Hyperglycemia is an independent risk factor and initiator for cardiac complications in diabetes, which has been reported to cause Cx43 down-regulation in various tissues, including endothelial cells and cardiomyocytes [7,22]. As a major gap junction protein, reduced Cx43 might affect GJIC function and cardiac conduction velocity, contributing to ventricular tachycardia and other arrhythmias. It would trigger mitochondrial shape change and cytochrome c release [23], finally leading to cell hypertrophy and apoptosis [24,25]. In our previous study, we also reported [6] that high glucose downregulated Cx43 protein level in neonatal rat cardiomyocytes. But, it still could not explain the discordant results of Cx43 expression in STZ-induced diabetic rat models [7–9]. There should be other mechanisms of elevating cardiac Cx43 expression in chronic diabetes. Recent evidence implicated that the accumulation of AGE in diabetes contribute to the pathogenesis of cardiovascular diseases [11,26]. Its effect on cardiac Cx43 expression, however, remains unknown.

In the present study, we found that Cx43 protein redistributed in AGE-infused rat heart and AGE significantly elevated Cx43 expression both in vivo and in cultured rat cardiomyocytes. The activation of PKC and Erk MAPK pathway might be involved in the mechanism. But, unexpectedly, the increased Cx43 did not enhance GJIC function.

Cx43 is the major gap junction protein in cardiomyocytes that mediates the function of cell coupling and leads to orderly electrical conduction [3]. The quantity alteration or redistribution of Cx43 contributes to various arrhythmias in diabetes. Its expression could be regulated at the transcriptional or post-transcriptional levels. Recently, Lin et al.[7] reported that in a three-week STZ-induced diabetic rat model, the total Cx43 amount and the immunoreactive particles for Cx43 at the intercalated disks were significantly decreased, but it could be alleviated by either lysosomal (NH4CL, Leupeptin) or proteasomal inhibitor (ALLN), which suggested that diabetes might cause proteolytic degradation of Cx43. It was consistent with another study [9] demonstrating the reduced Cx43 in a diabetic heart. Howarth et al.[8], however, found that in a 12-week diabetic rat model, total and phosphorylated Cx43 were elevated in ventricular myocytes, leading to prolongation of QRS and QT intervals. Therefore, the results of cardiac Cx43 remodeling still differed, which might be due to the miscellaneous factors in diabetic rat models. In diabetes, high glucose is regarded as a major initiator for cardiac complications. But, it also accelerates the formation and accumulation of AGE, which is the product of nonenzymatic glycoxidation of proteins and participates in various cardiovascular diseases. As high glucose has been previously recognized for its downregulation effect on cardiac Cx43 [6], the different results of Cx43 regulation might be partly due to the accumulation of AGE with the development of diabetes, which increased total Cx43 expression as our present study noticed.

We also demonstrated that AGE elevated both protein and mRNA levels of Cx43, indicating the transcriptional regulation, which differed from the possible post-transcriptional regulation of high glucose for not changing Cx43 mRNA level [6]. Our results suggested that besides high glucose, AGE might play an important role in cardiac electrical remodeling of diabetes. But, it was not consistent with the results of AGE on Cx43 in other cell lines. It has been demonstrated that in human hepatoma cell line (SKHep1) [27] and human aortic endothelial cells (HAEC) [28], the Cx43 protein was reduced by AGE treatment, while Rubenstein et al.[29] recently reported that AGE did not alter the Cx43 expression in endothelial cells. The different effect of AGE might be partly associated with the treated cell lines; even the proliferation rate differed in AGE-treated SKHep1 and HAEC.

We also noticed that Cx43 protein scattered and lost orderly distribution between intercalary discs in AGE-infused rat hearts. Recent studies [16,17] have already reported that disorganized connexins were associated with non-functional gap junctions and reduced cell excitability. And the increased lateral Cx43 was unlikely to enhance intercellular coupling, partly due to the possible dissociation of other junction proteins, such as N-cadherin and β-catenin. Therefore, the Cx43 upregulation by AGE might further contribute to conduction disturbance in a diabetic heart.

In our study, AGE did not affect the cell viability in neonatal rat cardiomyocytes determined by MTT assay, which was not consistent with the results in other cell lines. Sakata et al. reported that the low dosage of AGE did not produce any change in proliferation of smooth muscle cells (SMCs) at concentrations of up to 20 mg/L, but decreased that of SMCs at 40 mg/L [30]. In endothelial progenitor cells, co-culturing with AGEs (50–200 mg/L) also inhibited cell proliferation [31]. Another study, however, noticed that AGE accelerated cell proliferation assessed by MTT assay at 200 mg/L in cardiac fibroblast [32]. Therefore, the unremarkable change in cell viability in our present work might be partly due to the different cell type. But, AGE increased RAGE, the specific receptor for AGE, in a dose- and time-dependent manner and peaked at 200 mg/L for 24 h. In a 16-week diabetic rat model [33], the atria showed abundant expression of RAGE with diffuse interstitial fibrosis, which could be prevented by administration of an AGE formation inhibitor, OPB-9195. It was consistent with our results in AGE-infused rat heart, indicating RAGE protein upregulation by AGE treatment. But, in AGE-treated cardiomyocytes, RAGE did not show essential change during the first 12 h. In a recent report conducted by Sun et al., they also demonstrated the time- and dose-dependent manner of RAGE expression by AGE treatment in endothelial progenitor cells, while it was only slightly elevated after 12 h [20]. In another study [34] about the effect of H₂O₂ on pancreatic cells, RAGE expression did not change after 6 h of H₂O₂ treatment, but it increased significantly after 24 h. To date, however, few studies mentioned the mechanism of RAGE elevation. It has been reported [35] that the RAGE promoter region had three potential NF-kappaB-like and two SP1 binding sites, indicating that RAGE activation might be associated with activation of the NF-kappaB pathway. Inactivated NF-kappaB is a latent form in the cytoplasm, which masks nuclear localization signals. Recently, Zhang et al. demonstrated that exposure of rat neonatal cardiomyocytes to Glycated-BSA (400 mg/L) increased nuclear NF-kappaB expression consistently after 24 h, which suggested the enhanced translocation of NF-kappaB to nucleus [19]. Therefore, it could be a limitation for the lack of investigation for NF-kappaB expression and its activating time in our study. It might be possible that RAGE acted more as a link to AGE and engaged downstream signal transduction pathways, which in turn upregulated RAGE expression and amplified these cellular effects. Recently, many studies [12–14] focused on the pathogenic effect of AGE-RAGE system in cardiomyocytes. We used siRNA technology to knock down RAGE expression, and it abrogated the upregulation effect of AGE on Cx43. Therefore, AGE might show its effect on cardiac Cx43 by RAGE activation.

Unexpectedly, the cellular gap junction function was not enhanced by AGE treatment. It was believed that function of cardiac gap junction depended on multiple levels involving Cx43 expression, phosphorylation and degradation [36,37]. But, in recent studies, the effect of Cx43 phosphorylation on cell coupling still differed in various cell types [38,39]. Some findings demonstrated that phosphorylated Cx43 acted as the main functional form of Cx43 and dephosphorylation could impair the electrical coupling and signal transmission through gap junction [40–42]. In our study, AGE increased both non-phosphorylation and phosphorylation expression of Cx43, and the possible mechanism was the activation of the PKC and Erk MAPK pathways. Erk was reported to be critical for cardiac Cx43 upregulation and increased GJIC [43]. However, PKC-induced Cx43 Ser368 phosphorylation resulted in disassembly of the gap junction and reduced channel conductance [39,44], indicating the negative effect of GJIC by PKC activation. It might partly explain the unchanged GJIC function by AGE treatment in spite of Cx43 upregulation. It was also believed that the phosphorylation sites of Cx43 could affect their justified location and membrane association. A recent study [7] further pointed out that Cx43 was redistributed into cell periphery or cytoplasm in diabetes, leading to the suppression of intercellular impulse propagation. Therefore, the AGE-increased Cx43 proteins were possibly unable to assemble properly to exert their activity, which contributed to the unremarkable change of GJIC function.

The present study demonstrated that AGE-RAGE system elevated Cx43 expression both in vitro and in vivo. Further work will still be required to determine the phosphorylation sites of Cx43 and their corresponding effect in a diabetic heart.

AGE was prepared by incubating bovine serum albumin (BSA, 50 mg/mL) with D-glucose (0.5 M) in phosphate buffered saline (PBS, 0.2 M, pH 7.4) for 12 weeks at 37 °C. Unincorporated glucose was dialyzed extensively overnight against PBS. Unmodified BSA was prepared under the same conditions without glucose as a control. As other report described [45], AGE was identified by a fluorescence spectrophotometer (Hitachi, Tokyo, Japan) using a 440 nm emission wavelength on excitation at 370 nm, which confirmed the higher intensity of AGE in AGE-modified BSA than that in unmodified BSA. The AGE solution was filtered to be sterile by 0.22 μM Millex GP filter unit (Millipore, Billerica, MA, USA) and confirmed to be endotoxin free (<2.5 μL of endotoxin) by using an endotoxin testing kit (Chromogenic TAL Endpoint Assay Kit, Xiamen, China). Receptor for AGE (RAGE) polyclonal antibody was obtained from Chemicon (Temecula, CA, USA). Monoclonal rabbit antibodies, including anti-Erk, anti-phospho-Erk, anti-JNK, anti-phospho-JNK, anti-p38, anti-phospho-p38 and anti-phospho-PKC α/β2, were purchased from Cell Signaling Technology (Boston, MA, USA). PKC polyclonal antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The corresponding selective inhibitors, including PD98059, SP600125, SB203580 and GF109203X, were obtained from Merck (Darmstadt, Germany). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT), Lucifer yellow dye and Cx43 polyclonal antibody were obtained from Sigma (St. Louis, MO, USA). HRP-marked anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody was supplied by Kangchen (Shanghai, China).

Male SD rats aged eight weeks and weighed 200 g were obtained from the experimental animal center of Zhejiang University. Rats were injected intraperitoneally (i.p.) with BSA alone or AGE-BSA at a dosage of 40 mg/kg/d for 28 days (n = 5 per group) [46,47]. At the end of the experiments, rats were sacrificed, and hearts were collected for immunohistochemisty. Animal procedures were performed according to the Guide for the Care and Use of Laboratory Animals (NIH, revised 1996), and approval was granted by the university ethics review board.

Hearts were fixed and embedded in paraffin. Samples were then cut into 4-μM sections and used for immuohistochemisty study, including RAGE (1:500) and Cx43 (1:2000) staining, respectively. They were followed by incubation with specific HRP-marked second antibodies and DAB (Dojindo, Kumamoto, Japan). After PBS rinsing, samples were counterstained in hematoxylin, dehydrated and mounted. The Cx43 and RAGE expressions were analyzed by ImageJ 1.43u software (http://rsb.info.nih.gov/ij) (National institutes of Health, Bethesda, MD, USA).

Neonatal rat cardiomyocytes were cultured, as we previously described [6]. Briefly, the heart tissue isolated from the heart of 1–2 day old Sprague-Dawley rats (provided by the experimental animal center of Zhejiang University) were dissected and digested in a Ca²⁺ and Mg²⁺ free PBS containing 0.1% trypsin and 0.1% type II collagenase for 10 min at 37 °C. DMEM cell culture medium (glucose concentration: 5.5 mM) containing 10% fetal bovine serum (FBS) was added into the collected cell supernatant to stop the digestive effect of trypsin and collagenase. The above steps were repeated until the tissues were completely digested. The cell suspension was then centrifuged and suspended in DMEM cell culture medium with 10% FBS in a humidified 5% CO₂/95% air atmosphere at 37 °C. The fibroblast in the cell suspension was reduced by pre-plating for 1 h due to the differential cell adhesion [48] and the addition of 5′-BrdU (0.01mM) during the first three days to inhibit the growth of fibroblast.

Cells were seeded at 2 × 10⁴ cells/well in 96-well culture plates and incubated in serum-free medium for 24 h before experiment. The cardiomyocytes were then treated with AGE at various concentrations for 24 h or by various incubation periods. The MTT solution (final concentration, 0.5 mg/mL) was added to each well for an additional 4 h at 37 °C. The supernatants were then removed, followed by 200 μL dimethyl sulfoxide (DMSO) added into each well. The absorbance was read at a wavelength of 490 nm. Each independent experiment was repeated five times.

Cells in 6-well plates were harvested and washed with PBS for three times, followed by lysing on ice in 100 μL RIPA solution containing 50 mM Tris (pH 7.4), 150mM NaCl, 1% NP-40, 5% deoxycholic acid, 0.1% SDS, 1 mM EDTA, 10 mM NaF, 1 mM Na₃VO₄, 1 mM dithiothreitol, 1 mM PMSF, 2 μg/mL leupeptin for 30 min. Proteins were quantified with the Bio-Rad DC Protein Assay Kit II (Bio-Rad, Hercules, CA, USA). Samples containing equal amounts of protein were run on a 10% tris-glycine gradient gel, transferred to PVDF membranes and blocked for 1 h with 5% nonfat milk in TBST (Tris-buffered solution containing 0.1% Tween 20) at room temperature. Membranes were then soaked with primary antibodies overnight at 4 °C followed by secondary antibody incubation for 1 h at room temperature. Finally, the membranes were reacted with enhanced chemiluminescence (ECL) reagents and exposed by Image Quant LAS-4000 (Fujifilm, Tokyo, Japan). Band densities were determined by an image Multi-Gauge Software (Fujifilm, Tokyo, Japan). Each experiment was repeated at least 3 times.

TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) was used to isolate total RNA, which was converted into cDNA by murine leukemia virus reverse transcriptase (Promega, Mannheim, Germany). The primers for amplification of specific rat mRNA were as follows: Cx43 (116 bp), 5′-TCTGCCTTTCGCTGTAACACT-3′ and 5′-GGGCACAGACACGAATATGAT-3′; RAGE (139 bp), 5′-CAGGGTCACAGAAACCGG-3′ and 5′-ATTCAGCTCTGCACGTTCCT-3′; GAPDH (152 bp), 5′-ATTCTCCCACGGCAAGTT-3′ and 5′-CGCCAGTAGACTCCACGACATA-3′ (sense and antisense, respectively). Real-time PCR was performed on the ABI 7500 cycler (Applied Biosystems, CA, USA) using SYBR Premix Ex Taq reagent (Takara, Shiga, Japan). The relative amount of target gene expression was normalized by GAPDH as an internal control.

The function of gap junctional intercellular communication (GJIC) was assessed by scrape loading dye transfer assay as previously described. Cells cultured in 35 mm dishes were incubated in preheated (37 °C ) 1% Lucifer Yellow dye solution (dissolved in Ca²⁺ and Mg²⁺ free PBS) after they were rinsed with warm PBS (37 °C ). A surgical blade was used to make scrape lines through the cell monolayer. After 3 min, the dye solution was removed. The cells were rinsed with PBS and fixed by 4% paraformaldehyde for 1 min. The distance of the dye transferred from the scraped margin was observed with an inverted fluorescent microscope.

Small interfering RNA molecules (siRNA) were designed by GenePharma (Shanghai, China). The siRNA (Up: 5′-GCUAGAAUGGAAACUGAACTT-3′; Down: 5′-GUUCAGUUUCCAUUCUAGCTT-3′) targeted at RAGE mRNA were used to knock down the expression of RAGE. A scrambled siRNA (Up: 5′-UUCUCCGAACGUGUCACGUTT-3′; Down: 5′-ACGUGACACGUUCGGAGAATT-3′) was designed as a negative control. The siRNA diluted in Hiperfect transfection reagent (Qiagen, Hilden, Germany) was premixed with serum-and antibiotic-free DMEM medium at room temperature for 10 min. Then, the cells in 6-well plates were incubated with 500 μL mixed solution containing 100 nM siRNA, followed by additional 500 μL culture medium after 4 h. 24 h later, the culture medium was replaced by DMEM medium with penicillin/streptomycin and 10% FBS. Experiments were carried out 48 h after completion of the transfection.

Results were expressed as mean ± SEM. for at least three individual experiments. Statistical analysis between groups was performed by one-way ANOVA. The level of p < 0.05 was considered statistically significant.