PPARα (Peroxisome Proliferator-activated Receptor α) Activation Reduces Hepatic CEACAM1 Protein Expression to Regulate Fatty Acid Oxidation during Fasting-refeeding Transition

Abstract

Carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) is expressed at high levels in the hepatocyte, consistent with its role in promoting insulin clearance in liver. CEACAM1 also mediates a negative acute effect of insulin on fatty acid synthase activity. Western blot analysis reveals lower hepatic CEACAM1 expression during fasting. Treating of rat hepatoma FAO cells with Wy14,643, an agonist of peroxisome proliferator-activated receptor α (PPARα), rapidly reduces Ceacam1 mRNA and CEACAM1 protein levels within 1 and 2 h, respectively. Luciferase reporter assay shows a decrease in the promoter activity of both rat and mouse genes by Pparα activation, and 5′-deletion and block substitution analyses reveal that the Pparα response element between nucleotides −557 and −543 is required for regulation of the mouse promoter activity. Chromatin immunoprecipitation analysis demonstrates binding of liganded Pparα to Ceacam1 promoter in liver lysates of Pparα^+/+, but not Pparα^−/− mice fed a Wy14,643-supplemented chow diet. Consequently, Wy14,643 feeding reduces hepatic Ceacam1 mRNA and CEACAM1 protein levels, thus decreasing insulin clearance to compensate for compromised insulin secretion and maintain glucose homeostasis and insulin sensitivity in wild-type mice. Together, the data show that the low hepatic CEACAM1 expression at fasting is mediated by Pparα-dependent mechanisms. Changes in CEACAM1 expression contribute to the coordination of fatty acid oxidation and insulin action in the fasting-refeeding transition.

Introduction

Plasma insulin levels are determined by several factors, including insulin secretion from pancreatic β-cells and insulin clearance (1, 2). Carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1),⁴ a substrate of the insulin receptor tyrosine kinase (3), promotes insulin clearance by up-regulating receptor-mediated insulin endocytosis and degradation in a phosphorylation-dependent manner (4). Its specific inactivation in liver impairs insulin clearance to cause hyperinsulinemia and ensuing insulin resistance with increased hepatic steatosis (5).

CEACAM1 is ubiquitously produced with a predominant expression in liver by comparison to kidney and a limited expression in white adipose tissue and skeletal muscle (6). Physiologically, this is consistent with the major role of the liver in insulin clearance by comparison to kidney and the insignificant involvement of the other insulin target tissues in insulin extraction. The mechanistic underpinning of this hierarchal expression profile relates to the up-regulation of Ceacam1 promoter activity by insulin (7), the level of which is 2–3-fold higher in the portal relative to the systemic circulation (8).

In addition to CEACAM1, the hepatocyte is home to the highest level of fatty acid synthase, a lipogenic enzyme that catalyzes the conversion of malonyl-CoA to palmitic acid. Similar to CEACAM1, this high level of expression of fatty acid synthase is mediated by transcriptional up-regulation by insulin (9). However, despite the abundance of this enzyme, fatty acid synthase activity in liver is restricted even under stimulating feeding conditions, during which carbohydrates undergo glycolysis and their products are converted to fatty acids to contribute to energy sources at fasting. We have shown (10) that this counter-regulation of fatty acid synthase activity in liver is mediated by increased CEACAM1 phosphorylation and binding to fatty acid synthase in response to transient pulses of insulin (11).

At fasting, when insulin secretion is low, metabolism shifts from glycolysis to lipolysis, during which free fatty acids are released from white adipose tissue into the liver to undergo β-oxidation and supply energy to the heart and brain. This mechanism is largely supported by the activation of peroxisome proliferator-activated receptor α (PPARα), a nuclear transcription factor that induces the expression of genes involved in fatty acid transport to mitochondria to undergo β-oxidation (12–16). Acutely after refeeding (∼8 h), PPARα plays a significant role in maintaining fatty acid β-oxidation during the stepwise reversal of the fasting metabolic state by insulin surges until glycogen stores are replenished (17), at which point the levels of malonyl-CoA are restored, β-oxidation stops, and glycolysis resumes (18).

Because CEACAM1 plays an important regulatory role in insulin and fatty acid homeostasis, we herein examined whether it is itself metabolically regulated during the fasting-refeeding transition and identified the underlying mechanisms.

Experimental Procedures

Animal Husbandry

Male mice were kept in a 12-h dark/light cycle. All procedures were approved by the Institutional Animal Care and Utilization Committee. Wild-type Pparα^+/+ and Pparα^−/− null mice propagated on the C57BL/6 background were from Taconic Biosciences (Cambridge City, IN). Male mice (2–4 months of age) were fed ad libitum a standard chow (Harlan Teklad 2016; Harlan, Haslett, MI). In some experiments mice were fed for 3–7 days a chow diet powdered and mixed in a geometric proportion with 0.1% w/w of Wy14,643 (Enzo Life Sciences, Farmingdale, NY), a PPARα agonist.

Plasma Biochemistry

After an overnight fast, mice were anesthetized at 1100 h. Whole venous blood was drawn to measure the levels of glucose, plasma insulin, and C-peptide levels (19). Glycogen content was measured as described previously (20).

Insulin Secretion and Glucose and Insulin Tolerance Tests

Awake overnight-fasted mice were injected intraperitoneally with 1.5 g/kg body weight (BW) dextrose solution before measuring glucose in tail blood. For insulin tolerance, mice were fasted for 6 h and intraperitoneally injected with human regular insulin (Novo Nordisk, 0.75 units/kg BW), and their glucose was measured as previously described (21). For glucose-stimulated insulin secretion, mice were fed a Wy-supplemented diet for 4 days, fasted overnight, and injected intraperitoneally with 3 g/kg BW of dextrose solution. Retro-orbital blood was removed at 0–30-min post-injection to measure plasma insulin levels using RIA kit (Millipore, Temecula, CA).

Fatty Acid Synthase Activity

This was measured by the incorporation of radiolabeled malonyl-CoA into palmitate, as described previously (10). Briefly, 60–100 mg of liver tissue was homogenized in DTT-containing Tris-buffer, pH 7.5, and centrifuged at 4 °C, and the supernatant was incubated for 20 min at 37 °C with 0.1 μCi of [¹⁴C]malonyl-CoA and 25 nm malonyl-CoA in the absence or presence of 500 μm NADPH. The reaction was stopped with 1:1 chloroform/methanol solution, mixed, and centrifuged, and the supernatant vacuum-dried was resuspended in 200 μl of water-saturated butanol to be extracted. The butanol layer was counted, and values were expressed as relative cpm of ¹⁴C-incorporated/μg of protein.

Isolation of Primary Hepatocytes and Insulin Internalization

As previously described (21), hepatocytes were isolated by perfusing liver (1 ml/min) with buffer containing Collagenase Type II (1 mg/ml) (Worthington Biochem, Lakewood, NJ) (5). Cells were dispensed in Williams E complete media (Invitrogen) containing 10 mm lactate, 10 nm dexamethasone, 100 nm insulin, 10% FBS, and 1% penicillin-streptomycin, counted and plated onto 12-well cell culture plates at 2.5 × 10⁵/well density, and incubated at 37 °C for 48 h. Medium was changed 24 h after plating. [¹²⁵I]Insulin (Human [¹²⁵I]insulin, PerkinElmer Life Sciences) (30,000cpm) was allowed to bind at 4 °C for 5 h in cold KRP buffer (100 mm HEPES, pH 7.4, 120 mm NaCl, 1.2 mm MgSO₄, 1 mm EDTA, 15 mm CH₃COONa, 10 mm glucose, 1%BSA) before it was allowed to internalize at 37 °C for 0–90 min, as previously described (21).

Cloning of Mouse Ceacam1 Promoter

Functional mapping of the mouse Ceacam1 promoter revealed three potential peroxisome proliferator response elements (PPRE)/RXR at nucleotides (nts) −1056/−1037, −557/−543, −260/−248. In a PCR reaction, double-stranded genomic DNA spanning nt +30 to −1100 was synthesized and amplified in a polymerase chain reaction (PCR) using 100 ng of mouse genomic DNA as template and 1 μm concentrations of sense forward primer (nt −1100/−1086) containing an Xho1 flanking sequence (small letters in italics, 5′-ataccctcgagCCTAAGAAGCTTTAC-3′) and antisense primer (nt +30/+11) with BglII flanking sequence (small letters in italics, 5′-gaagatctTTTGTGGAGATGTGCTGAGG-3′). After the initial DNA denaturation at 94 °C for 5 min, 30 cycles of PCR were carried out as described previously (7). The 5′ deletion mutant was synthesized using the same PCR conditions but with a forward primer spanning nt −467 to −453 (5′-ataccctgctcgagTCAGTGACGATGGAT-3′). Amplified genomic DNA was subsequently purified and subcloned at the KpnI and BglII sites of pGL4.10 [luc2] BASIC promoterless firefly luciferase reporter plasmid (Promega Corp., Madison, WI).

Scanning Mutants of individual PPREs (nts −1056/−1037 (Δ1), nts −557/−543 (Δ2), and nts −260/−248 (Δ3)) were obtained in two sequential PCR reactions using the QuikChange II XL Site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA). The first PCR reaction utilized 100 ng of the PGL4.10 plasmid containing the genomic DNA sequence between nt −1100 and +30 as the template, with the reverse antisense nt +30/+11 primer and forward primers containing a sequence replacing each of the native PPRE site. The resulting PCR product was then used as the template to insert a mutation replacing the native RXR site downstream of each PPRE to mutate the PPRE/RXR site. The Δ1 mutations were: 5′-TAATCGA-3′ (for PPRE alone) and 5′-TAATCGAGCTAGT-3′ (for PPRE/RXR). The Δ2 mutations were 5′-CAATTCT-3′ (for PPRE alone) and 5′-CAATTCTATGAAATC-3′ (for PPRE/RXR). The Δ3 mutations were 5′-CTTTTCT-3′ (for PPRE alone) and 5′-CTTTTCTGTTATG-3′ (for PPRE/RXR). The resulting individual mutant products were used as templates to create any combinational mutations of the three PPRE/RXRs following a similar PCR-based strategy.

Cell Culture, Transfection, and Luciferase Assay

FAO rat hepatoma and human embryonic kidney cells (HEK293) were maintained at 5% CO₂ in Dulbecco-modified Eagle's medium (DMEM). Human liver carcinoma HepG2 cells were maintained in minimum essential medium α (α-MEM) (Cellgro, Manassas, VA) supplemented with 10% FBS, 1% l-glutamine, 100 units/ml penicillin (Sigma), and 10 mg/ml streptomycin (Sigma). Cells were serum-starved with phenol red-free medium (Invitrogen) supplemented with 25 mm HEPES and 0.1% BSA for 16 h treated with either ethanol (vehicle) or 30 μm Wy14, 643 (Enzo Life Sciences) for 24 h before they were subjected to Western and Northern blot analyses (below).

For luciferase assays cells were seeded at 4.4 × 10⁵ into 6-well plates and at ∼60–70% confluence, and transfection was performed with 500 ng of promoter constructs and 10 ng of Renilla luciferase (pRL-TK, Promega) using FuGENE 6 (Promega) per the manufacturer's instructions. Empty pGL4.10 vector was used as the negative control, and PPREx3-TK-luc (plasmid 1015; Addgene) (22) was used as a positive control. Twenty-four hours post-transfection, cells were serum-starved and treated with ethanol or 30 μm Wy14,643 for 24 h as above. Luciferase activity was assessed using the Dual-Luciferase Reporter Assay System (Promega).

To assay for luciferase activity of the rat promoter, FAO rat hepatoma and HEK293 cells were co-transfected with 300 ng of pGL3 containing the rat Ceacam1 promoter (nts −1609/−21) (7) and 30 ng of Renilla luciferase (pRL-TK) using Lipofectamine 2000 (Invitrogen) before they were treated for 24 h, and luciferase activity was assayed.

Chromatin Immunoprecipitation (ChIP)

As described (23), mice were fed a Wy14,643-supplemented diet for 7 days before liver extraction, grinding under liquid nitrogen, and cross-linking in 1% formaldehyde, 1× PBS at 37 °C for 20 min. Cross-linking was terminated with 125 mm glycine, and the cell pellet was washed with 1× PBS. Nuclei were lysed in SDS buffer (50 mm Tris-HCl, pH 8.1, 10 mm EDTA, 1% SDS, and protease inhibitors), chromatin was sheared by sonication, and the nuclei lysate was cleared by centrifugation at 50,000 × g for 30 min. The soluble chromatin was diluted 10-fold in 0.01% SDS, 1.1% Triton X-100, 1.2 mm EDTA, 16.7 mm Tris-HCl, pH 8.1, 167 mm NaCl and immunoprecipitated with an antibody against Pparα (H-98; Santa Cruz Biotechnology, Inc., Dallas, TX). The antibody-protein-DNA complex was isolated using magnetic beads conjugated with protein A (New England BioLabs, Inc, Ipswich, MA). The protein-DNA complex was washed, eluted in 50 mm NaHCO₃, 1% SDS, and left at 65 °C overnight to reverse cross-linking. After protein digestion with proteinase K for 1 h at 45 °C, DNA was purified by phenol/chloroform/isoamyl alcohol extraction and PCR-amplified using a proximal (forward, nts −562 to −541, 5′-TTTCCAGGACACACAGGTCACC-3′; reverse, nts −301 to −280, 5′-GCAGCAGGAGTTCCAGCATTTC-3′) and a distal primer set (forward, nts +6499 to +6520, 5′-GCTTCATTCCACTCTCTCCTTCC-3′; reverse, nt +6743 to +6764, 5′-CATACAGCCCAGCAGGTCTTTG-3′), which yielded a 282-bp and a 265-bp product, respectively. For the malic enzyme control, the primer set yielding a 190-bp was used (forward, 5′-TCTTGGTTTCCAGAACGCTCA-3′, and reverse, 5′-TCGTCCACTGTCCTCCCCTAAC-3′).

Western Blot Analysis

Proteins were analyzed by SDS-PAGE and immunoprobing with custom-made polyclonal antibodies against endogenous mouse CEACAM1 (α-CC1) (10) and rat CEACAM1 (α-rCC1; α-P3(Ex) raised in rabbit against three peptides (amino acids 49–64, 482–495, and 506–519)) and affinity-purified against the extracellular peptide (amino acids 49–64), and phospho-CEACAM1 (α-pCC1) against phosphorylated Tyr-488 and fatty acid synthase (α-Fasn) (10). Also used were antibodies against Pparα and CD36 (Santa Cruz). For normalization, membranes were reprobed with monoclonal antibodies against Actin, GAPDH, and tubulin (Sigma or Santa Cruz). Proteins were detected by enhanced chemiluminescence (ECL; Amersham Biosciences) and quantified by densitometry (Image J software) (National Institutes of Health, Bethesda, MD).

Northern Blot Analysis

mRNA was purified using TRIzol (Invitrogen) followed by the MicroPoly (A) Pure kit (Ambion, ThermoFisher Scientific) and analysis by probing with cDNAs for Pparα and Ceacam1 using the Random Primed DNA Labeling kit (Roche Applied Science) before reprobing and normalizing to β-actin.

Semi-quantitative Real-time RT-PCR

Total RNA was isolated with the PerfectPure RNA tissue kit (5 Prime) following the manufacturer's protocol. cDNA was synthesized with ImProm-II^TM reverse transcriptase (Promega) using 1 μg of total RNA and oligo(dT) primers (21). cDNA was evaluated with quantitative real-time-PCR (Step One Plus, Applied Biosystems, Waltham, MA). mRNA was normalized to Gapdh or 18S. Results are expressed in -fold change as the mean ± S.E.

Statistical Analysis

Data were analyzed with SPSS software by two-way analysis of variance or two-tailed Student's t test with GraphPad Prism 4 software. p < 0.05 was statistically significant.

Results

CEACAM1 Is Regulated by Fasting/Refeeding

After an overnight fast, mice were refed a regular chow diet for up to 24 h. As previously shown (10), insulin surged at 1, 4, and 7 h of refeeding (Fig. 1A). Consistent with its ability to increase Ceacam1 promoter activity (7), hepatic CEACAM1 protein content was induced in parallel to transient insulin surges (Fig. 1B). Insulin surge during refeeding also induced CEACAM1 phosphorylation, as Western blot analysis using an antibody against tyrosyl phosphorylated CEACAM1 (α-pCC1) shows (Fig. 1B). Moreover, the activity of Fasn is diminished in parallel to CEACAM1 phosphorylation (Fig. 1C) as previously reported (10). Of interest, the CEACAM1:Fasn protein ratio begins to drop at ∼8 h of refeeding (Fig. 1B), pointing to the possibility that the higher CEACAM1:Fasn ratio plays a role in maintaining low fatty acid synthase activity in the early hours of refeeding (24, 25). This could contribute to the regulation of Fasn substrate, malonyl-CoA, and hence fatty acid transport to mitochondria for β-oxidation, an essential step in glycogen repletion. As has been reported (26), replenishment of glycogen content in liver takes ∼8 h of refeeding (Fig. 1D), at which point CEACAM1 phosphorylation dropped (Fig. 1A) and Fasn enzymatic activity reciprocally increased (Fig. 1C).

FIGURE 1.

Changes in hepatic CEACAM1 protein content at fasting/refeeding and physiologic implications. A, mice were fasted (F) overnight and refed ad libitum for 1–24 h (Rfd) before assessing plasma insulin levels. B, liver lysates were subjected to Western analysis by immunoblotting (Ib) with antibodies against α-CEACAM1 (α-CC1) to assess changes in hepatic CEACAM1 protein levels, phospho-CEACAM1 (α-pCC1) to detect phosphorylated CEACAM1, and Fasn to detect protein content of fatty acid synthase. Gels were reimmunoblotted (reIb) with an antibody against actin to normalize for protein loading. Liver tissues were assayed for fatty acid synthase activity relative to microgram of proteins (C) and for glycogen content relative to wet tissue weight (D). Assays were carried out in triplicate and on more than three mice per each time point. Data are presented as the mean ± S.E. E, liver tissues were subjected to Northern analysis to assess hepatic Pparα mRNA levels followed by β-actin for normalization. Both Northern and Western gels represent more than three experiments done on more than three mice per each time point.

Pparα Activation Decreases Ceacam1 Expression

Because PPARα is activated at fasting to increase transcription of genes that are involved in fatty acid β-oxidation in liver (15), we then investigated whether it is implicated in the metabolically driven changes in hepatic CEACAM1 expression. Northern analysis shows a higher hepatic Pparα mRNA levels at fasting than the first few hours of refeeding (Fig. 1E).

To test the hypothesis that CEACAM1 is reduced by Pparα activation, we assessed the effect of PPARα agonist, Wy14,643 (Wy), on rat and mouse Ceacam1 promoter activity as well as on its mRNA and protein levels. Using a luciferase reporter assay (7), we show that Wy14,643 treatment reduced rat Ceacam1 promoter activity by ∼4-fold in rat hepatoma FAO cells (Fig. 2A, Wy-treated versus vehicle (Veh)-treated −1609pLuc). Similarly, Wy14,643 treatment decreased the promoter activity of the rat Ceacam1 promoter (−1609pLuc) in human embryonic kidney (HEK293) cells (not shown), with an expected lower potency than in rat cells (27). Furthermore, Wy14,643 decreased Ceacam1 mRNA and CEACAM1 protein content in rat hepatoma FAO cells beginning at 1 and 2 h, respectively (Fig. 2, B and C).

FIGURE 2.

Regulation of rat Ceacam1 expression by PPARα activation. A, a rat Ceacam1 promoter fragment spanning a genomic DNA sequence from nt −1609 to −21 was subcloned into a promoterless luciferase reporter plasmid in both sense and reverse orientations. The constructs were transiently co-transfected with Renilla luciferase (pRL) into rat FAO hepatoma cells and treated with 30 μm Wy,14,463 (Wy) before luciferase activity was assayed, determined relative to that in Renilla, and expressed as the mean ± S.D. of triplicate transfections in relative light units. The graph represents four separate experiments. B, rat FAO cells were treated with ethanol (−) or Wy (+) for 0–1 h and analyzed by Northern blot using a Gapdh cDNA probe for normalization. rCC1, rat CC1. C, Rat FAO cells were treated with ethanol (Veh) or Wy for 0–12 h and subjected to Western blot (Ib) to analyze CEACAM1 protein content followed by tubulin for normalization. Each gel represents more than three different experiments.

Like rat promoter, Wy14,643 treatment reduced mouse Ceacam1 promoter activity by ∼2-fold in human HepG2 hepatoma cells (Fig. 3B, Wy-treated versus Veh-treated −1100pLuc). 5′-Deletion analysis indicated that removing the genomic DNA region containing the two potential distal PPREs between nts −1100 and −467 abolished the repressive effect of Pparα activation on mouse Ceacam1 promoter activity (Fig. 3B, Wy-treated versus Veh-treated −467pLuc). To further identify the active PPRE in the mouse Ceacam1 promoter, we then carried out mutational mapping of the three potential PPRE/RXR sites in the promoter. As Fig. 3C indicates, mutating the sequence between nts −557 and −543 either individually (Δ2 construct) or together with the other two (Δ1,2; Δ2,3 and Δ1,2,3 constructs) completely abolished the repressive effect of Wy14,643 treatment on Ceacam1 promoter activity, as opposed to mutating the other two PPRE/RXR motifs alone (Δ1;Δ3 and Δ1,3 constructs). This points to the functional relevance of the PPRE/RXR sequence located between nts −557 and −543 in the regulation of mouse Ceacam1 promoter activity.

FIGURE 3.

The effect of PPARα activation on mouse Ceacam1 promoter activity. A, sequence analysis of the mouse Ceacam1 promoter revealed three potential PPRE/RXR elements for potential Pparα binding: nts −1056/−1037, −557/−543, and −260/−248. B, as above, 5′-deletion constructs from the mouse promoter were subcloned into pGL4.10 promoterless luciferase reporter plasmid and transfected in HepG2 human hepatoma cells. Luciferase activity was determined in transfected cells treated with (black bars) or without (white bars) 30 μm Wy14,643 (Wy). As the negative control, cells were transfected with the empty pGL4.10. As a positive control, cells were transfected with PPREx3-TK plasmid. C, a series of constructs from nt −1100 to +30 mouse promoter bearing block mutations on individual or combinational PPRE/RXR sites were generated and subcloned into the pGL4.10 promoterless plasmid before their luciferase activity in response to ethanol (Veh) or Wy was determined as above. Luciferase light units were expressed as the mean ± S.D. in relative light units. The graph represents typical results from four separate experiments.

A ChIP assay using a proximal primer (PP) set spanning the mouse Ceacam1 promoter between nts −280 and −562 showed that liganded Pparα binds to Ceacam1 gene in liver lysates derived from Wy14,643-fed to a higher extent than Veh-fed Pparα^+/+ mice or Wy14,643-fed Pparα^−/− mice (Fig. 4A). A similar observation is made for the positive control, malic enzyme, a known target of Pparα. The distal primer set amplifying a region between nts +6499 and +6764 in Ceacam1 gene did not detect any binding. Together, the data revealed that activated Pparα binds to Ceacam1 gene to down-regulate its transcription.

FIGURE 4.

Regulation of mouse Ceacam1 expression by PPARα. A, Pparα^+/+and Pparα^−/− mice were fed a Wy-supplemented diet for 7 days before liver extraction and ChIP analysis as described under “Experimental Procedures.” The relative location of the fragment amplified using the proximal (PP; −548/−282) and distal (DP; +6499/+6764) primers in the Ceacam1 (Cc1) gene are also shown. Malic enzyme, a positive target of PPARα, was used as the control. The gel represents experiments on more than three mice per treatment category per mouse group. Wy, Wy14,643. reIb, reimmunoblot. B, livers were removed from mice treated with Wy for 7 days to analyze mRNA levels of Ceacam1 and CD36, a transcriptional target gene of PPARα, by quantitative real-time-PCR analysis. Analysis of each mouse was done in triplicate, and values are represented as the mean ± S.E. C, as in B, livers were removed to determine protein content by Western blot (Ib) analysis. Analyses were done on more than five mice per treatment per mouse group. Two mice from each feeding category per mouse group are shown.

Metabolic Consequence of Pparα-dependent Down-regulation of CEACAM1

Feeding 2-month-old male mice a chow supplemented with Wy14,643 activated hepatic Pparα, as assessed by increased mRNA (Fig. 4B) and protein (Fig. 4C) levels of its target gene, CD36/FABP (16), in Pparα^+/+, but not Pparα^−/− mice. In parallel, this reduced Ceacam1 mRNA (Fig. 4B) and CEACAM1 protein levels (Fig. 4C) in Pparα^+/+ but not Pparα^−/− mice. Consistently, Pparα activation by Wy14,643 reduced insulin clearance in Pparα^+/+ wild-type mice, as assessed by the ∼3-fold decrease in C-peptide/insulin molar ratio relative to RD-fed mice (Table 1). This is further supported by a ∼2-fold decrease in CEACAM1 protein content (Fig. 5A) and in [¹²⁵I]insulin internalization (Fig. 5B) in Wy14,643-treated primary hepatocytes derived from Pparα^+/+ but not Pparα^−/− mice (dashed versus solid lines). Consistent with decreased insulin secretion as a result of enhanced fatty acid oxidation in β-cells upon Pparα activation (28, 29), C-peptide levels were ∼4–5-fold lower in Pparα^+/+ mice fed with a Wy14,643-supplemented diet compared with chow-fed (Table 1). In contrast, supplementing chow with Wy14,643 did not affect C-peptide levels in Pparα^−/− mutants (Table 1). Furthermore, insulin release in response to glucose is also markedly reduced in Wy14,643-fed by comparison to chow-fed wild-type mice but remained unaffected in the mutants (Fig. 6A; dashed versus solid lines). Given the observed lowering effect of Pparα activation on insulin secretion in response to glucose, it is likely that CEACAM1-dependent insulin clearance is reduced by Pparα activation to compensate for the diminished insulin secretion and limit insulin deficiency, as assessed by mildly reduced plasma insulin levels in Wy14,643-fed wild-type mice (Table 1). This maintains normal glucose homeostasis and insulin sensitivity, as demonstrated by normal fasting glucose levels (Table 1 and Fig. 6B) and by better tolerance to exogenous glucose (Fig. 6B, AUC 19,124 ± 1457 versus 27,290 ± 1113; p < 0.002) and insulin (Fig. 6C, AUC 9,206 ± 559 versus 13,938 ± 403, p < 0.0001) in Wy14,643-fed by comparison to RD-fed wild-type mice (dashed versus solid lines). In contrast, Wy14,643 treatment failed to affect insulin secretion (Fig. 6A), glucose tolerance (Fig. 6B, AUC 16,577 ± 858 versus 16,115 ± 784) and insulin tolerance in Pparα^−/− mice (Fig. 6C, AUC 11,738 ± 1,197 versus 11,528 ± 1,425).

TABLE 1.

Effect of Wy14,643 on biochemical parameters

Male mice (n > 8) were fed either RD alone or supplemented with Wy14,643 (Wy) for 7 days. Blood was drawn from overnight-fasted mice to measure blood glucose as well as plasma C-peptide/insulin molar ratio (insulin clearance). Values are expressed as the mean ± S.E.

Biochemical parameters	Pparα+/+		Pparα−/−
	RD	Wy	RD	Wy
Plasma insulin (pm)	44.2 ± 1.3	37.2 ± 1.5a	43.2 ± 1.5	45.0 ± 1.0
Plasma C-peptide (pm)	470 ± 75	106 ± 14a	337. ± 44	391. ± 17
C-peptide/insulin	10.2 ± 2.0	3.5 ± 0.8a	8.8 ± 1.3	9.2 ± 0.7
Fasting glucose (mg/dl)	90.0 ± 11.0	82.1 ± 0.9	85.2 ± 4.2	83.3 ± 3.2

^ap < 0.05 vs. RD in each mouse group.

FIGURE 5.

Effect of PPARα activation on hepatocytic insulin uptake. Primary hepatocytes were isolated from mice and treated with either ethanol (Veh) or Wy,14,463 (Wy) as above. A, cell lysates were subjected to Western analysis (Ib) with α-CEACAM1 antibody (α-CC1) followed by reimmunoblotting (reIb) with α-GAPDH antibody to examine CEACAM1 protein content normalized to GAPDH. AU, arbitrary units. B, [¹²⁵I]insulin internalization was determined in primary hepatocytes pretreated with either ethanol (Veh; solid lines) or Wy (dashed lines) as a first step in insulin degradation and clearance. After binding, [¹²⁵I]insulin was allowed to internalize at 37 °C for 0–90 min (horizontal axis). Internalized ligand was plotted on the vertical axis as the percent of specifically bound ligand. Experiments were done in triplicate. Values are expressed as the mean ± S.D. The graph is representative of more than three experiments.

FIGURE 6.

Metabolic effect of PPARα activation on insulin secretion and action. A, 2-month-old mice were fed a regular diet supplemented with Wy,14,463 (Wy) (dashed lines) or with vehicle alone (solid lines) for 3–7 days. A, acute-phase insulin release in response to glucose was assessed at 0–30 min post-injection. B, mice were challenged with an intraperitoneally injection of glucose to assess blood glucose levels at 0–120 min. C, mice were challenged with an intraperitoneally injection of insulin to assess blood glucose levels at 0–180 min. *, p < 0.05 Wy versus Veh in each mouse group (n > 6 per feeding group). i.p., intraperitoneal.

Discussion

The current studies demonstrate that hepatic CEACAM1 expression is lower at fasting than refeeding. The rise in CEACAM1 expression occurs in parallel to insulin surges in the early hours of refeeding, consistent with the ability of insulin to induce Ceacam1 promoter activity (7). Whereas it is possible that minimal basal insulin contributes to low hepatic CEACAM1 levels at fasting, the current studies provide evidence that this is largely mediated by a Pparα-dependent mechanism. That Pparα activation down-regulates CEACAM1 expression at fasting is demonstrated by 1) reduction in the promoter activity of rat and mouse Ceacam1 in response to Wy14,643 (Pparα agonist) in cultured hepatoma cells, 2) rapid decline of Ceacam1 mRNA and CEACAM1 protein content upon treatment of hepatoma cells with Wy14,643, 3) reduction of Ceacam1 mRNA and CEACAM1 protein levels by supplementing the diet with Wy14,643 in Pparα^+/+ but not Pparα^−/− mice, and 4) binding of liganded Pparα to Ceacam1 promoter in liver lysates derived from Pparα^+/+ but not Pparα^−/− mice. Moreover, the rapid down-regulation of Ceacam1 mRNA and CEACAM1 protein content by Wy14,643 in hepatoma cells suggests that the effect of Pparα activation on Ceacam1 expression in murine liver can occur independently of confounding metabolic factors. Further studies are needed to delineate the mechanisms and identify co-repressors/co-regulators of Ceacam1 expression by Pparα, but ChIP analysis suggests that liganded Pparα directly regulates Ceacam1 expression. Moreover, our observations are consistent with the reported decrease of Ceacam1 mRNA in mice treated with Pparα-selective piperidine agonists that are potent Pparα activators (30). Although Pparα is more commonly known to increase expression of genes, such as those involved in fatty acid catabolism, it has also been shown to repress the expression of many liver-specific genes involved in glucose metabolism, cell adhesion, the CYP2C family of steroid hydroxylases, and positive acute-phase response genes induced during inflammation (30, 31).

The opposing effect of Pparα and insulin on hepatic CEACAM1 expression is probably related to the well characterized role of CEACAM1 in regulating insulin action by promoting insulin clearance (5, 19, 32). Fasting promotes a shift from glycolytic to lipolytic metabolism and robust Pparα activation (16, 33). In the early hours of refeeding, Pparα maintains fatty acid β-oxidation during the stepwise metabolic recovery by insulin surges until glycogen stores are replenished (17) and glycolysis resumes (18). As summarized in Fig. 7 and reviewed in Hue and Taegtmeyer (34), long chain fatty acyl CoA is transported to the mitochondria at fasting to undergo fatty acid β-oxidation to produce acetyl-CoA followed by citrate. Inhibition of pyruvate dehydrogenase by acetyl-CoA leads to accumulation of citrate in the cytoplasm and its gradual inhibition of 6-phosphofructo-1-kinase. This elevates glucose 6-phosphate (G-6-P) levels and, subsequently, causes inhibition of hexokinase and rerouting of G-6-P to the glycogen synthesis pathways until glycogen is replenished, at which point fatty acid oxidation stops, mediated largely by the gradual increase in malonyl-CoA levels and its inhibition of carnitine palmitoyltransferase I-mediated transport of LCFAcyl-CoA (long chain fatty acyl-CoA) into the mitochondria. Contributing to the regulation of malonyl-CoA content is the activity of fatty acid synthase that catalyzes malonyl-CoA conversion to palmitic acid. Down-regulation of this step by insulin phosphorylation of CEACAM1 (10) leads to the gradual recovery of malonyl-CoA and, subsequently, gradual inhibition of carnitine palmitoyltransferase I (CPT1; Fig. 7, dashed lines). This paradigm positions CEACAM1 at the crossroads of the coordinated regulation of fatty acid oxidation by PPARα and insulin in the fasting-refeeding transition (35).

FIGURE 7.

Mechanism of regulation of fatty acid oxidation by CEACAM1 during fasting-refeeding transition. As the schematic diagram (modified from Ref. 34) depicts, fatty acids (FA) are transported to the mitochondria by carnitine palmitoyltransferase I (CPT1) to undergo β-oxidation and produce acetyl-CoA at fasting (left arm). Acetyl-CoA strongly inhibits pyruvate dehydrogenase (PDH; thick red line) to prevent glycolysis (right arm) and reroute pyruvate to gluconeogenesis. This also leads to accumulation of citrate in the cytoplasm and, subsequently, inhibition of 6-phosphofructo-1-kinase (PFK1) and the rise in glucose-6-phosphate (G-6-P), which in turn inhibits hexokinase (HK) and ultimately mediates glycogen repletion (green arrows). At the end of the completion of glycogen replenishment (within ∼8 h of refeeding), β-oxidation stops, mediated largely by the gradual recovery of malonyl-CoA levels and its inhibition of carnitine palmitoyltransferase I activity. The current studies propose that the pulsatile release of insulin in the early hours of refeeding elevates CEACAM1 expression and its tyrosine phosphorylation (pY) by the insulin receptor (IR) to cause its binding to fatty acid synthase (FAS) and reduction of its activity, thus, contributing to the gradual increase in malonyl-CoA levels and inhibition of carnitine palmitoyltransferase I (dashed lines). In this manner, reduction of CEACAM1 transcription by PPARα activation at fasting and its stimulation by insulin positions CEACAM1 to contribute to the regulation of fatty acid β-oxidation in the fasting-refeeding transition, as mediated by the coordinated action of PPARα and pulsatile insulin surges in the early hours of refeeding. LCFAcyl-CoA, long chain fatty acyl CoA.

Global Ceacam1 null mice and L-SACC1 mutants with liver-specific overexpression of the dominant negative phosphorylation-defective CEACAM1 mutant develop secondary insulin resistance due to chronic hyperinsulinemia (5, 19, 32). Moreover, reduced hepatic CEACAM1 expression by a high fat diet results in impaired insulin clearance and insulin insensitivity (21). This demonstrates that lowering CEACAM1 levels can cause insulin resistance. Yet activation of Pparα by synthetic ligands, such as fibrates, improves insulin sensitivity and reverses dyslipidemia. Hence, the reduction in hepatic CEACAM1 levels (and consequently, insulin extraction) by Pparα activators may appear contradictory to the well characterized negative role of impairing CEACAM1-dependent insulin clearance pathways on insulin sensitivity. But Wy14,643 activation of Pparα blunts, almost completely, acute-phase insulin secretion in response to glucose and markedly reduces plasma C-peptide levels in Pparα^+/+ mice, suggesting a negative effect of Pparα on insulin secretion. However, this occurs in the absence of insulin insufficiency and any adverse effect on glucose homeostasis, and Wy14,643 activation of Pparα increases tolerance to exogenous glucose and insulin in Pparα^+/+ wild-type but not in Pparα^−/− mutants. Thus, we postulate that reduction in CEACAM1-dependent hepatic insulin clearance pathways by Pparα activation mediates a compensatory mechanism for the decline in insulin secretion caused by Pparα activation in pancreatic β-cells to limit insulin insufficiency and prevent systemic insulin resistance.

By demonstrating changes in CEACAM1 expression during the fasting-refeeding transition, with hepatic CEACAM1 levels being lower during fasting via a Pparα-dependent mechanism, the current studies provide in vivo evidence for the physiologic regulation of hepatic CEACAM1. Moreover, they lend further credence to the critical role of CEACAM1 in promoting normal metabolism. The relevance of this regulation can be exploited for therapeutic purposes by targeting this protein to prevent and/or reverse metabolic derangements.

Author Contributions

S. K. R., S. S. K., and Q. Y. A. researched the data, designed and coordinated the experiments, and wrote the manuscript. L. R., S. L. A., P. R. P., G. H., H. T. M., and B. R. M. researched the data. A. M. O. researched and analyzed the data. Y. M. S. researched, analyzed the data, and reviewed the manuscript. E. R. S. analyzed the data and discussed and reviewed the manuscript. S. M. N. was responsible for the study design, conceptualization, data analysis, results interpretation, and review of the manuscript. S. M. N. has full access to all of the data of the study and takes responsibility for the integrity and accuracy of data analysis and for the decision to submit and publish the manuscript.