An intrinsic gut leptin-melanocortin pathway modulates intestinal microsomal triglyceride transfer protein and lipid absorption.

Fat is delivered to tissues by apoB-containing lipoproteins synthesized in the liver and intestine with the help of an intracellular chaperone, microsomal triglyceride transfer protein (MTP). Leptin, a hormone secreted by adipose tissue, acts in the brain and on peripheral tissues to regulate fat storage and metabolism. Our aim was to identify the role of leptin signaling in MTP regulation and lipid absorption using several mouse models deficient in leptin receptor (LEPR) signaling and downstream effectors. Mice with spontaneous LEPR B mutations or targeted ablation of LEPR B in proopiomelanocortin (POMC) or agouti gene related peptide (AGRP) expressing cells had increased triglyceride in plasma, liver, and intestine. Furthermore, melanocortin 4 receptor (MC4R) knockout mice expressed a similar triglyceride phenotype, suggesting that leptin might regulate intestinal MTP expression through the melanocortin pathway. Mechanistic studies revealed that the accumulation of triglyceride in the intestine might be secondary to decreased expression of MTP and lipid absorption in these mice. Surgical and chemical blockade of vagal efferent outflow to the intestine in wild-type mice failed to alter the triglyceride phenotype, demonstrating that central neural control mechanisms were likely not involved in the observed regulation of intestinal MTP. Instead, we found that enterocytes express LEPR, POMC, AGRP, and MC4R. We propose that a peripheral, local gut signaling mechanism involving LEPR B and MC4R regulates intestinal MTP and controls intestinal lipid absorption.

Dietary fat is a major source of energy. Biosynthesis of apolipoprotein B (apoB)-containing lipoproteins in the gut is required for the absorption of fat into the body. Biosynthesis of apoB-containing lipoproteins is critically dependent on the activity of microsomal triglyceride transfer protein (MTP) that mainly resides in the lumen of the endoplasmic reticulum. MTP physically binds to apoB and transfers lipids from membranes/lipid droplets to nascent apoB. These activities are believed to play critical roles in the lipidation of nascent apoB and in the formation of primordial lipoprotein particles ( 1 ). ApoB and MTP are also expressed in the liver to mobilize hepatic triacylglycerol by synthesizing very low density lipoproteins (2)(3)(4). Regulation of MTP expression at transcriptional and posttranscriptional levels is a major determinant of hepatic and intestinal fat mobilization ( 5 ), and recent evidence indicates that MTP expression is differentially regulated in the liver and intestine ( 6 ). For example, hepatic, but not intestinal, MTP is increased by peroxisome proliferator-activated receptor ␣ (PPAR ␣ ) agonists ( 7 ). In the intestine, but not in the liver, inositol-requiring enzyme 1 ␤ (IRE1 ␤ ) has been suggested to posttranscriptionally degrade MTP mRNA ( 6 ). Specifi c intestinal inhibition of MTP is considered a viable therapeutic approach to lower plasma Abstract Fat is delivered to tissues by apoB-containing lipoproteins synthesized in the liver and intestine with the help of an intracellular chaperone, microsomal triglyceride transfer protein (MTP). Leptin, a hormone secreted by adipose tissue, acts in the brain and on peripheral tissues to regulate fat storage and metabolism. Our aim was to identify the role of leptin signaling in MTP regulation and lipid absorption using several mouse models defi cient in leptin receptor (LEPR) signaling and downstream effectors. Mice with spontaneous LEPR B mutations or targeted ablation of LEPR B in proopiomelanocortin (POMC) or agouti gene related peptide (AGRP) expressing cells had increased triglyceride in plasma, liver, and intestine. Furthermore, melanocortin 4 receptor (MC4R) knockout mice expressed a similar triglyceride phenotype, suggesting that leptin might regulate intestinal MTP expression through the melanocortin pathway. Mechanistic studies revealed that the accumulation of triglyceride in the intestine might be secondary to decreased expression of MTP and lipid absorption in these mice. Surgical and chemical blockade of vagal efferent outfl ow to the intestine in wild-type mice failed to alter the triglyceride phenotype, demonstrating that central neural control mechanisms were likely not involved in the observed regulation of intestinal MTP. Instead, we found that enterocytes express LEPR, POMC, AGRP, and MC4R. We propose that a peripheral, local gut signaling mechanism involving LEPR B and MC4R regulates intestinal MTP and controls intestinal lipid absorption. important role in the transport of triacylglycerol with chylomicrons ( 44 ). Chylomicrons are the major apoB-containing lipoproteins synthesized and secreted with the help of MTP by enterocytes to transport dietary fat and fat-soluble vitamins into the circulation. Consequently, we studied the role of LEPR in the regulation of lipid absorption by the intestine. Our studies identifi ed local gut LEPR and MC4R signaling mechanisms in enterocytes that control MTP expression and lipid absorption.

Hepatic and celiac vagotomies
Hepatic and celiac vagotomies were performed as previously described ( 48 ). Briefl y, animals were anesthetized and laparotomy was performed. The stomach was exposed and hepatic and celiac branches of the abdominal vagus were identifi ed. Sham vagotomy mice were then sutured and treated with antibodies. The hepatic and celiac nerve bundles of mice in the vagotomy groups were isolated on a fi ne wire loop and rapidly heated until the loop cut through the nerve. Mice were then closed up and treated with antibiotics. All the mice were euthanized two weeks after the vagotomies.

Plasma lipid measurements
Total plasma cholesterol and triglyceride levels were measured using kits (Thermo Fisher Scientifi c). HDL lipid levels were measured after precipitating apoB lipoproteins ( 49 ). Lipids in apoB lipoproteins were determined by subtracting HDL lipids from total lipids. Plasma was pooled from all the animals in each group, and lipoproteins were separated by gel fi ltration (fl ow rate of 0.2 ml/min) using a Superose 6-column, and 200-µL fractions were collected.  ( 49 ). Two h after the gavage, plasma was collected and used to measure radioactivity and plasma lipids. In a separate experiment to study hepatic lipoprotein production, overnight fasted Pomc-CRE Lepr fl /fl , and Lepr fl /fl lipids ( 8 ), as hepatic MTP inhibition might promote hepatos teatosis and increased plasma transaminases. Therefore, new pharmacologic interventions to curb intestinal MTP activity are underway ( 9 ). Identifi cation of novel tissuespecifi c regulatory mechanisms might be benefi cial in targeting intestinal MTP to lower plasma lipids and avoid deleterious hepatic side effects associated with its inhibition. Leptin is released by adipose tissue into the circulation in proportion to the amount of stored lipid, and it plays an important role in the control of metabolism ( 10,11 ). Leptin acts on the central nervous system (12)(13)(14) by binding to its receptor (LEPR-B) to decrease food intake and increase energy expenditure ( 15 ), as well as to regulate body weight and adipose tissue mass ( 10,12,15,16 ). Two types of leptin-responsive neurons, orexigenic neuropeptide Y/agouti gene related peptide (NPY/AGRP) neurons and anorexigenic proopiomelanocortin/cocaine amphetamine regulated transcript (POMC/CART) neurons, modulate ingestive behavior and metabolism ( 17,18 ), and they are regulated in a reciprocal manner by leptin ( 17,19,20 ). Leptin reduces body weight by inhibiting AGRP neurons and stimulating anorexigenic POMC neurons. In the hypothalamic arcuate nucleus, leptin induces production and release of POMC, a polypeptide precursor that is cleaved to generate ␣ -and ␥ -melanocyte-stimulating hormone (MSH). Interaction of MSH with melanocortin 4 receptor (MC4R) in the hypothalamus decreases food intake and increases energy expenditure ( 21 ). In contrast, the MC4R antagonist AGRP increases food intake and decreases energy expenditure ( 22 ).

Short-term lipid absorption and hepatic lipoprotein production studies
Mice bearing mutations either in the leptin ( Lep ob/ob ) or leptin receptor ( Lepr db/db ) genes show high AGRP and low POMC neuronal activities; the combination of these changes leads to increased food intake and reduced energy expenditure, causing profound obesity (23)(24)(25). Decreased hypothalamic Pomc and increased Agrp mRNA levels in Lep ob/ob mice can be normalized by exogenous leptin administration (26)(27)(28)(29). Furthermore, transgenic complementation of the LEPR-B in neurons of Lepr db/db mice ( 30 ), viral-mediated expression of LEPRs into the arcuate nucleus of LEPR-defi cient rodents ( 31,32 ), or central leptin administration into the cerebral ventricles in Lep ob/ob and normal mice ( 33 ) reduces body weight and food intake. Similarly, a critical role for melanocortin signaling in the regulation of body weight has been established ( 34 ). Targeted deletion of Mc4r in mice results in obesity, hyperphagia, and reduced energy expenditure ( 35 ), and a similar phenotype is seen in humans with inactivating MC4R mutations ( 36 ).
Leptin also has a wide repertoire of peripheral effects, mediated indirectly through the central nervous system as well as via direct actions on target tissues. Direct leptin action suppresses insulin expression and secretion in pancreatic ␤ -cells ( 37,38 ); stimulates lipolysis and fatty acid oxidation in adipose tissue, skeletal muscle, and pancreas (39)(40)(41); decreases triglyceride content and secretion rates in liver ( 42 ); and inhibits apolipoprotein AIV transcription in the jejunum ( 43 ). The apolipoprotein AIV plays an tive RT-PCR. For semiquantitative RT-PCR, 50 ng of cDNA was used per 25 µl reaction volume. Transcripts were amplifi ed using the primers designed by using PrimerExpress 3.0 (Applied Biosystems). PCR was performed at 93°C for 3 min and then 35 cycles of 93°C for 30 s, 57°C for 30 s, and 72°C for 45 s. Final PCR amplifi cation products were subjected to agarose gel electrophoresis and ethidium bromide staining to confi rm product size. Quantitative RT-PCR was performed using SYBR Green to quantify changes in mRNA. The data were analyzed using the ⌬ ⌬ C T method and presented as arbitrary units.

Statistics
Data are presented as mean ± SD. Statistical signifi cance ( P < 0.05) was determined using the Student's t -test (GraphPad Prism).

Mutations in leptin receptor decrease intestinal MTP and increase tissue lipids
To evaluate the role of LEPR in lipid absorption, we fi rst measured steady-state plasma lipids in homozygous LEPR mutant ( Lepr db/db ) and heterozygous ( Lepr db/+ ) mice ( Table  1 ). Total plasma cholesterol and triglyceride were elevated signifi cantly in db/db mice relative to heterozygotes. Increased plasma cholesterol and triglycerides in db/db mice were due to 33% and 80% increases in HDL and non-HDL apoB lipoproteins ( Table 1 ), respectively. These augmentations in plasma lipoproteins are consistent with other studies and might be related to lower lipoprotein lipase activity in these mice ( 52 ). Besides increases in plasma lipids, we also observed that db/db mice accumulated significant amounts of triglycerides in the intestine and liver, without any signifi cant change in tissue cholesterol ( Table  2 ). Because mutations in Lepr increased intestinal and hepatic triglycerides, we wondered whether this increase was related to defects in apoB-lipoprotein assembly and secretion. MTP is an essential chaperone for the biosynthesis of these lipoproteins and is known to be modulated by various stimuli, such as dietary fat and carbohydrate ( 53 ), free mice were injected intraperitoneally with poloxamer 407 and plasma was collected after 3 h.

Determination of MTP activity
Samples from the liver (0.1 g) and proximal small intestine ( ‫ف‬ 1-2 cm) were homogenized in low-salt buffer, centrifuged, and supernatants were used in triplicate for protein determination and the MTP assay ( 50 ) using a kit (Chylos, Inc).

Western blot analyses
Tissues were homogenized in PBS containing 1% Triton-X 100. Proteins (50 g) were resolved on 4%-20% gradient gels (Biorad). A mouse IgG2a monoclonal anti-MTP antibody (BD), and a mouse anti-GAPDH (Santa Cruz) were used as primary antibodies to detect endogenous proteins. Alexa Fluor 633 (Invitrogen) anti-mouse or anti-rabbit secondary antibodies were used to visualize and quantify the blots using STORM 860 (Amersham).

Immunohistochemistry
For imaging of POMC and LEPR in the intestine, we performed immunohistochemistry on formaldehyde-fi xed tissues. LEPR was imaged using a Lepr-CRE Rosa26-fl oxed eGFP mouse line with GFP as the proxy for LEPR expression ( 51 ). Mice were initially perfused with saline followed by 10% formaldehyde. Tissues were embedded in Tissue-TEK OCT compound for cryostat sectioning. Sections (10 µm) were incubated with two primary antibodies (rabbit anti-ACTH [Phoenix Pharmaceuticals] and chicken anti-GFP [Aves Labs]) overnight. After washing, sections were incubated for two h with two secondary antibodies (Cy3 labeled goat anti-rabbit IgG and FITC labeled goat anti-chicken IgY; both from Invitrogen) and mounted in Vectashield (Vector Laboratories), a mounting medium containing DAPI. Antibody dilutions were empirically determined for optimal staining. Images were observed on a Zeiss Axioplan 2 fl uorescence microscope.

Semiquantitative and quantitative RT-PCR
Total RNA from tissues and cells were isolated using TriZol (Invitrogen). The purity was assessed by the A260/A280 ratio and RNAs with ratios more than 1.7 were used for cDNA synthesis. The fi rst strand cDNA was synthesized using Omniscript RT (Qiagen) kit, and then used for semiquantitative and quantita-  ( 54 ), and insulin ( 55 ). db/db mice had decreased intestinal MTP activity and mRNA ( Fig. 1A , B ), while these were unchanged in the liver compared with control Lepr db/+ mice ( Fig. 1C, D ). Therefore, the loss of LEPR signaling in Lepr db/db mice increased plasma apoB lipoproteins and tissue triglyceride while reducing intestinal MTP expression.

Ablation of leptin receptors in either POMC-or AGRP-expressing cells reduces intestinal MTP and increases tissue lipids
Leptin exerts its central effects on energy homeostasis in part by binding to its receptors expressed in POMC and   mination of MTP activity ( Fig. 2C ), mRNA ( Fig. 2D ), and protein ( Fig. 2G ) in the intestines of Pomc-Cre Lepr fl /fl and Agrp-Cre Lepr fl /fl mice revealed signifi cant decreases compared with control mice. In contrast, hepatic MTP activity ( Fig. 2E, F ), mRNA ( Fig. 2F ), and protein ( Fig. 2H ) did not differ as a function of genotype. Thus, deletion of LEPR in POMC and AGRP cells increases plasma lipoproteins, elevates intestinal and hepatic triglyceride, and reduces intestinal MTP.
due to increases in non-HDL apoB lipoproteins ( Table 1 ). Gel fi ltration also revealed that cholesterol in HDL lipoproteins ( Fig. 2A ) and triglycerides in non-HDL apoB lipoproteins ( Fig. 2B ) were increased in Pomc-Cre Lepr fl /fl and Agrp-Cre Lepr fl /fl mice. These cell type-specifi c LEPR knockout mice also accumulated more triglycerides in the intestine and liver compared with wild-type controls, but there were no signifi cant differences in intestinal and hepatic cholesterol between these groups ( Table 2 ). Deter-  (Panels A and B). Representative data from one mouse in each group are shown. Intestinal and hepatic samples were used to measure MTP activity (Panels C and E, respectively) and to isolate total RNA for the measurement in triplicate of MTP (Panels D and F, respectively) and ARPp0 mRNA. The ratio between gene of interest and ARPp0 in WT mice was used to normalize mRNA levels in all samples. Intestinal and hepatic samples were also used to measure MTP protein by western blotting (Panels G and H, respectively). Values are mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001 compared with WT. AGRP, agouti gene related peptide; LEPR, leptin receptor; MTP, microsomal triglyceride transfer protein; POMC, proopiomelanocortin; WT, wild-type.
We also studied lipoprotein production by the liver in wild-type and Pomc-Cre Lepr fl /fl mice after inhibiting lipoprotein lipase by poloxamer 407. There was a signifi cant increase (14.5%) in total plasma triglyceride ( Fig. 3G ) due to an increase in apoB lipoproteins ( Fig. 3H ) but not HDL lipoproteins ( Fig. 3I ), indicating that the Pomc-Cre Lepr fl /fl liver assembles and secretes higher amounts of lipoproteins than wild-type liver. Since MTP levels are not changed in Pomc-Cre Lepr fl /fl liver, we suggest that these increases in lipoprotein production might be secondary to increased lipid accumulation in the liver. In short, these studies demonstrate signifi cantly lower intestinal and higher hepatic lipoprotein production in LEPR-defi cient mice.

Partial intestinal defi ciency reduces triglyceride absorption
To test the hypothesis that reduced MTP leads to decreased intestinal lipid absorption and lipoprotein production, we obtained partial intestine-specifi c Mttp gene deletion by crossing Mttp fl /fl mice with villin-Cre mice. Analysis of MTP activity and mRNA in the intestine of villin-CRE Mttp fl /+ showed a decrease of 60% and 71%, respectively, compared with Mttp fl /fl ( Fig. 4A , B ), indicating partial gene ablation. There was no change in the MTP activity and mRNA levels in the livers of these mice ( Fig.  4C, D ). These studies indicate partial intestine-specifi c

Leptin receptor ablation in POMC neurons results in reduced fat absorption
Experiments were then performed to determine the consequences of reduced intestinal MTP expression on lipoprotein production and tissue triglyceride. Pomc-Cre Lepr fl /fl mice were fasted overnight and injected with poloxamer 407, a lipoprotein lipase inhibitor that blocks lipoprotein catabolism leading to accumulation of lipids in the plasma ( 57 ), and gavaged with radiolabeled triolein along with olive oil. Intestinal absorption of radiolabeled triolein was reduced in these mice, mainly due to a significant reduction ( ‫ف‬ 55%) in triglycerides present in apoB lipoproteins ( Fig. 3A -C ). Absorption of total lipids was also measured by quantifying plasma triglyceride and cholesterol mass. Total triglycerides were signifi cantly reduced ( ‫ف‬ 15%) in plasma and apoB lipoproteins ( Fig. 3D, E ) of Pomc-Cre Lepr fl /fl mice with no signifi cant difference in HDL triglycerides ( Fig. 3F ). Lesser percent reduction in total mass compared with those seen in radiolabeled triglycerides might be due to higher amounts of plasma triglycerides in these mice and secretion of apoB lipoproteins from the liver. No signifi cant differences were observed in plasma cholesterol (data not shown). These studies indicate that reduced expression of intestinal MTP in Pomc-Cre Lepr fl /fl mice is associated with reduced absorption of dietary triglycerides and reduced intestinal lipoprotein production. triglycerides were increased, intestinal cholesterol was reduced, and hepatic cholesterol was elevated ( Table 2 ). Further analyses revealed decreased MTP activity ( Fig. 5C ) and MTP mRNA levels in the intestine ( Fig. 5D ) but no signifi cant change in the liver ( Fig. 5E, F ). Hence, Mc4r Ϫ / Ϫ mice showed decreased intestinal MTP expression and increased intestinal and plasma triglyceride, similar to LEPRdefi cient mice. These studies indicate that reductions in leptin and melanocortin signaling diminish intestinal MTP expression.

The central nervous system is not required to regulate intestinal MTP
As central leptin signaling has been implicated in the control of multiple determinants of energy balance, including the control of adipose tissue, hepatic glucose production, and food intake, we next evaluated the role of neural communication between brain and gut in the regulation of intestinal MTP levels and activity. Mice were subjected to sham (Sham) surgical operations, hepatic vagotomy (HV), celiac vagotomy (CV), or combined hepatic and celiac vagotomy (HCV). Celiac branch vagotomy alone interrupts a large portion of the vagal efferent  ( Fig. 4E ); this decrease was mainly due to decrease in non-HDL lipoprotein radiolabel triolein counts ( Fig. 4F ). Thus, partial intestine-specifi c MTP defi ciency leads to substantial decrease in acute triglyceride absorption.

MC4R deletion increases plasma lipids and decreases intestinal MTP
Since many of leptin's actions are mediated by stimulation of the melanocortin system ( 26,(58)(59)(60), we wondered whether MC4R knockout ( Mc4r Ϫ / Ϫ ) mice would have intestinal and hepatic lipid profi les similar to LEPR knockout mice. Mc4r Ϫ / Ϫ mice showed a signifi cant increase in total plasma cholesterol and triglyceride levels ( Table 1 ). The increase in cholesterol was due to increases in both HDL and non-HDL lipoproteins. Fast-protein liquid chromatography (FPLC) analysis showed a broad distribution of cholesterol across different lipoproteins ( Fig. 5A ). Triglycerides were high in apoB lipoproteins but not in HDL ( Fig. 5B ). In Mc4r Ϫ / Ϫ mice intestinal and hepatic signaling but may instead refl ect the local actions of leptin at the level of enterocytes.

Expression of LEPR, POMC, AGRP, and MC4R in enterocytes
Although complete or tissue-specifi c deletions of LEPRs modulate intestinal lipid and MTP levels, the above data suggested that extrinsic innervation of the intestine is not required for this modulation. Hence, we hypothesized that leptin might control intestinal lipid levels and MTP expression via local signaling mechanisms intrinsic to the gut. Consequently, we examined the expression of different genes involved in leptin signaling ( Fig. 7 ). Semiquantitative PCR analysis showed that LEPR, POMC, AGRP, and MC4R were expressed in the intestine as well as in the hypothalamus ( Fig. 7A ). In the intestine, LEPRs might be expressed at nerve endings or in enterocytes. To identify the site of expression of these molecules, we isolated innervation of the small intestine, while hepatic vagotomy eliminates vagal hepatic efferent innervation as well as a small portion of the intestinal nerve supply. Therefore, the combined hepatic and celiac vagotomy condition produces a larger degree of denervation of the intestine than either celiac or hepatic vagotomy alone. Both MTP activity as well as mRNA levels remained unchanged in the intestine and liver of these mice compared with sham ( Fig.  6A -D ). These data indicate that vagal innervation is not required for the regulation of intestinal MTP. To further validate these observations, we used chemical inhibition to interfere with neuronal communication. Follow-up studies using peripheral atropine to block all vagal cholinergic outfl ow to the entire intestine also failed to affect intestinal and hepatic MTP activity ( Fig. 6E, G ) and mRNA ( Fig.  6F, H ). Taken together, these data suggest that our observations of reduced intestinal MTP in LEPR signalingdefi cient mice are not due to altered leptinergic brain-gut ) mice were subjected to gel fi ltration and fractions were analyzed for cholesterol (Panel A) and triglycerides (Panel B). Intestinal and liver samples from these animals (n = 6 per group) were used for MTP activity assays (Panels C and E). RNA was used to measure in triplicate MTP (Panels D and F) and ARPp0 mRNA. The ratio between MTP and ARPp0 in Mc4r +/+ mice was used to normalize mRNA levels. Values are mean ± SD. * P < 0.05, *** P < 0.001 compared with wild-type mice. MC4R, melanocortin 4 receptor; MTP, microsomal triglyceride transfer protein.
suggested that LEPR, POMC, AGRP, and MC4R are expressed in the enterocytes ( Fig. 7A ). To further study the expression of these genes, we used Lepr-CRE Rosa26fl oxed eGFP mouse line with GFP as the proxy for LEPR expression ( 51 ). Expression of POMC and LEPR was studied by immunohistochemistry using specifi c antibodies. Immunopositive labeling for POMC indicate that enterocytes enterocytes from the intestine and evaluated the remaining tissue as a source of smooth muscle harboring nerve endings. We used neuron-specifi c enolase ( 61 ) to document that isolated enterocytes were not contaminated with nerve cells ( Fig. 7A ). Neuronal enolase was not present in enterocyte preparations, indicating successful separation of enterocytes from nerve cells. Semiquantitative analysis Fig. 6. Surgical transection and chemical inhibition of extrinsic vagal gut neuronal communication has no effect on intestinal MTP. Panels A-D: Hepatic and celiac vagotomies in male C57BL6/J mice were performed as previously described ( 48 ). Mice were allowed to recover for 2 weeks and fasted overnight before euthanizing. Intestinal and hepatic tissues from hepatic vagotomy (HV, n = 13), celiac vagotomy (CV, n = 9), hepatic and celiac double vagotomy (HCV, n = 7), and sham (Sham, n = 21) operated mice were used for MTP activity assays (Panels A and C). RNA was used to measure in triplicate MTP (Panels B and D), and ARPp0 mRNA. The ratio between gene of interest and ARPp0 in sham mice was used to normalize mRNA levels in all samples. Panels E-H: Male C57BL6/J mice (5 per group) were injected intraperitoneally with atropine methyl nitrate (10 mg/kg). After 3 h of fasting, intestinal and liver samples were used to measure MTP activity (Panels E and G, respectively) and mRNA levels (Panels F and H, respectively). Values are mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.001 compared with Sham. HCV, hepatic and celiac vagotomy; HV, hepatic vagotomy; MTP, microsomal triglyceride transfer protein.
7E ) showed 57% and 59% reductions in LEPR and POMC, respectively. These studies indicate that enterocytes express several components of the LEPR signaling and that expression of Cre recombinase under the control of the POMC promoter results in partial ablation of POMC and LEPR-B. To begin to address why hepatic MTP expression was intact in LEPR-defi cient animals, we examined expression of leptin signaling pathway genes in the liver. While express POMC ( Fig. 7B ). Furthermore, expression of GFP in the enterocytes also suggests that these cells express LEPR-B ( Fig. 7C ).
Next, we looked at the expression of LEPRs and other signaling molecules in the enterocytes of Pomc-CRE Lepr fl /fl mice. Semiquantitative analyses showed reduced expression of LEPR and POMC with no effect on AGRP and MC4R in these mice ( Fig. 7D ). Quantitative analyses ( Fig.   Fig. 7. Enterocytes express LEPR, POMC, AGRP, and MC4R. Semiquantitative RT-PCR was used to determine the expression of LEPR-B, POMC, AGRP, and MC4R in the hypothalamus, intestine, enterocytes, and smooth muscle cells (Panel A). Immunohistochemistry to visualize expression of POMC (Panel B) and LEPR (Panel C) in Lepr-CRE Rosa26-fl oxed GFP mice was performed as described in "Materials and Methods." Expression of LEPR, POMC, AGRP, and MC4R was also determined by semiquantitative RT-PCR in the enterocytes isolated from WT and Pomc-Cre Lepr fl /fl mice (Panel D). Quantitative RT-PCR was used to measure changes in the expression of these genes in enterocytes (Panel E). Expression of LEPR, POMC, AGRP, and MC4R was also determined by semiquantitative RT-PCR in the liver isolated from WT mouse (Panel F). AGRP, agouti gene related peptide; LEPR-B, leptin receptor B; MC4R, melanocortin 4 receptor; POMC, proopiomelanocortin; WT, wild-type. pine blockade of all cholinergic vagal efferent transmission ( 69 ) failed to affect intestinal MTP expression. These data support the interpretation that parasympathetic neuronal communication between the brain and the intestine is not important in the regulation of intestinal MTP by leptin. Instead, our data point to a direct role of leptin receptors in enterocytes.
It has previously been shown that enterocytes express LEPRs and exhibit increased signal transducer and activator of transcription 3 (STAT3) phosphorylation in response to leptin ( 43 ). We extend these studies and show for the fi rst time that enterocytes also express MC4R, POMC, and AGRP. Furthermore, we show that expression of the Cre recombinase under POMC promoter results in partial ablation of fl oxed LEPR. These studies highlight the presence of a local, autocrine, regulatory loop where LEPRs regulate the expression of POMC and AGRP in enterocytes. The POMC products bind to MC4R and regulate MTP. It is tempting to speculate that LEPR-signaling mechanisms might have evolved in the gut and later diversifi ed into two different sets of neurons to regulate various aspects of energy metabolism.
POMC and AGRP are expressed in distinct hypothalamic neuronal populations that exert opposing effects on melanocortin receptor signaling and energy balance. Therefore, similar decreases in intestinal MTP activity in mice subjected to ablation of LEPRs after the expression of Cre-recombinase using either POMC or AGRP promoters were perplexing ( Fig. 2 ). Subsequent analyses of POMC and AGRP, however, revealed that enterocytes express both of these genes. Furthermore, activation of Cre-recombinase by POMC promoter resulted in a partial deletion of LEPR in enterocytes. We interpret these data to suggest that LEPR expression in enterocytes is infl uenced by both POMC and AGRP. Thus, defi cient LEPR signaling in both Pomc-CRE Lepr fl /fl and Agrp-CRE Lepr fl /fl mice leads to comparable intestinal MTP defi ciencies in these strains.
Most of the mouse models we tested in this study are also hyperinsulinemic. Although we did not rule out the effect of hyperinsulinemia on the intestinal MTP activity and mRNA levels, hyperinsulinemic mice have higher intestinal ( 70 ) and hepatic ( 71 ) MTP expression and plasma triglyceride. Similarly, diabetic men and rabbits have been shown to express signifi cantly higher intestinal MTP mRNA levels compared with nondiabetic subjects ( 72,73 ). Contrary to these results, we observed decreased intestinal expression of MTP, and this decreased activity of MTP may have led to increased accumulation of triglycerides in the intestine.

Differential regulation of intestinal and hepatic MTP by leptin receptor signaling
The data presented in this study highlight differences in the control of intestinal and hepatic MTP expression in distinct mouse models of LEPR defi ciency. LEPR deficiency decreases MTP expression in the intestine but has no effect on hepatic MTP expression. Both enterocytes and hepatocytes express LEPR and respond to leptin ( 43,74,75 ). Therefore, it was surprising that intestinal cells LEPR-B was expressed in liver, POMC, AGRP, and MC4R were not detectable ( Fig. 7F ). Therefore, ablation of LEPR had no signifi cant effect on hepatic MTP expression due to the absence or low-level expression of POMC, AGRP, and MC4R.

DISCUSSION
These studies demonstrate that LEPR regulates intestinal MTP in mouse enterocytes independently of extrinsic vagal innervation of the gut. Furthermore, melanocortins and MC4 receptors localized to enterocytes likely participate in this regulation. Such an enterocyte-specifi c regulatory mechanism could provide an additional peripheral control of fat mobilization independently of that determined by central nervous system leptin signaling.

Regulation of intestinal MTP and lipid absorption by leptin receptor signaling
Leptin is an adipokine secreted when adipose tissue receives dietary lipids to inhibit food intake. We had anticipated that high leptin concentrations would decrease intestinal MTP expression and thereby limit lipid absorption as part of an energy regulatory response to lower energy intake. Hence, we had predicted that LEPR defi ciency would increase MTP expression and enhance lipid absorption. Instead, we observed signifi cant decreases in intestinal MTP expression and lipid absorption. It is unclear whether the observed reductions in intestinal MTP and lipoprotein production are causally related to LEPR deficiency or are secondary to enhanced hyperlipidemia. Most likely, these changes represent a long-term adaptation to chronic LEPR defi ciency and plasma lipid accumulation. Such an adaptation would help lower exogenous lipid absorption and thereby attenuate hyperlipidemia and obesity. Alternatively, these observations point to novel mechanisms that might exist in the intestine to lower intestinal lipoprotein production during fasting to facilitate endogenous lipid mobilization by the liver. Another possibility is that intestinal MTP is regulated by gastric leptin. It is known that stomach is another major organ, besides adipocytes, that synthesizes and secretes leptin. Some stomach-derived leptin is not proteolytically degraded in the gastric juice ( 43,62 ) and may reach the intestine in an active form (63)(64)(65). Therefore, gut leptin, which is rapidly secreted in the lumen after a meal (in contrast to slow secretion of leptin by adipocytes as a satiety signal), may be a key molecule controlling intestinal nutrient absorption. Leptinergic regulation of intestinal MTP might represent a postprandial response of jejunal enterocytes to promote the absorption of dietary fat.
In addition to its well-established role in the hypothalamic control of food intake and energy balance ( 10,12,15,16,66 ), leptin has also been reported to play a role in peripheral tissue biology ( 38,67,68 ). Our study shows that MTP expression in the intestine is not affected by vagal efferent outfl ow, as both selective and combined hepatic and intestinal vagotomies had no effect on intestinal MTP mRNA or activity levels ( Fig. 6 ). Furthermore, atro-respond to LEPR defi ciency and regulate MTP expression and lipid absorption, but hepatocytes do not. This dichotomy could be explained by our observations that enterocytes express POMC, AGRP, and MC4R, while hepatocytes do not express these genes. It has previously been shown that liver does not express MC4R ( 76 ); we are unaware of studies describing POMC and AGRP expression in the liver. Therefore, we speculate that differential regulation of hepatic and intestinal MTP by LEPR might be secondary to the expression melanocortins and MC4R in these cells.

Leptin receptor defi ciency affects plasma and tissue lipid metabolism
LEPR defi ciency is known to cause hyperlipidemia in db/db mice. Here, we report for the fi rst time that ablation of LEPR in either POMC-or AGRP-expressing cells also causes hyperlipidemia. Furthermore, we observed that whole body or POMC/AGRP cell-specifi c defi ciency in LEPR results in signifi cant accumulation of hepatic and intestinal triglycerides. Therefore, LEPR defi ciency results in lipid accumulation in the intestine, liver, and plasma. Experiments to delineate molecular mechanisms for the accumulation of triglycerides in the intestine revealed that defi ciency in LEPR signaling decreases intestinal MTP expression, reduces assembly and secretion of lipoproteins, and elevates triglyceride accumulation. However, there were no signifi cant reductions in hepatic MTP activity and mRNA; therefore, the hepatic steatosis observed in LEPRdefi cient mice might not be due to MTP defi ciency. Instead, we speculate that excess delivery of lipoproteins secondary to enhanced accumulation of plasma lipoproteins might contribute to hepatic triglyceride accumulation in LEPR-defi cient mice. Another possibility is that these mice have reduced oxidation of triglycerides due to decreased leptin signaling ( 45 ).
Similarly, our proposed mechanisms for intestinal triglyceride accumulation do not explain enhanced assimilation of lipoproteins in the plasma; reduced lipid absorption and lipoprotein secretion by the intestine is expected to lower plasma lipoproteins. Three possibilities may account for the hyperlipidemias in these mice. First db/db mice have reduced lipoprotein lipase activity ( 52 ), thereby reducing lipoprotein catabolism, while promoting hyperlipidemia. Second, total lipid oxidation is reduced in db/db mice and other cell-specifi c LEPR defi ciencies, causing an accumulation of fat ( 45 ). Overall decreases in lipid utilization would promote lipid accumulations in serum and tissue. Third, we observed increases in hepatic lipoprotein production, suggesting that enhanced hepatic lipoprotein production might also contribute to hyperlipidemia. Therefore, different mechanisms might lead to lipid accumulation in plasma, liver, and intestine in LEPR-defi cient mice.

CONCLUSION
These studies identify intestine-specifi c regulatory mechanisms involving LEPR, POMC, and AGRP that control MTP expression and lipid absorption. Understanding signaling pathways that lead to intestine-specifi c MTP