4-HNE Increases Intracellular ADMA Levels in Cultured HUVECs: Evidence for miR-21-Dependent Mechanisms

Objective To investigate whether 4-hydroxynonenal (4-HNE) regulates asymmetric dimethylarginine (ADMA) metabolism through pathway independent of direct adduct formation with ADMA metabolizing enzyme and the involvement of microRNA (miRNA) miR-21 in human umbilical venous endothelial cells (HUVECs). Methods Cultured HUVECs were treated with 4-HNE (at concentrations of 1, 5, and 10 µM, respectively) or 1‰ DMSO (vehicle control) for 24 h. MiR-21 inhibitor (final concentration of 100 nM) was transfected at 1 h before 4-HNE treatment. HUVECs were also transfected with miR-21 (at concentrations of 50 nM and 100 nM) and cultured for 12, 24, and 48 h, respectively. DDAH mRNA and miR-21 expression in the HUVECs were determined by semi-quantitative real time PCR. DDAH1 and DDAH2 protein expression were analyzed by Western blot. ADMA in the cell medium and cell lysates were analyzed by ELISA. ADMA metabolizing activity of the cell lysates was also determined. Results MiR-21 decreased DDAH1 and DDAH2 expression and ADMA metabolic activity significantly, while increased intracellular ADMA accumulation significantly in HUVECs. 10 µM 4-HNE treatment for 24 h increased the expression of miR-21 and intracellular ADMA concentration, decreased the expression of DDAH1/2 mRNA and protein, decreased ADMA metabolizing activity of the cell lysates significantly. MiR-21 inhibitor reversed the inhibitory effects of 4-HNE on DDAH1 expression completely, and partially reversed the changes in ADMA metabolizing activity and intracellular ADMA accumulation challenged by 10 µM 4-HNE. Conclusion 4-HNE down-regulates DDAH1 expression and increases intracellular ADMA accumulation in HUVECs through a miR-21-dependent mechanism.


Introduction
The endogenous NO synthase (NOS) inhibitor asymmetric dimethylargiline (ADMA) can decrease NO production through inhibiting NOS competitively and is proved to play important roles in the development of diseases such as hypertension, atherosclerosis (AS), coronary arterial disease, stroke, insulin resistance, and diabetes mellitus [1,2,3]. ADMA has been regarded as a novel and independent predictor of cardiovascular events in recent years [4]. Dimethylarginine dimethylaminohydrases (DDAHs) are key enzymes involved in the inactivation of ADMA in the body. The DDAH/ADMA/NOS pathway has become a potential target of drug discovery for cardiovascular diseases. Two isoforms of the DDAHs, i.e DDAH1 and DDAH2, are found in human. Though there is evidence that DDAH2 is involved in ADMA metabolism [5], DDAH1 is proved to be the critical enzyme responsible for ADMA metabolism in vivo [6].
MicroRNAs (miRNAs) are a class of highly conserved short non-coding RNAs which regulate a majority of the human genome at post-transcriptional level [7]. It is recognized that miRNAs regulate not only normal cellular functions such as cell proliferation and differentiation, but also contribute to the development of human diseases such as cancer and cardiovascular diseases [8,9]. MiR-21 is one of the miRNAs that contributes to the development of cardiovascular diseases such as cardiac hypertrophy and cardiovascular remodeling. An increase in circulating miR-21 levels is associated with kidney fibrosis [10]. MiR-21 also plays important roles in the ischemic heart with resveratrol [11]. A 4.6-fold increase in miR-21 expression in AS plaque is observed [12]. By using microarray analyses, Fleissner and colleagues find that miR-21 is one of the 16 miRNAs upregulated by ADMA treatment in circulating angiogenic progenitor cells (APCs) [13]. Evidence shows that ADMA impaires the migratory capacity of APCs through increasing miR-21 expres-1 sion, repressing the expression of its target sprouty-2, as well as increasing reactive oxygen species (ROS) formation [13]. In addition, in patients with coronary artery disease demonstrating high ADMA plasma levels, an elevation in miR-21 expression in APCs is observed [13]. Interestingly, we observed that DDAH1 is predicted to be a potential miR-21 target (Figure 1). Therefore, we raised the hypothesis that there might be a positive feedback loop between ADMA and miR-21 in accelerating the progress of cardiovascular diseases.
4-Hydroxynonenal (4-HNE) is a major active product formed following lipid peroxidation. 4-HNE is highly lipophilic and can interfere with the functions of proteins by adduct-forming capacity with macromolecules [14]. Physiological levels of 4-HNE in human plasma ranged in 0.3,1.0 mM. However, under pathophysiological conditions, the levels may increase to 10 mM or even higher. Double-edged sword effects for 4-HNE are observed: induction of the expression of antioxidant enzymes at physiological levels [15]; inhibition or inactivation of enzymes, membrane proteins and cytoskeletal proteins by forming 4-HNE-protein adducts at pathophysiological levels [14]. The molecule can also affect signal transduction pathways such as the insulin-dependent Akt signaling [16,17] and the PI3K/Akt pathway [18]. Increased levels of 4-HNE in plasma and biological fluids are observed in many human diseases including atherosclerosis [19], and 4-HNE can also contribute to the development of atherosclerosis and related diseases. Evidence shows that 4-HNE can inhibit DDAH1 activity and decrease NO generation in cultured bovine aortic endothelial cells in a dose-dependent manner through formation of Michael adducts on His 173 in DDAH1 [20,21]. However, it is unknown whether 4-HNE regulates DDAH activity and ADMA levels through other mechanisms.
In this study, we determined whether 4-HNE could influence intracellular ADMA levels through regulating DDAH1 expression and the contribution of miR-21 in cultured HUVECs. We found that both 4-HNE and miR-21 decreased DDAH1 expression and ADMA metabolizing activity in HUVECs. Meanwhile, 4-HNE up-regulated miR-21 expression, and the inhibitory effects of 4-HNE on DDAH1 expression was reversed by miR-21 inhibitor.

4-HNE Treatment and miR-21 Mimic Transfection of HUVECs
The HUVECs cell line was purchased from ATCC CRL-1730. 4-HNE (10 mg/ml in 100% ethanol) was purchased from Cayman Chemical Co. (Ann Arbor, MI, USA), dried under a N 2 flow and redissolved in 1% DMSO. Cells (1610 5 /mL) were seeded into 6well plates and cultured in DMEM medium containing 10% fetal calf serum (FCS) in a humidified atmosphere under 5% CO 2 for 24 h. Confluent cells were synchronized with 1% FCS for 24 hours, and were then treated with various concentration of 4-HNE (1 mM, 5 mM, and 10 mM) or vehicle (1% DMSO) for 24 h. To investigate the effect of miR-21 on DDAH expression, HUVECs were seeded into 6-well plates and cultured to 95% confluence. Cells were then washed with PBS and cultured in DMEM medium containing 10% FCS. Has-miR-21 mimic (50 nM and 100 nM) and/or inhibitor (100 nM) (RiboBio, China) were transfected by using the lipofectamine 2000 reagents (Fermentas, R0531). Cells and the medium were harvested at 12 h, 24 h, and 48 h, respectively, after miR-21 transfection. To investigate whether miR-21 inhibitor can reverse the 4-HNE induced increase in DDAH1 expression, has-miR-21 inhibitor was also transfected at 1 h before the addition of 4-HNE (10 mM). The medium was collected to detect ADMA concentration. Cells were harvested and the RNA and protein samples were extracted.

Western Blot Analysis
HUVECs cells were collected and washed with PBS. Cells were treated with 200 mL of cell lysis buffer at 4uC for 30 min and then centrifuged at 1,2000 rpm for 10 min. The cell lysates were then stored at 280uC until analysis. An aliquot of 60 mg proteins from each sample were subjected to 10% SDS-PAGE and then transferred to PVDF membrane (Millipore Corp). Semi-dried transfer was carried out under 1 mA/cm2 flow in 50 min. The membrane was blocked with 5% non-fat dry milk for 2 h at room temperature, and then incubated with a rabbit monoclonal antibody (1:1000) against human DDAH1 (Abcam, UK), or a rabbit monoclonal antibody (1:1000) against human DDAH2 (sigma, USA), or a mouse monoclonal antibody (1:1000) against human GAPDH (sigma, USA) at room temperature for 1 h, followed by incubation at 4uC for 18 h. The membrane was washed thoroughly and then incubated with a peroxidaseconjugated mouse anti-rabbit or goat anti-mouse secondary antibody (1:5000) (Santa Cruz, USA) for 1 h at room temperature. The membrane was washed, and DAB Horseradish Peroxidase Color Development Kit was used for develop (beyotime, China). The band intensities were semi-quantified, DDAH1 or DDAH2

Determination of ADMA in Cell Medium and Cell Lysates by ELISA
Cells and cellular medium were collected after treatment. Cells were washed with PBS, and then treated with 200 mL of cell lysates buffer at 4uC for 30 min, centrifuged at 1,2000 rpm for 10 min. Protein concentration of the cell lysates were determined. Cell Lysates and cellular medium were then stored at 280uC. An aliquot of 10 mL cell lysates was used to detect intracellular ADMA concentration. Human ADMA ELISA kit was used to analysis ADMA concentration in cell medium and the lysates according to the manufacturer's protocol (CUSABIO, China). ADMA contents in the cell lysis were normalized by protein content.

Determination of ADMA Metabolizing Activity in Cell Lysates
DDAH activity was determined by measurement of the amount of ADMA metabolized by the cell lysates as described elsewhere [22]. An aliquot of 80 mL cell Lysates from each well was added into two separate Eppendorf tubes (on ice), followed by the addition of 40 mL ADMA (1 mmol/L). Then, 40 mL 30% 5sulfosalicylic acid was immediately added into one tube to inactivate DDAH enzymes; the other tube was incubated at 37uC for 2 hours before the addition of 5-sulfosalicylic acid. The tubes were then centrifuged and ADMA concentrations in the supernatant were measured by ELISA assay kit (CUSABIO, China). Difference in ADMA content between the 2 tubes indicated ADMA metabolized. ADMA metabolizing activity was expressed as amount of ADMA metabolized in 1 min by 1 g protein (mmol/min/g protein).

Data Analysis
Data were expressed as mean 6 SEM. All statistical analysis were performed using the software SPSS, version 11.5 (SPSS Inc., Chicago, IL, USA). Differences among groups were compared by using univariate analysis of variance (ANOVA), and comparisons between two groups were analyzed by the LSD test. Two-tailed p,0.05 was considered to be statistically significant.

MiR-21 Inhibited DDAH Expression and Increased Intracellular ADMA Concentration in Cultured HUVECs
Transfection of miR-21 at concentrations of both 50 nmol/L and 100 nmol/L increased miR-21 expression in HUVECs significantly since 12 h after transfection (p,0.01, respectively). The increases in intracellular miR-21 expression decreased gradually but were still significant at 24 h and 48 h, respectively, after transfection (p,0.01, respectively, Figure 2A). Both 50 and 100 nmol/L miR-21 decreased DDAH1 and DDAH2 mRNA expression significantly at all three time points after transfection (p,0.05, respectively, Figure 2B and 2C). When DDAH protein expression was analyzed, both concentrations of miR-21 decreased DDAH1 expression at 24 h and 48 h (p,0.05, respectively, Figure 2D and 2E) but not at 12 h after transfection ( Figure 2D and 2E). DDAH2 protein expression was decreased at all three time points after miR-21 transfection (p,0.05, respectively,  Figure 2D and 2F). As compared with the control group at 12 h and 24 h, both DDAH1 mRNA and protein expression were decreased significantly in the control group at 48 h (p,0.05, respectively).
ADMA metabolizing activity of the cell lysates was decreased significantly at all three time points after 100 nmol/L miR-21 transfection (p,0.05 for 24 h, and p,0.01 for 12 and 48 h, respectively, Figure 3A). Intracellular ADMA concentration was increased significantly at 24 and 48 h (p,0.05, respectively) but not at 12 h (p.0.05) after 100 nmol/L miR-21 treatment ( Figure 3B). 50 nmol/L miR-21 treatment did not affect either the intracellular ADMA concentration or ADMA metabolizing activity of the cell lysates ( Figure 3A and 3B). MiR-21 transfection showed no effects on ADMA concentration in cell medium. As compared with the control group at 12 h and 24 h, intracellular ADMA concentration was decreased significantly in the control group at 48 h (p,0.05, Figure 3B).

4-HNE Decreased DDAH1/2 Expression and Increased Intracellular ADMA Concentration in Cultured HUVECs
Treatment of HUVECs with various concentrations of 4-HNE for 24 hours decreased DDAH1 mRNA expression significantly (p,0.05 for 1 mmol/L; p,0.01 for 5 mmol/L and 10 mmol/L, respectively, Figure 4A). The degree of decrease in DDAH1 mRNA expression was more obvious in 5 mmol/L and 10 mmol/L 4-HNE treated as compared with 1 mmol/L 4-HNE treated cells (p,0.05, respectively, Figure 4A). Significant decrease in DDAH1 protein expression was observed in cells treated with 10 mmol/L 4-HNE (p,0.01, Figure 4B) but not with 1 mmol/L and 5 mmol/L 4-HNE. Both DDAH2 mRNA and protein expression were decreased by 4-HNE treatment at all three concentrations (p,0.01, Figure 5). DMSO (solvent control) or lipofectamine 2000 (transfection reagent control) alone had no effect on either DDAH1 mRNA or protein expression ( Figure 4). Intracellular ADMA metabolizing activity was decreased while ADMA concentration was increased significantly by 4-HNE treatment at all three concentrations (p,0.05, Figure 4C and 4D). The degree of decrease in ADMA metabolizing activity was more obvious for the 10 mmol/L as compared with both 1 mmol/ L and 5 mmol/L 4-HNE treated cells (p,0.01, Figure 4C). The increase in intracellular ADMA concentration was comparable among the different concentration of 4-HNE treated groups ( Figure 4D). ADMA concentration in cell culture was increased by 5 mmol/L and 10 mmol/L 4-HNE treatment (p,0.05, respectively, Figure 4C).

Discussion
In this study, we observed the influences of 4-HNE and miR-21 on DDAH/ADMA system for the first time. We found that high concentration 4-HNE increased intracellular miR-21 expression and ADMA accumulation, decreased DDAH1 expression and ADMA metabolizing activity in cultured HUVECs. We also observed that overexpression of miR-21 by miR-21 mimic transfection down-regulated both DDAH1 and DDAH2 expression, decreased intracellular ADMA metabolizing activity and increased intracellular ADMA concentration in HUVECs. In addition, miR-21 inhibitor reversed the 4-HNE-induced decrease in DDAH1 expression, and partially reversed the 4-HNE induced increase in intracellular ADMA accumulation and decrease in ADMA metabolic activity.
ADMA, miR-21 and 4-HNE are all factors contribute to the development of CVDs especially for atherosclerosis. Previous study has shown that ADMA impairs the migratory ability of APCs through up-regulating miR-21 expression [13]. In our study, we found a potential miR-21 binding element (position 344,364) in DDAH1 39-UTR, and observed that overexpression of miR-21 by lipofectamine mediated transfection at concentrations of both 50 nmol/L and 100 nmol/L down-regulated both DDAH1 and DDAH2 expression in cultured HUVECs. However, only 100 nmol/L miR-21 transfection affected intracellular ADMA metabolizing activity and intracellular ADMA level. The inhibition of DDAH1 mRNA expression occurred at 12 h while inhibition of DDAH1 protein expression was observed at 24 h after miR-21 transfection, suggesting a time delay in the inhibition of DDAH1 expression by miR-21. It is known that miRNAs act through translational repression or degradation of target mRNA at post-transcriptional level [23]. Though we failed to perform the reporter gene assay that may provide more direct evidence to confirm whether DDAH1 is the target for miR-21, our observation of the delay between the effects of miR-21 on DDAH1 mRNA and protein expression indicates that miR-21 may act through affecting DDAH1 mRNA stability. As ADMA can also upregulate miR-21 expression [13], our findings suggest a possible positive feedback regulation loop between miR-21 and the DDAH/ ADMA system, which may accelerate the development of AS and related human diseases (Figure 7). MiR-21 is supposed to decrease NO production by increasing intracellular ADMA levels in HUVECs according to our observations. However, a study by Weber et al has shown that miR-21 overexpression increases NO production by promoting eNOS phosphorylation in HUVECs [24]. This controversy should be further studied in future.
Obvious inhibition of DDAH2 mRNA and protein expression by miR-21 was also observed in our study. As no miR-21 recognition site in the DDAH2 mRNA was observed, the exact  mechanism for this inhibition remains unknown. Since the expression of DDAH2 is ROS sensitive [25], while miR-21 can regulate ROS formation by attenuating SOD3 or limiting TNFa production [26], there is the possibility that miR-21 downregulates DDAH2, and possibly partially for DDAH1, through ROS-dependent pathways. We also observed that miR-21 inhibitor did not reverse 4-HNE induced decrease in DDAH2 expression like DDAH1, which also supports our findings that miR-21 may inhibit DDAH2 through indirect mechanisms.
It is interesting that 50 nM miR-21 transfection reduced both mRNA and protein expression of DDAH1 and DDAH2, however, there was just a trend of increased intracellular ADMA level and decreased DMA metabolizing activity of the cell lysates challenged by 50 nM miR-21. The reason for this discrepancy is not known. The intracellular ADMA level is a result of the balance among ADMA production, metabolism and efflux from the intracellular stores. The proteins such as protein argiline methyltransferase (PRMTs) [27], alanine glyoxylate aminotransferase2 (AGXT2) [28], and the cationic amino acid transporter [29] are also involved in the regulation of intracellular ADMA concentration. Therefore, we assume that miR-21 might also regulate the expression or activity of these proteins, and thus compensates for the changes in intracellular ADMA level and metabolizing activity due to DDAH1 inhibition.
4-HNE is considered as a new atherogenic factor, and plasma 4-HNE level is suggested to be potential biomarker for AS related disease [30,31]. Few mechanisms by which 4-HNE acts through are reported [15,21]. Evidence has shown that 4-HNE can increase ADMA accumulation in HUVECs by formation of Michael adducts with DDAH1 [21]. In addition to this, we observed for the first time that 4-HNE increases intracellular ADMA accumulation by inhibiting DDAH1 expression, while miR-21 inhibitor completely reversed the inhibitory effects of 4-HNE (10 mM) on DDAH1 expression. This indicates that 4-HNE can also affect DDAH1/ADMA system in a miR-21 dependent pathway. However, it is still not known as to how 4-HNE upregulates miR-21 expression. It is reported that 4-HNE increases the generation of reactive oxygen species (ROS) by promoting the activity of NADPH oxidase [32]. Activation of the ROS-NF-kB/ ERK pathway by 4-HNE is also observed [33]. Interestingly, the same ROS-NF-kB/ERK pathway also mediates up-regulation of miR-21 expression under oxidative stress [34]. It is possible that the 4-HNE induced up-regulation of miR-21 expression observed in our study is mediated by the ROS-NF-kB/ERK pathway. Of course, further studies are needed to verify this hypothesis.
We observed that miR-21 inhibitor only partially reversed ADMA metabolizing activity of the cell lysates and intracellular accumulation of ADMA by 10 mM 4-HNE treatment. These findings indicate that miR-21-independent mechanisms are also involved in the regulation of ADMA metabolism under 4-HNE exposure. Direct inhibition of DDAH1 activity by 4-HNE may account for this phenomenom [21]. We also observed that miR-21 inhibitor further increased ADMA levels in cell medium after 10 mM 4-HNE challenge. There is still no reasonable explanation for this observation. As it is known, the transmembrane transportation (intake and extrusion) of ADMA is mediated by transporters such as cationic amino acid transporter 1 [35,36]. There is the possibility that miR-21 may also act through affecting ADMA transportation in HUVECs when challenged by 4-HNE.

Conclusion
In conclusion, we find in this study that 4-HNE can dysregulate intracellular ADMA in HUVECs through miR-21 dependent pathway and involving inhibition of DDAH1 expression. We also suggest a positive regulation loop between miR-21 and the DDAH1/ADMA system. Our findings suggest that intervention with miR-21 expression may provide a new therapeutic target for the treatment of 4-HNE and ADMA related diseases such as atherosclerosis.

Author Contributions
Conceived and designed the experiments: XPC. Performed the experiments: LC JPZ DBK. Analyzed the data: LC YJL. Contributed reagents/ materials/analysis tools: JT. Wrote the paper: LC XPC.