Department of Biochemistry and Molecular Biology, Faculty of Medicine, Dalhousie University, Tupper Medical Building, 1850 College Street, Halifax, Nova Scotia, Canada B3H 1X5. E-mail: email@example.com
Objective: To determine whether adipocyte enhancer binding protein (AEBP) 1, a transcriptional repressor that is down-regulated during adipogenesis, functions as a critical regulator of adipose tissue homeostasis through modulation of phosphatase and tensin homolog deleted on chromosome ten (PTEN) tumor suppressor activity and mitogen-activated protein kinase (MAPK) activation.
Research Methods and Procedures: We examined whether AEBP1 physically interacts with PTEN in 3T3-L1 cells by coimmunoprecipitation analysis. We generated AEBP1-null mice and examined the physiological role of AEBP1 as a key modulator of in vivo adiposity. Using adipose tissue from wild-type and AEBP1-null animals, we examined whether AEBP1 affects PTEN protein level.
Results: AEBP1 interacts with PTEN, and deficiency of AEBP1 increases adipose tissue PTEN mass. AEBP1-null mice have reduced adipose tissue mass and enhanced apoptosis with suppressed survival signal. Primary pre-adipocytes from AEBP1-null adipose tissues exhibit lower basal MAPK activity with defective proliferative potential. AEBP1-null mice are also resistant to diet-induced obesity, suggesting a regulatory role for AEBP1 in energy homeostasis.
Discussion: Our results suggest that AEBP1 negatively regulates adipose tissue PTEN levels, in conjunction with its role in proliferation and differentiation of pre-adipocytes, as a key functional role in modulation of in vivo adiposity.
A major focus for therapy in obesity has been directed toward appetite suppression and increased energy expenditure through interventions that target the central nervous system (1). A complementary approach for the treatment of obesity is one that targets the peripheral adipose tissue, where the actual obese phenotype arises (2, 3). Multipotent stem cells become committed to the adipocyte lineage, and post-natally develop into white adipose tissue (WAT).1 Adult WAT is a complex mixture of cells, including proliferative pre-adipocytes and terminally differentiated, nonproliferative adipocytes. In response to an obesity-inducing stimulus, adipose tissue increases in mass by adipocyte hyperplasia due to the proliferation of pre-adipocytes and their subsequent differentiation to adipocytes and by adipocyte hypertrophy (4). In children, obesity typically involves both hyperplasia and hypertrophy of adipocytes (5), whereas in the adult, obesity is mainly caused by adipocyte hypertrophy. The increased adipocyte number in obese children may predispose them to lasting obesity or may even be causative because childhood obesity frequently persists into adulthood. In extreme forms of adult-onset obesity, adipocyte hyperplasia can occur secondary to adipocyte hypertrophy, possibly due to a critical increase in the adipocyte size leading to production of growth factors that induce adipocyte hyperplasia (6). Clinical studies indicate that extreme obesity is rarely treated successfully, suggesting that the adipocyte hyperplasia in the severely obese may strongly predispose them to continued obesity. Also, the rising incidence and severity of obesity seen in older individuals may be due to increased adipocyte number during aging (7). Thus, various lines of evidence indicate that adipocyte hyperplasia could be a factor in the development and subsequent clinical manifestations of some types of obesity. However, little is known about the mechanisms for this hyperplasia or for the regulation of fat cell number in general. It is postulated that paracrine factors, released from existing adipocytes, influence pre-adipocyte proliferation (5, 8). Adipocyte number may also be modulated by apoptosis of pre-adipocytes and adipocytes (9, 10, 11, 12), which is controlled by the death receptor-mediated apoptotic signaling pathway and the insulin-like growth factor 1-mediated cell-survival signaling pathway of phosphoinositide 3-kinase (PI3K)/Akt.
The tumor suppressor phosphatase and tensin homolog deleted on chromosome ten (PTEN) plays a key role in mammalian growth control (13, 14) and cell migration (15, 16, 17). It antagonizes the activity of growth factor-stimulated PI3K that controls many downstream cellular processes including cell growth, apoptosis, and cell motility. Through its specific lipid phosphatase activity, responsible for removal of the phosphate in the D3-phosphate group of the second messenger phosphatidylinositol 3,4,5-triphosphate, PTEN down-regulates the prosurvival serine-threonine kinase Akt. Although much is known about the mechanism by which PTEN regulates cellular processes, less is known about the regulatory mechanisms that modulate PTEN activities. Recently, a number of studies have revealed that the activity of PTEN is regulated by subcellular localization, phosphorylation, and protein turnover. PTEN has been shown to bind to several proteins through the C-tail region or the PDZ domain binding sequence at its C terminus (18, 19, 20, 21, 22). It was shown recently that neutral endopeptidase constitutively recruits PTEN to the plasma membrane and enhances its stability and phosphatase activity (23). A protein referred to as PICT-1 was identified that binds to the C terminus of PTEN and promotes the phosphorylation and stabilization of PTEN (24).
Recently, using the yeast two-hybrid system, adipocyte enhancer binding protein (AEBP) 1 was identified as an interacting partner of PTEN (25). AEBP1 is a multifunctional protein, originally shown to be an important player in adipogenesis through its transcriptional repression of the aP2 gene, encoding the adipocyte lipid binding protein (26). Since this original characterization, AEBP1 has been found to interact with the γ5 subunit of a heterotrimeric G protein to modulate the differentiation of pre-adipocytes (27). AEBP1 has also been found to complex with and protect the mitogen-activated protein kinase (MAPK) extracellular signal-regulated kinase (ERK) 1/2 from the MAPK-specific phosphatase 3 (28). Cell lines overexpressing AEBP1 exhibited a much higher proliferation rate than control cell lines, and antisense cell lines with reduced expression of AEBP1 showed lower basal MAPK activity than control cell lines. Recent results suggest that MAPK regulates the transcriptional activity of AEBP1 by a novel dual mechanism, in which MAPK interaction enhances and subsequent phosphorylation decreases the DNA-binding ability of AEBP1 (29). AEBP1 expression is down-regulated during adipocyte differentiation (26), although it is abolished only at the terminal stage of adipocyte differentiation (28). Similarly, AEBP1 is highly expressed in the stromal compartment of adipose tissue, including proliferative preadipocytes, but its expression is abolished in terminally differentiated, nonproliferative adipocytes (30). We have recently presented evidence for the role of AEBP1 in the control of in vivo adiposity by transgenic overexpression of AEBP1 during adipogenesis (31).
We report here that AEBP1 physically interacts with PTEN in mammalian cells and may negatively regulate PTEN function by promoting PTEN degradation. By in vivo disruption of AEBP1 gene, we provide evidence for a physiological role of AEBP1 in adipose tissue homeostasis through its negative regulation of PTEN function and its modulatory effect on MAPK activation. Our results suggest that AEBP1 is a critical regulator of energy metabolism.
Research Methods and Procedures
Generation of AEBP1-Null Mice
A replacement targeting vector was prepared in which a segment of the AEBP1 gene between the 6th and 12th introns, including the start codon of AEBP1 in exon 10, was removed and replaced with a PGK-NEO-poly(A) expression cassette (32) in the opposite orientation (see Figure 1A). The AEBP1 targeting vector was constructed by inserting a 5.1-kb XbaI/SalI fragment lying in the middle of intron 6, a 1.8-kb SalI/XbaI fragment containing a PGK-NEO-poly(A) expression cassette, and a 5.7-kb XbaI/EcoRI fragment beginning at the 3′ end of intron 12 into the pGEM-11Zf vector (Promega, Madison, WI). The targeting plasmid was linearized with NotI and electroporated into J1 embryonic stem cells provided by E. Li and R. Jaenisch (33). A homologous recombination event between the targeting vector and the endogenous AEBP1 locus resulted in a modified AEBP1 locus in which a 2.1-kb fragment, including all of exons and introns 7 to 12 and the 3′ one-half of intron 6, were removed. Selection with G418 was performed as described (32). Drug-resistant clones were isolated and expanded followed by genomic DNA extraction for Southern blot analysis. A 1.8-kb XhoI/BamHI fragment lying outside of the targeting construct was used as a 3′ probe for Southern blot analysis of DNA from embryonic stem cells and mice. A 19-kb fragment, corresponding to the wild-type allele, and a 10-kb fragment, corresponding to the disrupted allele, were detected on Southern blots (see Figure 1A). Successful targeting of the AEBP1 locus was achieved in 12 of the 92 G418-resistant clones. Four independently targeted embryonic stem cell clones were used to produce male chimeras in the C57BL/6J background, and germ-line transmission of the disrupted allele was obtained. The resulting heterozygous (F1) progeny were interbred to produce F2 offspring.
Cell Culture, Transfection, and PTEN Half-Life Determination
3T3-L1 pre-adipocytes and cell line stably transfected with an AEBP1 antisense construct were cultured and differentiated as described previously (28). Stable cells expressing antisense-AEBP1 were also differentiated by treating the confluent cells with DMEM containing 10% cosmic calf serum, 5 μg/mL insulin, and various amounts of BRL49653 every 2 days for 7 days. HCll cells were cultured in 90-mm dishes and then transfected at 70% to 80% confluence using Polyfect (QIAGEN, Mississauga, Ontario, Canada) or at 90% to 95% confluence using Lipofectamine 2000 (Invitrogen, Burlington, Ontario, Canada). Neo-1 or AS/Neo-11 cells (1.0 × 106) were treated with 100 μg/mL cycloheximide (in ethanol) for the indicated time. For control (0 hours) samples, cells were treated with vehicle (ethanol) for 8 hours. After treatment, cells were lysed in radioimmunoprecipitation assay (RIPA) lysis buffer [150 mM NaCl, 50 mM Tris-Hcl (pH 7.4), 0.25% sodium deoxycholate, 0.1% NP-40], and clear cell lysates were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and immunoblotting.
All animal experiments were performed according to procedures approved by the institutional Animal Care Committee. AEBP1-null and wild-type mice were weaned on a standard rodent chow or high-fat diet (HFD; 45% of total calories from fat; D12451; Research Diets, New Brunswick, NJ). Mice had access to food and water ad libitum and were weighed every week for up to 35 weeks. The amount of food consumed was determined by weighing the food from each cage weekly, and values are presented as the average amount of food consumed per mouse. Mice were sacrificed by cervical dislocation, and the visceral fat pads were dissected and weighed.
After completion of the 25-week HFD studies, the mice were deeply anesthetized by injection of sodium pentobarbital (Somnitol), and blood was obtained by intra-cardiac puncture. Blood samples were allowed to clot overnight at 4 °C, and the serum fraction was collected after centrifugation at 2000g for 20 minutes. Cholesterol and triglyceride concentrations in plasma samples were determined by enzymatic assays from Roche Diagnostics or Sigma-Aldrich, respectively, adapted to microtiter plate format as described previously (34).
Insulin and Glucose Tolerance Tests
Mice were fasted for 16 hours before insulin and glucose tolerance testing. Blood samples were obtained from the tail vein, and glucose levels were measured using a Bayer Elite XL glucometer (Toronto, ON). After collection of the fasting blood sample, mice received an injection of insulin (0.75 U/kg mouse; Sigma). Blood samples were collected at 15, 30, 45, and 60 minutes post-injection. After 2 days of recovery, the mice were fasted again for 16 hours and, after collection of a fasting blood sample, each mouse received an injection of glucose (1 mg/g mouse, Invitrogen). Glucose values were averaged for each group and reported as millimolar.
In Vivo Insulin Stimulation
Mice were fasted for 16 hours before testing. An intra-peritoneal injection of 40 to 50 mg/kg mouse Somnitol was given to each mouse as anesthetic. When mice were sedated, the surgical area was cleaned with sterile saline, and a lateral incision was made to open the abdomen. A surgical blade was used to separate dermis from the intra-peritoneal cavity, and the samples of mammary gland and WAT were removed before stimulation. Dissection sites were cleaned with sterile saline, and an injection of insulin (5 U; Sigma) was given directly to the inferior vena cava. The surgical area was soaked with saline and then covered with gauze during the 15-minute stimulation period. Post-insulin tissues were recovered after 15 minutes (previously determined as the optimal stimulation time-point), and protein lysates were prepared and used for protein analysis by immunoblotting. Antibodies against AKT and phospho-AKT (Thr308) were purchased from Cell Signaling Technology Inc. (Danvers, MA) and used to determine activation of PI3K pathway by insulin stimulation.
RNA and Protein Analysis
Total RNA from mouse tissues was prepared using the RNA STAT-60 Solution (Tel-Test Inc., Friendswood, TX) according to the manufacturer's protocol. Twenty micrograms of total RNA from each tissue was loaded onto a formaldehyde denaturating 1% agarose gel and transferred onto MSI nylon transfer membrane (GE Osmonics, Inc., Minnetonka, MN) as described (35). The filter was hybridized in QuikHyb solution (Stratagene, La Jolla, CA) with 32P-labeled AEBP1 cDNA probe in a hybridization oven (Thermo Electron Corporation, Waltham, MA) for 3 hours at 65 °C. The filter was washed twice with 2× standard saline citrate/0.1% SDS for 15 minutes at room temperature and twice with 0.2× standard saline citrate/0.1% SDS for 30 minutes at 65 °C before exposure to X-ray film overnight at −70 °C.
For immunoprecipitation (IP) analysis, mouse tissues were homogenized in cold RIPA buffer containing 1 mM phenylmethysulfonyl fluoride, 50 mM sodium pyrophosphate, 1 mM Na3VO4, 1 mM NaF, 5 mM EDTA, 5 mM EGTA, and a protease inhibitor cocktail (Roche Applied Science, Laval, Quebec, Canada). The tissue lysates (1 mg) were incubated with preimmune serum for 1 hour, then with protein A-agarose for 30 minutes. The beads were discarded, and the supernatants were incubated with antiAEBP1 antibody (2.5 μg/mL affinity-purified antiAEBP1 antibody) for 1 hour, then with protein A-agarose overnight. Samples were collected and washed four times with RIPA buffer. For coIP experiments, cell lysates were first incubated with protein A/G Plus agarose for 2 hours. The pre-cleared supernatants were incubated with antiPTEN antibody (1:150; Cell Signaling Technology) overnight at 4 °C, and then with protein A/G Plus agarose for 3 hours. Beads were collected and washed five times with RIPA buffer. Mouse tissue extracts were also prepared by homogenization in high-salt buffer [500 mM NaCl, 10 mM Tris (pH 7.4), 1% Triton X-100, 2 mM EDTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na3VO4, 1 mM sodium molybdate, and protease inhibitor cocktail]. Liver extracts were prepared by homogenization in phosphate-buffered saline (pH 7.4) containing 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and protease inhibitor cocktail. The immunoprecipitated samples or total protein extracts were separated on SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes. The membranes were incubated with blocking buffer for 1 hour at room temperature and then incubated with specific primary antibodies for 1 hour at room temperature or overnight at 4 °C. Membranes were incubated with horseradish peroxidase-conjugated secondary antibody (1:3000) for 1 hour at room temperature and revealed with an ECL detection system (GE Healthcare, Baie d'Urfé, Quebec, Canada).
Results and Discussion
AEBP1 Is a Critical Modulator of in Vivo Adiposity
AEBP1 expression persists during 3T3-L1 pre-adipocyte differentiation and then is abolished only at the terminal stage of adipocyte differentiation (28). In vivo, AEBP1 is abundantly expressed in WAT, most likely from nonadipocyte cells including pre-adipocytes, within adipose tissue stromal-vascular fraction. However, AEBP1 expression is not detected when the mature fat cells are purified from WAT (30). To substantiate the role of AEBP1 as a key modulator of in vivo adiposity and to gain further insight into the physiological role of AEBP1, we generated AEBP1-deficient mice by homologous recombination. We anticipated that the lack of AEBP1 would diminish the proliferative expansion of pre-adipocytes and result in reduced adiposity in the AEBP1-null mice.
The AEBP1 gene extends over 10 kb, has 21 exons, and gives rise to two mRNAs [AEBP1 and aortic carboxypeptidase-like protein (ACLP)] by alternative splicing, during which intron 9 is retained in the mature AEBP1 transcript (30). ACLP encodes an additional 380 amino acids N-terminal to the first ATG codon of AEBP1, which are encoded by exon 10 (Figure 1A). Western blot analysis has shown that ACLP is expressed predominantly in the smooth muscle cells of the adult mouse aorta but not in the adventitia, heart, liver, skeletal muscle, or kidney. In contrast, in situ hybridization experiments have shown that ACLP is expressed in the smooth muscle cells of the aorta, but not in skeletal muscle cells. ACLP was shown to be up-regulated during vascular smooth muscle cell differentiation (36). Northern blot analysis showed a single AEBP1/ACLP-related transcript of ∼4.0 kb with varying abundance in all mouse tissues examined (30). Therefore, AEBP1 knockout (KO) mice, in theory, should be identical to the ACLP KO mice because both animals have disruption of the AEBP1/ACLP gene, which should affect the expression of both proteins. Mice lacking ACLP die perinatally due to gastroschisis, and only ∼6% of KO mice survive to adulthood (37).
Using Southern blot analysis (Figure 1A), we genotyped over 350 F2 offspring from heterozygote mating and found a genotypic distribution of 42 homozygous wild-type (+/+):96 heterozygotes (+/−):20 homozygous null (−/−) for males and 54 +/+:110 +/−:32 −/− for females (combined, 96 +/+:206 +/−:52 −/−, 27%:58%:15%), which deviated slightly from the expected Mendelian ratio (25%:50%:25%). The ratio of the three genotypes shows that the frequency of AEBP1-null mice is about one-half of that expected. These results suggest that AEBP1 gene disruption may affect embryonic development. Northern blot analysis of RNA derived from WAT of these mice revealed that heterozygous AEBP1-null mice produced about one-half the amount of AEBP1 mRNA as did wild-type mice. The homozygous-null mice failed to produce any detectable AEBP1 mRNA (Figure 1B, top). When Western blots of antiAEBP1 immunoprecipitates from heart homogenates were probed with the antiAEBP1 antibody, we readily detected a band corresponding to the 82-kDa AEBP1 protein in wild-type and heterozygous-null homogenates, but there was no AEBP1 protein in homozygous AEBP1-null homogenates (Figure 1B, lower). These results demonstrate that disruption of the AEBP1 locus results in a complete loss of the 4-kb AEBP1 mRNA and 82-kDa AEBP1 protein in the homozygous-null mice.
Analysis of body weight changes over a 20-week post-natal period showed that up to ∼5 weeks of age both male and female homozygous AEBP1-null mice grew at rates similar to those of their wild-type and heterozygous litter mates. However, between 5 and 6 weeks of age, the male and female homozygous AEBP1-null mice began to grow more slowly than their wild-type and heterozygous litter mates, and between 14 and 15 weeks of age, the growth rate for both homozygous-null offspring dramatically diminished (Figure 1C). The difference in body weight for homozygous-null males and females and their litter mates was ∼22% and ∼24%, respectively, at 20 weeks post-partum. Thus, homozygous disruption of the AEBP1 gene slows growth, resulting in smaller animals. Comparison of the homozygous AEBP1-deficient mice with their normal and heterozygous litter mates indicated no other macroscopic physical aberrations. There were no apparent differences in the structures or sizes of internal organs when normalized to body weight. However, the homozygous AEBP1-null females have defects in mammary gland development (premature involution) at late pregnancy, and they are unable to nurse despite production of milk (Reidy SP, Zhang L, Ma H, Wu X, Webber C, Hall BK, Ro H-S, unpublished data). Interestingly, the homozygous AEBP1-null females carrying the mouse AEBP1 transgene, driven by the aP2 promoter (31), had pups that suckled successfully, indicating that stromal restoration of AEBP1 expression can rescue the premature involution phenotype (Reidy SP, Zhang L, Ma H, Wu X, Webber C, Hall BK, Ro H-S, unpublished data). These results imply that loss of ACLP, in the AEBP1-null mice, does not result in an obvious phenotype.
The discrepancy in phenotypes between homozygous ACLP- and AEBP1-null mice is dramatic. The more severe embryonic lethal phenotype in the ACLP KO mice is, perhaps, due to a compound effect of AEBP1 disruption and deregulation of a neighboring gene. Layne et al. (37) replaced exons 7 to 16 (inclusive), whereas we replaced exons 7 to 12 (inclusive) of the AEBP1/ACLP gene with a PGK-NEO cassette in antisense orientation, thus, disrupting ACLP and AEBP1 transcripts (30). AEBP1/ACLP gene is only 102 base pairs upstream of the Pold2 gene that encodes the small subunit of DNA polymerase δ (polδ), the principal DNA replicase in eukaryotes. polδ also participates in several DNA repair pathways, including nucleotide excision repair, mismatch repair, and long-patch base excision repair (38, 39). Given the closer proximity of the PGK promoter to the Pold2 gene in the ACLP KO construct, it is possible that expression of the downstream Pold2 gene was affected. The ACLP KO phenotype may, therefore, be a result of AEBP1 loss and deregulated polδ function. It might be interesting to examine the expression of polδ in ACLP KO mice.
Homozygous AEBP1-null mice have smaller white fat pads than wild-type animals. The total fat pad mass in the male AEBP1-null mice (5 to 6 months) was drastically reduced to 35% that of normal mice (Table 1), corresponding to ∼45% of the wild-type mouse fat pad mass when normalized to body weight. Despite markedly reduced WAT, they do not develop hyperlipidemia or fatty liver. The total fat pad mass was also reduced in the heterozygous male mice to 69% of that wild-type mice. However, we did not observe any significant reduction of total fat pad mass in 5- to 6-month-old homozygous AEBP1-null females when normalized to body weight. A significant reduction of total WAT to 66% of wild-type was observed when the mice were ∼7 to 9 months old. These results indicate that the gain of adipose tissue mass was much more delayed in male AEBP1-null mice than in females. The molecular mechanism underlying the gender-specific effect on adiposity, in which female mice are resistant to the reduction of WAT by disruption of AEBP1, is not understood. AEBP1 may be mediating the gender-specific effect by controlling energy metabolism. It has been postulated for many years based on indirect evidence that estrogen may act to increase energy expenditure in rodents and humans (40, 41, 42, 43). We have shown recently that AEBP1 expression is modulated by estrogen (31). AEBP1 may be a critical mediator of the estrogen effect on energy expenditure, and disruption of AEBP1 may cause a blunting of this effect. Moreover, we have shown that when coupled with an HFD during adipogenesis, transgenic overexpression of AEBP1, but not ACLP, resulted in massive obesity in female transgenic mice (AEBP1TG) by adipocyte hyperplasia (31). These results suggest that AEBP1, not ACLP, has an important role in the regulation of adiposity.
Table 1. AEBP1−/− mice display reduced total (gonadal and visceral) WAT mass (grams)
1.5 to 3 Months
5 to 6 Months
7 to 9 Months
0.39 ± 0.16 (n = 8), 100%
1.13 ± 0.38 (n = 6), 100%
0.34 ± 0.19 (n = 19), 88%
0.78 ± 0.43 (n = 16), 69%
0.29 ± 0.13 (n = 13), 75%
0.40 ± 0.21 (n = 11), 35%
0.16 ± 0.09 (n = 5), 100%
0.65 ± 0.24 (n = 5), 100%
1.73 ± 0.51 (n = 5), 100%
0.17 ± 0.10 (n = 6), 110%
0.77 ± 0.32 (n = 4), 118%
1.79 ± 0.69 (n = 4), 103%
0.13 ± 0.02 (n = 3) 83%
0.52 ± 0.21 (n = 5) 79%
0.86 ± 0.41 (n = 5) 50%
Layne et al. (37) concluded that ACLP is an extracellular matrix protein essential for normal embryonic development and dermal wound healing. The hypoproliferation phenotype of ACLP−/− fibroblasts is most likely due to AEBP1 disruption. We have shown previously that AEBP1 stimulates cell growth and survival and inhibits apoptosis through negative regulation of MAPK (28) and PTEN (31), see below). Cell fractionation and immunofluorescent staining experiments revealed that ACLP is excluded from the nucleus and cytosol and is localized in the perinuclear space, indicative of its entry into the secretory pathway (36). Unlike ACLP, AEBP1 can be detected in the nucleus and cytosol (27). Because ACLP is not localized to the cytosol, it is inconceivable that ACLP can have any effect on PTEN function by direct protein-protein interaction (see below), despite the fact that both AEBP1 and ACLP share the exact region that mediates AEBP1-PTEN interaction. It is also important to emphasize that the sex-specific diet-induced obesity (DIO) phenotype observed in AEBP1TG mice (31) can only be attributed to AEBP1, not ACLP. Importantly, AEBP1 overexpression has no effect on ACLP expression in WAT (data not shown).
AEBP1 Promotes PTEN Protein Degradation Possibly by Protein-Protein Interaction
PTEN plays an important role in the cell survival signaling pathway by antagonizing the activity of PI3K. Because AEBP1 is an interacting partner of PTEN (25), we asked whether the modulation of in vivo adiposity by AEBP1 is mediated through control of PTEN function. First, we assessed whether AEBP1 and PTEN interact in mammalian cells by co-IP experiments. Extracts from 3T3-L1 pre-adipocytes were subjected to IP with either normal IgG or antiPTEN antibody and immunoprecipitates probed with either antiAEBP1 (Figure 2A, left) or antiPTEN (Figure 2, middle) antibodies. The results show that AEBP1 is detected in the PTEN immunoprecipitates (Figure 2, lane 2, left) but not in the control IgG immunoprecipitates (Figure 2, lane 1, left). Figure 2, right, shows similar levels of AEBP1 and PTEN in the total cell lysates used for IP. To define the domain of AEBP1 responsible for the interaction, expression plasmids encoding different deletion mutants of AEBP1 were used (Figure 2B). It has been previously demonstrated by yeast two-hybrid analysis that the N-terminal 235 amino acid residues of AEBP1 are involved in the interaction between AEBP1 and PTEN (25), which includes the entire discoidin-like domain (DLD) and the N-terminal 70 amino acid residues of the carboxypeptidase-like domain (CP). The mutant derivative Δ serine, threonine, proline-rich domain (STP) A has a deletion of the C-terminal STP and A regions, whereas the mutant derivative ΔC lacks the entire C-terminal domain containing basic amino acid-rich domain, STP, and A regions. The mutant derivative ΔN lacks the entire DLD domain, and the mutant derivative ΔNCP contains an internal deletion of 120 amino acid residues from the N terminus of the CP domain. These constructs were individually transfected into HC11 cells and analyzed for their ability to interact with PTEN by co-IP analysis. Figure 2C (top) shows the expression of AEBP1 and its mutant derivatives in HC11 cells transfected with the respective plasmids. The expression level of ΔN was always much lower than other mutant derivatives, which suggests that the N terminus of AEBP1 may be important for its stability. Cell lysates were immunoprecipitated with either normal rabbit IgG (lane 1) or antiPTEN antibody (lanes 2 to 6), and the immunoprecipitates were then immunoblotted (IB) with antiAEBP1 (Figure 2, middle) or antiPTEN (Figure 2, lower) antibody. The results show that all of the mutant derivatives, except ΔNCP, were able to interact with PTEN. Although the expression level of ΔN is relatively low, it was still able to coprecipitate with PTEN. In contrast, ΔNCP was not able to interact with PTEN despite its higher expression than ΔN (compare lanes 5 and 6). Taken together, these results suggest that the N-terminal 70 amino acid residues of the CP domain of AEBP1 are important for interaction with PTEN.
Because the PTEN C-tail has been implicated in the stability of the molecule (44) and the C-terminal region of PTEN is involved in AEBP1 interaction (25), we examined whether AEBP1 participates in the regulation of PTEN protein turnover. We examined the possible impact of the AEBP1-PTEN association on PTEN stability by analyzing PTEN protein expression level. Cell lysates from transfected HC11 cells expressing increasing amounts of AEBP1 were examined for PTEN level by immunoblotting analysis. Figure 2D shows a representative blot of cell extracts from transfected cells, in which the amount of PTEN decreased in a dose-dependent manner as AEBP1 levels increased. In contrast, PTEN levels did not change when the mutant derivative ΔNCP was transfected (Figure 2E). These results suggest that the negative regulation of PTEN protein levels by AEBP1 is mediated by protein-protein interaction. To further confirm that endogenous AEBP1 controls the levels of PTEN in cells, we determined the PTEN level in a stable antisense cell line, AS/Neo-11, expressing reduced levels of endogenous AEBP1 in comparison with the levels in control cell lines established with empty vector (28). Figure 2F shows that the PTEN level is significantly increased in AS/Neo-11 cells in comparison with the level in the control cells, Neo-1. To convincingly demonstrate an AEBP1-dependent alteration in PTEN protein stability, we determined the half-life of PTEN in AS/Neo-11 cells. Figure 2G shows that the PTEN half-life is significantly increased in AS/Neo-11 cells (>8 hours) in comparison with the half-life in the control cells, Neo-1 (∼4 hours). Together, our results suggest that in cells where PTEN associates with AEBP1, PTEN levels are decreased due to increased degradation of the protein and that PTEN activity is negatively correlated with the amount of AEBP1 protein.
Additionally, our results suggest that the negative modulation of PTEN by AEBP1 would further promote cell growth from the sustaining effect on MAPK activation by AEBP1. We have shown previously that cell lines overexpressing AEBP1 exhibited a much higher basal MAPK activity and proliferation rate than control cell lines (28). These results suggest that AEBP1 may be a critical regulator of cell growth and survival in adipose tissue.
AEBP1−/− Mice Display Suppressed Survival Signal with Proapoptotic Adipose Tissue
Adipocyte number may be modulated by apoptosis of pre-adipocytes and adipocytes (9, 10, 11, 12), which is controlled by the PTEN-PI3K signaling pathway. As anticipated from the negative modulation of PTEN by AEBP1, the steady-state level of PTEN in AEBP1−/− WAT was about 2-fold higher than the level in wild-type WAT (Figure 3A). The increased level of PTEN would have proapoptotic and antisurvival effects in WAT, which may eventually contribute to the change in adiposity. Figure 3B shows that AEBP1−/− mice produced an increased amount of cleaved caspase 9, the activated form of an initiator of apoptosis, in WAT. To evaluate the mechanisms involved in PTEN and PI3K activation by AEBP1, AEBP1−/− mice were injected with insulin directly to the inferior vena cava to stimulate PI3K activation. The status of Akt protein phosphorylation in WAT was then determined by immunoblotting. Figure 3C shows that AEBP1−/− animals seem to display suppressed insulin-induced Akt phosphorylation (lane 2) compared with that of wild-type mice (lane 4). This would be anticipated from the lower steady-state level of PTEN in wild-type WAT compared with the level in the KO WAT (Figure 3A). Furthermore, the amount of PTEN in the insulin-stimulated wild-type WAT is significantly lower than the level in the insulin-stimulated KO WAT (Figure 3, compare lanes 2 and 4 in D). These results suggest a potential mechanism by which the abrogation of AEBP1 may lead to reduced WAT. The absence of AEBP1 appears to diminish PTEN protein turnover, which would consequently promote apoptosis and suppress survival signal in WAT explaining the decreased adipose tissue mass in the KO animals.
It has been demonstrated recently that PTEN is a negative regulator of insulin signaling and insulin sensitivity in adipose tissue, in which adipose-specific PTEN deletion in mice resulted in improved systemic glucose tolerance and insulin sensitivity (45). However, AEBP1-deficient mice displayed no glucose intolerance despite increased PTEN levels and suppressed Akt activation in WAT. Moreover, the hypoglycemic response to insulin in AEBP1-deficient mice was similar to that of wild-type control. The lack of a glucose homeostasis defect in AEBP1-null mice may be due to a compensatory effect in skeletal muscles, in which PTEN levels and Akt activation are not altered from that of wild type (data not shown). These results suggest that AEBP1 may control PTEN stability in a tissue- or lineage-specific manner.
Abrogation of AEBP1 Produces Hypoproliferative and Hyperdifferentiative Pre-adipocytes
To understand the specific biological functions that AEBP1 exerts on adipose tissue growth, we took advantage of the known actions of this transcription factor during adipogenesis. We have shown previously that AEBP1 expression is down-regulated during adipocyte differentiation (26), and constitutive overexpression of AEBP1 in pre-adipocytes prevents their differentiation into adipocytes (28). Moreover, AEBP1 has been found to control adipocyte differentiation through its modulatory effect on MAPK activation. The growth rate and basal MAPK activity were markedly increased in a cell line overexpressing AEBP1; conversely, the AEBP1 antisense cell lines, with reduced levels of endogenous AEBP1, were hypoproliferative and exhibited lower basal MAPK activity than control cell lines (28); data not shown). In vivo, AEBP1 is abundantly expressed in the stromal compartment of adipose tissues, including proliferative pre-adipocytes, but is abolished in mature fat cells (30). We speculated that the contribution of AEBP1 to decreased adiposity might also involve the suppression of the proliferative expansion of pre-adipocytes. Figure 4A shows that the growth rate of primary pre-adipocytes from AEBP1−/− WAT was markedly reduced compared with the growth rate of primary pre-adipocytes from wild-type WAT. The hypoproliferative phenotype was also observed in ACLP−/− fibroblasts (37). The basal MAPK activity (phospho-ERK1/2) in the primary pre-adipocytes from AEBP1−/− WAT was only ∼27% of the activity in wild-type pre-adipocytes (Figure 4B). These results suggest that AEBP1 modulates MAPK activation, leading to a control of the proliferative potential of pre-adipocytes in WAT and of in vivo adiposity.
Next, we asked whether the decreased adiposity in AEBP1-null mice was due to accelerated differentiation of pre-adipocytes, which would limit their proliferative potential. We attempted to assess the differentiation capacity of primary pre-adipocytes from AEBP1-null mice but were unable to grow and maintain these pre-adipocytes at confluence because some of the cells spontaneously underwent apoptosis. We previously established two stable pre-adipocyte cell lines (AS/Neo-7 and −11), which showed decreased levels of AEBP1 in comparison with that of the control cell line (Neo-12) established with the empty vector. Of the two clones, AS/Neo-11 exhibited lower levels of AEBP1 expression and MAPK activity (24% vs. 61% of the control level) (28). The AEBP1 antisense pre-adipocyte cells grew more slowly than the control cells, but unlike the primary pre-adipocytes from AEBP1-null mice, they were able to reach confluence and survive for an extended period of time (data not shown). Figure 4C shows that the antisense pre-adipocyte cell lines differentiated much more efficiently than the control cells, and differentiation efficiency was inversely correlated with AEBP1 level and MAPK activity. We looked for a possible molecular mechanism to explain the enhanced adipocyte differentiation in the antisense cells by examining the expression pattern of genes, such as CCAAT/enhancer-binding protein (C/EBP) α and peroxisome proliferator-activated receptor (PPAR) γ, that play a vital role in inducing adipocyte differentiation (46). The C/EBP family of proteins is an important group of transcription factors involved in adipogenesis. There are at least three C/EBP isoforms expressed during different stages of adipocyte differentiation. PPARγ belongs to the peroxisome proliferator-activated receptor subfamily of nuclear hormone receptors and exists as two isoforms (γ1 and γ2) generated by alternative promoter usage and mRNA splicing (47). These proteins play a vital role in the initiation of adipogenesis and throughout the differentiation process (48, 49). PPARγ1 directly mediates the increase of C/EBPα protein in the early phase of adipogenic induction (50). PPARγ1 and C/EBPα, along with adipocyte determination- and differentiation-dependent factor 1/sterol regulatory element binding protein 1 (ADD1/SREBP1), mediate the delayed formation of PPARγ2; together, these factors mediate the insulin-induced transcription of lipogenic genes (51, 52, 53, 54, 55, 56). PPARγ is found to form a heterodimer with retinoid X receptor, and it is able to activate fat-specific genes, which also mediate enhanced adipogenesis through ligand activation (57, 58, 59).
Marked induction of C/EBPα mRNA was observed in the antisense cells before differentiation, and the induction level was again inversely correlated with AEBP1 levels (28, Figure 4D). These results suggest that the accelerated differentiation of the antisense cells may be due to the induction of C/EBPα expression. Expression of C/EBPα at higher than physiological levels also induces adipogenesis in fibroblasts (53), whereas expression of antisense C/EBPα RNA suppresses adipogenesis in 3T3-L1 cells (60). Moreover, the marked induction of C/EBPα expression may be responsible, in part, for the decreased proliferation rate in the antisense cells. In both cultured cells and animals, C/EBPα inhibits cell proliferation, in part, by stabilization of the p21 protein, an inhibitor of cyclin-dependent kinases (56, 61). These results suggest that AEBP1 may have a critical role in cell cycle control. Interestingly, we also observed a significant induction of PPARγ mRNA expression in the antisense cells before differentiation (Figure 4D). Sequence analysis of the promoter upstream regions of the PPARγ gene revealed that the PPARγ1, but not PPARγ2, promoter upstream region contains a sequence very similar to that of the AEBP1-binding sequence adipocyte enhancer 1 (62). In addition, the reporter constructs containing the PPARγ1, but not PPARγ2, promoter region were trans-repressed by AEBP1 (63). These results suggest that PPARγ1 expression is directly regulated by AEBP1, and the induced PPARγ1 protein mediates the increase of C/EBPα expression in the antisense cells. Taken together, the above results suggest that the reduced adiposity in the AEBP1−/− mice is due to a decreased proliferative potential and an accelerated capacity to differentiation in pre-adipocytes and additionally increased apoptosis in WAT through the combined modulatory effects of AEBP1 on MAPK and PTEN, C/EBPα, and PPARγ gene expression. Our results indicate that AEBP1 modulates in vivo adiposity through multiple effects on proliferation, apoptosis, and differentiation.
AEBP1−/− Mice Are Resistant to DIO
To further substantiate the gender-specific role of AEBP1 in energy balance, AEBP1−/− mice were challenged with an HFD. Mice were fed an HFD beginning at 10 weeks, and their growth pattern followed for the next 25 weeks. During the 6 months that the mice were on an HFD, male and female wild-type litter mates rapidly gained weight, whereas AEBP1−/− mice were protected from diet-induced weight gain (Figure 5A). Despite the apparent increased energy intake in AEBP1−/− males (Figure 5B), the weight gain was substantially suppressed during HFD feeding. The energy intake was not significantly different between AEBP1−/− and wild-type females. Although the energy intake was not reduced in AEBP1−/− mice, the feed efficiency was significantly reduced in both genders (Figure 5C). The relative mass gained per amount of food consumed in AEBP1−/− males was markedly decreased to ∼40% of wild-type males. The feed efficiency reduction in AEBP1−/− females was less pronounced and was reduced to ∼70% that of wild-type females. These results indicate that males are more resistant to DIO when AEBP1 is disrupted. Consistent with protection from DIO in AEBP1-null mice, the AEBP1 disruption markedly diminished the hypertriglyceridemia caused by high-fat feeding. The plasma triglyceride concentration was markedly reduced (55%) in male AEBP1-null mice compared with control wild-type mice (Figure 5D). Somewhat surprisingly, the plasma triglyceride concentration of HFD-fed female wild-type mice was less than one-half of the level determined in HFD-fed male wild-type mice, and its level was further reduced (45%) when AEBP1 was disrupted (Figure 5D). Similarly, total plasma cholesterol concentration of HFD-fed female wild-type mice was less than one-half of the level in HFD-fed male wild-type mice (Figure 4E). However, the level of plasma cholesterol was only markedly decreased (60%) in males when AEBP1 was disrupted (Figure 5E). The gender-specific effects on lipid and energy metabolism are in agreement with the gender-specific effect on adiposity, in which the gain of WAT mass is much more delayed in male KO mice than in females. The molecular mechanisms underlying the gender-specific regulatory role of AEBP1 on energy homeostasis are elusive. AEBP1 is expressed in many tissues, including the brain (30), but the relative importance of AEBP1's effects on central vs. peripheral sites has not been resolved. The increased energy intake in male AEBP1-null mice suggests that AEBP1 may be involved in hypothalamic energy homeostasis. Development of transgenic mice harboring a tissue-specific ablation of AEBP1 in WAT or brain would extend our understanding on the role of AEBP1 in energy balance.
Human obesity is a polygenic disease controlled by many genetic loci. Because the prevalence of obesity is increasing in industrialized societies, it is apparent that many of these genes must confer susceptibility to environmental factors, such as availability of food and composition of diets. Genes involved may be part of already-identified pathways or may be unknown components of known pathways. Our results raise an intriguing possibility that AEBP1 may be one of the genes that confer susceptibility to DIO. Overall, our results provide the first evidence for the expression, regulation, and functional role of AEBP1 in mouse adipose tissue. We suggest that AEBP1 may negatively regulate PTEN, and this may be a key to understanding the functional role of the modulatory effect of AEBP1 on adiposity. AEBP1 may, therefore, have potential therapeutic benefit in the treatment and management of obesity.
We thank Bradford Lowell for support in the work, which was initiated during H.S.R.'s sabbatical leave at the Department of Medicine, Division of Endocrinology, Beth Israel Deaconess Medical Center and Harvard Medical School. This work was supported by the Canadian Diabetes Association, by the Heart and Stroke Foundation (Nova Scotia) of Canada, by the Natural Sciences and Engineering Research Council of Canada, and by the Canadian Institute of Health Research to H.S.R. We acknowledge the support of the Heart and Stroke Foundation (Nova Scotia) of Canada Visiting Scientist Award to H.S.R. and the Nova Scotia Health Research Foundation graduate studentship to P.J.L.
The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.