Effect of Antifibrotic MicroRNAs Crosstalk on the Action of N-acetyl-seryl-aspartyl-lysyl-proline in Diabetes-related Kidney Fibrosis

N-acetyl-seryl-aspartyl-lysyl-proline (AcSDKP) is an endogenous antifibrotic peptide. We found that suppression of AcSDKP and induction of dipeptidyl peptidase-4 (DPP-4), which is associated with insufficient levels of antifibrotic microRNA (miR)s in kidneys, were imperative to understand the mechanisms of fibrosis in the diabetic kidneys. Analyzing streptozotocin (STZ)-induced diabetic mouse strains, diabetic CD-1 mice with fibrotic kidneys could be differentiated from less-fibrotic diabetic 129Sv mice by suppressing AcSDKP and antifibrotic miRs (miR-29s and miR-let-7s), as well as by the prominent induction of DPP-4 protein expression/activity and endothelial to mesenchymal transition. In diabetic CD-1 mice, these alterations were all reversed by AcSDKP treatment. Transfection studies in culture endothelial cells demonstrated crosstalk regulation of miR-29s and miR-let-7s against mesenchymal activation program; such bidirectional regulation could play an essential role in maintaining the antifibrotic program of AcSDKP. Finally, we observed that AcSDKP suppression in fibrotic mice was associated with induction of both interferon-γ and transforming growth factor-β signaling, crucial molecular pathways that disrupt antifibrotic miRs crosstalk. The present study provides insight into the physiologically relevant antifibrotic actions of AcSDKP via antifibrotic miRs; restoring such antifibrotic programs could demonstrate potential utility in combating kidney fibrosis in diabetes.

Diabetic nephropathy is a renal complication of diabetes and the leading cause of end-stage renal disease worldwide 1 . The final consequence of kidney diseases, including diabetic nephropathy, is kidney fibrosis 2 . The source and heterogeneity of kidney fibroblasts remain unclear 3 . N-acetyl-seryl-aspartyl-lysyl-proline (AcSDKP) is a tetrapeptide that is normally present in human plasma, and it is exclusively hydrolyzed by angiotensin-converting enzyme (ACE). ACE-I treatment increases the plasma level of AcSDKP fivefold 4 . Earlier studies supported the antifibrotic activity of AcSDKP (e.g., by inhibiting transforming growth factor (TGF)-β -induced expression of plasminogen activator inhibitor-1 and α 2 collagen type I in human mesangial cells by inhibiting the smad 2/3 signaling pathway 5 . AcSDKP prevented mesangial matrix expansion in diabetic db/db mice 6 , and AcSDKP demonstrated protective effects on organs fibrosis in various experimental animal models with fibrosis [7][8][9][10][11][12][13][14] . Recently, we reported that AcSDKP prevents renal fibrosis by counteracting the endothelial mesenchymal transition (EndMT) 15 . Matrix-producing mesenchymal cells derived from EndMT contribute to the pathogenesis of kidney fibrosis [16][17][18][19][20] .
The enzyme activity of dipeptidyl peptidase-4 (DPP-4) is highest in the kidney per organ weight 21 . Such high level of DPP-4 in kidney could be involved in renal metabolism, physiology and disease development [21][22][23] . We have reported that DPP-4 could be a key molecule associated with kidney fibrosis, as well as with the inhibition of DPP-4 through linagliptin ameliorated kidney fibrosis associated with the inhibition of EndMT and TGF-β / smad signaling 18 .

Results
Fibrotic kidney mouse strain exhibited the induction of DPP-4, the suppression of AcSDKP and low antifibrotic miRs levels in diabetes. In mice, the kidney fibrosis phenotype is largely dependent upon the strain specificity 26 . We analyzed comparative renal histopathology in non-diabetic and streptozotocin (STZ)-induced diabetic CD-1 and 129Sv mice. At the time of sacrifice, the diabetic CD-1 and diabetic 129Sv mice had similar blood glucose levels (Fig. S1A), reduced body weight (Fig. S1B), and higher kidney weight per body weight (Fig. S1C) compared with those of the non-diabetic animals. The kidneys of the CD-1 mice showed fibrosis, with excessive deposition of extracellular matrix while the 129Sv strain displayed minor fibrotic alterations (Fig. 1A). PAS staining revealed increased mesangial areas (Fig. 1B) in the kidneys of diabetic CD-1 mice compared with those of the diabetic 129Sv mice. The plasma of diabetic CD-1 mice exhibited elevated level of cystatin C whereas such alteration in cystatic C level was not found in diabetic 129Sv when compared to control (Fig. S1D). Western blot analysis revealed that the protein expression of DPP-4 and TGFβ R1 and smad3 phosphorylation were elevated in the kidneys of the diabetic CD-1 mice while such kind of alteration was not found in the kidneys of diabetic 129Sv mice when compared with those of the non-diabetic mice (Fig. 1C,D). In addition, the diabetic CD-1 mice displayed higher DPP-4 gene expression levels (Fig. 1E). DPP-4 enzymatic activity in either the kidney or plasma was significantly higher in the diabetic CD-1 mice than in the control mice; minor induction of DPP-4 activity in diabetes was also observed in the 129Sv strain, although it was significantly lower than in the CD-1 mice (Fig. 1F,G). When analyzing the antifibrotic peptide AcSDKP levels in the urine, diabetic CD-1 mice displayed significantly lower AcSDKP levels in the urine than the controls, whereas such diabetes-induced alterations in the AcSDKP levels were not found in the 129Sv mice (Fig. 1H). Such difference in AcSDKP level in strains was not explained by known mechanisms (Fig. S1E-I).
To evaluate the miR expression signature in the fibrotic kidney of STZ-induced diabetic mice, we performed miRNA microarray analysis in the kidneys of CD-1 and 129Sv in both non-diabetic and diabetic mice. Sixty-two miRs were downregulated, and 30 were upregulated in diabetic CD-1 mice, whereas 26 miRs were downregulated, and 61 miRs were upregulated in the diabetic 129Sv mice strains compared with the non-diabetic control mice (data shown in excel file). Among these miRs, we found that 27 miRs were downregulated, and 12 were upregulated in the diabetic CD-1 mice compared with the diabetic 129Sv mice ( Fig. 2A). MiR-29s and let-7s emerged as major altered miRs and showed a remarkable trend of suppression in the diabetic CD-1 mice; such alteration was not found in the case of the 129Sv diabetic mice ( Fig. 2A). Validation of miR-29s and miR-let-7s by qPCR revealed similar trends of suppression in the diabetic CD-1 mice, and there were no such alterations in the diabetic 129Sv mice compared with the corresponding non-diabetic mice ( Fig. 2B-N).
AcSDKP inhibited TGF-β-stimulated DPP-4 and induced anti-fibrotic miRome. High DPP-4 levels have shown to be associated with kidney fibrosis in diabetic CD-1 mice and also with the induction of mesenchymal activation program 18,27 . AcSDKP administration significantly suppressed CD31/DPP-4 ( Fig. 3A-C), α -SMA/DPP (Fig. S2A), FSP-1/DPP-4 ( Fig. S2B), TGFβ R1/DPP-4 ( Fig. S2C) and protein/mRNA levels of DPP-4 in the diabetic kidneys of CD-1 mice (Fig. 3D-H). AcSDKP treatment in the diabetic mice also showed reduced protein levels of TGFβ R1 and smad3 phosphorylation compared with the kidneys of the diabetic mice ( Fig. 3D,F,G). DPP-4 activities in both plasma and kidneys exhibited higher levels in diabetic mice; AcSDKP treatment significantly reduced the DPP-4 activity ( Fig. 3I-J). AcSDKP administration also suppressed the plasma cystatin C levels in diabetic CD-1 mice (Fig. S2D). In the Human Dermal Microvascular Endothelial Cells (HMVEC), the TGF-β 2 treatment increased the DPP-4, TGFβ R1 protein levels and the smad3 phosphorylation; AcSDKP treatment decreased these levels ( Fig. S3A-D). The AcSDKP treatment also reduced the DPP-4 mRNA levels in the TGF-β 2-stimulated HMVECs (Fig. S3E). In our previous studies we have shown that miR-29s target DPP-4 by binding on the 3′ UTR of DPP-4 mRNA 18 . In this study we analyzed the luciferase activity in transfected HMVECs carrying the construct containing the 3′ UTR of the DPP-4 mRNA 18 ; we observed that the TGF-β 2 treatment increased the luciferase activity, whereas AcSDKP treatment significantly reduced the TGF-β 2-induced luciferase activity in the HMVECs (Fig. S3F), suggesting that AcSDKP inhibited DPP-4 transcription in a 3′ -UTR dependent manner. Therefore, we analyzed the miR-expression profile of the kidneys of the AcSDKP-treated diabetic mice and found that the AcSDKP-treated kidneys of the diabetic mice exhibited significant restoration of miR-29s and miR-let-7s expression levels compared with the diabetic mice ( Fig. 4A-K). In vitro, the stimulation by TGF-β 2 on the HMVECs decreased the miR-29s expression level, and the AcSDKP treatment markedly increased the miR-29s expression levels ( Fig. S4A-C).

DPP-4 inhibition also induced anti-fibrotic miRome.
In the series of attempts studying the effect of DPP-4 inhibition on anti-fibrotic miRs, we also analyzed the expression levels of miR-29s (miR-29a, miR-29b and miR-29c) and miR-let-7s (miR-let-7b, miR-let-7c, miR-let-7f, miR-let-7g and miR-let-7i) in the kidneys of diabetic mice and linagliptin-treated diabetic mice. We found that miR-29s and miR-let-7s were significantly reduced in the diabetic group compared with those in the control group and that linagliptin treatment caused upregulation in miR-29s and miR-let-7s (miR-let-7b, miR-let-7c, miR-let-7f, miR-let-7g and miR-let-7i) compared with the kidneys of diabetic mice ( Fig. 4L-U). Linagliptin treatment in the TGF-β 2-induced HMVECs also showed upregulation in the miR-let-7s ( Fig. S4D-I). We have recently shown that miR-29s negatively regulate DPP-4 and that inhibition of DPP-4 with linagliptin triggered upregulated miR-29s gene expression, which was associated with the inhibition of mesenchymal activation program and renal fibrosis 18 .

Bidirectional regulation of miR-29s and miR-let-7s in HMVECs. Antifibrotic interventions, such
as AcSDKP and DPP-4 inhibitor linagliptin, induced similar anti-fibrotic miRome in vivo and in vitro. These data led us to confirm the crosstalk of anti-fibrotic miRs. MiR-let-7b and miR-let-7c were predominant miR were analyzed in each data set. The data are expressed as the means ± s.e.m and are included in the graph. Hs_RNU6 was used as internal control. Diabetic mice are designated as DM in the figure.
Interferon-γ disrupts anti-fibrotic miRs crosstalk. Inflammation is the key factor during the fibroblast activation process [29][30][31][32] . The plasma of diabetic CD-1 mice exhibited elevated levels of interferon (IFN)-γ , the key inflammatory cytokine in fibrotic pathogenesis 32 , compared with the plasma of non-diabetic CD-1 mice; such an aberration was not found in the diabetic 129Sv strains (Fig. 6A). The IFN-γ gene expression was also induced in the kidneys of the diabetic CD-1 mice compared with those of the non-diabetic controls; such an aberration was not found in the diabetic kidneys of the 129Sv mice (Fig. 6B). AcSDKP treatment in the diabetic CD-1 mice significantly reduced the plasma IFN-γ (Fig. 6C) and mRNA levels in the kidneys (Fig. 6D). A similar suppression of the IFN-γ levels was observed in the kidneys of the diabetic CD-1 mice treated with DPP-4 inhibitor linagliptin (Fig. 6E,F).
MiR-29s target 3′ UTR of IFN-γ mRNA and suppress IFN-γ levels 33 . Interestingly, IFN-γ has been shown to suppress both fibroblast growth factor receptor (FGFR1) and FGFR1-dependent miR-let-7s production 28 . Such suppression in miR-let-7s resulted in the induction of TGFβ R1 and the promotion of EndMT 28 . Therefore, we analyzed whether FGFR1 expression levels affected the control of miR-29s. First, we reconfirmed the inhibitory role of miR-29s on IFN-γ expression in HMVECs (Fig. 6G). TGF-β 2-suppressed both FGFR1 protein levels and phosphorylation of FGFR1; mim-miR-29s inhibited such TGF-β 2-mediated alterations on FGFR1 (Fig. 6H). Finally, we analyzed the FGFR1 protein levels, phosphorylation of FGFR1, and miR-let-7s gene expression levels in the antagomirs of the miR-29s (combination of a, b, c)-treated HMVEC, with or without IFN-γ neutralizing antibody incubation. Interestingly, we observed that the suppression of FGFR1 protein levels and phosphorylation of FGFR1 induced by the antagomirs of miR-29s were restored by the IFN-γ neutralizing antibody incubation (Fig. 6I). In addition, the IFN-γ neutralizing antibody restored the levels of miR-29s antagomir-suppressed miR-let-7s levels in HMVEC (Fig. 6J-L). These data demonstrated that IFN-γ is a key profibrotic cytokine through the disruption of anti-fibrotic miRs crosstalk between miR-29s and miR-let-7s.

Discussion
The physiological importance of AcSDKP has been reported elsewhere 15,[34][35][36] . Suppression of AcSDKP in the fibrotic strain is associated with the induction of DPP-4, activation of TGF-β /smad signaling and down regulation of antifibrotic miRs (miR-29s and miR-let-7s), whereas such alterations were not prominent in the less-fibrotic diabetic 129Sv mice or in the diabetic CD-1 mice treated with an intervention of AcSDKP, thereby demonstrating the physiological significance of AcSDKP in fibrogenic programs in diabetic kidneys.
AcSDKP is degraded by the N-terminal catalytic domain of ACE. N-terminal truncation mutations of ACE in mice demonstrate protection from bleomycin-induced lung fibrosis and such anti-fibrotic effects phenotype in N-terminal truncation mutant mice was abolished when treated with S-17092, a prolyl-oligopeptidase inhibitor that inhibits the formation of AcSDKP 37 . Macconi et al. 36 , reported that high levels of miR-324-3p and the associated suppression of POP expression led to the suppression of low AcSDKP levels in urine and progressive renal damage in Munich Wistar Frömter (MWF) rats. Thymosin β 4, the precursor peptide of AcSDKP, also exhibited anti-fibrotic and tissue repair effects 35 . POP inhibitor treatment combined with thymosin β 4 administration abolished such beneficial effects of thymosin β 4 and even displayed a pro-fibrotic phenotype 35 , demonstrating that in the metabolism of AcSDKP, POP is the critical enzyme for production and that ACE is the enzyme responsible for the degradation of AcSDKP. In our analysis, however, the suppression of AcSDKP in fibrotic diabetic CD-1 mice cannot be explained by either pathway because we found that POP activity was significantly higher and ACE activity/protein levels were reduced in fibrotic diabetic CD-1 mice. The levels of thymosin β 4 were robust in both strains; therefore, there must be altered regulatory mechanisms other than either POP or ACE. However, currently there is no explanation available in the literature or from our experimental data set. Further studies would be required to investigate the detailed molecular regulations of AcSDKP.
Importantly, suppression of AcSDKP in diabetic CD-1 mice was associated with the induction of DPP-4 protein and enzyme activity. A profibrotic role of DPP-4 has been described by many researchers 18,[21][22][23] . Elevated levels of DPP-4, TGF-β /smad signaling and downregulation of anti-fibrotic miRs in the fibrotic strain have been associated with increased EndMT in diabetes. After the restoration of AcSDKP levels in diabetic CD-1 mice, the mice displayed suppressed AcSDKP levels, inhibited renal fibrosis in fibrotic kidney involved concomitant reduction in the DPP-4-associated increase in the mesenchymal activation program, suggesting that the suppression of the endogenous antifibrotic peptide AcSDKP is a therapeutic target for combating kidney fibrosis.  (J-L) qPCR analysis of miR-let-7s after transfection of the antagomir treatment of miR-29s in the HMVECs, with or without neutralizing IFN-γ antibody, n = 6 were analyzed in each data set. Hs_RNU6 was used as internal control. The data are expressed as the means ± s.e.m and are included in the graph.
Several reports have described the regulatory role of miRs in kidney disease and progression 19,24,[38][39][40][41][42] . Recently TGF-β 1 regulated crosstalk of microRNAs were found deregulated in type 1 diabetes patients who were at the risk of rapid progression to ESRD 43 . Interestingly, the miR-29s and miR-let-7s displayed crosstalk regulation, and AcSDKP induced such crosstalk of miR-29s and miR-let-7s in HMVECs. MiR-let-7s has been shown to inhibit TGFβ R1 28 , and TGF-β /smad signaling has been well described as an inhibitory pathway of miR-29s 18,44-46 ; therefore, it was not surprising to find that 'miR-let-7s induced miR-29s' . However, a contrary pathway, molecular mechanisms of 'miR-29s induced miR-let-7s expression' , could not be explained by a known direct molecular target of miR-29s. Thus, we focused on the involvement of the inflammatory cytokine IFN-γ . IFN-γ mRNA is the known target for miR-29s 33 , and we re-confirmed that observation. In addition, IFN-γ has been shown to inhibit FGFR1, a molecule that exhibits essential roles in the induction of miR-let-7s 28 . The suppressed miR-29s in the fibrotic kidneys resulted in the elevation of IFN-γ ; subsequently, synthesized IFN-γ inhibited both FGFR1-and FGFR1-dependent miR-let-7s expression levels in diabetes. This suppression of miR-let-7 could induce TGFβ R1 expression; again, activated TGF-β /smad signaling inhibited miR-29s 18 . These series of events could limit the crosstalk regulation between miR-29s and miR-let-7s during the fibrotic process in diabetic kidneys. Figure 7 depicts the involvement of AcSDKP in the homeostasis of anti-fibrotic milieu in the diabetic kidney. In short, elevated levels of miR-let-7s suppressed TGF-β -associated signaling, which concomitantly induced the miR-29s gene expression. Elevated levels of miR-29s inhibit IFN-γ , which in turn activates FGFR1 phosphorylation, finally suppressed TGF-β and associated signaling. These series of events finally leads to upregulation of miR-let-7s (Fig. 7). Therefore, miR-29s and miR-let-7s showed bidirectional crosstalk regulation. These data support the idea that AcSDKP is the key endogenous anti-fibrotic peptide responsible for the homeostasis of anti-fibrotic miRs in the kidneys.
In conclusion, AcSDKP-mediated antifibrotic program were contributing to the maintenance of the kidney homeostasis at least in part. The importance of anti-fibrotic peptide AcSDKP has been underestimated even when considering the biological significance of ACE inhibition on organ fibrosis. This study, therefore, provides novel insights into the understanding of miR-mediated alterations, which might be critical in the regulation of the anti-fibrotic program of AcSDKP.

Materials and Methods
Reagents and antibodies. AcSDKP was a gift from Dr. Omata from Asabio Bio Technology (Osaka, Japan).
Linagliptin was provided by Boehringer Ingelheim (Ingelheim, Germany), with an MTA. The rat polyclonal antimouse CD31 antibody was purchased from EMFRET Analytics GmbH & Co. KG (Eibelstadt, Germany). The mouse monoclonal anti-β -actin, rabbit polyclonal anti-TGFβ -receptor I and goat polyclonal DPP-4 antibodies were obtained from Sigma (St. Louis, MO, USA). A rabbit polyclonal anti-phospho smad3 (s423 and s425) antibody was purchased from Rockland immunochemicals (Gilbertsville, PA). Mouse monoclonal anti-human CD31 and a goat polyclonal anti-mouse DPP-4 (for tissue labeling) were purchased from R&D System (Minneapolis, MN). A rabbit polyclonal anti-α SMA antibody was purchased from Gene Tex (Irvine, CA). A rabbit monoclonal anti-interferon gamma antibody and anti-TGFβ -receptor-II antibody were purchased from Abcam (Cambridge, UK). Anti-human interferon gamma antibody (neutralizing) was purchased from Affymetrix (Santa Clara, CA, US). Fluorescein-, rhodamine-, and Alexa 647-conjugated secondary antibodies were obtained from Jackson ImmunoResearch (West Grove, PA, USA). Anti-fibroblast specific proteins (FSP1, sometimes displayed as S100A4) and the HRP-conjugated secondary antibodies for western blot detection were purchased from Cell Signaling Technology (  AcSDKP could induce both miR-29s and miR-let-7s. Elevated level of miR-let-7s suppressed TGF-β associated signaling, which concomitantly induced the miR-29s gene expression. Elevated levels of miR-29s inhibit of IFN-γ , which in turn activates FGFR1 phosphorylation, which was essential for miR-let-7 induction. AcSDKP could regulate such anti-fibrotic miRs crosstalk. in accordance to Japanese guidelines and regulations for scientific and ethical experimentation. The induction of diabetes in the CD-1 and 129Sv mice was performed according to the previously established experimental protocol 26 . In brief, 8-week-old CD-1 mice were induced diabetes with the single intraperitoneal injection of STZ at 200 mg/kg in 10 mmol/l citrate buffer (pH 4.5) 26 . For 129Sv mice, we injected 50 mg/kg of STZ via intraperitoneally for 5 consecutive days 26 . As the control set, citrate buffer without STZ was injected. Cystatin C levels in plasma were analyzed using the Mouse/Rat cystatin C kit (R&D System).
For the interventional study, we utilized a fibrotic diabetic kidney disease model (STZ-treated CD-1 mice). Sixteen weeks after the induction of diabetes, the diabetic mice were divided into the following two groups: a control group and an AcSDKP treatment group (500 μ g/kg BW/day using an osmotic mini-pump) for 8 weeks 15 . In the second set of experiments, we utilized two groups of diabetic animals: a diabetic control (DM), and a linagliptin treatment group, which was treated with linagliptin (5 mg/kg BW/day) for 4 weeks as previously reported 18 . In vitro experiment. Human dermal microvascular endothelial cells (HMVECs, Lonza, Basel, Switzerland) cultured in EGM medium were used in this experiment. When the HMVECs on the adhesion reagent (Kurabo medical, Osaka, Japan) reached 70% confluence, 5 ng/ml recombinant human TGF-β 2 for 48 h was placed in the experimental medium (HuMedia-MVG in serum-free RPMI at a 1:3 ratio), with or without AcSDKP (100 nM) preincubation for 2 h.
Western blot analysis. Protein lysates were denatured in a SDS sample buffer at 100 °C for 5 min and were separated on SDS-polyacrylamide gels and blotted onto PVDF membranes (Pall Corporation, Pensacola, FL, USA) using the semidry method. The immunoreactive bands were developed using an enhanced chemiluminescence (ECL) detection system (Pierce Biotechnology, Rockford, IL, USA) and detected using an ImageQuant LAS 400 digital biomolecular imaging system (GE Healthcare Life Sciences, Uppsala, Sweden).

DPP-4 activity detection.
A DPP-4 activity Fluorometric assay kit was used for DPP-4 activity detection (Biovision, Milpitas, CA, USA). Activity was determined as per the manufacturer's instructions. Enzyme activity was expressed in pmol/min/ml (μ U/mL).

AcSDKP measurements.
We analyzed urine AcSDKP concentrations using a competitive enzyme immunoassay kit (SPI-BIO, Massy, France), according to the manufacturer's instructions. Urine AcSDKP was normalized with the urine creatinine level 15 . RNA extraction and miR array analysis. Frozen kidney tissues were first placed on the RNAlater ® -ICE (Life technologies) for 16 h at − 20 °C before the subsequent homogenization process. Total RNA was isolated using the miRNeasy kit (Qiagen) following the manufacturer's instructions and was quantified with a Nanodrop spectrophotometer (ND-1000, Nano drop Technologies, DE, USA). The input for the Agilent miR labeling system was 100 ng total RNA. Dephosphorylated and denatured total RNA was labeled with cyanine 3-pCp and subsequently hybridized to the Agilent mouse miR microarray release version 15 using the miR Complete Labeling and Hyb Kit (Agilent Technologies, Inc., Santa Clara, CA, USA). Following hybridization for 20 h, the slides were washed with a Gene Expression Wash Buffer Kit (Agilent) and measured using an Agilent Scanner G2565BA. Agilent Feature Extraction Software (version 9.5.1) and GeneSpring GX software (version 12.5, Agilent) were used for data processing, analysis, and monitoring.
Scientific RepoRts | 6:29884 | DOI: 10.1038/srep29884 Luciferase assay. For the luciferase assay, to analyze the activity of the 3′ UTR in human DPP-4, we cloned the fragment of human DPP-4 3′ UTR sequence by PCR with the primer set (Fw: ATAGAGCTCAATAGCTAGCAGCACAGCACACCAAC Rev: ATATCTAGA GTGTCCATATGCCAGTGC GGTTTAGG) and BAC clone human RP11 178A14 as a template. The purified PCR fragment and pmirGLO Dual-Luciferase miR Target Expression Vector (Promega) were enzyme digested (Sac-1 and Xba-1), purified, and ligated (Thermo Fisher Scientific, Waltham, MA). The sequence of DPP-4 3′ UTR was confirmed, and the amplified vector DNA (300 ng/well in a 12-well plate) was transfected into HMVEC cells. In the presence and absence of TGF-β 2, AcSDKP was stimulated (100 nM in final concentration); the transcript in the transcriptional activity was evaluated with the Dual-Luciferase ® Reporter (DLR ™ ) Assay System (Promega) in triplicate samples.
Statistical analysis. The data are expressed as the means ± s.e.m. The one-way ANOVA Tukey test was performed to analyze the level of significance, which was defined as P < 0.05, if not specifically mentioned. GraphPad Prism software (Ver 3.0f) was used for the statistical analysis.