Antiobesity Action of ACAM by Modulating the Dynamics of Cell Adhesion and Actin Polymerization in Adipocytes

Quite a few adhesion molecules have been identified by molecular genetics as well as gene expression profile studies in adipocytes derived from experimental models and human studies. For instance, only cadherins were reported to be expressed in premature adipocytes. In the cell lines, such as C3H10T1/2 and 3T3-L1 cells, N-cadherin and cadherin-11 are expressed, and they are prominently suppressed by the induction of adipocyte differentiation and downregulated to very low levels after the full differentiation (1). Transgenic (Tg) expression of dominant-negative N-cadherin decreased bone formation, delayed acquisition of peak bone mass, and increased body fat (2). Although the information of adhesion molecules in adipocyte biology is limited, we identified adipocyte adhesion molecule (ACAM) from the visceral adipose tissues of OLETF (Otsuka Long-Evans Tokushima fatty) rats by PCR-based cDNA suppressive subtraction methods (3). Mouse ACAM was independently identified as adipocyte-specific protein 5 (ASP5) from 3T3-L1 cells by using signal sequence trap by a retrovirus-mediated expression screening method (4). Human ACAM had been identified as coxsackie virus and adenovirus receptor-like membrane protein (CLMP) by bioinformatics approaches, and Raschperger et al. (5) demonstrated that CLMP is a component of the tight junction of epithelial cells and colocalized with zonula occludens-1. ACAM/CLMP belongs to CTX (cortical thymocyte marker in Xenopus), and they are characterized by extracellular variable (V-type) and constant (C2-type) immunoglobulin domains, which are involved in the homophilic adhesion and aggregation of the cells. Although we reported the expression of ACAM increased in mature adipocytes in genetically obese db/db and diet-induced obesity mice, and also in adipose tissues in the subjects with obesity, the functional role of ACAM in mature adipocytes and in obesity remained totally unknown. Here, we generated aP2 promoter–driven ACAM Tg mice, and they are protected from obesity and diabetes with reduced accumulation of white and brown adipose tissues. In ACAM Tg mice fed high-fat, high-sucrose (HFHS), chow, ACAM is abundantly expressed on plasma membranes of mature adipocytes, where the zonula adherens–like structure is formed. The adhesion process and formation of zonula adherens are associated with the formation of actin polymerization on the surface of the mature adipocytes. The formation of zonula adherens may increase the mechanical strength, inhibit the adipocyte hypertrophy, alter the signaling events, and improve the insulin sensitivities. Our finding provides the new therapeutic modalities targeting the processes of cell adhesion and actin polymerization of adipocytes in the treatment of obesity and diabetes.

β-Globin intron and human growth hormone poly-A signal were ligated into BamHI-EcoRI and EcoRI-XhoI sites of pcDNA3.1Zeo vector (Invitrogen), respectively (Supplementary Fig. 1A). Coding region of mouse ACAM cDNA was inserted into EcoRI site by blunt end ligation after filling-in reaction. BamHI and XhoI fragment was subjected to blunt end ligation into SmaI site of pBluescript SKII(+) (Stratagene), in which mouse aP2 promotor was inserted into EcoRV-PstI site by blunt end ligation. The insert was excised with HindIII and NotI, and transgene was generated. Microinjected C57BL/6JJcl one-cell stage zygotes were oviduct transferred and permitted to develop to term. Three Tg founders were obtained, and Southern blot analysis was performed using EcoRI and SmaI fragment of the transgene. Genotyping of Tg mice was performed by PCR using primers 5′-GACATTGAATGGCTGCTCACCG-3′, 5′-GCTCTGCACATACTGTACAGTC-3′, 5′-GTTGGAACGCTGGGAACTCACACTGAGATC-3′, and 5′-GGTTCAGAACCTCTCACTTCCGGTCCTATG-3′.

Male C57BL/6JJcl mice were housed in cages and maintained on a 12-h light-dark cycle. For the animal experiments with mice, standard chow (NMF; Oriental Yeast) and HFHS (D12331; Research Diet) diets were used and the mice were killed at 30 weeks of age. Oxygen consumption was measured using an O₂/CO₂ metabolism–measuring system (MK-5000; Muromachi). All animal experiments were approved by the Animal Care and Use Committee of the Department of Animal Resources, Advanced Science Research Center, Okayama University. Liver triglyceride was measured by the Folch method (Skylight Biotech, Tokyo, Japan).

Mouse 3T3-L1 fibroblasts (American Type Culture Collection) were cultured in vitro and differentiated into mature adipocytes. 3T3-L1 cells were grown in DMEM (Sigma-Aldrich) supplemented with 10% calf serum (Hyclone), 100 units/mL penicillin, and 0.1 mg/mL streptomycin (Sigma-Aldrich). Confluent cells were induced to differentiate by the addition of 0.5 mmol/L 1-methyl-3-isobutylxanthine (IBMX), 0.25 μmol/L dexamethasone (DEX), and 10 μg/mL insulin (Sigma-Aldrich). After 48 h, induction medium was removed and cells were cultured with DMEM supplemented with 10% FBS (Invitrogen) and 10 μg/mL insulin for 14 days. For determination of the major inducer for ACAM expression, the confluent preadipocytes were treated with DEX/IBMX/insulin, DEX, IBMX, insulin, DEX/IBMX, IBMX/insulin, DEX/insulin, FBS, pioglitazone, DEX/IBMX/insulin/pioglitazone, 8-bromoadenosine–cAMP (Calbiochem), or forskolin (Sigma-Aldrich) at the concentrations indicated. In separate experiments, confluent preadipocytes were preincubated with H89 (Sigma-Aldrich), SB203580 (Calbiochem), PD98059 (Calbiochem), and LY294002 (Sigma-Aldrich) for 30 min, respectively, in prior to induction of differentiation.

For quantitative real-time PCR analysis in various tissues in mice, cDNA synthesized from 2 μg total RNA was analyzed in a Sequence Detector (model 7900; Applied Biosystems) with specific primers and SYBR Green PCR Master (Applied Biosystems). The relative abundance of mRNAs was standardized with 36B4 mRNA as the invariant control. The mRNA expression was determined by SYBR Green. Gene-specific primers are indicated in Supplementary Table 1.

Total RNA was isolated from 3T3-L1 cells by using RNeasy Lipid Tissue Mini kit (Qiagen). For Northern blot analysis, 10 µg total RNA was subjected to 2.2 mol/L formaldehyde 1% agarose gel electrophoresis and capillary transferred to the Hybond XL nylon membranes (GE Healthcare Life Sciences). The membranes were hybridized with [α-³²P]dCTP-radiolabeled mouse ACAM, mouse C/EBPβ, mouse C/EBPδ, mouse C/EBPα, mouse peroxisome proliferator–activated receptor γ, mouse LPL (lipoprotein lipase), and mouse 18S ribosomal RNA (American Type Culture Collection) cDNAs at 68°C in ExpressHyb Hybridization Solution (Clontech) for 1 h. Filters were washed in high-stringency conditions, i.e., four times in 1 × sodium chloride–sodium citrate/0.1% SDS at 20°C, followed by two times at 50°C in 0.1 × sodium chloride–sodium citrate/0.1% SDS.

Sheep polyclonal anti-ASAM (adipocyte-specific adhesion molecule) (ACAM) (R&D Systems), rat monoclonal anti-F4/80 (Cl:A3-1) (AbD Serotec), rabbit polyclonal anti–nonmuscle myosin heavy-chain II-A (Covance), and mouse monoclonal anti–γ-actin (Sigma-Aldrich) antibodies were used for immunoblotting, immunofluorescence, and immunogold studies. For secondary antibodies, rabbit anti-sheep IgG, horseradish peroxidase–conjugated (Millipore), Alexa Fluor 594 donkey anti-sheep IgG (H+L), Alexa Fluor 488 goat anti-rat IgG (H+L), and Alexa Fluor 488 chicken anti-rabbit IgG (H+L) (Invitrogen) were used.

Recombinant full-length mouse ACAM tagged with calmodulin-binding peptide (CBP) and streptavidin-binding peptide (SBP) (Adeno-pCTAP-mACAM) (InterPlay Mammalian TAP System; Stratagene) were prepared by adenovirus expression kit (Takara). Adeno-pCTAP-mACAM was introduced to 3T3-L1 cells. Soluble proteins were purified by CBP and SBP binding resin (InterPlay Mammalian TAP System; Stratagene) and subjected to SDS-PAGE and Coomassie blue staining. Visible bands were excised and in-gel digested with trypsin and analyzed with liquid chromatography–tandem mass spectrometry (LC-MS/MS) and matrix-assisted laser desorption/ionization–top of flight mass spectrometry (MALDI-TOF/MS).

3T3-L1 cells at 24 h after the induction were fractioned. Cells were washed in PBS and resuspended in suspension buffer (20 mmol/L Tris-HCl, pH 7.4; 1 mmol/L EDTA; and 255 mmol/L sucrose) containing protease inhibitor cocktail (Sigma-Aldrich) and homogenized with a precooled motor-driven Potter-Elvehjem grinder with 20 strokes at 1,400 rpm. Nuclei were pelleted by centrifugation at 1,000g for 10 min. The postnuclear supernatant was centrifuged at 16,000g for 20 min. The pellet containing mitochondria, peroxisomes, and plasma membrane fractions was resuspended in 5 mL resuspension buffer and layered onto 5 mL sucrose cushion (1.12 mol/L sucrose, 1 mmol/L EDTA, and 20 mmol/L Tris-HCl, pH 7.4) and centrifuged at 101,000g for 25 min. The mitochondria and peroxisomes were collected as a pellet and resuspended in suspension buffer. The plasma membrane was collected at the interface, resuspended in 10 mL suspension buffer, centrifuged at 16,000g for 15 min, and resuspended in suspension buffer. The postnuclear supernatant was centrifuged at 48,000g for 20 min, and the pellet containing high-density microsomes (HDMs) was obtained. The supernatant was centrifuged at 212,000g for 70 min, separating the pellet containing low-density microsomes (LDMs) and the supernatant containing the cytosol. Pellets from fractions containing HDMs and LDMs were resuspended in suspension buffer.

Adipose tissues and 3T3-L1 cells were fixed with 4% paraformaldehyde and 0.05% glutaraldehyde in 0.01 mol/L phosphate buffer, pH 7.4, for 30 min and then scraped from the dish, and a pellet was prepared with a microfuge. The tissues and cell pellets were then dehydrated in a graded series of ethanol, and they were embedded and polymerized in LR-White (Polysciences, Inc.) at 50°C for 48 h. Approximately 60-nm-thin sections were picked up with a 300 mesh nickel grid. The sections were pretreated with 20 mmol/L Tris-buffered saline, pH 8.2, containing 0.25% BSA. They were incubated with anti-ACAM antibody and then with anti-sheep IgG conjugated with 10 nm colloidal gold particles for 4 h at room temperature. They were washed with Tris-buffered saline and stained with lead citrate and uranyl acetate and examined by an electron microscope at accelerating voltage of 75 kv. Transmission and scanning electron microscopy was performed as previously described (6).

The adenoviral vector encoding a constitutively active form of C/EBPβ (liver-enriched activator protein [LAP]) was generously provided by Professor Hiroshi Sakaue (Tokushima University, Tokushima, Japan). For adenoviral infection, 3T3-L1 preadipocytes were cultured to 90% confluency and infected at a multiplicity of infection of 10 plaque-forming units/cell 48 h before induction of differentiation (7). The cells were collected at indicated time points and were subjected to Northern blot analysis. As a control, the adenoviral vector encoding LacZ was used.

The ACAM 1,869–base pair (−1,741 to 128) luciferase plasmid was generated by ligating into the cloning site of the promoterless luciferase reporter plasmid pGL3-Basic (Promega). Various 5′-deletion constructs −1,341, −584, −464, −235, −97, −79, and −57 were made. The mutant constructs of promoter sequence using −79 deletion construct—wild type (WT) (−72 to −59), 5′-AGCTAACCCCAAAC-3′; mutant (MT)1, 5′-AGCTAACAATGGCC-3′; and mutant 2 (MT2), 5′-AGCTAAACCCGAAC-3′—were generated by PCR. 3T3-L1 preadipocytes were transfected using a Neon electroporation transfection system (Invitrogen). Transfections were performed using 10 μg pBIND (Renilla luciferase)-SV40 (internal control) along with 10 μg pGL3-Basic plasmids containing the ACAM promoter or with pGL3-Basic plasmid. Forty-eight hours after transfection, cells were incubated with 0.5 mmol/L IBMX for 6 h, and luciferase reporter assays were performed using the Dual Luciferase Reporter Assay System (Promega). Transfection efficiencies were normalized to the Renilla luciferase activity. For the cytomegalovirus promoter–driven transient overexpression of Kruppel-like factor (KLF)4, GATA-binding protein 1 and GATA-binding protein 6, EX-Mm19461-M02 (pReceiver-M02), EX-Mm02659-M02 (pEZ-M02), and EX-Mm02664-M02 (pEZ-M02) (OmicsLink Expression Clone, Genecopia) were used, respectively. For negative control, EX-NEG-M02 (pReceiver-M02CT) was used.

Data are expressed as the mean ± SE, and the multiple comparisons were performed by a one-way ANOVA with Bonferroni and Tukey corrections. A value of P < 0.05 was regarded as statistically significant. The data were analyzed using the IBM SPSS Statistics software program.

Three independent aP2-driven ACAM Tg lines of C57BL/6JJcl mice were established (Supplementary Fig. 1A). The copy numbers of transgenes differed in L11 (high-), L22 (intermediate-), and L4 (low-expression) lines, and the protein expression of ACAM in both epididymal and subdermal fat tissues corresponded to the copy numbers of transgene (Supplementary Fig. 1B). On HFHS chow, mRNA expression of ACAM was significantly augmented in epididymal, subdermal, and brown adipose tissues and not in liver and skeletal muscle (Supplementary Fig. 1C). The body weight gains on HFHS chow were prominently reduced in ACAM Tg mice in parallel with the expression of ACAM (Supplementary Fig. 1D–F). Throughout the following experiments, intermediate-expression (L22) line was used, since the mice derived from high-expression (L11) line were completely protected from obesity and the reduction of body weight may dominantly influence the phenotype such as improvement of glucose metabolism. In ACAM Tg mice, the body and fat pad weight was significantly reduced on HFHS chow compared with WT mice (Fig. 1A and B), and the adipocyte size was also significantly reduced (Fig. 1C–E). In ACAM Tg mice, the immunoreactivity of ACAM was enhanced on the cell surface of adipocytes in both immunofluorescence and immunoperoxidase studies (Fig. 1F). The staining of ACAM merged with F4/80, and the adipose tissue macrophages abundantly expressed ACAM (Fig. 1G). Both glucose tolerance and insulin sensitivity were significantly improved in ACAM Tg mice on HFHS chow (Fig. 2A–C), and serum small-sized LDL cholesterol was significantly reduced compared with WT mice (Fig. 2D). In contrast, the lipid droplets in liver tissue, liver weight, and triglyceride content were not altered in ACAM Tg mice on HFHS chow (Fig. 2E–G). The data suggested that ACAM Tg mice are protected from obesity and insulin resistance.

Although the locomotor activity, food intake, and respiratory quotient were not altered in ACAM Tg mice on HFHS chow, the weight of brown adipose tissues was significantly reduced in ACAM Tg mice on HFHS chow compared with WT mice (Fig. 3A–D). The light and electron microscopy demonstrated that lipid droplets were prominently reduced in ACAM Tg mice on HFHS chow in brown adipose tissues (Fig. 3D). The oxygen consumption rate was significantly elevated in both dark and light periods in ACAM Tg mice on HFHS chow (Fig. 3E–G). Quantitative RT-PCR demonstrated that lipid metabolism–related genes (Fabp4, Acaca, Fasn, Lpl, Hsd11b1, and Slc27a1), inflammation-related genes (Adipoq, Cd36, and Retn), and glucose metabolism–related genes (Hk1 and Slc2a4) significantly increased in ACAM Tg mice compared with WT mice (Fig. 4A). In brown adipose tissues, the gene expression of Ucp1 and Cpt1a significantly increased in ACAM Tg mice compared with WT (Fig. 4C). In contrast, most of the genes expressed in subdermal adipose tissues, skeletal muscle, and liver were not altered in ACAM Tg mice on HFHS chow (Fig. 4B, D, and E). Taken together, the overexpression of ACAM in adipocytes directly altered the biological functions of adipose tissues in the status of obesity and it increased brown fat activity. We also checked conversion of white to bright/beige adipocyte in subdermal WATs; however, we did not observe such browning in the tissues (Supplementary Fig. 1G).

To explore the function of ACAM in adipocytes, we next investigated the mRNA expression during the 3T3-L1 adipocyte differentiation. ACAM mRNA revealed the first peak at 6 h after induction, i.e., the early stage of differentiation, and was downregulated within 48 h, which coincided with the induction of Cebpb and Cebpd (Fig. 5A and Supplementary Fig. 2A). At 14 days, ACAM mRNA demonstrated the second peak after the appearance of adipocyte maturation markers, such as Cebpa, Pparg, and Lpl (Fig. 5B and Supplementary Fig. 2B). By the fractionation of the 3T3-L1 cells, ACAM localized mainly in the plasma membrane fraction, to a lesser extent in LDM and HDM fractions (Fig. 5C and Supplementary Fig. 2C). Western blot analyses demonstrated the upregulation of ACAM after the induction of differentiation (Supplementary Fig. 2D). Adeno-pCTAP-mACAM was introduced to 3T3-L1 cells, and soluble proteins were purified by CBP and SBP binding resin and finally subjected to SDS-PAGE and Coomassie blue staining (Fig. 5D). Visible bands were excised and in-gel digested with trypsin and analyzed with LC-MS/MS and MALDI-TOF/MS, by which myosin II-A and γ-actin were identified, respectively. We next confirmed the complex formation of ACAM and γ-actin by the immunoprecipitation using 3T3-L1 cells (Fig. 5E). Before the induction and at the early stage of differentiation, ACAM expressed mainly on the cell processes of 3T3-L1 cells revealed by immunofluorescence and immunogold studies (Fig. 5F–H). Before the induction, ACAM did not colocalized with myosin II-A; however, it merged with myosin II-A concentrated on the cell surface as well as perinuclear cytosol when 3T3-L1 cells fully differentiated to the mature adipocytes (Fig. 5G).

We then searched the critical inducers for the upregulation of ACAM mRNA expression. The induction with IBMX or its combination with other inducers upregulated ACAM mRNA levels; the effects were almost comparable with those from DEX/IBMX/insulin stimulation (Supplementary Fig. 3A). IBMX induced ACAM mRNA expression in a dose- and time-dependent manner, and it peaked at 6 h in the presence of 500 nmol/L IBMX (Supplementary Fig. 3B and C). IBMX is known to inhibit phosphodiesterases, stimulate adenyl cyclase activity, and increase the intracellular cAMP accumulation (8). Both 8-bromoadenosine–cAMP and forskolin, a cAMP-elevating agent, upregulated ACAM mRNA levels in a dose-dependent manner (Supplementary Fig. 3D and E). It is also known that cAMP stimulates protein kinase A (PKA) activity (9) and ACAM mRNA expression was reduced by specific PKA inhibitor (H89) in a dose-dependent manner (Supplementary Fig. 3F). In contrast, SB203580 (p38 kinase inhibitor), PD98050 (MEK [mitogen-activated protein kinase kinase 1] inhibitor), and LY294002 (phosphatidylinositol 3-kinase inhibitor) demonstrated no effects on IBMX-induced expression of ACAM mRNA (data not shown).

The transcriptional regulation of ACAM mRNA seems to be dependent on the IBMX-PKA pathway, which is a known inducer of C/EBPβ (10). We found 13 consensus C/EBPβ binding sites by analyzing the −1,741–base pair ACAM promoter region (Supplementary Fig. 4A) and observed approximately eightfold induction of luciferase activities in −584, −464, −235, −97, and −79 ACAM promoter constructs by the treatments of IBMX or adenoviral vector encoding active-form C/EBPβ (LAP) (Supplementary Fig. 4B and C). A possible C/EBPβ binding consensus site between −72 and −59, AGCTAACCCCAAAC, was mutated by using −79 ACAM promoter construct to produce −79 (MT1) and −79 (MT2) ACAM promoter constructs (Supplementary Fig. 4A), which resulted in a loss of transactivation induced by IBMX and LAP. KLF4 has been reported to induce adipocyte differentiation by activating C/EBPβ. KLF4-expressing plasmids in the combination with GATA1- and/or GATA6-expressing plasmids synergistically enhanced luciferase activity using −79 ACAM promoter construct (Supplementary Fig. 4D).

We investigated the expression of ACAM and polymerized form of actin (F-actin) by phalloidin staining. In WT mice on HFHS chow, both F-actin and ACAM were faintly stained surrounding the adipocytes in epididymal adipose tissues, while they colocalized and accentuated in a patchy fashion on the adipocytes (Fig. 6A and B). In WT mice, the distance of plasma membranes between adjacent adipocytes was ∼100–200 nm in WT mice on HFHS chow (Fig. 6C). In ACAM Tg mice on HFHS chow, we observed the zone with an intercellular space of ∼10–20 nm, the structure of zonula adherens (Fig. 6C and Supplementary Figs. 5A and 7, black arrows). The zonula adherens was continuous and characterized by strict parallelism of the adjoining cell membranes over distances of 1–20 μm (Supplementary Figs. 6 and 7). The intercellular space was occupied by homogeneous and amorphous material of low density and conspicuous bands of dense material located in the subjacent cytoplasmic matrix. In the zonula adherens, two layers of plasma membrane joined in a linear and parallel pattern, but they also demonstrated invagination or engulfment (Fig. 6C and Supplementary Figs. 5A and 6 , white arrows). The zonula adherens was associated with immunogold particles when the sections were stained with γ-actin– or ACAM-specific antibodies (Fig. 6D and Supplementary Fig. 5B, arrows). In brown adipose tissues in ACAM Tg mice on HFHS chow, similar structures of zonula adherens were observed and immunogold particles associated with ACAM were demonstrated along the zonula adherens (Supplementary Fig. 8). The ACAM-mediated homophilic adhesion of adipocytes and formation of zonula adherens promoted the actin polymerization. The formation of zonula adherens inhibited the adipocyte hypertrophy by actin polymerization, which resulted in the improvement of obesity and insulin resistance.

In 3T3-L1 adipocytes, IBMX is a major inducer of ACAM mRNA expression, and it is a known inducer of C/EBPβ. We identified C/EBPβ binding sites in the promoter regions of the ACAM gene, and luciferase-reporter gene assay demonstrated that the promoter activities were enhanced by C/EBPβ expressed by LAP. KLF4 functions as an immediate early regulator of adipogenesis to induce C/EBPβ (11), and KLF4, GATA1, and GATA6 have been reported to enhance the transcriptional activity of ACAM gene in Sertoli cells (12). In the current study, we demonstrated that the promoter activities in luciferase assay were further enhanced by the expression of KLF4, GATA1, and GATA6 and they are the critical coactivator of the transcription of the ACAM gene. Immunofluorescence and immunogold studies demonstrated the presence of ACAM at the cell processes of the 3T3-L1 cells, and we speculate that ACAM may serve the heterophilic binding by interacting with certain extracellular matrix (ECM) glycoproteins and also with cytoskeletal components such as γ-actin and myosin II-A complex.

The mature adipocytes are surrounded by ECM glycoproteins such as collagens type I, IV, V, and VI; fibronectin; and thrombospondin, and the ECM environment regulates the adipogenesis and adipocyte function (13). The remodeling of ECM surrounding the adipocytes takes place and the basal lamina are thickened in omental adipose tissues of the rats fed with a high-fat diet (14). The administration of matrix metalloproteinase inhibitor tolylsam into the mice fed with a high-fat diet resulted in lower body weight and lower subcutaneous and gonadal adipose tissue mass (15). Thus, the periadipocyte ECM remodeling is tightly related to the adipogenesis, tissue inflammation, fibrosis, insulin sensitivity, and cardiovascular diseases (13). Since periadipocyte ECM plays an important role in adipocyte biology and the close cell-cell contact of the adipocytes has not been visualized in EM (electron microscopy) observation, the role of ACAM in mature adipocytes remains elusive. ACAM/CLMP is expressed in tight junction of epithelial cells and involved in the development and maintenance of epithelial cells. Recently, we have developed ACAM knockout mice, and ACAM^−/− mice demonstrate the elongation of small intestine, dilatation of bronchi, and formation of huge renal cysts.

The presence of functional gap junction in cultured osteoblasts and adipocytes was demonstrated; the inhibition of gap junctional communication by 18-α-glycyrrhetinic acid blocks the maturation of preosteoblastic cells and converts to an adipogenic phenotype (16). A later study further confirmed that the application of gap junction inhibitor and small interfering RNA targeting connexin 43 inhibit the mitotic clonal expansion, expression of Cebpb, and adipocyte differentiation (17). Connexin 43 is a membrane phosphoprotein forming gap junction, and phosphorylation of connexin 43 is downregulated after the induction of adipocyte differentiation. After the downregulation of phosphorylation, connexin 43 is displaced from the plasma membrane and degraded by a proteosomal pathway (18). Thus, gap-junctional intercellular communication is required in mitotic clonal expansion and adipocyte differentiation. Historically, Farquhar and Palade (19) observed EMs of various epithelial cells and provided the identification of the zonula occludens (tight junction), zonula adherens (intermediary junction), and the macula adherens (desmosome). Gap junction was not delineated from tight junction until the work of Revel and Karnovsky (20), since extracellular or intermembrane gap is quite similar in tight and gap junctions (21). In later study, the treatment of samples with uranyl acetate instead of lanthanum established the presence of a 1.8-nm gap between the outer leaflets of the apposed membranes of gap junction (21). Although the presence of functional gap junction in adipocytes has been proved in previous studies, the morphological investigations by EM in adipose tissues have not been published. We also extensively searched the adipose tissues of WT and ACAM Tg mice on HFHS chow; however, we could not find the gap junction–like structures by EM. Instead, we readily recognized zonula adherens with 10–20 nm intercellular space in ACAM Tg mice on HFHS chow, in which Tg overexpression of ACAM facilitated the formation of zonula adherens. Originally, Farquhar and Palade (19) defined zonula adherens by the presence of an intercellular space (∼20 nm) occupied by homogeneous, apparently amorphous material of low density; by strict parallelism of the adjoining cell membranes over distances of 0.2–0.5 μm and by conspicuous bands of dense material located in the subjacent cytoplasmic matrix (19), the observed structures of ACAM Tg mice on HFHS chow exactly corresponded with the definition of zonula adherens.

To further characterize the molecular organization of zonula adherens in adipocytes, we performed TAP purification and immunoprecipitation studies and demonstrated that ACAM interacts with myosin II-A and γ-actin in 3T3-L1 cells. We further confirmed that ACAM and γ-actin colocalized with zonula adherens observed in adipose tissues of ACAM Tg mice by immunoelectron microscopy. Coxsackie and adenovirus receptor (CAR) is a homolog of ACAM with 35% identity, and it also interacts with and binds to actin (22). CAR is strongly expressed in the developing nervous system, it uniformly expresses on all neural cells at an initial stage, and it downregulates and then restricts to axonal and dendritic surface at more advanced stages (23). The membrane proximal Ig domain of CAR binds to a fibronectin fragment and is involved in the heterophilic interactions, while two extracellular Ig domains are involved in the homophilic interactions of CAR (23). Similarly, ACAM is localized on the cell processes of preadipocytes, downregulates after differentiation, and redistributes on the plasma membrane of 3T3-L1 cells. In mature adipocytes in ACAM Tg mice, ACAM colocalized with polymerized actin concentrated on the area of zonula adherens. Recently, the actin cytoskeleton dynamics of polymerization and depolymerization cycles have been reported to drive the adipocyte differentiation. Adipogenic stimuli downregulate RhoA–Rho-associated protein kinase signaling, induce the disruption of actin stress fibers, and result in the conversion to the monomeric globular-actin. The binding globular-actin to MKL1 (megakaryoblastic leukemia 1) inhibits the nuclear translocation of MKL1, the reduction of MKL1 in nuclei activates the transcriptional activity of Pparg gene, and it results in adipocyte differentiation (24). In fully differentiated adipocytes, cortical F-actin is again formed and it regulates the insulin-simulated translocation of GLUT4 from intracellular pool to the plasma membrane. The overexpression of ACAM in Tg mice prominently facilitates the formation of cortical F-actin on the HFHS chow. One can speculate that the enhanced formation of cortical F-actin may increase the mechanical strength per unit area, and it inhibits the adipocyte hypertrophy in the status of obesity induced by HFHS chow. In addition, the formation of cortical F-actin enhances the translocation of GLUT4, and it may improve the insulin sensitivity in ACAM Tg mice. We have also generated ACAM knockout mice; however, they were small for age and died around 14 weeks, and we were unable to analyze the size of adipocytes.

In 3T3-L1 preadipocytes, the expression of ACAM is induced by the cAMP-PKA-C/EBPβ pathway, and it concentrates on the cellular processes. ACAM expression declines and again appears on cell surface of the fully differentiated 3T3-L1 adipocytes. ACAM Tg mice are protected from obesity and insulin resistance. The current investigation also provides evidence that the mature adipocytes develop the zonula adherens with the molecular organization of ACAM and cortical F-actin in ACAM Tg mice on HFHS chow. The promotion and maintenance of cortical F-actin by targeting ACAM constitute a candidate for the new therapeutic modalities in the treatment of obesity and diabetes (Supplementary Fig. 9).

Funding. This work was supported by Japan Society for the Promotion of Science Grant-in-Aid for Scientific Research (25126716, 26293218, 26461361, 26461362, and 40620753). K.M. is supported by the Okayama Medical Foundation and is a recipient of the Biological Study Award for Encouragement (Ryobi Teien Memory Foundation).

Duality of Interest. J.W. receives speaker honoraria from Astellas, Boehringer Ingelheim, Novartis, Novo Nordisk, and Tanabe Mitsubishi and receives grant support from Bayer, Daiichi Sankyo, Kyowa Hakko Kirin, Merck Sharp & Dohme, Novo Nordisk, Otsuka, Torii, Pfizer, Takeda, Taisho Toyama, and Tanabe Mitsubishi. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. K.M., J.E., K.H., and J.W. participated in the design of the whole study. K.M., J.E., K.H., A.N., and J.W. participated in the generation of ACAM Tg mice. Cell culture studies were performed by K.M., J.E., A.N., A.K., M.S., and H.C. Electron microscopy was performed by M.F. and J.W. Western blot analyses were performed by D.O., K.T., and F.O. K.M., K.T., F.O., and J.W. conceived of the study, participated in coordination, performed the statistical analyses, and helped to draft the manuscript. All authors read and approved the final manuscript. J.W. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.