.13 Further, these mice had elevated liver triglyceride and cholesterol levels,13 indicating an alteration in hepatic lipid metabolism. We have now discovered that mice lacking hepatic leptin signaling have larger apolipoprotein B (apoB)-containing lipoproteins and elevated triglyceride levels in very low density lipoprotein (VLDL) particles. This is accompanied by decreased plasma apoB, higher lipoprotein lipase (LPL) activity in the liver, and lower non-LPL activity compared with controls. Taken together, these data reveal a novel role for hepatic leptin signaling in regulating triglyceride metabolism.

Abbreviations

Ad-β-gal, adenovirus expressing β-galactosidase; Ad-Lepr-b, adenovirus expressing isoform b of the leptin receptor; apoB, apolipoprotein B; HL, hepatic lipase; LPL, lipoprotein lipase; mRNA, messenger RNA; VLDL, very low density lipoprotein.

Materials and Methods

Animals.

Leprflox/flox AlbCre and Leprflox/flox AlbCre ob/ob mice were generated as described.13, 16 Leprflox/flox AlbCre ob/ob mice were treated with 0.6 μg/day mouse recombinant leptin (National Hormone and Peptide Program, Torrance, CA) via mini-osmotic pumps (Alzet, Palo Alto, CA). Db/db mice were treated intravenously with 1 × 109 pfu of an adenovirus expressing either the long signaling isoform of the mouse leptin receptor (Ad-Lepr-b) or β-galactosidase (Ad-β-gal) as a control. Ob/ob mice were treated with 1.5 μg/g leptin via intraperitoneal injections or 0.6 μg/day leptin via miniosmotic pumps. Procedures were performed in accordance with the University of British Columbia Animal Care Committee guidelines.

Triglyceride Secretion.

Four-hour fasted mice were injected intraperitoneally with 1 g/kg of poloxamer-407 (Sigma-Aldrich, Oakville, Ontario, Canada) followed by an intraperitoneal injection of 0.6 U/kg or 0.725 U/kg insulin (Novolin; Novo Nordisk, Mississauga, Ontario, Canada). Plasma samples were taken throughout the experiment for triglyceride measurements.

Plasma Analytes.

Cholesterol levels were measured using a Cholesterol E kit (Wako Chemicals USA, Richmond, VA), free fatty acids using a HR Series NEFA-HR(2) kit (Wako Chemicals USA), triglycerides using a Serum Triglyceride Determination kit (Sigma-Aldrich), and insulin using a Ultrasensitive Mouse Insulin ELISA (ALPCO Diagnostics, Salem, NH).

Lipoprotein Profiles.

Equal volumes of plasma from mice of the same genotype were pooled and 200 μL of the pooled plasma was applied to a Superose 6L HR 10/30 column (GE Healthcare, Baie d'Urfe, Quebec, Canada) with 154 mM NaCl, 1 mM ethylene diamine tetraacetic acid (pH 8). Fractions were assayed using modified protocols of the Cholesterol E kit and Serum Triglyceride Determination kit.

Western Blotting.

Sodium dodecyl sulfate–polyacrylamide gel electrophoresis was performed on 1 μL of plasma or 15 μL of fast protein liquid chromatography eluate using a 4%-15% gradient gel. Polyvinylidene fluoride membranes were probed with an apoB antibody that detects both apoB48 and apoB100 (K23300R; Meridian Life Science, Saco, ME).

Quantitative Polymerase Chain Reaction.

Quantitative polymerase chain reaction for apoB, hepatic lipase, and LPL is described in detail in the Supporting Information.

Lipase Activity.

Liver lysates were prepared and assessed for LPL and non-LPL activity as described in the Supporting Information.

Microscopy.

Livers were fixed in 4% paraformaldehyde overnight and stored in 70% ethanol. Sections (5 μm) were stained with hematoxylin and eosin and visualized under oil immersion.

Oral Lipid Tolerance.

Four-hour fasted mice were given 5 μL/g olive oil via oral gavage. Plasma samples were taken over 5 hours and assayed for triglycerides.

Results

We first determined whether triglyceride output from the liver was altered in the fasting state. To evaluate VLDL triglyceride secretion from the liver, we injected fasted mice with poloxamer-407. Poloxamer-407 was a potent inhibitor of triglyceride uptake (Fig. 1A), but the accumulation in plasma triglycerides occurred at similar rates in Leprflox/flox AlbCre+ mice and their Leprflox/flox AlbCre− littermate controls. Because insulin suppresses VLDL triglyceride secretion17 and the livers of Leprflox/flox AlbCre+ mice are more sensitive to the effects of insulin,13 we examined whether a bolus of insulin could differentially affect VLDL triglyceride secretion in these mice. In response to insulin, there was a decreased rate of plasma triglyceride accumulation in both Leprflox/flox AlbCre+ mice and littermate controls (Figs. 1B,C). Surprisingly, in Leprflox/flox AlbCre+ mice, there were elevated levels of plasma triglycerides after insulin injection compared with controls (Figs. 1B,C), suggesting that insulin mediated suppression of triglyceride secretion is muted in mice lacking hepatic leptin signaling.

Figure 1.

Mice lacking hepatic leptin signaling have increased hepatic triglyceride secretion after an insulin bolus. Mice were fasted for 4 hours, injected with poloxamer-407 at t = 0, and then were either (A) not injected (n = 5 for Leprflox/flox AlbCre−, n = 4 for Leprflox/flox AlbCre+), (B) injected with 0.6 U/kg insulin (n = 6 for Leprflox/flox AlbCre−, n = 8 for Leprflox/flox AlbCre+), or (C) injected with 0.725 U/kg insulin (n = 5 for Leprflox/flox AlbCre−, n = 9 for Leprflox/flox AlbCre+) at t = 2 hours. Plasma triglyceride levels were monitored throughout the experiment. Data are expressed as the mean ± SEM. P values were determined via two-way analysis of variance with a Holm-Sidak post hoc test.

We next investigated the effects of hepatic leptin signaling on fasting plasma triglycerides under more strenuous metabolic conditions. Leprflox/flox AlbCre mice were crossed onto an obese, hyperinsulinemic ob/ob background to generate ob/ob mice lacking functional hepatic leptin receptors (Leprflox/flox AlbCre ob/ob mice). Unlike Leprflox/flox AlbCre+ mice, which could have developmental differences compared with their littermate controls due to a life-long loss of hepatic leptin signaling, Leprflox/flox AlbCre+ ob/ob mice are equivalent to their littermate controls until treated with exogenous leptin. At 7 weeks of age, Leprflox/flox AlbCre ob/ob mice were treated with low dose leptin so as to maintain obesity and hyperinsulinemia (Supporting Fig. 1). This dose of leptin lowered plasma cholesterol and triglycerides in ob/ob mice with and without hepatic leptin signaling (Fig. 2A-C). However, plasma triglyceride levels in Leprflox/flox AlbCre+ ob/ob mice did not decrease as much as in their Leprflox/flox AlbCre- ob/ob controls (Fig. 2C). By the last day of leptin treatment, the Leprflox/flox AlbCre+ ob/ob mice had 36% higher plasma triglycerides than their littermate controls (Fig. 2C). The effects of leptin treatment persisted even after leptin therapy ceased, with plasma triglyceride levels in both groups only returning to near pre-leptin levels 50 days after the leptin pump was removed (Fig. 2C), indicating leptin treatment in ob/ob mice has long-term effects on lipid metabolism.

Figure 2.

Hepatic leptin signaling has a subtle but consistent effect on plasma triglycerides in obese, hyperinsulinemic mice. (A-C) Seven-week-old male Leprflox/flox AlbCre ob/ob mice were treated with 0.6 μg/day mouse recombinant leptin for 28 days via mini-osmotic pump starting on day 1. Mice were tracked for 4-hour-fasted plasma cholesterol (A), free fatty acids (B), and triglycerides (C) (n ≥ 5 for both groups). (D-F) Male db/db mice were treated with Ad-β-gal or Ad-Lepr-b at 12 weeks of age and tracked for 4-hour-fasted plasma cholesterol (D), free fatty acids (E), and triglycerides (F) (n ≥ 6 for both groups). Data are shown as the mean ± SEM. P values were determined via two-way analysis of variance with a Holm-Sidak post hoc test.

Because the effect on plasma triglycerides in Leprflox/flox AlbCre ob/ob mice was subtle, we sought to reproduce these results in a complementary mouse model. We treated leptin receptor-deficient db/db mice with Ad-Lepr-b, which confers liver-selective expression18 and restores phospho-STAT3 signaling in the liver.16 Upon treatment with Ad-Lepr-b, the db/db mice remained obese and hyperinsulinemic (Supporting Fig. 2). Also, db/db mice treated with Ad-Lepr-b and control db/db mice treated with Ad-β-gal both had a response to the virus itself independent of the Lepr-b or β-gal constructs. We attribute this to an acute phase immune response to the virus, which has been shown to have effects on lipid metabolism.19 Nonetheless, we observed no differences in plasma cholesterol and free fatty acids between Ad-Lepr-b– and Ad-β-gal–treated db/db mice (Fig. 2D-E). Although both virus-treated groups had an increase in plasma triglycerides, db/db mice treated with Ad-Lepr-b had lower fasting plasma triglycerides than the Ad-β-gal–treated controls between 1 and 3 weeks postinfection, with Ad-Lepr-b treated mice reaching 31% lower plasma triglycerides 12 days postinfection (Fig. 2F). These data are similar to those of Lee et al.,14 who treated fa/fa rats with an adenovirus expressing β-gal or Lepr-b and also saw a marked increase in plasma triglycerides in the β-gal–treated animals compared with the Lepr-b–treated animals. Collectively, the data show that under obese, hyperinsulinemic conditions, hepatic leptin signaling is required for maintaining normal plasma triglyceride levels.

Because leptin has been implicated in regulating the amount of triglyceride incorporation into VLDL,17 we evaluated lipoprotein profiles in Leprflox/flox AlbCre+ mice and their littermate controls. Mice lacking hepatic leptin signaling had no alterations in the distribution or amount of cholesterol (Fig. 3A). Interestingly, Leprflox/flox AlbCre+ mice had elevated triglycerides in fractions consistent in size with VLDL particles (Fig. 3B). We next performed western blots for apoB, since each VLDL particle is associated with one apoB molecule.20 The western blots showed that particles containing apoB in the plasma of Leprflox/flox AlbCre+ mice eluted earlier than apoB-containing particles in Leprflox/flox AlbCre− mice, suggesting larger apoB-containing particles (Fig. 3C-F). Therefore, mice lacking hepatic leptin signaling have more triglyceride-rich VLDL particles and larger apoB-containing lipoprotein particles.

Figure 3.

Attenuation of hepatic leptin signaling results in increased VLDL triglycerides and larger apoB-containing lipoprotein particles. Pooled 4-hour-fasted plasma samples from 17- to 30-week-old male Leprflox/flox AlbCre+ mice and littermate controls (n ≥ 3 for each trial) were subjected to fast protein liquid chromatography, and the fractions were assayed for cholesterol (A) and triglycerides (B). Data represent the average of two trials in (A) with plasma from a total of eight Leprflox/flox AlbCre− mice and 11 Leprflox/flox AlbCre+ mice. In (B), the average of three trials is shown, representing plasma from a total of 14 Leprflox/flox AlbCre− mice and 14 Leprflox/flox AlbCre+ mice. (C) Four-hour-fasted plasma samples (n ≥ 3) were pooled from 18- to 27-week-old Leprflox/flox AlbCre+ mice and their littermate controls. The pooled samples were fractionated and then western blots for apoB were performed on fractions 6-37. Mean pixel intensity of the apoB100 band (D), apoB48 band (E), and combined apoB100 and apoB48 bands (F) are shown for each fraction. Data are expressed as the mean ± SEM. *P ≤ 0.05 (Student t test).

Because there appeared to be a slight decrease in total apoB levels in mice lacking hepatic leptin signaling (Fig. 3F), we measured total apoB levels in whole plasma from individual mice. Indeed, plasma apoB100 levels were 18% lower in Leprflox/flox AlbCre+ mice compared with controls (Figs. 4A,B), with a similar but nonsignificant trend for plasma apoB48 levels (Figs. 4A,C). Because apoB can come from the small intestine as well as the liver, we measured hepatic apoB transcript levels to see whether changes in the liver could account for the decreased plasma apoB levels. Hepatic apoB messenger RNA (mRNA) levels were 24% lower in Leprflox/flox AlbCre+ mice, suggesting that decreased plasma apoB can be accounted for by decreased hepatic expression of apoB (Fig. 4D). Accordingly, hepatic apoB mRNA levels in db/db mice were 26% lower than C57BL/6 controls, and upon re-expression of functional leptin receptors in the liver, hepatic apoB transcript levels returned to wild-type levels (Fig. 4E). Thus, functional hepatic leptin signaling is positively correlated with plasma apoB levels.

Figure 4.

Attenuation of hepatic leptin signaling decreases plasma and hepatic apoB levels. (A) Western blots for apoB were performed on 4-hour-fasted plasma samples from 25- to 30-week-old Leprflox/flox AlbCre mice (n ≥ 6). Quantification of apoB100 and apoB48 for all samples by densitometry is shown in (B) and (C) (n ≥ 6). (D) ApoB transcript levels were measured in the livers of 22-week-old male Leprflox/flox AlbCre mice following a 4-hour fast (n = 9). (E) Hepatic apoB transcript levels were measured in C57BL/6 controls and db/db mice 4 weeks postinfection with either Ad-β-gal or Ad-Lepr-b (n ≥ 5). Data are expressed as the mean ± SEM. P values were determined via Student t test.

Our data indicate that mice specifically lacking hepatic leptin signaling have less total plasma apoB, larger apoB-containing lipoprotein particles, and increased amounts of triglycerides in VLDL-sized particles. It is possible that a reduction in lipase activity could explain some of these observations, since patients with hepatic lipase (HL) deficiency display abnormally large lipoprotein particles.21 Indeed, mice lacking leptin signaling in the liver had 23% lower HL mRNA (Fig. 5A) and a trend toward lower non-LPL activity levels in the liver (Fig. 6A) compared with controls. However, there was a substantial 4.5-fold increase in LPL activity in the liver of mice lacking liver leptin signaling (Fig. 6B). This was surprising given that LPL is not normally expressed in adult mouse liver.22, 23 To determine whether a loss of hepatic leptin signaling induces the liver to produce LPL, we measured hepatic LPL mRNA levels and found no difference in transcript levels between Leprflox/flox AlbCre+ mice and their littermate controls (Fig. 5B). The contribution of hepatic LPL to total triglyceride lipase activity in the liver increased from 17% in control mice to 57% in mice lacking hepatic leptin signaling (Fig. 6C). Overall, these alterations to LPL activity resulted in increased total triglyceride lipase activity in the livers of Leprflox/flox AlbCre+ mice (Fig. 6C). Because overexpression of LPL in different tissues can cause lipid accumulation,22 we performed a histological examination of livers from Leprflox/flox AlbCre mice and observed that loss of hepatic leptin signaling led to enlarged lipid droplets (Fig. 7). These data clearly reveal a role for hepatic leptin signaling in regulating lipase activity in the liver.

Figure 5.

Hepatic leptin signaling can regulate hepatic lipase and lipoprotein lipase mRNA levels in the liver. (A, B) Hepatic lipase and lipoprotein lipase transcript levels were measured in the livers of 22-week-old male Leprflox/flox AlbCre mice following a 4-hour fast (n ≥ 8). (C, D) Hepatic lipase and lipoprotein lipase transcript levels were measured in the livers of 16-week-old C57BL/6 controls (n = 5) and db/db males 4 weeks postinfection with Ad-β-gal or Ad-Lepr-b (n ≥ 7). Data are expressed as the mean ± SEM. P values were determined via Student t test.

Figure 6.

Hepatic leptin signaling regulates lipase activity levels in the liver. (A, B) Non-LPL (A) and LPL (B) activity levels were assessed in liver lysates from 22-week-old male Leprflox/flox AlbCre mice (n = 6 per group). (C) Total triglyceride lipase activity in the liver. (D-F) Non-LPL (D), LPL (E), and total lipase activity (F) levels were measured in liver lysates from C57BL/6 controls (n = 5) and db/db mice 4 weeks postinfection with either Ad-β-gal or Ad-Lepr-b (n = 7). Data are expressed as the mean ± SEM. P values were obtained by Student's t-test.

Figure 7.

Loss of hepatic leptin signaling results in hepatic lipid accumulation. (A-C) Representative images from three male Leprflox/flox AlbCre− littermate controls. (D-F) Representative images from three male Leprflox/flox AlbCre+ mice. (Original magnification ×100; hematoxylin and eosin stain.)

Similar to mice that have a liver-specific loss of leptin signaling, Ad-β-gal-treated db/db mice also had a ∼30% decrease in non-LPL activity in the liver compared with C57BL/6 controls (Fig. 6D), and this correlated with a decrease in hepatic HL mRNA (Fig. 5C). When functional leptin receptors were overexpressed in the livers of db/db mice, non-LPL activity increased even beyond levels seen in wild-type mice (Fig. 6D). Furthermore, control db/db mice had a two-fold increase in LPL activity levels, and when db/db mice were treated with Ad-Lepr-b, LPL activity returned to wild-type levels (Fig. 6E). We also observed that in the total lack of leptin signaling, hepatic LPL activity contributed to 60% of total triglyceride lipase activity in the liver, and when leptin signaling was selectively restored to the liver, hepatic LPL activity contributed only 20% to total triglyceride lipase activity, which is similar to wild-type levels (Fig. 6F). These data from two complementary models reveal a novel role for hepatic leptin signaling in modulating lipase activity in the liver. However, the manner (transcriptional versus posttranscriptional) by which lipase activity in the liver is regulated in mice with a life-long loss of hepatic leptin signaling and mice with an induced gain of hepatic leptin signaling is different (Figs. 5 and 6). Nonetheless, the functional end result is that with a loss of hepatic leptin signaling, non-LPL lipase activity is decreased and LPL activity is increased.

To determine whether these effects of leptin on apoB transcription and lipase activity in the liver are due to direct or indirect actions of leptin, we treated ob/ob mice with acute leptin injections as well as chronic leptin infusions, which restored leptin signaling to all tissues. Acute leptin injections increased apoB mRNA in the liver by nearly 60%, but chronic low-dose leptin treatment had no effect (Supporting Fig. 3A). Further, while liver-selective restoration of leptin signaling in db/db mice decreased hepatic LPL expression back toward wild-type levels (Fig. 5D), acute leptin injections into ob/ob mice increased hepatic LPL mRNA (Supporting Fig. 3C). Therefore, the increase in hepatic LPL mRNA in ob/ob mice after acute leptin treatment is likely a result of leptin action outside of the liver. Interestingly, we previously observed that a whole body loss of leptin signaling has distinct effects, in fact opposite, from a liver specific loss of leptin signaling with respect to glucose homeostasis.13 Notably, chronic low-dose leptin did not change hepatic LPL mRNA expression in ob/ob mice (Supporting Fig. 3C), further highlighting the differences between acute versus chronic actions of leptin that have also been observed by others.24 Collectively, these results suggest that the direct and indirect effects as well as acute versus chronic effects of leptin on liver lipid metabolism are distinct.

Similar to Ad-β-gal–treated db/db mice, which showed decreased hepatic HL mRNA levels compared with wild-type controls (Fig. 5C), ob/ob mice also had decreased hepatic HL transcript levels (Supporting Fig. 3B). Liver HL mRNA levels were restored almost to wild-type levels by acute leptin injections as well as chronic low-dose leptin to ob/ob mice (Supporting Fig. 3B). However, these effects of leptin on hepatic HL transcript levels appear to be independent of direct hepatic leptin signaling, because restoration of functional leptin signaling selectively in the livers of db/db mice did not restore wild-type hepatic HL mRNA levels (Fig. 5C).

Interestingly, these changes in hepatic LPL and HL mRNA levels in leptin-treated ob/ob mice did not translate into corresponding changes in hepatic LPL or non-LPL activity levels (Supporting Fig. 4). Ob/ob mice had decreased LPL activity in the liver despite elevated LPL mRNA. Furthermore, wild-type LPL activity levels were unable to be restored by leptin in ob/ob mice despite a marked increase in LPL mRNA expression after acute leptin injections (Supporting Fig. 4B). Similarly, despite changes in HL mRNA levels, non-LPL activity levels in the liver were largely unchanged by loss of leptin signaling in the ob/ob mice (Supporting Fig. 4A). Thus, the regulation of lipase activity in the liver by leptin seems to involve both transcriptional and posttranscriptional mechanisms.

Because altered lipase activity can affect triglyceride clearance and leptin may act on the liver to promote postprandial triglyceride clearance,25 we performed an oral lipid tolerance test on mice with a loss of leptin signaling in the liver. These mice had no alterations in lipid tolerance compared with controls (Supporting Fig. 5). However, when we treated obese, hyperinsulinemic Leprflox/flox AlbCre+ ob/ob mice with leptin, lipid tolerance was not improved to the same extent as in their littermate controls (Fig. 8A). Interestingly, the effects of leptin on lipid tolerance seemed to persist even after leptin therapy was ceased, indicating again that leptin treatment in ob/ob mice has long-term effects on lipid metabolism (Fig. 8B). We also assessed lipid tolerance in db/db mice treated with Ad-Lepr-b or Ad-β-gal. Lipid tolerance in the mice that received Ad-Lepr-b was improved compared with control mice that received Ad-β-gal (Figs. 8D,E). These data further suggest that lipid metabolism is differentially affected by a loss of hepatic leptin signaling in lean mice compared with hepatic leptin signaling in obese, hyperinsulinemic mice.

Figure 8.

Hepatic leptin signaling acutely improves lipid tolerance in ob/ob and db/db mice. (A-C) Seven-week-old male Leprflox/flox AlbCre ob/ob mice were treated with 0.6 μg/day mouse recombinant leptin for 28 days via mini-osmotic pump starting on day 1. Oral lipid tolerance tests were performed by fasting the mice for 4 hours and then administering an oral gavage of 5 μL/g olive oil. Mice were subjected to an oral lipid tolerance test after 7 days of leptin treatment (A), 9 days after pump removal (B), and 75 days after pump removal (C) (n ≥ 5). (D, E) Male db/db mice were subjected to an oral lipid tolerance test (D) 9 days before and (E) 7 days after treatment with either Ad-β-gal or Ad-Lepr-b (n ≥ 6). Data are expressed as the mean ± SEM. P values were determined via two-way analysis of variance with a Holm-Sidak post hoc test.

Discussion

It is well-established that leptin affects lipid metabolism, but whether these effects are a result of direct leptin action on the liver has not been fully addressed. We investigated lipid metabolism in four complementary mouse models and found a previously unreported role for hepatic leptin action in modulating lipase activity in the liver. Although it has been reported that HL mRNA levels in the liver are decreased in ob/ob mice and restored with whole body leptin treatment,12 we now report that this is not associated with increased non-LPL lipase activity in the liver. However, in mice with a liver-specific loss or gain of leptin signaling, our data do support a role for leptin signaling specifically in the liver to positively regulate non-LPL activity. Further, we also report a novel finding that leptin resistance specifically in the liver leads to a marked increase in hepatic LPL activity. Because an overexpression of LPL in tissues can cause increased lipid uptake and lipid accumulation,22 we speculate that the elevation of LPL activity in Leprflox/flox AlbCre+ mice contributes to their elevated hepatic triglycerides.13

LPL activity has a complex mechanism of regulation, including transcriptional, posttranscriptional, translational, and/or posttranslational mechanisms depending on nutrient status and tissue.26 Adding to this complexity, our data show that in db/db mice, the loss of leptin signaling caused an elevation of hepatic LPL activity through transcriptional changes, but in Leprflox/flox AlbCre+ mice, the increased LPL activity was mediated through posttranscriptional mechanisms. Furthermore, because insulin can regulate LPL activity in adipose and muscle,26 leptin regulation of hepatic LPL activity may be indirect through the effects of leptin on hepatic insulin signaling. Additionally, because leptin treatment in ob/ob mice was unable to fully normalize lipase activity in the liver, secondary extrahepatic effects of leptin signaling also appear to contribute to the regulation of lipase activity in the liver. Although it is clear that loss of hepatic leptin signaling can increase hepatic LPL mRNA, the exact mechanism by which leptin regulates lipase activity in the liver remains to be determined.

Leprflox/flox AlbCre+ mice have increased hepatic insulin sensitivity,13 and insulin is an important regulator of lipid metabolism in the liver as evidenced by its role in decreasing plasma apoB levels.17 Consistent with this, our data show that in mice lacking hepatic leptin signaling, increased hepatic insulin sensitivity is associated with decreased plasma apoB levels even in the fasting state. Although it is possible that this effect on apoB is mediated directly by leptin signaling independent of insulin, we speculate that it is actually the effect of leptin on insulin signaling that mediates changes in apoB, since leptin itself does not affect plasma apoB levels.17, 27 Interestingly, in ob/ob mice, leptin was able to increase hepatic apoB mRNA only when leptin was administered at a dose and route of administration that has been shown to lower plasma insulin levels12, 28 and not at a dose that did not affect plasma insulin levels (Supporting Fig. 1D). In association with decreased plasma apoB levels, Leprflox/flox AlbCre+ mice had increased triglycerides in VLDL particles, suggesting that these mice may have fewer VLDL particles in total but more triglycerides per VLDL particle. We hypothesize that the increased incorporation of triglycerides in these mice is due in part to elevated liver triglycerides, leading to increased substrate availability. This can result in more triglyceride incorporation into each VLDL particle,17, 29 leading to enlarged, more triglyceride-rich VLDL particles.29 Intriguingly, overexpression of HL in a rat liver cell line resulted in secretion of triglyceride-poor VLDL,30 and patients with HL deficiency have been shown to have triglyceride-rich lipoproteins that are also larger in size.21 Therefore, although we cannot rule out the involvement of other non-LPL lipases in the liver, decreased HL activity in Leprflox/flox AlbCre+ mice likely contributes to altered lipid loading, leading to enlarged, triglyceride-rich VLDL particles.

Our observations are compatible with the theory that insulin is responsible for modulating the number of VLDL particles, while hepatic leptin action can modulate the amount of triglycerides available for incorporation into each VLDL particle through the effects of leptin on increasing fatty acid oxidation.17 In our model of increased hepatic insulin sensitivity with extreme hepatic leptin resistance, there is less plasma apoB, indicating fewer VLDL particles, and increased hepatic triglycerides, which may lead to larger, more triglyceride-rich VLDL particles. According to this model, one might expect that hepatic triglyceride secretion would be suppressed more in Leprflox/flox AlbCre+ mice because they are more insulin-sensitive than controls.13 Surprisingly, while we did observe insulin-mediated suppression of hepatic triglyceride secretion in both groups of mice, mice lacking hepatic leptin signaling had higher plasma triglycerides than controls after insulin (Fig. 1). This may be due to enhanced lipogenic effects of insulin in the Leprflox/flox AlbCre+ mice, which together with the lack of leptin signaling can result in even more substrate availability during hyperinsulinemic conditions, allowing for more triglycerides per VLDL particle. Interestingly, liver insulin receptor knockout mice have increased plasma apoB yet decreased plasma triglycerides and no alterations in liver triglycerides.31 Thus, insulin signaling in the liver may also have effects on triglyceride loading onto VLDL independent of its effects on substrate availability and our data suggest that leptin signaling in the liver may serve to counter this effect.

Despite evidence of major changes in hepatic lipid metabolism genes upon leptin treatment in models of leptin deficiency,6, 8, 12 our data suggest that indirect effects are involved. Consistent with the literature,14, 32 our data do support the fact that direct leptin action on the liver plays an antisteatotic role, but the level of hepatic steatosis in Leprflox/flox AlbCre+ mice was not as severe as in livers of db/db mice. Furthermore, unlike db/db mice with a global loss of leptin signaling, lean mice with a liver-specific loss of leptin signaling have normal total fasting plasma triglycerides and cholesterol levels.13 It appears that although mice with a hepatocyte-specific loss of leptin signaling have increased incorporation of triglycerides into VLDL particles, they do not develop hypertriglyceridemia due to their concurrent reduction in hepatic apoB production. However, in more metabolically stressed obese, hyperinsulinemic mice, we did observe a more pronounced perturbation in fasting plasma triglycerides and lipid tolerance. Interestingly, patients with metabolic syndrome have a higher proportion of large VLDL than healthy patients, even in patients with normal plasma triglyceride levels.33 Therefore, subtle effects of hepatic leptin resistance on lipid metabolism could have a major impact on health.

Our model of hepatic leptin resistance shows that loss of leptin signaling in the liver can contribute to the development of hepatic steatosis and large, triglyceride-rich lipoproteins. Given that obese humans are leptin-resistant, our data suggest that defects in lipid metabolism seen in obesity may stem in part from resistance to leptin action in the liver. Although the effects of liver leptin signaling on lipid metabolism appear subtle, our data show that these effects are more pronounced in obese and hyperinsulinemic states. Intriguingly, polymorphisms in the LEPR,34 HL,35 and LPL36 genes have been linked with familial combined hyperlipidemia, the most common genetically linked hyperlipidemia in humans. Thus, alterations to HL and LPL activity in the liver due to hepatic leptin resistance may result in increased risk of dyslipidemia and perhaps contribute to the development of metabolic syndrome.

Acknowledgements

We thank Streamson C. Chua (Albert Einstein College of Medicine) for his generous contribution of the Leprflox/flox mice and A. F. Parlow (National Hormone and Peptide Program) for providing mouse recombinant leptin. We also thank Martin G. Myers (University of Michigan) and Christopher J. Rhodes (University of Chicago) for providing the Ad-Lepr-b virus.

Ancillary

Advertisement