SREBP-1 activation by glucose mediates TGF-β upregulation in mesangial cells

Lalita Uttarwar, Bo Gao, Alistair J. Ingram, Joan C. Krepinsky

Abstract

Glomerular matrix accumulation is a hallmark of diabetic nephropathy. Recent studies showed that overexpression of the transcription factor sterol-responsive element-binding protein (SREBP)-1 induces pathology reminiscent of diabetic nephropathy, and SREBP-1 upregulation was observed in diabetic kidneys. We thus studied whether SREBP-1 is activated by high glucose (HG) and mediates its profibrogenic responses. In primary rat mesangial cells, HG activated SREBP-1 by 30 min, seen by the appearance of its cleaved nuclear form (nSREBP-1), EMSA, and by activation of an SREBP-1 response element (SRE)-driven green fluorescent protein construct. Activation was dose dependent and not induced by an osmotic control. Site 1 protease was required, since its inhibition by AEBSF prevented SREBP-1 activation. SCAP, the ER-associated chaperone for SREBP-1, was also necessary since its inhibitor fatostatin also blocked SREBP-1 activation. Signaling through the EGFR/phosphatidylinositol 3-kinase (PI3K) pathway, which we previously showed mediates HG-induced TGF-β1 upregulation, and through RhoA, were upstream of SREBP-1 activation (Wu D, Peng F, Zhang B, Ingram AJ, Gao B, Krepinsky JC. Diabetologia 50: 2008–2018, 2007; Wu D, Peng F, Zhang B, Ingram AJ, Kelly DJ, Gilbert RE, Gao B, Krepinsky JC. J Am Soc Nephrol 20: 554–566, 2009). Fatostatin and AEBSF prevented HG-induced TGF-β1 upregulation by Northern blot analysis, and HG-induced TGF-β1 promoter activation was inhibited by both fatostatin and dominant negative SREBP-1a. Chromatin immunoprecipitation analysis confirmed that HG led to SREBP-1 binding to the TGF-β1 promoter in a region containing a putative SREBP-1 binding site (SRE). Thus HG-induced SREBP-1 activation requires EGFR/PI3K/RhoA signaling and SCAP-mediated transport to the Golgi for its proteolytic cleavage. Activated SREBP-1 binds to the TGF-β promoter, resulting in TGF-β1 upregulation in response to HG. SREBP-1 thus provides a potential novel therapeutic target for the treatment of diabetic nephropathy.

  • epidermal growth factor receptor
  • transforming growth factor-β1 promoter
  • phosphoinositide 3-kinase

the kidney is an important site of diabetic microvascular complications, and hyperglycemia is central to glomerular matrix accumulation. Although strict glucose control and inhibition of the renin-angiotensin system are effective in delaying the development of nephropathy, disease progression often occurs (31). A better understanding of the pathogenesis, leading to the development of new treatment approaches, is thus required.

Glomerulosclerosis, the pathological hallmark of diabetic nephropathy (13, 31), is triggered by a complex interplay of numerous factors. In high glucose (HG) concentrations, mesangial cells (MC) synthesize extracellular matrix proteins (33). The prosclerotic cytokine TGF-β is an important mediator of HG-induced matrix protein synthesis in MC (20). Recent studies indicate a role for the transcription factor sterol-responsive element-binding protein (SREBP)-1 in glomerular matrix protein and TGF-β1 upregulation. Increased renal expression or activation of SREBP has also been noted in models of diabetic nephropathy (24, 42, 49). Importantly, overexpression of SREBP isoforms (1a or 1c) leads to a phenotype similar to that of diabetic nephropathy, marked by albuminuria, glomerulosclerosis, and elevated expression of TGF-β, fibronectin, and collagen IV in glomeruli (21, 45).

SREBPs are transcription factors most extensively studied in lipid homeostasis. Three isoforms exist, i.e., SREBP-1a, -1c and 2, with 1a and 1c generated from alternate transcription start sites from a single gene (11). SREBP-1a and 1c are expressed in MC (22). In its inactive precursor form of 120 kDa, SREBP resides in the endoplasmic reticulum (ER) membrane. The activation paradigm is that in the setting of low intracellular cholesterol, SREBP cleavage-activating protein (SCAP) binds to and escorts SREBP from the ER to the Golgi. Here, SREBP is processed in sequence by two membrane-associated proteases, site 1 protease (S1P) and S2P, releasing the mature, N-terminal form of the protein (∼68 kDa). This translocates to the nucleus as a dimer and binds promoters of SREBP target genes at SRE (sterol-regulatory element) sequences (2, 11). However, increasing appreciation for non-sterol-mediated SREBP activation is now emerging. For example, shear stress led to proteolytic cleavage and nuclear translocation of SREBP-1 and -2 by 1–2 h in endothelial cells (28, 30). Growth factors including insulin, EGF, and PDGF were also shown to activate SREBP (17, 52). Whether HG activates SREBP-1 in MC, however, is unknown.

In general, SREBPs are weak transcription factors requiring cooperation with coactivators, most commonly Sp1 (43, 44). The recruitment of distinct coactivators renders specificity to gene transcription by SREBPs (4, 34). SREBPs have been well described to control the regulation of genes related to lipid synthesis (11). Although in vivo overexpression of SREBP-1 was associated with matrix protein accumulation and increased TGF-β1 production (21, 45), direct regulation of matrix gene transcription by SREBP has only been demonstrated for collagen 6a1 in 3T3 cells (12). Furthermore, while overexpression of SREBP-1c in MC increased TGF-β expression, it is not known whether SREBPs have a direct transcriptional effect on TGF-β1.

In this work, we thus studied whether HG can activate SREBP-1, and the potential signaling mediators involved. We further assessed whether HG-induced TGF-β1 upregulation requires SREBP-1 activation, and whether SREBP-1 directly regulates activation of the TGF-β1 promoter.

MATERIALS AND METHODS

Cell culture.

Primary MC were obtained from glomeruli of Sprague-Dawley rats by differential sieving and cultured in DMEM with 20% fetal calf serum (Invitrogen), streptomycin (100 μg/ml), and penicillin (100 U/ml). Experiments used cells between passages 6 and 15.

The medium contained 5.6 mmol/l d-glucose. Either 24.4 mmol/l d-glucose (final 30 mmol/l) or mannitol or l-glucose was added for HG or osmotic controls, respectively. MC were serum deprived for 24 h before treatment. Inhibitors were added before HG as follows: 500 μM AEBSF, 1 h (Calbiochem); 20 μM fatostatin, 4 h (Chembridge); 1 μM AG1478, 30 min (Sigma); 0.5 μM PD168393, 30 min (Calbiochem); 20 μM LY294002, 30 min (Sigma); 100 nM wortmannin, 1 h (Sigma); 10 μM Akt inhibitor VII, 1 h (EMD); 2 μg/ml C3 transferase, 4 h (Cytoskeleton); and 10 μM Y-27632, 30 min (Calbiochem).

Protein analysis.

Cells were lysed and protein was extracted as published, with the addition of ALLN to the lysis buffer at 25 μg/ml (Calbiochem) (27). Lysates were centrifuged at 4°C, 14,000 rpm for 10 min. The supernatant (50 μg) was separated on SDS-PAGE, and Western blotting was performed. Antibodies used were monoclonal anti-RhoA (1:500, Santa Cruz Biotechnology), monoclonal SREBP-1 (1:500, Santa Cruz Biotechnology), polyclonal cAMP response element binding protein (CREB; 1:1,000, Cell Signaling), and goat polyclonal connective tissue growth factor (CTGF; L-20, 1:2,000, Santa Cruz Biotechnology). Nuclear lysates were prepared as published and briefly described below under EMSA (26). Equality of loading of nuclear preparations was assessed by immunoblotting for CREB.

A RhoA pull-down assay was performed as described previously (27). Briefly, cells were lysed in hypertonic buffer and GTP-bound RhoA was immunoprecipitated from cleared lysate with 30 μg of glutathione-S-transferase-tagged Rhotekin RhoA binding domain (Cytoskeleton). The beads were washed, and the immunoprecipitate was resolved on 15% SDS-PAGE. Membranes were probed with anti-RhoA. The lysate (40 μg) was also probed for RhoA to ensure equality across conditions.

EMSA.

After treatment, nuclear extracts were prepared as published (26). Cells were lysed in hypotonic buffer, homogenized and sedimented at 16,000 g for 20 min at 4°C. Pelleted nuclei were resuspended in hypotonic buffer containing 0.42 M NaCl and 20% glycerol, rotated for 30 min at 4°C, centrifuged as above, and the supernatant containing nuclear proteins was collected. Nuclear proteins (3 μg) were incubated for 5 min at room temperature with a biotin-labeled SREBP consensus oligonucleotide (sense ATCACCCCAC, antisense GTGGGGTGAT, Sigma) as per the manufacturer's instructions (Pierce). In some cases, either 100× excess cold probe or a monoclonal antibody to SREBP-1 (2A4, Santa Cruz Biotechnology) were also included. Reaction mixtures were electrophoresed in a 6% polyacrylamide gel, transferred, and DNA cross-linked to a nylon membrane (Amersham), then probed with horseradish peroxidase-conjugated streptavidin antibodies (1:300, Pierce).

Green fluorescent protein quantification.

MC plated to 85% subconfluence were transfected with 1 μg of an SRE-green fluorescent protein (GFP) plasmid (kindly provided by Dr. R. Austin) (9) in six-well plates using Effectene (Qiagen). MC were serum-deprived overnight 24 h after transfection, then exposed to treatment including glucose for 1 h. GFP was quantified as described elsewhere (5): cells were then washed in cold PBS, lysed in 0.2 N HCl, and centrifuged to remove cell debris. GFP fluorescence in the supernatant was measured in a fluorometer (Gemini EM, Molecular Devices) at λ = 475 excitation and λ = 510 emission. Readings were normalized to protein concentration.

Luciferase assay.

MC plated to 85% subconfluence were transfected with 0.5 μg of a TGF-β1 promoter-luciferase construct (kindly provided by Dr. N. Kato) and 0.05 μg pCMV-β-galactosidase (β-gal, Clontech) using Effectene (Qiagen). Where indicated, cells were also transfected with empty vector pcDNA or dominant negative SREBP-1a (Y335A; kindly provided by Dr. A. Schulze) (39). MC were serum-deprived overnight 24 h after transfection, then exposed to glucose for 24 h. Lysis was achieved with reporter lysis buffer (Promega, Madison, WI) using one freeze-thaw cycle, and luciferase and β-gal activities were measured on clarified lysate using specific kits (Promega) with a Berthold luminometer and a plate reader (420 nm), respectively. β-gal activity was used to adjust for transfection efficiency.

Northern blot analysis.

Total RNA (20 μg), extracted using TRIzol (Invitrogen), was separated on a formaldehyde-agarose gel and transferred to a nylon membrane (Hybond, Amersham Biosciences). Hybridization was performed with random primed digoxygenin-11-dUTP-labeled cDNA probes prepared from TGF-β cDNA amplified by PCR. Hybridized probes were detected using alkaline phosphatase-labeled anti-digoxygenin antibodies and CDP-star as a substrate. Kits and reagents were from Roche Applied Science (Laval).

Chromatin immunoprecipitation.

After treatment, MC were cross-linked with 1% formaldehyde, then washed and scraped into cold PBS with protease inhibitors. After centrifugation, the cell pellet was resuspended in buffer (20 mM HEPES, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, protease inhibitors), incubated on ice for 20 min, and centrifuged. The pellet (nucleus) was resuspended in breaking buffer (50 mM Tris·HCl pH 8.0, 1 mM EDTA, 150 mM NaCl, 1% SDS, 2% Triton X-100, protease inhibitors), sonicated 5 × 10 s, and Triton buffer was added (50 mM Tris·HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl, 0.1% Triton X-100). Samples were precleared with blocked protein G-Sepharose beads, an aliquot was reserved as the input, and the remainder was divided to immunoprecipitate with control rabbit IgG (Jackson ImmunoResearch) or SREBP-1 (sc-8984X, Santa Cruz Biotechnology) antibodies followed by incubation with protein G beads. Samples were washed three times in Triton buffer, and SDS buffer was added [62.5 mM Tris·HCl, pH 6.8, 200 mM NaCl, 2% SDS, 10 mM DTT, 2 μl of proteinase K (40 mg/ml)], then vortexed and incubated at 65°C overnight to reverse cross-linking. DNA was isolated using phenol/chloroform extraction and resuspended in dH2O. PCR was performed using the rat TGF-β1 promoter primers 5′-ATCCCGGTGGCATACTGAG (F) and 5′-CACGGAACTTCGGAGAGC (R) at 60°C annealing temperature for 35 cycles, or the fatty acid synthase (FAS) promoter primers as published (15).

Animal studies.

Animal studies were carried out in accordance with McMaster University and Canadian Council on Animal Care guidelines. Diabetes was induced in male 7- to 8-wk-old B6129SF1/J mice (Jackson Laboratory) by daily injection of 50 mg/kg streptozotocin (STZ; Sigma) intraperitoneally for 5 days. Control mice received 0.1 M citrate buffer, pH 4.5. Mice with a glucose of ⩾17 mmol/l measured by a glucometer (Precision Xtra, Medisense) were followed. Monitoring was weekly for weight and blood glucose and monthly for blood pressure by tail-cuff plethysmography. Mice were euthanized after 10 mo of diabetes. For immunohistochemistry, cortical sections stored in OCT were processed as previously described (50). Primary antisera used were polyclonal rabbit SREBP-1 (1:50, sc-367, Santa Cruz Biotechnology) and polyclonal sheep von Willebrand factor to identify glomeruli (1:100, Cedarlane).

Statistical analyses.

Statistical analyses were performed with SPSS14 for Windows using one-way ANOVA and Tukey's honest significant difference for post hoc analysis. A t-test was used for experiments with only two groups. A P value <0.05 (2-tailed) was considered significant. Data are presented as means ± SE, and the number of repetitions are denoted as n.

RESULTS

SREBP-1 is activated by glucose.

Increased SREBP-1 levels have been observed in diabetic kidneys and in MC exposed to longer-term HG (48 h) (24, 45, 49). Whether HG activates SREBP-1 in MC, however, is unknown. To assess SREBP-1 activation, we first detected appearance of the cleaved, mature 68 kDa form by immunoblotting. As seen in Fig. 1A, SREBP-1 is activated as early as 30 min by 30 mM of HG, and activation persists up to 6 h. There is some dose dependence to its activation, with glucose concentrations <25 mM unable to efficiently activate SREBP-1 (Fig. 1B). Two bands running close together are generally observed in immunoblots, corresponding to the isoforms SREBP-1a and -1c. To exclude osmotic activation of SREBP-1, responses to equimolar concentrations of mannitol or l-glucose were assessed. Figure 1C shows a lack of SREBP-1 activation by these osmotic controls. Here, SREBP-1 activation was assessed by appearance of the active form in nuclear preparations, with CREB used as a loading control. For all subsequent experiments, nuclear isolates were prepared as this provided more consistent and reproducible detection of SREBP-1 activation. Finally, we assessed the response of SREBP-2 to HG. Figure 1D shows that this isoform is not activated by glucose.

Fig. 1.

Sterol-responsive element-binding protein (SREBP)-1 is activated by glucose. A: mesangial cells (MC) were treated with high glucose (HG) for the indicated times, and SREBP-1 activation was assessed by immunoblotting for the cleaved, mature form (mSREBP-1). *P < 0.05 HG vs. control; n = 6. B: MC were treated with the indicated concentrations of glucose for 1 h, and SREBP-1 activation was assessed by immunoblotting. *P < 0.05, ‡P < 0.001 vs. control; n = 5. Immunoblotting for actin confirms equal loading across conditions for A and B. C: MC were treated with 30 mM HG or equimolar concentrations of mannitol or l-glucose for 1 h. SREBP-1 activation was assessed by appearance of its mature form in nuclear extracts. Immunoblotting for cAMP response element binding protein (CREB) was used to confirm equal loading. D: MC were treated with HG for the indicated times, and SREBP-2 activation was assessed by appearance of its mature form in nuclear extracts. E: immunofluorescence of sections of type 1 diabetic mouse kidney cortex demonstrates an increase in SREBP-1 (green) in glomeruli, which are identified by staining for the endothelial cell marker von Willebrand factor (VWF; red). STZ, streptozotocin. The boxed diabetic glomerulus is shown as an inset in higher power. F: periodic acid-Schiff (PAS) staining of control and diabetic mouse kidney cortex demonstrates increased glomerular matrix deposition in the diabetic mouse.

To determine whether SREBP-1 was also activated in diabetic glomeruli in vivo, we used immunofluorescence to detect SREBP-1 in mice after 10 mo of diabetes. Figure 1E shows increased SREBP-1 in diabetic glomeruli, identified by the endothelial cell marker von Willebrand factor. No colocalization is apparent, even at higher power (inset), and the distribution is characteristic of the mesangium, suggesting that SREBP-1 is primarily increased in MC. Figure 1F shows the increased matrix deposition seen by periodic acid-Schiff staining in diabetic glomeruli.

Glucose-activated SREBP-1 is transcriptionally active.

To confirm increased SREBP-1 DNA binding, EMSA was performed using a probe consisting of an SREBP consensus binding sequence. Figure 2A shows increased DNA binding by 1 h of HG. This was eliminated by concurrent incubation with excess cold probe and significantly decreased by preincubation with an SREBP-1 antibody (Fig. 2B). A lack of supershift with antibody incubation may be attributed to masking of the DNA-binding site by the antibody. In MC treated with the osmotic control mannitol, no increase in SREBP-1 DNA binding was seen, confirming a glucose-specific effect on its activation (Fig. 2C). The functionality of SREBP-1 DNA binding was next confirmed using a GFP reporter plasmid under the control of the SRE (9). After exposure of MC to HG for 1 h, GFP fluorescence was quantified and normalized to protein concentration as described in materials and methods. Figure 2D confirms HG-induced SREBP-1 transcriptional activation.

Fig. 2.

Glucose-activated SREBP-1 is transcriptionally active. A: MC were treated for the indicated times with HG, and SREBP-1 DNA binding was assessed by EMSA. *P < 0.05 HG vs. control; n = 4. B: nuclear samples from cells treated for 1 h with HG were incubated in the presence or absence of 100-fold excess cold probe or an antibody to SREBP-1. C: EMSA was conducted on nuclear extracts from MC treated for 1 h with HG or equimolar concentrations of mannitol. D: MC were transfected with the SREBP-responsive plasmid SRE-green fluorescent protein (GFP) and treated with HG for 1 h. GFP production, driven by SREBP activation, was quantified by fluorescence measurement. †P < 0.01 HG vs. control; n = 5.

Glucose-induced SREBP-1 activation requires serine proteases and SCAP.

Cholesterol-regulated activation of SREBP-1 is dependent on SCAP-mediated translocation to the Golgi for generation of a transcriptionally active factor by a two-step cleavage process involving S1P and S2P (2). However, SREBP cleavage and hence activation may also be mediated by other enzymes independently of serine protease activity and SCAP . These include caspases-3, -4 ,and -12, as well as CPP32, a cysteine protease (19, 36, 47). We thus first determined whether SCAP and S1P/S2P are required for glucose-induced SREBP-1 activation. Figure 3A shows that the serine protease inhibitor AEBSF completely prevents HG-induced SREBP-1 activation, demonstrating a requirement for S1P, a serine protease, in cleavage. We next used the recently described inhibitor of SCAP function, fatostatin (25), to confirm that SCAP-to-Golgi cleavage is also required. Figure 3B clearly shows that SREBP-1 activation by glucose is dependent on SCAP. Furthermore, SREBP-1 induction of GFP expression in cells transfected with pSRE-GFP was also blocked by fatostatin (Fig. 3C).

Fig. 3.

SREBP-1 activation requires serine proteases and SREBP cleavage-activating protein (SCAP). A: MC were treated with the S1P protease inhibitor AEBSF (500 μM, 1 h), then treated with HG for 1 h. The mature form of SREBP-1 in nuclear preparations was assessed by immunoblotting. †P < 0.01 HG vs. others; n = 5. B: SREBP-1 activation was similarly assessed after pretreatment with the SCAP inhibitor fatostatin (20 μM, 4 h). C: MC transfected with SRE-GFP were treated with fatostatin followed by HG for 1 h, and GFP fluorescence was quantified. †P < 0.01 HG vs. others; n = 4.

EGF receptor/phosphatidylinositide 3-kinase/Akt signaling mediates SREBP-1 activation by glucose.

We have previously shown that transactivation of the EGF receptor (EGFR) mediates glucose-induced matrix (collagen I) and TGF-β upregulation in MC (50, 51). We had observed EGFR transactivation within 10 min of HG exposure (50). Furthermore, EGF itself has been shown to activate SREBP-1 in glioblastoma cells (17). We thus determined whether EGFR transactivation was required for SREBP-1 activation. Figure 4, A and B, shows that two different inhibitors of EGFR, AG1478 and PD168393, both prevented SREBP-1 cleavage and nuclear translocation. Functionality was confirmed using the pSRE-GFP plasmid. As can be seen in Fig. 4C, SREBP-1-driven GFP induction by HG was prevented by AG1478.

Fig. 4.

The EGF receptor (EGFR) is required for SREBP-1 activation by glucose. MC were treated with EGFR inhibitors AG1478 (1 μM, 30 min; A) or PD168393 (0.5 μM, 30 min; B) before 1 h of HG. Appearance of the mature form of SREBP-1 in nuclear preparations was assessed by immunoblotting. ‡P < 0.001 HG vs. others; n = 4 (A). †P < 0.01 HG vs. others; n = 5 (B). C: MC transfected with SRE-GFP were treated with AG1478 followed by HG for 1 h, and GFP fluorescence was quantified. ‡P < 0.01 HG vs. others; n = 5.

We have also previously established a requirement for PI3K/Akt signaling downstream of the EGFR in mediating matrix and TGF-β upregulation in MC (50, 51). We thus next tested whether these were involved in SREBP-1 activation. Figure 5A shows that two distinct phosphatidylinositide 3-kinase (PI3K) inhibitors, LY294002 and wortmannin, both prevented HG-induced SREBP-1 activation as assessed by appearance of the mature form in the nucleus. Similarly, inhibition of Akt using Akt inhibitor VII also prevented HG-induced SREBP-1 activation (Fig. 5B).

Fig. 5.

SREBP-1 activation requires phosphoinositide 3-kinase (PI3K)/Akt signaling. A: MC were treated with the structurally distinct PI3K inhibitors LY294002 (20 μM, 30 min) or wortmannin (100 nM, 1 h) before 1 h of HG. Appearance of the mature form of SREBP-1 in nuclear preparations was assessed by immunoblotting. *P < 0.05 HG vs. others; n = 5. B: SREBP-1 activation was assessed as in A after pretreatment with Akt inhibitor VII (10 μM, 1 h). *P < 0.05 HG vs. others; n = 5.

RhoA/Rho-kinase activation regulates glucose-induced SREBP-1 activation.

RhoA and its downstream kinase Rho-kinase have been shown to enable SREBP-2 activation in endothelial cells exposed to shear stress by enhancing its transport to the Golgi for cleavage (28). Furthermore, we have shown that this pathway is an important mediator of glucose-induced matrix upregulation in MC (37), and in some cases PI3K signaling mediates RhoA activation (32, 38). We thus assessed whether RhoA/Rho-kinase might be involved in SREBP-1 activation by HG. As can be seen in Fig. 6A, the cell-permeable RhoA inhibitor C3-transferase prevented HG-induced SREBP-1 activation. Inhibition of Rho-kinase with Y-27632 similarly prevented HG-induced SREBP-1 activation (Fig. 6B). To establish whether PI3K mediated RhoA activation by 1 h of HG, we tested the effects of the PI3K inhibitor LY294002 on RhoA activation. Figure 6C shows that RhoA is activated by HG at 1 h and that this does require PI3K.

Fig. 6.

RhoA/Rho-kinase activation regulates glucose-induced SREBP-1 activation. A: MC were treated with the cell-permeable RhoA inhibitor C3 transferase (2 μg/ml, 4 h) before 1 h of HG. Appearance of the mature form of SREBP-1 in nuclear preparations was assessed by immunoblotting. *P < 0.05 HG vs. others; n = 4. B: SREBP-1 activation was assessed as in A after pretreatment with the Rho-kinase inhibitor Y-27632 (10 μM, 30 min). †P < 0.01 HG vs. others; n = 3. C: MC were treated with the PI3K inhibitor LY294002 (20 μM, 30 min) followed by 1 h of HG. GTP-bound (active) RhoA was assessed by immunoprecipitation with glutathione-bound glutathione-S-transferase (GST)-Rhotekin as described in materials and methods. Total RhoA was detected in supernatants. *P < 0.05 HG vs. others; n = 4.

SREBP-1 mediates glucose-induced TGF-β1 upregulation.

Increased renal TGF-β1 expression has been observed in vivo with SREBP-1a or -1c overexpression (21, 45). Since our data show that SREBP-1 is activated by HG, and upregulation of TGF-β1 in response to HG has been well established (55), we asked whether SREBP-1 mediated glucose-induced TGF-β1 upregulation. We first performed Northern blot analysis to assess TGF-β1 transcript levels. Since longer HG exposure is required for TGF-β upregulation (35), 24-h treatment was used for these studies. Figure 7A shows that the upregulation of TGF-β1 transcript by HG was blocked by both the serine protease inhibitor AEBSF and the SCAP inhibitor fatostatin, supportive of a role for SREBP-1 in transcriptional upregulation of this cytokine. To confirm this, we assessed TGF-β promoter activation using a promoter-luciferase construct. Fatostatin prevented TGF-β1 promoter activation by 24 h of HG, as seen in Fig. 7B. Furthermore, overexpression of dominant negative SREBP-1a Y335A also prevented TGF-β1 promoter activation by glucose, while transfection of the empty vector pcDNA had no inhibitory effect (Fig. 7C). These data confirm a direct effect of SREBP-1 on TGF-β1 upregulation.

Fig. 7.

SREBP-1 mediates glucose-induced TGF-β1 upregulation. A: MC were treated with HG for 24 h in the presence or absence of the S1P inhibitor AEBSF (500 μM, 1 h) or SCAP inhibitor fatostatin (20 μM, 4 h). TGF-β1 transcript was detected by Northern blot analysis, and densitometry was normalized to 28S values. †P < 0.01 HG vs. others; n = 5. B: MC were transfected with a TGF-β1 promoter-luciferase construct, and promoter activity was assessed after 24 h of HG ± fatostatin. *P < 0.05 HG vs. others; n = 4. C: MC were transfected with a TGF-β1 promoter-luciferase construct concurrently with either empty vector pcDNA or dominant negative SREBP-1a Y335A (dnSREBP1a), and promoter activation was assessed after 24 h of HG. ‡P < 0.001 HG pcDNA vs. others; n = 6.

SREBP-1 is a relatively weak transcription factor, functioning in coordination with other transcription factors, largely Sp1, to effect gene transcription (43, 44). Sp1 was previously shown to be activated by glucose in MC (7, 14). Analysis of the human, mouse, and rat TGF-β promoters (J04431, M57902, NM021578) revealed a potential SREBP binding sequence in close proximity to an Sp1 site within −100 base pairs of the start codon. We thus performed chromatin immunoprecipitation (ChIP) analysis using primers encompassing the region −230 to 0 to identify whether SREBP-1 directly binds to the TGF-β1 promoter. Figure 8A shows that, after HG exposure for 24 h, SREBP-1 does bind to this region of the TGF-β1 promoter. This is inhibited by fatostatin. No promoter binding was detected in control lanes in which IgG was used for immunoprecipitation. Data are quantified in Fig. 8B. Thus SREBP-1 activated by HG directly binds to the TGF-β1 promoter, enabling promoter activation and TGF-β upregulation. To further confirm functional activation of SREBP-1 and as a positive control for ChIP studies, we assessed binding of SREBP-1 to its well-known downstream target FAS. As expected, Fig. 8C shows that after HG treatment for 24 h, SREBP-1 binds to the FAS promoter. Immunoprecipitation with IgG alone did not result in promoter amplification. Data are quantified in Fig. 8D.

Fig. 8.

Glucose induces SREBP-1 binding to the TGF-β1 promoter. A: chromatin immunoprecipitation (ChIP) was performed as described in materials and methods to detect association of SREBP-1 with the TGF-β1 promoter after treatment of MC with HG for 24 h in the presence or absence of fatostatin. B: quantification of the lanes from SREBP-1 immunoprecipitates is shown. *P < 0.05 HG vs. others; n = 4. C: MC were treated with HG for 24 h with or without fatostatin, after which ChIP was performed to detect SREBP-1 association with the fatty acid synthase (FAS) promoter. D: quantification of the lanes from SREBP-1 immunoprecipitates is shown. †P < 0.01 HG vs. others; n = 3.

TGF-β upregulation leads to deposition of matrix proteins. CTGF is a well-known regulator of the matrix effects of TGF-β (29). We thus assessed the consequences of SREBP-1 inhibition on glucose-induced upregulation of CTGF. After HG exposure for 48 h, a significant increase in CTGF protein levels was seen. This was blocked by fatostatin (Fig. 9). Thus SREBP-1 inhibition prevents upregulation of TGF-β1 and its downstream target CTGF.

Fig. 9.

Connective tissue growth factor (CTGF) upregulation by glucose requires SREBP activation. CTGF protein levels were assessed by immunoblotting in MC treated with HG for 48 h with or without the SCAP inhibitor fatostatin (20 μM, 4 h). ‡P < 0.01 HG vs. others; n = 6.

DISCUSSION

SREBP-1 activation has been observed in diabetic rodent kidneys, and its overexpression is associated with matrix and TGF-β1 upregulation. Our studies now define, for the first time, early SREBP-1 activation by glucose in MC, a key cell source of glomerular matrix accumulation. We further identify that SREBP-1 activation occurs through the canonical cholesterol-sensitive pathway involving SCAP and serine protease cleavage and identify an upstream signaling pathway leading to its activation (Fig. 10). Our studies also now establish a novel role for SREBP-1 in the direct activation of the TGF-β1 promoter and provide a basis for further exploration of the therapeutic potential of targeting SREBP-1 in diabetic kidney disease.

Fig. 10.

Schematic showing the proposed pathway for SREBP-1 activation by HG. Exposure of MC to glucose leads to early transactivation of the EGFR and PI3K/Akt activation. These mediate SREBP-1 activation, potentially through facilitating translocation of its precursor form from the ER to the Golgi for cleavage to the mature transcription factor. Akt-induced stabilization of mature SREBP-1 may further contribute to its increased transcriptional activity. SREBP-1 activation additionally requires RhoA/Rho-kinase signaling downstream of PI3K. Activation of this pathway may also facilitate SREBP-1 translocation to the Golgi for processing. Upon activation, SREBP-1 directly binds to the TGF-β1 promoter and increases TGF-β1 transcription, thereby enhancing the profibrotic response to glucose.

Overexpression of either the transcriptionally active 1a or 1c isoform of SREBP-1 in the kidney leads to a phenotype similar to that of diabetic nephropathy. This is characterized by albuminuria, glomerulosclerosis, and elevated expression of TGF-β, fibronectin, and collagen IV in glomeruli (21, 45). Indeed, increased nuclear (active) SREBP-1 and -2 have been observed in whole kidneys in rodent models of diabetic nephropathy including Akita type 1 diabetic and db/db type 2 diabetic mice (24, 42, 49). Our data now extend this to show increased SREBP-1 activation in glomeruli of type 1 diabetic kidneys, supporting a potential role for this transcription factor in the pathogenesis of diabetic glomerular lesions.

Glucose has been shown to activate SREBP-1 in limited cell types, including myotubes and tubular epithelial and hepatic cell lines (16, 18, 24). In primary rat hepatocytes, however, glucose did not activate SREBP (1). Whether HG activates SREBP-1 in MC has not as yet been described. Our data now show increased activation of SREBP-1 in MC within 1 h of HG exposure. SREBP-1 has also been found in other glomerular cell types, including podocytes and endothelial cells (23, 30). The activation of SREBP-1 by glucose in these cells has yet to be studied. However, factors known to be involved in the pathogenesis of diabetic nephropathy have been shown to activate SREBP-1 in these cells. These include mechanical stress and VEGF (30, 54). It is thus quite possible that glucose-induced SREBP-1 activation would additionally be seen in these other glomerular cell types.

The most well-described paradigm for SREBP activation is that of SCAP-assisted transport of SREBP-1 to the Golgi for cleavage by S1P/S2P to produce the mature transcription factor. This pathway is classically regulated by cholesterol levels, such that a decrease in cellular cholesterol enables SCAP/SREBP release from the ER membrane (2). However, increasing appreciation for non-sterol-mediated regulation of SREBP activation is now emerging. Some of these stimuli, including shear stress and insulin, also activate SREBP through this pathway (28, 53). Others, however, utilize different proteases for SREBP-1 cleavage. These include caspase-3, -4, and -12, as well as the cysteine protease CPP23 (19, 36, 47). Our studies using AEBSF and the SCAP inhibitor fatostatin indicate involvement of the SCAP/S1P-S2P pathway for SREBP-1 activation by HG.

The signaling leading to non-sterol-mediated activation of SREBPs is not fully defined and is dependent on the cell stimulus. In endothelial cells exposed to shear stress, SREBP-1 activation required signaling through β1-integrin, FAK, and c-Src (30). Similar activation of SREBP-2 was observed and this was dependent on RhoA/Rho-kinase signaling, which facilitated SREBP translocation to the Golgi for processing (28). We have previously shown that RhoA is activated by HG in MC (37). Our data now show that RhoA activation, and its downstream kinase Rho-kinase, are also required for HG-induced SREBP-1 activation. Whether the function of RhoA signaling in this setting is to facilitate SREBP-1 movement from the ER to the Golgi for processing requires further study.

Our data further show that PI3K activation mediates both RhoA and SREBP-1 activation by glucose. SREBP-1 activation in response to various growth factors, the most studied of which is insulin, was also shown to require PI3K (6, 10, 52). We have previously shown that PI3K and its downstream kinase Akt are activated by HG in MC (50), and our data now show that both are required for SREBP-1 activation. Akt signaling may affect SREBP at different levels, including its translocation from the ER to the Golgi and the stability of the active transcription factor. The mature (nuclear) form of SREBP is rapidly degraded by the ubiquitin-proteasome system in a manner regulated by glycogen synthase kinase 3 (GSK3β) phosphorylation, which permits its interaction with the ubiquitin ligase SCFFbw7 and subsequent degradation. Akt phosphorylation of GSK3 is inhibitory, thereby increasing the stability of nuclear SREBP-1 and -2 (3, 46). A requirement for Akt in the translocation of the SCAP/SREBP complex from the ER to the Golgi has also been shown in response to insulin in rat hepatocytes. Here, Akt phosphorylated SREBP-1c on serine residues, which increased affinity of the SCAP/SREBP-1c complex for proteins on COPII transport vesicles, enhancing transport to the Golgi and thus SREBP cleavage (52). In some settings, Akt activation of the mammalian target of rapamycin (mTOR) complex 1 was required for SREBP-1 activation, but not stabilization of the nuclear form (40, 41). However, mTOR was not required for insulin-induced SREBP-1c activation in rat hepatocytes despite being PI3K dependent (1). Taken together, PI3K/Akt signaling may affect SREBP-1 activation through at least two mediators, GSK3 and mTORC1, as well as through direct phosphorylation of the precursor SREBP protein. Which downstream factors are involved is dependent on cell type and stimulus. How PI3K/Akt enable glucose-induced SREBP-1 activation requires further study.

Our previous studies have shown that transactivation of the EGFR is an important mediator of glucose-induced matrix and TGF-β1 upregulation in MC through the activation of PI3K/Akt (50, 51). With our data now showing that PI3K/Akt signaling is required for SREBP-1 activation, this suggested a role for the EGFR in activation of SREBP-1. Indeed, our studies do show a role for the EGFR, with two separate inhibitors blocking glucose-induced SREBP-1 activation. Although growth factors including EGF were shown to activate SREBP-1 (10, 17, 52, 54), this is the first description of a role for EGFR transactivation in regulation of SREBP-1 signaling.

The observed elevation of TGF-β and matrix protein expression in vivo with overexpression of active SREBP-1a or -1c was attributed to cellular accumulation of excess lipids (21, 45). Adenoviral overexpression of nuclear SREBP-1c in transformed MC, however, also increased TGF-β expression, while basal TGF-β expression was decreased by dominant negative SREBP-1c (21). Since HG is known to increase TGF-β upregulation (8), and our data showed that HG increased SREBP-1 activation, this suggested a potential direct role for SREBP-1 in mediating HG-induced fibrotic responses in MC. Indeed, analysis of the TGF-β1 promoter demonstrated a potential SRE site within the first 100 base pairs of the start codon. Of particular interest, this is in close proximity to an Sp1 site, with Sp1 being a well-recognized coactivator for SREBP (43, 44). Furthermore, in MC, HG-induced activation of Sp1 (7, 14) was shown to be required for TGF-β1 promoter activation (7). We thus analyzed whether TGF-β1 upregulation was dependent on SREBP-1. Through both Northern blot analysis and assessment of TGF-β promoter activation, we confirmed a regulatory role for SREBP-1 in TGF-β induction and transcription in response to HG. Furthermore, using ChIP analysis, we demonstrated that SREBP-1 directly binds to the TGF-β1 promoter in response to HG. Our data thus show a novel role for SREBP-1 in the direct upregulation of TGF-β.

In aggregate, our studies show an important and direct role for SREBP-1 in glucose-induced TGF-β1 upregulation, suggesting the potential therapeutic benefit of its blockade in diabetic nephropathy. One study has thus far shown decreased albumin excretion in SREBP-1c null mice compared with wild-type after 2 mo of diabetes (21), although this duration of diabetes in mice is insufficient for observing effects on renal function or histology. Agonists of the farnesoid X receptor (FXR), a negative regulator of SREBP-1c, have also been shown to decrease not only SREBP-1c but also fibronectin and TGF-β expression after 1 and 12 wk of treatment in db/db type 2 diabetic mice. However, since treatment also decreased glucose levels, it was difficult to attribute results entirely to the treatment's effect (22). Conversely, deletion of FXR increased susceptibility of mice to diabetic nephropathy, although mice were also placed on a high-fat diet to accelerate disease (48). Long-term studies will be required to assess whether inhibition of SREBP-1 activation will prevent or delay the development of diabetic nephropathy.

GRANTS

J. Krepinsky gratefully acknowledges the support of the Canadian Diabetes Association (CDA) and Canadian Institutes of Health Research (CIHR), as well as St. Joseph's Healthcare for support of nephrology research.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: L.U. and J.C.K. provided conception and design of research; L.U. and B.G. performed experiments; L.U. and J.C.K. interpreted results of experiments; A.J.I. and J.C.K. edited and revised manuscript; J.C.K. analyzed data; J.C.K. prepared figures; J.C.K. drafted manuscript; J.C.K. approved final version of manuscript.

ACKNOWLEDGMENTS

We thank Dr. N. Kato (University of Tokyo) for providing the TGF-β1 promoter-luciferase construct, Dr. R. Austin (McMaster University, Hamilton) for providing the pSRE-GFP plasmid, and Dr. A. Schulze (London Research Institute) for providing the SREBP-1a-Y335A construct.

REFERENCES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
View Abstract