Matrix metalloproteinase 9 induces endothelial-mesenchymal transition via Notch activation in human kidney glomerular endothelial cells
Endothelial-mesenchymal transition (EndoMT) is a major source of myofibroblast formation in kidney fibrosis. Our previous study showed a profibrotic role for matrix metalloproteinase 9 (MMP-9) in kidney fibrosis via induction of epithelial-mesenchymal transition (EMT). Inhibition of MMP-9 activity reduced kidney fibrosis in murine unilateral ureteral obstruction. This study investigated whether MMP-9 also plays a role in EndoMT in human glomerular endothelial cells.
TGF-β1 (10 or 20 ng/ml) induced EndoMT in HKGECs as shown by morphological changes. In addition, VE-cadherin and CD31 were significantly downregulated, whereas α-SMA, vimentin, and N-cadherin were upregulated. RT-PCR revealed that Snail, a known inducer of EMT, was upregulated. The MMP inhibitor GM6001 abrogated TGF-β1-induced EndoMT. Zymography indicated that MMP-9 was also upregulated in TGF-β1-treated HKGECs. Recombinant MMP-9 (2 μg/ml) induced EndoMT in HKGECs via Notch signaling, as evidenced by increased formation of the Notch intracellular domain (NICD) and decreased Notch 1. Inhibition of MMP-9 activity by its inhibitor showed a dose-dependent response in preventing TGF-β1-induced α-SMA and NICD in HKGECs, whereas inhibition of Notch signaling by γ-secretase inhibitor (GSI) blocked rMMP-9-induced EndoMT.
Taken together, our results demonstrate that MMP-9 plays an important role in TGF-β1-induced EndoMT via upregulation of Notch signaling in HKGECs.
KeywordsMatrix metalloproteinase 9 Endothelial-mesenchymal transition Human glomerular endothelial cells TGF-β1 Notch
one way analysis of variance
chronic kidney diseases
endothelial cell growth supplement
fetal bovine serum
focal segmental glomerulosclerosis
fibroblast specific protein-1
gamma secretase inhibitor
matrix metalloproteinase 9
Notch intracellular domains
transforming growth factor-beta1
unilateral ureteral obstruction
vascular endothelial growth factor
α-smooth muscle actin
Kidney fibrosis is an inevitable consequence of a wide variety of progressive chronic kidney diseases (CKD) that progress to end-stage kidney failure, a devastating disorder that requires kidney replacement therapies such as dialysis or kidney transplantation. Kidney fibrosis is characterized by tubular atrophy/dilation, interstitial leukocyte infiltration, fibroblast accumulation, and increased interstitial matrix deposition [1, 2]. Although different cells are involved in kidney fibrosis, fibroblasts or myofibroblasts are considered to play a pivotal role . However, the cellular origins of myofibroblasts are diverse. Resident fibroblasts , fibrocytes from bone marrow [4, 5], pericytes and perivascular fibroblasts [5, 6], tubular epithelial cells [4, 5, 7], podocytes  and endothelial cells [7, 9] have been identified as contributing to the myofibroblast population.
Endothelial-mesenchymal transition (EndoMT) is process similar to that of epithelial-mesenchymal transition (EMT). During EndoMT, endothelial cell markers are downregulated whereas mesenchymal markers are upregulated. EndoMT is involved in organ development and various types of fibrosis. For example, EndoMT contributes to cardiac fibrosis [10, 11, 12], pulmonary fibrosis [10, 13], corneal fibrosis , radiation-induced pelvic disease  and inflammatory bowel disease-associated fibrosis . EndoMT was also found in the early development of kidney interstitial fibrosis in the streptozocin (STZ)-induced diabetic nephropathy (DN) model [17, 18]. In addition, EndoMT also contributes to carcinoma-associated fibroblasts , atherogenesis, inflammation and hypertension [20, 21]. LeBleu et al. showed that ~10 % of the interstitial myofibroblasts co-stained with markers of endothelial cells and activated fibroblasts in unilateral ureteral obstruction (UUO) mice . Zeisberg and colleagues  have demonstrated a role for EndoMT in several models of kidney disease. These studies demonstrated that activated fibroblasts co-express the endothelial marker CD31 as well as fibroblast markers such as fibroblast specific protein-1 (FSP-1) and α-smooth muscle actin (α-SMA). To demonstrate the presence of EndoMT-derived fibroblasts, these authors also used lineage tagged transgenic mice to trace endothelial lineage. Taken together, these findings suggest EndoMT plays a critical role in myofibroblast formation. EndoMT inhibition or reversal might be a potential target for treatment and prevention of kidney fibrosis.
Notch family is involved in podocyte and kidney tubular cell differentiation . Abnormal Notch pathway activation can lead to glomerulonephritis (GN) and focal segmental glomerulosclerosis (FSGS) [23, 24]. Notch signaling is typically activated upon binding of ligands (such as Dll1, Dll3, Dll4, Jag1, and Jag2) with Notch 1–4 receptors. Then intramembrane proteolysis such as ectodomain shedding of both the ligand and the receptor, releasing the intracellular domains (ICD) of the ligand and receptor, thereby allowing Notch ICD (NICD) nuclear translocation to regulate gene expression.
We previously found that matrix metalloproteinase 9 (MMP-9) is capable of inducing tubular cell EMT and contribute to tubulointerstitial fibrosis [25, 26]. Although endothelial cells are capable of expressing MMP-9 , whether MMP-9 plays a role in EndoMT was unknown. In the current study, we defined a role for MMP-9 in EndoMT via Notch signaling.
Cell culture and treatment
Human kidney glomerular endothelial cells (HKGECs) were cultured in endothelial cell media (ScienCell; Carlsbad, CA, USA) containing vascular endothelial growth factor (VEGF; Sigma-Aldrich; St. Louis, MO, USA; 2.5–5 μg/ml) in basal medium (ScienCell), 5 % fetal bovine serum (FBS; ScienCell), 1 % endothelial cell growth supplement (ECGS; ScienCell) and 1 % penicillin/streptomycin (P/S; ScienCell). Cells were maintained at 37 °C with 5 % CO2. For treatment, HKGECs were cultured for 24 h at low density in flasks or plates pre-coated with fibronectin and washed in PBS. Cells were treated with TGF-β1 (10 ng/ml or 20 ng/ml; Sigma-Aldrich) alone, TGF-β1 plus 10 μM GM6001 (Calbiochem; Darmstadt, Germany), TGF-β1 plus the MMP-9 inhibitor I (Merck Chemicals; Darmstadt, Germany) at different dosages (0.05 nmol/ml, 0.25 nmol/ml, and 0.5 nmol/ml), recombinant MMP-9 (rMMP-9; 2 μg/ml; Biomol International; Plymouth Meeting, PA, USA), or rMMP-9 (2 μg/ml) plus the gamma secretase inhibitor (GSI; 5–10 μM; Merck Millipore; Billerica, MA, USA). Our study does not require any human or animal ethics approval.
For indirect immunofluorescence, HKGECs were cultured on glass coverslips, washed twice in PBS, fixed with absolute methanol for ten minutes at −20 °C, and blocked for 1 h with 2 % BSA (Sigma) at room temperature. Cells were then incubated for 1 h at room temperature with primary antibodies against endothelial markers rabbit polyclonal anti-VE-cadherin (1:200; Alexis Biochemicals; Farmingdale, NY, USA), mouse monoclonal anti-CD31 (1:100; Cell Signaling Technology; Boston, MA, USA) and mesenchymal markers mouse monoclonal anti-α-SMA (1:200; Sigma Chemical Co), rabbit monoclonal anti-vimentin (1:200; Cell Signaling Technology) and rabbit monoclonal anti-N-cadherin (1:100, BD Bioscience; San Jose, CA, USA) in 2 % BSA. The following secondary antibodies were used: goat anti-mouse IgG2a/2b phycoerythrin (PE)-conjugated antibody (1:400; Invitrogen; Carlsbad, CA, USA) for CD31, α-SMA and goat anti-rabbit IgG2a/2b FITC-conjugated antibody (1:400; Invitrogen) for VE-cadherin, vimentin and N-cadherin. Cells were washed twice with PBS, counterstained with DAPI for 5 min, and washed twice with PBS. Samples were mounted using fluorescence mounting media. For negative isotype controls staining, rat IgG2a κ purified (eBioscience; San Diego, CA, USA) was used for VE-cadherin, vimentin and N-cadherin mouse IgG2a κ (Biolegend; San Diego, CA, USA) was used for CD31 and α-SMA, and their corresponding secondary antibodies were applied.
RNA extraction, purification, and quantitation
Total RNA was extracted from cultured cells using 350 μl of RLT buffer and homogenized by shredding through a 0.5 ml insulin syringe 5 times. Extracted RNA was purified using RNeasy Mini Kit (Qiagen; Hilden, Germany) following the manufacturer’s instruction and resuspended in 30 μl of RNAse-free water. The yield and purity of RNA was measured spectrophotometrically by absorption at 260 nm (A260) and 280 nm (A280) using a Beckman-Coulter DU800 spectrophotometer (CA, USA).
Real-time RT-PCR analysis
cDNAs were synthesized using 200 ng of extracted RNA in 20 μl reaction buffer by reverse transcription using SuperscriptTM First Strand Synthesis System (Invitrogen) and random hexamer primers at 50 °C for 50 min. Designed primers and established primers from published papers were used for Real-time RT-PCR. The sequences of hes-1 primers used for this analysis are as follows: forward: 5′-GAC AGC ATC TGA GCA CAG AAA TG-3′ and reverse: 5′- GTC ATG GCA TTG ATC TGG GTC AT-3′ . Housekeeping gene β-actin was used as the internal control. For Real-time RT-PCR, PCR mixture contained 0.5 μl of cDNA and 10 ρmol/μl of each primer in a 20 μl final volume of SYBR mastermix (Invitrogen). Amplification was performed using the Rotogene-6000 Real-Time cycler thermos.
Western blot analysis
Equal volumes or quantities of protein were loaded in 12-well NuPAGE 4–15 % Bis-Tris gels (Bio-Rad; Hercules, CA, USA) and electrophoresed under reducing conditions. After electrophoresis, proteins were transferred for 2 h to PVDF membranes using a Mini Trans-Blot Electrophoretic Transfer Cell apparatus (Bio-Rad). For immunodetection, membranes were blocked overnight at 4 °C in 5 % skim milk and incubated for 2 h at room temperature with the primary antibody against mouse monoclonal anti-α-SMA (1:300, Sigma Chemical Co.), rabbit polyclonal anti-VE-cadherin (1:1000; Alexis Biochemicals), mouse monoclonal anti-CD31 (1:1000; Cell Signaling Technology), mouse polyclonal anti-NICD (1;1000; Merck Millipore, Cat. # 07–1232) , β-actin (1:3000, Sigma), rabbit monoclonal anti-Notch1 (1:1000; Cell Signaling Technology), rabbit monoclonal anti-vimentin (1:1000; Cell Signaling Technology), rabbit monoclonal anti-N-cadherin (1:500, BD BioSciences), rabbit polyclonal anti-Hes-1 (1:500, Abcam; Cambridge, UK) and rabbit polyclonal anti-Hey-1 (1:500, Abcam) prepared in blocking buffer. Membranes were washed three times (10 min per wash on a rocking platform) and incubated for 1 h with their respective horseradish peroxidase (HRP)-conjugated secondary antibodies; goat anti mouse HRP and goat anti-rabbit HRP (1:5000; Cell Signaling Technology) prepared in blocking buffer. Membranes were again washed three times (10 min per wash on a rocking platform). Bands were visualized using an enhanced chemiluminescence detection kit.
MMP-9 activity in medium derived from TGF-β treated HKGECs was determined by gelatin zymography. Briefly, medium was mixed with Tris-Glycine SDS Native Sample Buffer (1:1; Invitrogen) and electrophoresed through 10 % Novex Zymogram Gelatin Gels (Invitrogen) with Tris-Glycine SDS Running Buffer (Invitrogen) under constant voltage of 125 V for 120 min. After electrophoresis, gels was incubated in Zymogram Renaturing Buffer (Invitrogen) for 30 min at room temperature with gentle agitation and washed with developing buffer (Invitrogen) for 30 min. The gel was further incubated for 24 h in fresh developing buffer at 37 °C. After developing, the gel was stained with 0.5 % (w/v) Coomassie Blue R-250 (Bio-Rad) in 50 % (v/v) methanol, 10 % (v/v) acetic acid for 30 min at room temperature, and destained as described previously . Gelatinolytic activity of MMP-9 was visualized as a clear band on a blue background. Band intensity was quantified by densitometry using ImageJ software. Briefly, zymogram gels were scanned using Kodak gel logic 100 imaging system and processed into gray scale. Gray scale images were quantified densitometrically by the measurement of the mean intensity of positive band multiplied by its corresponding area. The optical band intensity was then corrected by subtracting background intensity of equal area.
Results from at least three independent experiments are expressed as mean ± SEM. Statistical significance was evaluated using two-tail t-test for comparison between two groups, whereas the one way analysis of variance (ANOVA) was used for comparison of multiple groups. P < 0.05 was considered significant.
TGF-β1 induces EndoMT in HKGECs
The MMP inhibitor GM6001 inhibits TGF-β1-induced EndoMT in HKGECs
MMP-9 is involved in TGF-β1-induced EndoMT in HKGECs
To further examine whether MMP-9 is involved in EndoMT, HKGECs were treated with TGF-β1 in presence of MMP-9 inhibitor. Our results demonstrated that the MMP-9 inhibitor reduced α-SMA expression in the TGF-β1-treated samples. In addition, when different dosages of MMP-9 inhibitor were used, the reduction in α-SMA and recovery of VE-cadherin expression was dose dependent (Fig. 3b). To confirm that the MMP-9 inhibitor was functional, zymography was performed to evaluate MMP-9 and MMP-2 activity. The results showed that the MMP-9 inhibitor successfully inhibited MMP-9 activity (Fig. 3c).
rMMP-9 induces EndoMT in HKGECs
Notch signaling is activated in TGF-β1-induced EndoMT but is inhibited by MMP-9 inhibitor
rMMP-9 activates Notch signaling
Inhibition of Notch signaling by γ-secretase inhibitor (GSI) blocks rMMP-9-induced EndoMT
Kidney EndoMT is a major source of fibroblast formation in kidney fibrosis, and has emerged as a potentially important mechanism in development and progression of kidney fibrosis . In a recent study, LeBleu et al. found that endothelial cells play a more important role than epithelial cells and pericytes in myofibroblast formation via EndoMT . Therefore, understanding and halting EndoMT is an important clinical challenge.
Kidney endothelial cells consist of glomerular, vascular and peritubular endothelial cells. Although it has been generally accepted that endothelial cells contribute to fibroblast formation in kidney, whether glomerular endothelial cells contribute to fibroblast formation remains unclear. In our recent study, we showed that mice peritubular endothelial cells have an important role in kidney EndoMT. Moreover, we found that MMP-9 was involved in peritubular EndoMT, likely via the Notch pathway. Human glomerular endothelial cells are likely to be another instructive model for studying kidney fibrosis.
Here, we demonstrate that TGF-β1 and rMMP-9 induced EndoMT in HKGECs via Notch signaling. The Notch signaling pathway is a highly conserved cascade in mammals that regulates multiple cellular processes, including proliferation, differentiation, and apoptosis [30, 31]. Notch signaling has been shown to be downstream of VEGF to regulate endothelial cell morphogenesis . In addition, Notch activation induces endothelial cell morphological, phenotypic, and functional changes consistent with mesenchymal transformation . Notch signaling has been found to initiate EndoMT in atrioventricular endothelial cells . Inhibition of Notch signaling ameliorates EMT and tubulointerstitial fibrosis in mouse . Importantly, Notch inhibition reversed podocyte injury (EMT) and kidney failure . All these findings suggest a critical role for Notch signaling in both EMT and EndoMT which are pivotal processes in kidney fibrosis. Our results demonstrate that MMP-9 activation of Notch signaling in glomerular endothelial cells is downstream of TGF-β1. Combining with observations with previous studies from our laboratory and others, a strong profibrotic role for MMP-9 is evident. It promotes EMT of tubular epithelial cells, EndoMT of peritubular endothelial cells and EndoMT of glomerular endothelial cells, thereby, leading to kidney fibrosis in both tubulointerstial and glomeruli compartments.
Our data demonstrate that MMP-9 plays an important role in TGF-β1-induced EndoMT in HKGECs, via upregulation of Notch signaling. Thus, inhibition of MMP-9 or Notch signaling could be therapeutic strategies for treatment for kidney fibrosis in CKD.
This work was supported by the National Health and Medical Research Council (NHMRC) Project Grants 632688 and 1046647.
- 11.Ghosh AK, Bradham WS, Gleaves LA, De Taeye B, Murphy SB, Covington JW, Vaughan DE. Genetic deficiency of plasminogen activator inhibitor-1 promotes cardiac fibrosis in aged mice: involvement of constitutive transforming growth factor-beta signaling and endothelial-to-mesenchymal transition. Circulation. 2010;122(12):1200–9.CrossRefPubMedGoogle Scholar
- 12.Widyantoro B, Emoto N, Nakayama K, Anggrahini DW, Adiarto S, Iwasa N, Yagi K, Miyagawa K, Rikitake Y, Suzuki T, et al. Endothelial cell-derived endothelin-1 promotes cardiac fibrosis in diabetic hearts through stimulation of endothelial-to-mesenchymal transition. Circulation. 2010;121(22):2407–18.CrossRefPubMedGoogle Scholar
- 21.Kitao A, Sato Y, Sawada-Kitamura S, Harada K, Sasaki M, Morikawa H, Shiomi S, Honda M, Matsui O, Nakanuma Y. Endothelial to mesenchymal transition via transforming growth factor-beta1/Smad activation is associated with portal venous stenosis in idiopathic portal hypertension. Am J Pathol. 2009;175(2):616–26.CrossRefPubMedPubMedCentralGoogle Scholar
- 26.Zheng G, Lyons JG, Tan TK, Wang Y, Hsu TT, Min D, Succar L, Rangan GK, Hu M, Henderson BR, et al. Disruption of E-cadherin by matrix metalloproteinase directly mediates epithelial-mesenchymal transition downstream of transforming growth factor-beta1 in renal tubular epithelial cells. Am J Pathol. 2009;175(2):580–91.CrossRefPubMedPubMedCentralGoogle Scholar
- 34.Chang AC, Fu Y, Garside VC, Niessen K, Chang L, Fuller M, Setiadi A, Smrz J, Kyle A, Minchinton A, et al. Notch initiates the endothelial-to-mesenchymal transition in the atrioventricular canal through autocrine activation of soluble guanylyl cyclase. Dev Cell. 2011;21(2):288–300.CrossRefPubMedGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.