AMP-activated protein kinase (AMPK) activation inhibits nuclear translocation of Smad4 in mesangial cells and diabetic kidneys

Abstract

Diabetic nephropathy is characterized by diffuse mesangial matrix expansion and is largely dependent on the TGF-β/Smad signaling pathway. Smad4 is required for TGF-β signaling; however, its regulation has not been well characterized in diabetic kidney disease. Here, we report that high glucose is sufficient to stimulate nuclear translocation of Smad4 in mesangial cells and that stimulation of the major energy sensor AMP-activated protein kinase (AMPK) has a potent effect to block Smad4 nuclear translocation. Activation of AMPK by 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) inhibited high glucose-induced and TGF-β stimulation of nuclear Smad4. To identify which of the catalytic α-subunits may be involved, small interfering (si) RNA-based inhibition of AMPK α1- or α2-subunit was employed. Inhibition of either subunit reduced overall AMPK activity and contributed to Smad4 nuclear accumulation. In an animal model of early diabetic kidney disease, induction of diabetes was found to markedly stimulate Smad4 protein levels and enhance nuclear accumulation. AMPK activation with AICAR completely prevented the upregulation of Smad4 and reduced mesangial matrix accumulation. We conclude that stimulation of Smad4 in cell culture and in in vivo models of early diabetic kidney disease is dependent on AMPK.

with the increasing number of diabetic patients, the prevalence of diabetic nephropathy continues to increase worldwide (1). Renal matrix accumulation plays a crucial role in the pathogenesis of diabetic nephropathy. Exposure to elevated levels of glucose induces remodeling of glomerular structure through accumulation of extracellular matrix (ECM), which contributes to a progressive decline in renal function (31, 32, 35). It is now well recognized that TGF-β is a key mediator in the onset and progression of diabetic nephropathy and is closely associated with mesangial matrix accumulation (10, 47). In the progression of glomerulosclerosis, Smad proteins are central components of intracellular TGF-β signaling pathways (41). Each ligand of the TGF-β family exerts its diverse effects by binding to two types of receptors, which are known as the type I and type II receptors. The activated type I receptor phosphorylates the receptor-regulated Smads (R-Smads; Smad1, 2, 3, 5, 8), followed by translocation into the nucleus with Smad4 (common pathway Smad, Co-Smad). These complexes regulate transcription of their target genes via direct binding to DNA sequences or cofactors (19). Recently, localization of Smad4 to the nucleus was found to be critical for gene transcriptional events stimulated by TGF-β (42). In addition, high glucose regulates renal and vascular cell function via the renin-angiotensin system and formation of advanced glycation end products (AGEs). Recently, activation of the Smad2/3 pathway was also found to mediate the effects of angiotensin II (ANG II) and AGEs in vascular smooth muscle and renal cells via both TGF-β-dependent and TGF-β-independent mechanisms (17, 40, 41). As Smad4 is a necessary factor for all Smad2/3-mediated target gene regulation, a better understanding of Smad4 activation is of major importance.

5′-AMP-activated protein kinase (AMPK) is an enzyme that participates in the cellular response to energetic changes (14). AMPK consists of three subunits, designated α, β, and γ. The α-subunit of AMPK contains the catalytic domain and has two isoforms, α1 and α2, that are phosphorylated at threonine-172 (Thr-172) upon enzyme activation. AMPK is strongly regulated by the cellular AMP/ATP ratio, oxidative status, LKB, CAMkinase II, and treatment with 5-aminoimidazole-4-carboxamide-1-β-d-ribonucleoside (AICAR) (11, 24, 30, 33). Our prior studies with the adiponectin knockout mouse and mouse models of obesity-related kidney disease identified that AMPK and adiponectin play major roles in the early renal inflammation associated with obesity (30, 32). In addition, recent studies by our group and others have demonstrated that AMPK activity is reduced in models of type 1 diabetic kidney disease as well as in chronic kidney disease (5, 8, 30). We have also demonstrated that activation of AMPK has a potent role in reducing TGF-β production in type 1 diabetic kidney disease (6).

The pathway by which AMPK may inhibit the TGF-β system has not been clarified. We previously found that a transcription factor, upstream stimulatory factor 1 (USF1), mediates glucose-induced stimulation of TGF-β, and AMPK inhibits nuclear translocation of USF1 (27). A prior report indicated that AMPK inhibits TGF-β-induced matrix stimulation; however, Smad2/3 phosphorylation was not affected by AMPK activation (18), and Smad4 was not evaluated in their study. The present study focused on the relationship between the AMPK pathway and the common Smad mediator, Smad4, in in vitro and in vivo models of diabetic kidney disease.

METHODS

Materials.

Cell culture reagents were purchased from GIBCO/Invitrogen (Carlsbad, CA), AICAR and lactacystin were obtained from Sigma-Aldrich (St. Louis, MO). pAMPK^T172 and AMPKα antibodies were purchased from Cell Signaling Technology (Danvers, MA). Smad4, Smad2/3, pSmad2, pSmad3, histone H3, and β-actin antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). An AMPKα1 antibody was obtained from Abcam (San Francisco, CA). For Western blotting, an AMPKα2 antibody was obtained from Calbiochem (La Jolla, CA), and for immunohistochemistry an AMPKα2 antibody was obtained from Abcam. All small interfering (si) RNA and transfection regents were purchased from Thermo Scientific Dharmacon (Chicago, IL).

Cell culture.

A murine mesangial cell line (MMC) was used in cell culture studies, as previously described (48). Cells were incubated at 37°C with 5% CO₂-95% air and propagated in DMEM containing 10 mM d-glucose, 10% FBS, 100 U/ml penicillin, 100 mg/ml streptomycin, and 2 mM supplemental glutamine. Cells were passaged every 72 h. MMCs were plated in 100- or 60-mm cell plates; after 50% confluence, the cell culture media were replaced with serum-free DMEM with normal (5.5 mM d-glucose) or high glucose (30 mM d-glucose) for 24 h. In separate plates, MMCs were treated with 0.5 mM AICAR or vehicle from 1 h before modulation with glucose. After incubation with high glucose for 24 h, whole cell and nuclear protein were isolated as previously described (29). In another experiment, MMCs were treated with 0.5 mM AICAR 1 h before incubation with 10 ng/ml TGF-β1 (R&D Systems). After incubation with TGF-β1 for 24 h, nuclear protein was isolated.

Immunoblotting and proteosomal activity.

Whole cell and nuclear protein were isolated from MMCs, as reported previously (45). The concentration of protein was determined using protein assay kits (Pierce Biotechnology). Forty micrograms protein of whole cell lysates and 30 μg of nuclear protein were loaded per well on a 4–12% SDS-PAGE and transferred to a nitrocellulose membrane. The blot was blocked with 5% nonfat dry milk and incubated with primary antibodies at 4°C overnight. Then, secondary antibodies were applied and the signals were developed with ECL. To evaluate the role of proteasomal activity, MMCs were pretreated with lactacystin (10 μM) or vehicle for 3 h, and incubated under normal or high glucose for 24 h. AICAR or vehicle was added 1 h before incubation with normal or high glucose. As primary antibodies, rabbit polyclonal anti-pAMPKα, AMPKα, AMPKα1, fibronectin, α-smooth muscle actin (SMA), pSmad2, pSmad3, and histone H3 (dilution 1:1,000), mouse monoclonal anti-Smad4 (1:1,000), rabbit polyclonal anti-AMPKα2 (1:500), and mouse monoclonal anti-β-actin (1:3,000) were used.

Immunofluorescence assay.

For immunofluorescence assay, MMCs were grown on coverslips to 30% confluence. After incubation with 0.5 mM AICAR or vehicle for 1 h, the cells were incubated with serum-free DMEM under normal or high glucose for 24 h. After incubation, cells were washed in PBS, fixed in 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and blocked with 3% FBS in PBS. The cells were then incubated with Smad4 antibody (1:200) at 4°C overnight and Alexa Fluor 488 goat anti-mouse IgG-conjugated FITC (1:1,000; Invitrogen Molecular Probes, Eugene, OR) secondary antibody for 1 h in a dark and humidified chamber at room temperature. After washing in PBS, nuclei were stained by ProLong Gold Antifade Reagent with 4,6-diamidino-2-phenylindole (Invitrogen Molecular Probes), and cells were visualized using a confocal LSM 510 microscope (Zeiss).

Small interfering RNA transfection of MMCs.

Transfection of MMCs with small interfering (si) RNA AMPKα1 and siRNA AMPKα2 was accomplished with DharmaFECT transfection according to the manufacturer's instructions. Briefly, the cells at 30–40% confluence were transfected with different siRNAs. After incubation with each siRNA (100 nM) solution for 24 h, cells were treated with AICAR (0.5 mM) or vehicle for 1 h, and then incubated with DMEM under normal or high glucose conditions for an additional 24 h. A positive control siRNA GAPDH and a negative control siRNA pool were used simultaneously. After siRNA transfection, immunoblotting was performed to test the efficiency of transfection and pAMPKα, Smad4 expression.

Animal study.

Male C57BL/6J (wild-type; WT) mice were purchased from Jackson Laboratories (Bar Harbor, ME). Mice were given standard rodent chow and water ad libitum. WT mice aged 8 wk were divided into the following four groups (n = 6 each): 1) nondiabetic group, 2) AICAR-treated nondiabetic group, 3) streptozotocin (STZ)-induced diabetic group, and 4) AICAR-treated diabetic group. Mice in the diabetic groups received intraperitoneal injection of STZ (each 60 mg/kg) for 5 consecutive days according to the guideline from the Animal Models of Diabetic Complications Consortium (3). Blood glucose levels were confirmed 14 days after STZ injection, and only mice with blood glucose concentrations >16 mmol/l were used in the study. Mice in the nondiabetic and AICAR-treated nondiabetic groups received citrate buffer injections only. STZ was purchased from Sigma-Aldrich. In the interventional study, AICAR (500 mg/kg ip daily) was given to the mice for a period of 2 wk (from 14 to 16 wk of age). Blood glucose, urinary albumin, and urinary creatinine were measured as described previously (43). Two months after the induction of diabetes, all mice were killed, and the kidneys were harvested as described previously (43). Kidney cortex was homogenized by sonication in lysis buffer (120 mM NaCl, 1 mM EGTA in PBS, pH 7.4), lysed for 20 min on ice, and centrifuged (15,000 rpm for 30 min). Then, proteins were immunoblotted with antibodies for Smad4, and Smad4 protein levels were standardized by β-actin. F1 Akita mice were derived from a cross between male Akita C57BL/6J-Ins2^Akita and female DBA/2J (Jackson Laboratories) as described previously (6), and age-matched C57BL/6J mice were used for controls. AICAR (500 mg/kg ip daily) was given to the F1 Akita diabetic mice and controls for a period of 2 wk before harvesting of kidneys (from 10 to 12 wk of age). All animal studies were approved by the Institutional Animal Care and Use Committee of the University of California (La Jolla, CA). The urine albumin and creatinine were measured with a mouse Albuwell ELISA kit and a Creatinine Companion kit (Exocell, Philadelphia, PA) according to the manufacturer's protocol.

Histology and immunohistochemistry.

Periodic acid-Schiff (PAS)-stained sections were analyzed as described previously with slight modifications (21). To evaluate the glomerular size and mesangial matrix area, 15 randomly selected glomeruli were analyzed using i-solution software (Advanced Imaging Concepts, Princeton, NJ). Quantitative analysis was performed in a blinded manner. AMPKα1 and AMPKα2 immunoperoxidase staining were performed using the method as described previously (21). As primary antibodies, a polyclonal antibody against AMPKα1 (1:500) or AMPKα2 (1:500) were applied for 12 h at 4°C, and biotin-XX goat anti-rabbit IgG (H+L; 1:200; Invitrogen Molecular Probes) secondary antibody was applied for 1 h at room temperature. Smad4 staining was performed with modifications (21). The mouse on mouse (M.O.M.) kit (Vector Laboratories, Burlingame, CA) was used for additional blocking according to the manufacturer's protocol. The primary antibody was mouse monoclonal antibody against Smad4 (1:100) and was applied for 30 min at room temperature. The secondary antibody was M.O.M Biotinylated Anti-mouse IgG Regent (Vector Laboratories) applied for 10 min at room temperature. To evaluate intraglomerular AMPKα1- or AMPK α2-positive cells, we examined 20 randomly selected glomeruli/mouse (n = 3/group) under high magnification (×400). To evaluate intraglomerular Smad4-positive cells, we examined 15 randomly selected glomeruli/mouse (n = 3/group) under high magnification (×400), and we counted the number of Smad4-positive cells per glomerulus as follows: 1) the total number of Smad4-positive cells and 2) the number of Smad4-positive cells stained only in the nucleus. Quantitative analysis for all staining was performed in a blinded manner.

Statistics.

Data are expressed as means ± SE. ANOVA or Student's t-test was used to analyze data. P < 0.05 was considered statistically significant.

RESULTS

High glucose induced both upregulation and nuclear translocation of Smad4 in MMCs.

After incubation with high glucose for 24 h, the Smad4 protein level in MMCs was significantly increased not only in whole cell lysate (Fig. 1A) but also in nuclei (Fig. 1B). However, the Smad4 protein expression in cytoplasm was similar among groups (Fig. 1C). Mannitol added to normal glucose as an osmotic control had no effect on Smad4 nuclear translocation (data not shown). Localization of Smad4 protein in MMCs was examined by immunofluorescence staining and confocal microscopy. In normal glucose-treated MMCs, Smad4 protein was predominantly distributed in the cytoplasm. Nuclear translocation of Smad4 was observed in high glucose-stimulated MMCs (Fig. 1D).

Fig. 1.

5-Aminoimidazole-4-carboxamide-1-β-d-ribofuranoside (AICAR) could reverse high glucose (HG)-induced upregulation and nuclear translocation of Smad4 in murine mesangial cells (MMCs). Smad4 protein expression in whole cell lysate (A) and the nucleus (B) was increased markedly after incubation with HG (30 mM) for 24 h. AICAR pretreatment attenuated the effect of HG on Smad4. C: there was no effect of HG on cytoplasmic protein expression of Smad4. Protein levels were standardized by β-actin. D: immunofluorescence labeling of Smad4 in MMCs. Green = Smad4; blue = 4,6-diamidino-2-phenylindole (DAPI) nuclear stain. Magnification ×60. E: Smad4 protein expression in the nucleus was increased markedly after incubation with transforming growth factor (TGF)-β1 (10 ng/ml) for 24 h, but AICAR pretreatment attenuated the effect of TGF-β1. F: AICAR had no effect on the activation of Smad2 and Smad3. Values are means ± SE from 3–4 independent experiments. ns, Not significant. *P < 0.05. **P < 0.01. ***P < 0.001.

AMPK activation reversed the effect of high glucose on Smad4 nuclear translocation and expression of fibrosis-related proteins.

After incubation with high glucose for 24 h, levels of pAMPK decreased markedly in MMCs, and this effect was reversed by pretreatment of MMCs with AICAR (Fig. 2A). The levels of each of the α-catalytic subunits of AMPK (AMPKα1 and AMPKα2) were evaluated and found not be different among groups (Fig. 2B). To evaluate the role of AMPK activation with respect to Smad4 nuclear translocation, AICAR was administered to cells in normal and high glucose. There was a complete inhibition of Smad4 nuclear translocation in AICAR-treated cells exposed to high glucose (Fig. 1, B and D). AICAR also suppressed high glucose-induced expression of α-SMA and fibronectin (Fig. 2C). In a similar manner, TGF-β-induced Smad4 nuclear protein levels were completely prevented with AMPK activation (Fig. 1E), but no significant difference was observed in the activation of Smad2 and Smad3 (Fig. 1F).

Fig. 2.

AMP-activated protein kinase α-subunit (AMPKα) phosphorylation was inhibited by HG and restored by AICAR in MMCs. Western blot analysis of pAMPKα (A), AMPKα1, and AMPKα2 (B) is shown. NG, normal glucose. MMCs were cultured under different conditions for 24 h, and whole cell lysates were used for analysis. Protein expression of pAMPK (A), but neither AMPKα1 nor AMPKα2 (B), was suppressed by HG and restored by AICAR treatment. C: expression of α-smooth muscle actin (SMA) and fibronectin were increased by HG and suppressed by AICAR treatment. Protein expression was standardized by AMPKα (A) or β-actin (B and C). Values are means ± SE from 3–4 independent experiments. *P < 0.05. ***P < 0.001.

AICAR inhibited high glucose-induced upregulation of Smad4 through increasing ubiquitin-dependent protein degradation.

As shown in Fig. 3, pretreatment of the MMCs with lactacystin (10 μM), a specific proteasome inhibitor, eliminated the difference in protein level of Smad4 in each treatment group. The stimulation of high glucose-induced nuclear translocation was mimicked by lactacystin under normal-glucose conditions, and there was no further stimulation with high glucose. Lactacystin also blocked the effect of AICAR to inhibit nuclear translocation under high-glucose conditions. These results suggest that the ubiquitin-proteasome protein degradation of Smad4 is regulated via the AMPK pathway under normal-glucose conditions.

Fig. 3.

AICAR inhibited HG-induced protein expression of Smad4 via ubiquitin-dependent protein degradation. MMCs were treated with or without specific proteasome inhibitor lactacystin (10 μM) for 2 h and cultured under different conditions. Mean value (mean ± SE) of Smad4 protein expression was obtained from densitometric analysis and standardized by β-actin. In the absence of lactacystin, Smad4 expression in whole cell lysate was increased by HG and reversed by AICAR. In contrast, in the presence of lactacystin, there are no significant difference in Smad4 expression between the HG-stimulated group and AICAR; n = 3. **P < 0.01 vs. untreated NG. ^##P < 0.01 vs. untreated HG.

AICAR attenuated high glucose-induced Smad4 expression through either AMPKα subunit.

To determine the role of AMPKα1 and AMPKα2 in the high glucose-induced upregulation of Smad4, we evaluated the relative roles of AMPKα1 and AMPKα2. Each AMPKα subunit was specifically blocked with silencer siRNAs targeting AMPKα1 and AMPKα2 (Fig. 4, A and B). Transfection with siRNAs targeting AMPKα1 or AMPKα2 significantly downregulated the expression level of AMPKα1 and AMPKα2 (>90%, respectively), as measured by immunoblotting (Fig. 4, A and B). Moreover, transfection of AMPKα1-specific siRNA suppressed AMPKα levels by 67%, whereas AMPKα2 siRNA reduced AMPKα levels by 37% (Fig. 4C). In addition, after incubation with each siRNA, cells were treated with AICAR (0.5 mM) or vehicle for 1 h, and then incubated with 5.5 mM d-glucose or high glucose (30 mM) for 24 h. As shown in Fig. 4D, high glucose stimulated Smad4 expression in whole cell lysates equivalently whether transfected with control siRNA, AMPKα1 siRNA, or AMPKα2 siRNA. However, activation of AMPK with AICAR inhibited Smad4 expression to the same level as the control only with control siRNA but not with siRNA to AMPKα1 or AMPKα2.

Fig. 4.

AMPKα1 and AMPKα2 siRNA reduced AICAR effects on HG-induced Smad4 expression. A–C: Western blot analysis indicated that AMPKα1, AMPKα2, pAMPKα, and AMPKα expression were downregulated by AMPKα1 or AMPKα2 small interfering (si) RNA. D: suppressive effect of AICAR on Smad4 protein expression in whole cell lysate was partially abrogated by AMPKα1 and AMPKα2 siRNA. The expression of pAMPKα and AMPKα were examined in a separate membrane for reference. Each image was originally captured from the same membrane, and unused lanes were removed. Values are means ± SE from 3–4 independent experiments. **P < 0.01, *** P < 0.001 vs. nontransfection control. ^#P < 0.05 vs. AMPKα1 siRNA+HG group. ^##P < 0.01 vs. AMPKα2 siRNA+HG group. ^###P < 0.001 vs. Con siRNA+HG group. ^ΔP < 0.05. ^ΔΔP < 0.01.

Regulation of glomerular matrix accumulation and Smad4 protein in the diabetic kidney.

There was no significant difference in blood glucose and body weights between the untreated and AICAR-treated diabetic groups (Table 1). Kidney weight per body weight was increased in the STZ group; however, there was no difference between two diabetic groups (Table 1). The urinary albumin/creatinine ratio was significantly increased in the STZ group and was significantly suppressed in the STZ+AICAR group (Fig. 5A). Glomerular hypertrophy was observed in only the STZ group compared with the nondiabetic control group (Fig. 5, B and C). Mesangial matrix expansion was observed in the STZ group; however, AICAR treatment significantly reduced mesangial matrix accumulation compared with the STZ group (Fig. 5, B and D). AMPK activation was reduced in the diabetic kidneys and increased with AICAR (Fig. 6A). Overall levels of AMPKα1 and AMPKα2 did not change, as AMPKα1- or AMPKα2-positive cells in glomeruli were similar between nondiabetic mice and diabetic mice (Fig. 6B). In accordance with the in vitro study, Smad4 protein levels were increased in STZ mice kidneys, whereas AICAR significantly inhibited Smad4 protein compared with the STZ group (Fig. 6C). By immunostaining, the number of Smad4-positive cells per glomerulus was significantly increased in the STZ group compared with the nondiabetic groups, and decreased by AICAR treatment (Fig. 6, D and E). In addition, the ratio of the number of nuclear Smad4-positive cells was significantly increased in the STZ group, and decreased by AICAR treatment (Fig. 6, D and F). To further validate whether the Smad4-AMPK axis also plays a role in a different mouse model of diabetes, Smad4 immunostaining was performed in the F1 cross of DBA2J and C57BL/6J with Akita mice. Similar to the diabetic models with C57BL/6J, the F1Akita model exhibited increased Smad4-positive cells in glomeruli compared with controls and AICAR reduced the number of Smad4-positive cells in glomeruli in the F1 diabetic mice (Fig. 7).

Table 1.

Metabolic characteristics of mice (2 mo after induction of diabetes)

Group	n	Body Weight, g	Blood Glucose, mg/dl	Kidney Weight, g	Kidney Weight, mg/g body weight
Normal (Nor)	6	29.7 ± 1.5	208 ± 57	0.36 ± 0.04	12.1 ± 0.6
Nor+AICAR	6	26.7 ± 2.1	208 ± 41	0.29 ± 0.01	10.8 ± 0.8
STZ	6	23.0 ± 0.4*	542 ± 28†	0.33 ± 0.03	14.3 ± 0.3‡
STZ+AICAR	6	26.7 ± 1.5	564 ± 46†	0.36 ± 0.04	13.5 ± 0.8

Values are means ± SE. STZ, streptozotocin-induced diabetes; AICAR, 5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside.

^*P < 0.05 vs. normal.

^†P < 0.001 vs. normal and Nor+AICAR.

^‡P < 0.01 vs. Nor+AICAR.

Fig. 5.

Regulation of diabetes-induced glomerular matrix accumulation by AICAR. A: urinary albumin/creatinine ratio (UACR). The UACR of the streptozotocin (STZ)-induced diabetic group was markedly increased compared with that of the normal group, whereas it was suppressed in AICAR-treated diabetic mice (n = 6/group). B: representative images of periodic acid-Schiff (PAS)-stained kidney sections (×400 magnification). C: glomerular hypertrophy was observed in the STZ group compared with nondiabetic control. D: PAS-positive mesangial matrix was increased in the STZ group and suppressed by AICAR treatment. Fifteen randomly selected glomeruli per mouse were examined (n = 4/group). Values are the means ± SE. *P < 0.05. **P < 0.01. ***P < 0.001.

Fig. 6.

In vivo, regulation of kidney Smad4 expression by AICAR. A: Western blot analysis indicated that pAMPK expression was markedly reduced in kidney cortex in the STZ group and which was significantly attenuated by AICAR treatment. Data were standardized by AMPK and expressed as means ± SE (n = 3/ group). B: representative images of AMPKα1 and AMPKα2 immunostaining (×400 magnification) and the ratio of positively stained cells to the total cells in the glomeruli. Twenty randomly selected glomeruli per mouse were examined (n = 3/group). C: kidney protein level for Smad4 and β-actin were analyzed by Western blotting. Smad4 protein expression in kidney cortex was increased in the STZ group and which was significantly inhibited by AICAR treatment. Data were standardized by β-actin and expressed as means ± SE (n = 3/group). D: representative images of Smad4-stained kidney sections in normal or diabetic mice (×400 magnification). E: average number of Smad4-positive cells in glomeruli. Smad4-positive cells were reduced in AICAR-treated diabetic mice compared with untreated diabetic mice. F: ratio of the number of Smad4-positive cells stained only in the nucleus to the total number of Smad4-positive cells. The nuclear Smad4-positive cells were markedly increased in STZ-induced diabetic mice and decreased by AICAR treatment. Fifteen randomly selected glomeruli per mouse were examined (n = 3/group). Values are the means ± SE. *P < 0.05. **P < 0.01. ***P < 0.001.

Fig. 7.

AICAR suppressed Smad4 expression in the glomeruli of F1Akita diabetic mice. Representative images of Smad4 immunostaining in Normal (Nor), Nor+AICAR, F1Akita, and F1Akita+AICAR groups (×400 magnification; n = 3/group).

DISCUSSION

Hyperglycemia plays a pivotal role in the pathogenesis of diabetic nephropathy, and TGF-β1 has been recognized as a key mediator (46). Previous studies have demonstrated that TGF-β activity and its downstream Smad signaling pathway play a critical role. High glucose has been shown to activate the Smad signaling pathway, as demonstrated by Smad2 and Smad3 phosphorylation and nuclear translocation, and this pathway has been demonstrated to induce type I collagen α1(I) synthesis in tubular epithelial cells and mesangial cells (12, 13, 36). Smad4 has been found to play a potent role in TGF-β-induced stimulation of collagen α1(I), PAI-1, and fibronectin in mesangial cells (36). A recent study indicated that deletion of Smad4 inhibits progressive renal fibrosis in the obstructive nephropathy model and TGF-β1-induced collagen I expression by fibroblasts (20). Apart from TGF-β signaling, Smad signaling also mediates ANG II- and AGE-induced matrix stimulation under control of high glucose (16).

AMPK activity has been found to be reduced in the diabetic kidney (4, 15, 47). Moreover, both metformin and the thiazolidenediones (TZDs), which are currently used as hypoglycemic drugs, activate AMPK independently of serum glucose levels (9). In type 1 diabetic rats, metformin increases renal AMPK phosphorylation and attenuates renal hypertrophy without affecting hyperglycemia (15). In the present study, AICAR activation of AMPK reduced glomerular volume increase but did not affect diabetic renal hypertrophy. Recently, activation of AMPK with AICAR was found to prevent or reduce parameters of diabetic nephropathy in experimental models (5–8, 30). Accumulating evidence also suggests that AMPK plays an important role in the fibrogenesis of many diseases, including heart failure, liver disease, and chronic asthma (18, 23, 34, 44). Sasaki et al. (28) recently reported that both metformin and AICAR treatment significantly inhibit myocardial fibrosis and contractile dysfunction in a canine heart failure model. In the present study, we found that AMPK activation inhibited high glucose-induced expression of extracellular matrix-related proteins and PAS-positive mesangial matrix in diabetic mice.

The mechanism of how AMPK activation inhibits matrix production is unclear. Lim et al. (18) reported that AMPK activation can inhibit TGF-β-induced liver fibrosis, but not affect TGF-β-stimulated phosphorylation, nuclear localization, or DNA-binding activity of Smad2/3. In the present study, we also confirmed that AMPK activation did not affect TGF-β-induced phosphorylation of pSmad2 or pSmad3. Lim et al. also found that AICAR attenuates TGF-β-induced Smad3 interaction with transcriptional coactivator p300, along with a reduction of Smad3 acetylation (18). However, little is known about the relationship between Smad4 and AMPK. Our data are the first to implicate Smad4 as a target of AMPK activation. In mesangial cells, AMPK activation inhibits TGF-β- or high glucose-induced nuclear translocation of Smad4. Furthermore, our data are relevant to multiple upstream mediators of Smad signaling apart from TGF-β. Inhibiting Smad4 nuclear translocation would be beneficial to prevent ANG II- and AGE-induced matrix regulation as the Smad signaling pathway has been implicated with these stimuli in models of diabetic complications (17, 40). Our recent data demonstrating that USF1 nuclear translocation is also inhibited by AMPK activation suggest that there may be a common pathway by which AMPK may regulate several transcription factors (27). It has also been established that AMPK activation inhibits NF-κB activation and nuclear localization of p65 (26, 39).

It is well known that the stability of Smads is regulated by ubiquitin-mediated degradation in proteasomes and lysosomes. The size of the Smad pool in both unstimulated and stimulated (with TGF-β) cells is controlled by ubiquitination (22). For instance, the E3 ligases JAB1/CSN5 (37) and SCF^β-TrCP1 have been demonstrated to promote degradation of wild-type Smad4 (38). Moreover, the overexpression of oncogenic Ras has been shown to decrease Smad4 protein levels, which was partially reversed by an inhibitor of the ubiquitin-proteasome pathway (25). On the other hand, Baskin et al. (2) reported that AMPK enhances the ubiquitin-proteasome pathway in mouse heart via transcriptional regulation of E3 ligases, including atrogin-1 and muscle RING finger protein 1 (MuRF1). In our study, pretreatment of MMCs with lactacystin significantly attenuated the effect of AICAR on the high glucose-induced increase in Smad4 expression, suggesting ubiquitin-proteasome protein degradation of Smad4 may be enhanced by AICAR through activating the AMPK pathway.

To confirm which isoform of the α catalytic subunit of AMPK is primarily involved in the regulation of Smad4, the isoform-specific siRNA of AMPKα1 and AMPKα2 were used in MMCs. Both AMPKα1 and -α2 are present in mesangial cells and contribute to AMPK activation. The levels of AMPK phosphorylation were reduced after either AMPKα1 or AMPKα2 was knocked down, and there was an accompanying reduction in total AMPKα expression. Furthermore, the effect of AICAR on the high glucose-induced increase in Smad4 expression was reduced after either AMPKα1 or AMPKα2 was knocked down. Thus a partial reduction of either AMPKα isoform is sufficient to reduce the effect of AICAR.

In conclusion, we have demonstrated an important role of the AMPK pathway in high glucose-enhanced Smad4 expression. AMPK phosphorylation can reduce the Smad4 protein level stimulated by high glucose in mesangial cells and by hyperglycemia in diabetic kidneys. Therefore, AMPK activation may become an effective therapeutic target for matrix accumulation via inhibiting Smad4 in diabetic nephropathy and possibly other progressive renal fibrotic disorders.

GRANTS

Our projects are supported by grants from a Veterans Affairs MERIT Award (to K. Sharma) and National Institute of Diabetes and Digestive and Kidney Diseases Grants U01 DK060995, DP3 DK094352–01, and DK083142 (to K. Sharma), the JSPS international Training Program (ITP) and the Uehara Memorial Foundation (to S. Miyamoto).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: J.Z., S.M., and Y.-H.Y. performed experiments; J.Z. and S.M. analyzed data; J.Z. and S.M. prepared figures; J.Z. and S.M. drafted manuscript; S.M. and K.S. interpreted results of experiments; S.M. and K.S. edited and revised manuscript; K.S. provided conception and design of research; K.S. approved final version of manuscript.