ERK5 Regulates Glucose-Induced Increased Fibronectin Production in the Endothelial Cells and in the Retina in Diabetes

Abstract

Purpose.: Fibronectin (FN) production and deposition in the tissue is a characteristic feature of diabetic retinopathy. ERK5 is a recent member of the mitogen activated protein kinase (MAPK) family, which plays a critical role in cardiovascular development and maintaining endothelial cell integrity. The aim of this study was to investigate the role of ERK5 signaling in glucose-induced FN overproduction.

Methods.: Dermal-derived human microvascular endothelial cells (HMVECs) and human retinal microvascular endothelial cells (HRMECs) were used in this study. FN mRNA levels and secreted FN protein levels were measured using real-time PCR and ELISA, respectively. Constitutively active MAPK/ERK kinase 5 (MEK5 [CAMEK5]) adenovirus was used to upregulate ERK5. Dominant negative MEK5 (DNMEK5) and ERK5 siRNA (siERK5) were used to downregulate ERK5. Parallel retinal tissues of diabetic rats were examined.

Results.: A significant decrease of FN was observed at both protein and mRNA levels following CAMEK5 transduction in basal as well as in high glucose. DNMEK5 transduction led to further enhancement of glucose-induced increased FN expression. siERK5 treatment led to an increase of FN synthesis. Retinal tissues of diabetic rats showed FN upregulation and ERK5 downregulation. TGFβ1 mRNA and phosphorylated Smad2 were markedly suppressed by CAMEK5 transduction with and without glucose treatment. On the other hand, siERK5 transfection enhanced TGFβ1 mRNA expression. Exogenous nerve growth factor supplementation resulted in elevated phosphorylated and total ERK5 with and without glucose treatment.

Conclusions.: Our experiments demonstrated a novel mechanism of glucose-induced increased FN production in diabetic retinopathy, which is mediated through decreased ERK5 signaling.

Introduction

Extracellular matrix (ECM) protein overproduction is a characteristic feature of diabetic retinopathy. Studies in our lab and others have shown that the synthesis of fibronectin (FN), one of the abundant ECM proteins is upregulated in diabetes as well as in endothelial cells (ECs) treated with high level of glucose. ^1–4 FN, a glycoprotein of 250 kilodalton (KD), is an important component of the extracellular matrix. FN regulates various functions of vascular ECs, such as adhesion, migration, growth, and proliferation. ^5–7 FN and its splice variants also play roles in new vessel formation, by providing outside-in signaling between the matrix and the ECs. ^8,9  

Increased FN expression has been demonstrated in several tissues in diabetes. We, and others, have shown increased FN mRNA and protein expression in the retina of diabetic animals. ^3,4 We have demonstrated that endothelin-1 (ET-1) mediates glucose-induced increased FN synthesis via activation of nuclear factor-κB (NF-κB), activating protein-1 (AP-1), and transcriptional co-activator p300. ^1–3,10 We have further shown that ET-1 and TGFβ1 are major regulators of FN production. ¹¹ Multiple signaling pathways are involved in molecular changes in the ECs. Several studies have shown that glucose-induced protein kinase C (PKC) and mitogen activated protein kinase (MAPK) activation mediate FN mRNA and protein expression in the ECs. ^12,13 ERK5 is a new member of MAPK family. ^14,15 It remains unknown whether ERK5 regulates FN overproduction in diabetic retinopathy. 

ERK5, also known as big MAPK1 (BMK1), is the most recently identified member of MAPK family. ^14,15 ERK5 shares high homology in the N-terminal kinase domain with ERK1/2, but has a unique long C-terminal, which has a transcriptional activation domain. ^14,15 MAPK/ERK kinase 5 (MEK5) is the specific MAPK kinase for ERK5 activation, which works through phosphorylation of the Thr-Glu-Tyr (TEY) activation motif in N-terminal. ^14–16 It has been shown that ECs express high levels of ERK5. ¹⁷ Studies on ERK5 knockout mice have shown that the ERK5 pathway is critical for endothelial function and for maintaining blood vessel integrity. ¹⁸ Our recent studies have demonstrated that ERK5 negatively regulates ET-1 and VEGF expression in diabetic retinopathy. ^19,20 FN interacts with ET-1 and VEGF in the pathogenesis of chronic diabetic complications. ^8,9 Thus, it is possible that FN is also regulated by ERK5 signaling. Moreover, ERK5 can be activated by neurotrophins in neuronal cells and other cell types. ^21–24 Whether neurotrophins regulate ERK5 signaling in the ECs in hyperglycemia remains to be explored. Hence, in this study, we investigated whether ERK5 plays any role in glucose-induced FN overproduction and the mechanisms of such regulation. We used microvascular ECs, as microvessels are a major target in diabetic retinopathy. We further expanded the investigation to study whether such changes are important in the retina of diabetic animals. 

Materials and Methods

Cell Culture

Dermal-derived human microvascular endothelial cells (HMVECs) and human retinal microvascular endothelial cells (HRMEC) were used in the study. HMVECs were obtained from Lonza (Walkersville, MD). HRMECs were purchased from Olaf Pharmaceuticals (Worcester, MA). Both HMVECs and HRMECs were grown in endothelial cell basal medium 2 (EBM-2; Lonza) containing human epidermal growth factor (hEGF), 0.1%; Hydrocortisone, 0.04%; gentamycin, 0.1%; fetal bovine serum (FBS), 10%; VEGF, 0.1%; human basic fibroblast growth factor, 0.4%; long R3 insulin-like growth factor, 0.1%; Ascorbic Acid, 0.1%. In EBM-2, the glucose concentration was 5 mM/L. Cells at 80% confluence were growth arrested by incubation in serum-free medium overnight prior to incubation with high glucose (HG; 25 mM/L D-glucose; Sigma-Aldrich, Oakville, Ontario) or osmotic control (OC; 25 mM/L L-glucose; Sigma-Aldrich) of the same concentration. To determine the effects of nerve growth factor (NGF) on ERK5 activation, a dose-dependent (10 and 100 ng/mL) study of recombinant human NGF (R&D Systems, Minneapolis, MN) was performed. Control cells were cultured with vehicle (sterile PBS containing 0.1% BSA). 100 ng/mL NGF was shown to be the optimal dose. Cells were seeded in 6-well plates, cultured overnight, and then treated with or without NGF for 24 hours. All experiments were performed at least in triplicates. Samples were collected for protein and RNA extractions. 

Viral Gene Transfer

ERK5 is activated through the phosphorylation of TEY motif by MEK5. Constitutively active human recombinant MEK5 (CAMEK5) adenovirus and dominant negative human recombinant MEK5 (DNMEK5) adenovirus (Cell Biolabs, San Diego, CA) were used to upregulate or downregulate ERK5 signaling, respectively. The dual phosphorylation site S311/S315 in the CAMEK5 adenovirus has been changed to D311/D315, a constitutively active form of human MEK5 sequence. DNMEK5 contains dominant negative form of human MEK5 sequence, which can not be phosphorylated, because the dual phosphorylation site S311/S315 has been changed to A311/A315. HMVECs and HRMEC were seeded in 6-well plate, cultured overnight, and infected with adenovirus for 48 hours as described before. ¹⁹ A nonspecific green fluorescent protein (GFP) adenovirus (Cell Biolabs) with the same multiplicity of infection was used as a negative control. Transduction efficacy was measured by Western blot analysis. 

Transfection of siRNA

The ERK5 siRNA (ON-TARGETplus siRNA) was purchased from Dharmacon Inc. (Lafayette, CO). A nontargeting siRNA (siGENOME Non-Targeting Pool; Dharmacon Inc.) was used as a negative control. The ECs were transiently transfected with 100 nM/L of control RNA or ERK5 siRNA using siRNA transfection reagent (DharmaFECT 4; Dharmacon Inc.) according to the manufacturer instructions. ²⁵ The cells were harvested 48 hours after siRNA transfection. siRNA knockdown efficiency was measured by real time RT-PCR and Western blot. 

RNA Isolation and cDNA Synthesis

Total RNA from ECs and rat retinal tissues was isolated using TRIzol reagent (Invitrogen, Burlington, ON) as previously described. ¹ RNA was quantified by measuring ultraviolet absorbance at 260 nm. cDNA was synthesized using 2 to 4 μg total RNA using High Capacity cDNA Reverse Transcription Kit (Applied Biosystem, Foster City, CA). The resulting cDNA products were stored at −20°C. 

Real Time RT-PCR

Real time RT-PCR was performed using LightCycler (Roche Diagnostics Canada, Laval, Canada) as described before. ²⁶ For a final reaction volume of 20 μL, the following reagents were added: 10 μL SYBR Advantage qPCR Premix (Clontech, Mountain View, CA), 1 μL of each forward and reverse 10 μmol/L primers (Table), 7 μL H₂O, and 1 μL cDNA template. To optimize the amplification of the genes, melting curve analysis (MCA) was used to differentiate melting temperature (Tm) of specific products from that of primer-dimers. Serially diluted standard template was used to construct a standard curve to quantify mRNA levels. The data were normalized to housekeeping gene 18S ribosomal RNA or β-actin to account for differences in reverse transcription efficiencies and the amount of template in the reaction mixtures. 

Table. 

 

Oligonucleotide Sequences for Real Time RT-PCR

Gene	Sequence 5′→3′
ERK5 (human)	CTGGCTGTCCAGATGTGAA
	ATGGCACCATCTTTCTTTGG
FN (human and rat)	GATAAATCAACAGTGGGAGC
	CCCAGATCATGGAGTCTTTA
TGFβ1 (human)	GCCCACTGCTCCTGTGACA
	CGGTAGTGAACCCGTTGATGT
NGF (human)	GGGTGCCGGGGCATTGACTC
	GAGCGTGTCGGCAGGTCAGG
BDNF (human)	GCGCCACTCTGACCCTGCC
	TCCCGCCCGACATGTCCACT
NTRK1 (human)	GGGAGGGCGCCTTTGGGAAG
	ACGCCGAAGAAGCGCACGAT
NTRK2 (human)	GGCTGGCACTGGCTGCTAGG
	ACGTGGGACAGGCGAAAGCG
β-actin (human)	CCTCTATGCCAACACAGTGC
	CATCGTACTCCTGCTTGCTG
18S (human and rat)	GTAACCCGTTGAACCCCATT
	CCATCCAACGGTAGTAGCG

Protein Extraction

Endothelial cells were cultured in 25-cm² tissue culture flasks. Subconfluent cells were growth arrested by incubation in serum-free medium overnight prior to stimulation. After stimulation, the cells were treated with 0.2 mL of lysis buffer including 25 mM/L Tris·HCl, pH7.5, 150 mM/L NaCl, 5 mM/L MgCl₂, 1% NP-40, 1 mM/L Dithiothreitol (DTT), 5% glycerol, protease inhibitor (complete Mini tablet; Roche Diagnostics Canada) and phosphatase inhibitor cocktail 1 and 2 (Sigma-Aldrich, St. Louis, MO). Cells were scraped off the flask and sonicated on ice with Kontes micro-ultrasonic cell disrupter (Kontes, Vineland, NJ). The samples were centrifuged at 15,000g at 4°C for 15 minutes. Supernatants were stored at −70°C. Protein concentrations were determined by bicinchoninic acid (BCA) protein assay kit (Pierce, Rockford, IL). 

Western Blot Analysis

Protein extract (30 μg) was boiled for 5 minutes and resolved by 10% SDS-PAGE. The protein extracts in the gels were then transferred to a polyvinyl difluoride (PVDF) membrane (Bio-Rad Laboratories, Hercules, CA). The membrane was blocked for 30 minutes at room temperature with 5% BSA in Phosphate Buffered Saline - Tween 20 (PBS-T). The blots were incubated overnight at 4°C with rabbit phospho-ERK5, rabbit ERK5, rabbit phospho-Smad2, mouse Smad2, rabbit Smad4 antibody (1:1000; Cell Signaling Technology, Beverly, MA), or mouse β-actin antibody (1:400; Santa Cruz Biotechnology, Santa Cruz, CA), followed by incubation with horseradish peroxidase conjugated anti-rabbit antibody (1:10,000; Upstate Biotechnology, Charlottesville, VA) or anti-mouse antibody (1:10,000; Vector Laboratories, Burlingame, CA) for 1 hour at room temperature as described before. ¹⁹ Visualization of immunoreactive bands were performed using an ECL plus chemiluminescence detection kit (Amersham Pharmacia Biotechnology, Buckinghamshire, UK). Blots were quantified by densitometry using Mocha software (SPSS, Chicago, IL). 

Enzyme-Linked Immunosorbent Assay for FN

As FN is a secreted protein, its measurement in the cell culture media gives accurate assessment of its protein level as described by our lab and others. ^27,28 To examine the amount of FN present in the medium during cell culture, culture media were collected after HG treatment, CAMEK5 and DNMEK5 infection, as well as siERK5 transfection and stored at −70°C. FN in cell culture media was measured using a human FN ELISA kit (Millipore Upstate, Temecula, CA) with a sensitivity of 10 to 20 ng/mL, according to the manufacturer's instructions. The developed color was measured at 450 nm wavelength with the Bio-Rad microplate reader (Bio-Rad Laboratories). The FN value in each sample was calculated according to a FN standard curve constructed using FN standards provided in the kit. Data were expressed in nanograms per milliliter. 

Animal Experiments

All procedures were conducted in accordance with the ARVO statement for the Use of Animals in Ophthalmic and Vision Research. Male Sprague-Dawley rats were purchased from Charles River (Wilmington, MA). Diabetes was induced by a single intravenous injection of streptozotocin (65 mg/kg, in citrate buffer, pH 5.6). Age- and sex-matched controls rats were given equal volume of citrate buffer. ³ The animals were fed standard laboratory chow and monitored for glucosuria and ketonuria (Uriscan Gluketo; Yeong Dong, Seoul, South Korea). They were sacrificed after 1 month of diabetes. The retinas were separated as described before ^19,20 and stored at −70°C. 

Statistical Analysis

Data were presented as the mean ± SEM. Statistical significance of difference between groups was tested using Student's t-test or two way ANOVA followed by post hoc analysis. A P value less than 0.05 was considered to be significant. 

Results

Glucose-Induced FN Upregulation Is Regulated by ERK5

A time course study of HG (25 mM/L) treatment was performed in HMVECs to establish glucose-induced increased FN production and its association with ERK5. HMVECs were treated with low glucose (LG, D-glucose, 5 mM/L), OC (L-glucose, 25 mM/L) or HG (D-glucose, 25 mM/L) for 24, 48 and 72 hours. HG (25 mM/L) treatment resulted in initial increase followed by decreased ERK5 signaling in HMVECs, demonstrated by real time PCR and Western blot (Figs. 1A, 1B). On the contrary FN mRNA and protein expression were elevated following glucose treatment in HMVECs (Figs. 1C, 1D). Result indicated that OC has no effect on ERK5 activation in the endothelial cells. It was noted that FN expression was also significantly increased in control groups with duration (Fig. 1D). This may represent changes due to aging; however, possibility of accumulation of secreted FN in the media cannot be completely excluded. We further explored the expression of ERK5 and FN in retinal tissues of diabetic rats. Real time RT-PCR showed reduced ERK5 mRNA in retinal tissues of diabetic animals (Fig. 1E), in keeping with previous studies. ^19,20 In contrast, FN mRNA and protein in retinal tissue of diabetic rats were elevated (Figs. 1F, 1G). 

Figure 1. 

 

Expression of ERK5 and FN in HMVECs and rat retinal tissues. (A–D) HMVECs were treated with LG (D-glucose, 5 mM/L), OC (L-glucose, 25 mM/L), or HG (D-glucose, 25 mM/L) for 24, 48, and 72 hours. (A) ERK5 mRNA was transiently increased after 24 hours HG treatment, and then decreased after 48 and 72 hours in HMVECs. (B) Western blot also showed initial increase followed by decrease of pERK5. (C) Real-time PCR showed increased FN mRNA expression in HMVECs following glucose treatment. (D) Such changes paralleled secreted FN protein expression as measured by ELISA. (E–G) expression of ERK5 and FN were also examined in retinal tissues of diabetic rats. (E) ERK5 mRNA expression was decreased and (F) FN mRNA and (G) FN protein expression were significantly increased in retinal tissues of diabetic rats compared with control. ([A–D], * = significantly difference from LG of the same time point, [E–G], * = significantly different from controls, P < 0.05, n = 3–6/group.)

To study the regulation of ERK5 on FN, gain of function experiments were performed in HMVECs. MEK5 specifically activates ERK5 without altering other MAPK. CAMEK5 recombinant adenovirus transduction upregulated ERK5 activity through the phosphorylation of the TEY activation motif by constitutively activating MEK5, as demonstrated by increased phophorylated ERK5 in Western Blot (Fig. 2A). ERK5 mRNA was also increased by CAMEK5 transduction (Fig. 2B). To study the effect of ERK5 upregulation on glucose-induced increased FN production, HMVECs were transduced with CAMEK5 and then cultured with and without HG (25 mM/L). Both basal and HG-induced increased FN mRNA and protein expressions were reduced by CAMEK5 transduction (Figs. 2C, 2D). Similar results were observed in HRMECs (Fig. 2E). 

Figure 2. 

 

ERK5 overexpression by CAMEK5 inhibited basal and glucose-induced FN expression in HMVECs and HRMECs. ECs were transduced with GFP control or CAMEK5 adenovirus and then cultured in LG (5 mM/L) or HG (25 mM/L) for 48 hours. Following CAMEK5 transduction, (A) increased pERK5 (by Western blot) and increased (B) ERK5 transcripts (by real time RT-PCR) were seen in HMVECs compared with GFP control. (C) CAMEK5 transduction significantly inhibited FN protein production in both basal and HG levels in HMVECs (ELISA). (D, E) CAMEK5 transduction reduced basal and glucose-induced increased FN mRNA expression in (D) HMVECs and in (E) HRMECs. (* = significantly different from LG group, # = significantly different from GFP LG, † = significantly different from GFP HG treatment, P < 0.05, n = 6/group. LG = 5 mM/L glucose, HG = 25 mM/L glucose.)

Transcription factor Krueppel-like factor 2 (KLF2) is a downstream target of ERK5 ²⁹ ; hence, we measured KLF2 by real time RT-PCR. In keeping with our previous study, ^19,20 CAMEK5 increased KLF2, whereas ERK5 siRNA reduced it (data not shown). 

Inhibition of ERK5 Caused FN Upregulation

To further characterize regulation of FN production by ERK5, loss of function studies were performed using siERK5 and DNMEK5. siERK5 knocked down ERK5 gene expression and resulted in a reduction of ERK5 protein, as demonstrated by Western blot (Fig. 3A). In contrast, ELISA demonstrated increased FN production following siERK5 transfection (Fig. 3B), further establishing that ERK5 negatively regulates FN synthesis. Moreover, we used another method (i.e., DNMEK5 to inhibit ERK5 activation). As the dual phosporylation site was changed, neither DNMEK5 nor ERK5 can be phosphorylated in this system. Western blot analysis showed reduced pERK5 after DNMEK5 transduction (Fig. 3C). Although DNMEK5 didn't increase FN expression in basal glucose, it further enhanced HG-induced FN overexpression (Fig. 3D). Two-way ANOVA confirmed that DNMEK5 interacted with HG in stimulating FN synthesis. Furthermore, enhanced glucose-induced FN synthesis following DNMEK5 transduction was observed in HRMECs (Fig. 3E). 

Figure 3. 

 

ERK5 downregulation by siERK5 and DNMEK5 increased FN levels in HMVECs and HRMECs. (A, B) HMVECs were transfected with siERK5 and control siRNA (ctrlsiRNA). (A) Reduced ERK5 protein (by Western blot) and (B) increased FN protein (ELISA) were seen in HMVECs following siERK5 transfection. (C–E) DNMEK5 was transduced to endothelial cells to inhibit the phosphorylation of ERK5 and then cultured in HG for 48 hours. DNMEK5 transduction resulted in (C) reduced pERK5 (Western blot) and enhanced glucose-induced increased FN mRNA levels (real time RT-PCR) in (D) HMVECs and in (E) HRMECs compared with GFP control. (* = significantly different from LG , † = significantly different from HG DNMEK5. P < 0.05, n = 3–6/group. LG = 5 mM/L glucose, HG = 25 mM/L glucose.)

Erk5 Regulated FN Production through Tgfβ1 Signaling

It has been well established that HG causes upregulation of TGFβ1 expression in association with Smad2 phosphorylation. To investigate whether TGFβ1 mediates ERK5 signaling on FN synthesis, we examined TGFβ1 expression in gain and loss of function experiments of ERK5. CAMEK5 transduction reduced TGFβ1 mRNA expression (Fig. 4A). On the other hand, ERK5 siRNA transfection upregulated TGFβ1 mRNA expression (Fig. 4B). As Smad family mediates TGFβ1 signaling, we examined pSmad2, total Smad2, and Smad4 using Western blot. pSmad2 was diminished by CAMEK5 transduction, while total Smad2 and Smad4 were not significantly altered (Fig. 4C). Two-way ANOVA demonstrated that both CAMEK5 transduction and glucose treatment affected on TGFβ1 signaling. In addition, glucose-induced increased pSmad2 expression was decreased after CAMEK5 transduction (Fig. 4D). Taken together, these results suggest that TGFβ1 signaling may mediate the regulation of ERK5 on FN. 

Figure 4. 

 

TGFβ1 signaling mediates the effect of ERK5 on FN in HMVECs. (A, B) ERK5 pathway was activated by CAMEK5 transduction or inhibited by ERK5 siRNA in HMVECs. Real time RT-PCR showed (A) diminished TGFβ1 mRNA following CAMEK5 and (B) increased TGFβ1 mRNA expression following ERK5 siRNA transfection (* = significant difference from GFP control or control siRNA. n = 6/group). (C, D) HMVECs were transduced with GFP control or CAMEK5 adenovirus and then cultured in LG or HG for 48 hours. (C) Representative Western blot showed diminished pSmad2 after CAMEK5 transduction in both LG (5 mM/L) and HG (25 mM/L) conditions. (D) Densitometric quantification of pSmad2 expression. (Data are expressed as a ratio to total Smad2, * = significantly different from LG group, # = significantly different from GFP LG, † = significantly different from GFP HG, P < 0.05, n = 3/group. LG = 5 mM/L glucose, HG = 25 mM/L glucose.)

It has been demonstrated that hyperglycemia-induced augmented ET-1 production causes increased FN synthesis ^1,30,31 ; hence, we also investigated whether ET-1 mediates the effect of ERK5 on FN. However, exogenous human ET-1 peptide failed to rescue the inhibition of CAMEK5 on FN synthesis (data not shown), excluding the possibility that ET-1 participates in the regulation of ERK5 on FN. 

NGF Regulated ERK5 Signaling under High Glucose Conditions

Neurotrophins, such as NGF and brain-derived neurotrophic factor (BDNF), are known to regulate ERK5 in the neuronal cells. ^32,33 ECs also express neurotrophins. ^21,34 To investigate whether neurotrophins are upstream regulators of ERK5, we examined expressions of NGF, BDNF, and their receptors in ECs using real time RT-PCR. We observed an initial increase and then a decrease of neurotrophins and their receptors following 25 mM/L glucose treatment (Figs. 5A–D). These findings paralleled the alteration of ERK5 (Figs. 1A, 1B). To further characterize the role of NGF, ECs were incubated with exogenous human recombinant NGF. Exogenous NGF stimulated ERK5 phosphorylation (Figs. 6A, 6B), which were further augmented by a 24 hour glucose treatment (Figs. 6A, 6B). ERK5 mRNA production was also increased by NGF treatment (Fig. 6C), while FN mRNA expression was reduced (Fig. 6D). 

Figure 5. 

 

Alterations of neurotrophins and their receptors parallel ERK5 after HG treatment in HMVECs. HMVECs were treated with 25 mM/L glucose (HG) for up to 48 hours. (A, B) Real time RT-PCR showed that NGF (A) and BDNF (B) mRNA expression were increased initially (after 24 hours) and then decreased (after 48 hours). (C, D) Similar results were observed in receptors of neurotrophins NTRK1 (C) and NTRK2 (D). (mRNAs are normalized to 24 hours LG, * = significant difference from LG at the same time point, P < 0.05, n = 3/group. LG = 5 mM/L glucose, HG = 25 mM/L glucose.)

Figure 6. 

 

NGF treatment increased ERK5 phosphorylation and ERK5 expression in HMVECs. (A, B) HMVECs were treated with exogenous human recombinant NGF and HG for 24 hours. (A) Representative Western blot and (B) quantitative densitometric analysis showing exogenous NGF treatment led to significant increase of pERK5. HG treatment further enhanced NGF induced ERK5 phosphorylation. (C, D) HMVECs were treated with exogenous NGF for 24 hours. Real time RT-PCR demonstrated increased (C) ERK5 mRNA and reduced (D) FN mRNA following NGF treatment. (* = significantly different from LG control, # = significant difference from HG group. P < 0.05, n = 3/group. LG = 5 mM/L glucose, HG = 25 mM/L glucose.)

Discussion

FN deposition is an important characteristic of endothelial injury in several chronic diabetic complications including diabetic retinopathy. ^1–4 In this study, we investigated the role of ERK5 signaling in glucose-induced increased FN production. We demonstrated that ERK5 negatively regulates basal and glucose-induced increased FN expression in ECs. We also observed increased FN expression and decreased ERK5 activation (Fig. 1) in retinal tissues of diabetic rats, indicating that attenuated ERK5 signaling may be of importance in the pathogenesis diabetic retinopathy. 

Our study demonstrated that constitutive activation of ERK5 signaling inhibited FN production in both basal and HG conditions (Fig. 2). The alteration of FN after CAMEK5 transduction was much stronger in comparison with that induced by siERK5 or DNMEK5. This may due to the fact both ERK5 phosphorylation and transcripation activation are involved in glucose-induced ERK5 alteration, as described in previous studies. ^19,20 ERK5 is different from other MAPKs because of its unique C-terminal, which contains transcriptional activation domain. Studies have shown that transcriptional activation of ERK5 greatly enhanced the effect of ERK5. ³⁵ CAMEK5 increased both ERK5 phosphorylation (Fig. 2A) and ERK5 transcript (Fig. 2B). In contrast, ERK5 siRNA decreased ERK5 gene expression and DNMEK5 inhibited ERK5 phosphorylation, respectively. Similar phenomenon was also observed by others, in cell migration study of ERK5. ³⁶ FN plays a major role in cell adhesion and migration. ³⁷ Since ERK5 negatively regulates FN, it is possible that ERK5 inhibits cell migration by suppressing FN. 

We also explored the possible mediator of the inhibition of ERK5 on FN. It has been established that TGFβ1 regulates the expression of FN. ^38,39 In keeping with such a notion, HG induced TGFβ1 mRNA expression and downstream signaling in the ECs, ^40,41 we found a significant inhibition of TGFβ1 signaling following CAMEK5 transfection, and an increase of TGFβ1 mRNA after siERK5 transfection (Fig. 4). KLF2 may mediate the inhibition of ERK5 signaling on TGFβ1. KLF2 is a downstream target of ERK5 signaling. ^42,43 In keeping with our previous reports, ^19,20 such process was also seen in this study. It has been shown that KLF2 inhibits TGFβ1 signaling via upregulating Smad7, therefore suppressing phosphorylated Smad2, which results in decreased transcriptional activity. ⁴⁴ Our study also demonstrated diminished pSmad2 following ERK5 upregulation, indicating that TGFβ1 signaling mediates inhibition of ERK5 on FN. Further studies are required to investigate whether Smad7 mediates the inhibition of pSmad2 by CAMEK5 transduction. Based on our experiments it appears that TGFβ1 and FN are downstream targets of ERK5 signaling pathway. Hyperglycemia leads to reduced ERK5 signaling, which causes increased FN through TGFβ1. Another study in the literature is supportive of this mechanism. ⁴⁴ However, whether similar pathways are in place with respect to other ECM proteins remains to be examined. 

Neurotrophins, such as NGF and BDNF, are growth factors that promote survival and function of neurons. ⁴⁵ ERK5 is activated by neurotrophins in neuronal cells and mediates a survival response. ^32,33 Neurotrophins can also activate ERK5 in other cell types. ^22,24 In ECs, ERK5 has been demonstrated to act as a mediator of neurotrophins' effect. ³⁴ It has been shown that neurotrophins expression is diminished in diabetes. ^21,23 Here, we showed that alteration of NGF, BDNF, and their receptors NTRK1 and NTRK2 paralleled ERK5 expression in ECs treated with HG (Fig. 5). Further, human recombinant NGF treatment stimulated both ERK5 phosphorylation and transcription in LG and in HG (Fig. 6), indicating its role in this process. However, ERK1/2 was also activated by exogenous NGF treatment in endothelial cells (data not shown). NGF activates both ERK5 and ERK1/2 in many cell types. ^22,24,46 Further studies are needed to explore the effect of differential activation of ERK1/2 and ERK5. 

In this study, however, there is a discrepancy between FN expression and ERK5 activation. After a 24 hour glucose treatment, although ERK5 activation was seen, FN production was also elevated (Fig. 4). It seems that even if ERK5 is activated by glucose, it can not completely prevent endothelial injury and FN accumulation. This may be associated with several other pathways that increase FN production in HG, including TGFβ1, ERK1/2, the serine-threonine kinase Akt (also known as protein kinase B), and so forth. ^12,13,41 Another possibility is ERK5 SUMOylation. Woo and Shishido's research has found that advanced glycation end-product (AGE) and H₂O₂, induced by HG, promote ERK5 SUMOylation and inhibit transcriptional activity of ERK5. ^47,48 Recent studies also demonstrated phosphorylation of atypical sites of ERK5 molecule, which inhibits ERK5 function. ⁴⁹ However, it remains unknown, whether HG can also activate atypical inhibitory phosphorylation of ERK5. 

In contrast to diabetic retinopathy, ERK5 may play a different role in diabetic nephropathy. It has been reported that HG activates ERK5 and stimulates mesangial cell proliferation and extracellular matrix (collagen I) accumulation. ^50,51 The differences between mesangial cells and ECs indicate that ERK5 signaling may regulate FN production in a cell type-specific manner. Effects of ERK5 in cell migration and cytoskeleton remodeling are also cell type-specific. ³⁶  

In summary, this study demonstrated a novel pathway that regulates glucose-induced increased FN production. The deposition of FN is involved in the pathogenesis of diabetic retinopathy. The inhibitory effect of ERK5 on FN sheds a light on developing a new, adjuvant treatment for diabetic retinopathy. However, the majority of our experiments were performed in vitro using ECs. Studies in both type 1 and type 2 diabetic animal models, are needed in the future to establish such pathogenic mechanisms and their effects on DR. Such experiments will provide valuable data to examine potential use of this target to prevent diabetic retinopathy (DR) or other diabetic complications. 

Footnotes

 Supported by a grant from the Canadian Diabetes Association.

Footnotes

 Disclosure: Y. Wu, None; B. Feng, None; S. Chen, None; S. Chakrabarti, None