The metalloproteinase ADAM15 is upregulated by shear stress and promotes survival of endothelial cells

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

Contents lists available at ScienceDirect

Journal of Molecular and Cellular Cardiology journal homepage: www.elsevier.com/locate/yjmcc

Original article

The metalloproteinase ADAM15 is upregulated by shear stress and promotes survival of endothelial cells

T

Aaron Babendreyera, , Lisa Mollsa, Indra M. Simonsa, Daniela Dreymuellera,b, Kristina Billera, Holger Jahrc,d, Bernd Deneckee, Reinier A. Boonf, Sebastian Betteg, Uwe Schnakenbergg, ⁎ Andreas Ludwiga, ⁎

a

Institute of Pharmacology and Toxicology, RWTH Aachen University, Aachen, Germany Institute of Experimental and Clinical Pharmacology and Toxicology, PZMS, ZHMB, Saarland University, Homburg, Germany c Institute of Anatomy and Cell Biology, RWTH Aachen University, Aachen, Germany d Department of Orthopaedic Surgery, Maastricht University Medical Centre+, Maastricht, the Netherlands e Interdisciplinary Center for Clinical Research, RWTH Aachen University, Aachen, Germany f Institute for Cardiovascular Regeneration, Center of Molecular Medicine, Goethe University, Frankfurt am Main, Germany g Institute of Materials in Electrical Engineering 1, RWTH Aachen University, Aachen, Germany b

ARTICLE INFO

ABSTRACT

Keywords: Shear stress Blood flow Endothelium Metalloproteinase Cell death Transcriptional regulation Vascular inflammation

Reduced shear stress resulting from disturbed blood flow can impair endothelial integrity and drive the development of vascular inflammatory lesions. Metalloproteinases of the ADAM family have been implicated in the regulation of cell survival and inflammatory responses. Here we investigate the mechanism and function of ADAM15 upregulation in primary flow cultured endothelial cells. Transcriptomic analysis indicated that within the ADAM family ADAM15 mRNA is most prominently upregulated (4-fold) when endothelial cells are exposed to physiologic shear stress. This induction was confirmed in venous, arterial and microvascular endothelial cells and is associated with increased presence of ADAM15 protein in the cell lysates (5.6-fold) and on the surface (3.1-fold). The ADAM15 promoter contains several consensus sites for the transcription factor KLF2 which is also upregulated by shear stress. Induction of endothelial KLF2 by simvastatin treatment is associated with ADAM15 upregulation (1.8-fold) which is suppressed by counteracting simvastatin with geranylgeranyl pyrophosphate. KLF2 overexpression promotes ADAM15 expression (2.1-fold) under static conditions whereas KLF2 siRNA knockdown prevents ADAM15 induction by shear stress. Functionally, ADAM15 promotes survival of endothelial cells challenged by growth factor depletion or TNF stimulation as shown by ADAM15 shRNA knockdown (1.6-fold). Exposure to shear stress increases endothelial survival while additional knockdown of ADAM15 reduces survival (6.7-fold) under flow conditions. Thus, physiologic shear stress resulting from laminar flow promotes KLF2 induced ADAM15 expression which contributes to endothelial survival. The absence of ADAM15 at low shear stress or static conditions may therefore lead to increased endothelial damage and promote vascular inflammation.

1. Introduction Endothelial cells are constantly exposed to blood flow which results in laminar shear stress on the endothelial surface. Exposure to laminar flow typically leads to an alignment of endothelial cells in the direction

of flow, cytoskeletal rearrangements, formation of tight endothelial cell to cell contacts, reduced permeability, arrest of cell proliferation and prolonged cell survival [1–5]. Shear stress on the endothelial surface can clearly differ in large and small arteries or veins, respectively [6,7]. In large arteries an average of

Abbreviations: ADAM, a disintegrin and metalloproteinase; EDN1, endothelin-1; FAK, focal adhesion kinase; GGPP, geranylgeranyl pyrophosphate; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HPMECs, human pulmonary microvascular endothelial cells; HUVECs, human umbilical vein endothelial cells; HUAECs, human umbilical artery endothelial cells; KLF2, krüppel-like factor 2; NOS3, endothelial nitric oxide synthase; PARP-1, poly [ADP-ribose] polymerase-1; TNF, Tumor necrosis factor; ICAM1, intercellular adhesion molecule 1; VCAM1, vascular cell adhesion molecule 1; CX3CL1, C-X3-C motif chemokine ligand 1; MCP1, monocyte chemoattractant protein 1; IL8, interleukin 8 ⁎ Corresponding authors at: Institute of Pharmacology and Toxicology, RWTH Aachen University, Wendlingweg 2, 52074 Aachen, Germany. E-mail addresses: [email protected] (A. Babendreyer), [email protected] (A. Ludwig). https://doi.org/10.1016/j.yjmcc.2019.06.017 Received 29 March 2018; Received in revised form 18 June 2019; Accepted 28 June 2019 Available online 01 July 2019 0022-2828/ © 2019 Published by Elsevier Ltd.

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

10 dyn/cm2 (=1 Pa) has been reported while in the human microvasculature the shear stress may be up to ten times higher [6–8]. Under pathological conditions shear stress can be significantly influenced. On the one hand, narrowing of the blood vessel leads to increased shear stress directly at the site of confinement. On the other hand, shear stress can be reduced below 1 dyn/cm2 or even absent. This can occur after ischemic events or in vascular beds that are prone to atherosclerosis, where branch points and curved areas lead to disturbed flow including flow reversal or turbulence [8,9]. In these areas a disturbed endothelial cell function can be observed. This includes the loss of endothelial alignment, increased endothelial permeability and the induction of several proinflammatory effector molecules such as chemokines and adhesion molecules, which promote binding and recruitment of inflammatory leukocytes to these sites [4,9–13]. Under pathological conditions, endothelial cell survival is impaired and increasing loss of endothelial cells can be observed. In a situation of endothelial denudation, platelets become activated and thrombus formation can lead to major cardiovascular complications. Pathological responses of endothelial cells to reduced flow conditions in the vasculature are associated with transcriptional changes resulting in the altered expression of vascular tone regulators, endothelial surface molecules, soluble mediators and cell survival genes. Typically, endothelial nitric oxide synthase (NOS3) is downregulated by reduction of shear stress, whereas endothelin-1 (EDN1) is upregulated [14,15]. Krüppel-like factor 2 (KLF2) is a transcription factor that is expressed in endothelial cells exposed to flow conditions but clearly less abundant when shear stress is reduced or absent [15]. Inhibition of the mevalonate pathway by statins leads to inhibition of Rho GTPase activation followed by induction of KLF2 and thereby statins can partially reverse endothelial responses to reduced shear stress [16]. A disintegrin and metalloproteinase (ADAM) family members have been implicated in various endothelial cell functions [17]. ADAMs are type-I transmembrane molecules consisting of several domains including a metalloproteinase and a disintegrin domain. They are maturated in the golgi apparatus by removal of a prodomain preceding the metalloproteinase domain. ADAM10 and ADAM17 are well known to contribute to increased endothelial permeability and leukocyte transmigration by cleaving endothelial surface molecules involved in junction formation and leukocyte adhesion. ADAM8 and ADAM15 also hold proteolytic functions but they can additionally act as adhesion molecules by interaction of their disintegrin domain with integrins. ADAM15 has been implicated in pathological neovascularization, endothelial permeability induction, transendothelial neutrophil migration and edema formation [18–22]. So far, the regulation of ADAM expression under flow conditions and its functional implications has not been addressed. In the present study, we investigate transcriptional regulation of ADAM proteases in endothelial cells exposed to increasing shear stress. We identify ADAM15 as one of the most prominently induced ADAMs in response to shear stress. By pharmacological inhibition, gene knockdown and overexpression experiments we provide multiple evidence that this induction of ADAM15 is mediated via KLF2 and that ADAM15 promotes endothelial survival. These findings imply that loss of ADAM15 expression in regions of low flow may contribute to increased endothelial damage in the course of atherosclerotic lesion development.

USA). Mouse monoclonal antibody against human β-actin (clone mAbcam 8226) was from Abcam (Cambridge, UK). Allophycocyanin (APC)-conjugated goat anti-mouse antibody, horse radish peroxidase (HRP)-conjugated goat anti-mouse and rabbit anti-goat antibody were from Jackson ImmunoResearch Laboratories, Inc. (West Grove, USA). Fluorescein isothiocyanate (FITC)-labelled annexin V was from ImmunoTools GmbH (Friesoythe, Germany). Tumor necrosis factor (TNF) was from PreproTech (Rocky Hill, USA). Simvastatin prodrug and TAPI-1 were from Merck KGaA (Darmstadt, Germany). Bleomycin (BLEO-cell) was from Cell Pharm GmbH (Hannover, Germany). Puromycin was from Carl Roth (Karlsruhe, Germany). Mitomycin was from Medac GmbH (Wedel, Germany). Geranylgeranyl pyrophosphate (GGPP) was from Sigma-Aldrich (St. Louis, USA). 2.2. Cell culture Human pulmonary microvascular endothelial cells (HPMECs) were from Promocell (Heidelberg, Germany) and cultured in endothelial cell growth medium MV2 (Promocell) as recommended by the supplier. Human umbilical vein endothelial cells (HUVECs) and human umbilical artery endothelial cells (HUAECs) were isolated from the umbilical cord of caesarean sections in our laboratory as described [23] and cultured in Endopan-3 from PAN-Biotech (Aidenbach, Germany). Each experiment with either HUVECs or HUAECs was performed with cells that were prepared from a different donor. 2.3. siRNA transfection Directly before transfection, HUVECs were seeded in a 24-well plate with a density of 3.5 × 104 cells/well in basal Endopan-3 medium. Subsequently, transfection solution of 50 μl Opti-MEM® reduced serum medium (Thermo Fisher Scientific, Waltham, USA), 1 μl Lipofectamin™ RNAiMAX (Invitrogen, Carlsbad, USA) and 30 nM small interfering RNA (siRNA) targeting human KLF2 (stealth RNAi™ HSS145585) or unspecific control siRNA (Invitrogen) was added. After 24 h, the medium was replaced and transfected cells were used for stimulation and flow experiments. 2.4. Lentiviral transduction Lentiviral transduction of short hairpin RNA (shRNA) was performed with the MISSION® shRNA system from Sigma-Aldrich (St. Louis, USA). For ADAM15 knockdown the pLKO.1-puro plasmid TRC number 0000050549 (sequence: CCGGGCTGGTGACTGGTACTTCATTC TCGAGAATGAAGTACCAGTCACCAGCTTTTTG; binding at position: 1021 to 1043 of transcript variants 1–6 and 8, at position 1051 to 1073 of transcript variant 7 and at position 973 to 995 of transcript variant 9) was used. In one experiment an additional sequence was used for ADAM15 knockdown (TRC number 0000050548, sequence: CCGGCG ATCCCTGCTGTGATTCTTTCTCGAGAAAGAATCACAGCAGGGATCGTT TTTG; binding at position: 1447 to 1467 of transcript variant 1–6 and 8, at position 1477 to 1497 of transcript variant 7 and at position 1399 to 1419 of transcript variant 9). The pLKO.1-puro non-mammalian shRNA control plasmid DNA (SHC002) served as control. Overexpression of human KLF2 was performed with a hKLF2-PGK plasmid, using the corresponding empty plasmid as control. For production of recombinant lentiviruses, subconfluent HEK293T cells in a 25 cm cell culture dish were cotransfected with 12.5 μg of the specific pLKO.1puro plasmid, pLVX-neo plasmid or PGK plasmid, 8.13 μg of psPAX2 (plasmid 12260, Addgene) and 4.375 μg of pMD2.G (plasmid 12259, Addgene) using 50 μl jetPEI® DNA transfection reagent (Polyplustransfection, Illkirch France). Medium was changed after 24 h, and lentivirus containing supernatants were harvested after another 48 h. Lentiviral particles were concentrated 500-times by ultracentrifugation at 50,000 ×g for 2 h and resuspended in PBS. Lentivirus concentrate (5 μl) was added to 1.5 × 105 cells in culture medium supplemented

2. Material and methods 2.1. Antibodies, recombinant proteins and chemical compounds Mouse monoclonal antibody against the mature form of human ADAM15 (clone 23G9), goat polyclonal antibody against poly [ADPribose] polymerase-1 (PARP-1) and mouse IgG1 isotype control were from R&D Systems (Wiesbaden, Germany). Mouse monoclonal antibody against human GAPDH (clone GA1R) was from Invitrogen (Carlsbad, 52

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

with polybrene (4 μg/ml, Sigma-Aldrich, St. Louis, USA). Cells transduced with pLKO.1-puro or pLVX-neo plasmids were selected with puromycin or geneticin, respectively. Transduction efficiency with PGK plasmids was controlled by monitoring GFP expression.

2.9. Western blotting Cultured HUVECs were resuspended at 1 × 106 cells per ml in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 1% TritonX-100, 1 mM EDTA, 1 mM Na3VO4, 1 mM PMSF, 10 mM 1,10-phenanthroline monohydrate) supplemented with Complete Inhibitor (Roche) and SDS sample buffer (with final concentrations of 2% SDS, 50 mM Tris-HCl, 10% glycerol and 0.02% bromophenol blue) and incubated for 10 min. After centrifugation at 16,000g for 5 min, supernatants were investigated by reducing SDS-PAGE. Proteins were transferred onto polyvinylidene difluoride membranes (Hybond-P, Amersham) and probed with primary monoclonal antibody against ADAM15 (2 μg/ml), β-actin (0.1 μg/ml), GAPDH (1 μg/ml) or polyclonal antibody against PARP (0.5 μg/ml) over night at 4 °C followed by HRP-coupled polyclonal secondary antibody (30 ng/ml in PBS-T with 2% non-fat dry milk) for 1 h. After addition of enhanced chemiluminescence substrate (ECL, Thermo Fisher Scientific, Waltham, USA), signals were recorded and quantified using the LAS 3000 Image Analyzer® (Fujifilm, Tokyo, Japan). ADAM15 was detected as a single protein representing the mature form of approximately 90 kDa [26] without further bands resulting from unspecific binding. Densitometric analysis was performed with Image Studio Lite (Li-Cor, Lincoln, NB, USA) and protein levels were expressed as signal density ratio of the protein of interest and GAPDH.

2.5. Simvastatin treatment For KLF2 induction, HUVECs were treated for 24 h (RT-qPCR) or 48 h (Western Blot and flow cytometry) with 1 μM simvastatin. To suppress the simvastatin mediated KLF2 induction, treatment of HUVECs with 1 μM simvastatin was performed in the presence of 10 μM GGPP for 48 h [24]. 2.6. Flow experiments For flow experiments cells were used in passage four to six and seeded with a density of 100.000 cells/cm2 in ibidi μ-Slides of different type (0.1, 0.2, 0.4 and 0.8 mm μ-slides, Martinsried, Germany) or a 24well plate for the static control. The flow in the μ-Slides was created with the ibidi pump system. Depending on the geometry of the μ-slide type, the flow rate was adjusted as specified by the manufacturer to yield the desired laminar shear stress.

2.10. Flow cytometric analysis

2.7. GeneChip array

PBS supplemented with 0.2% BSA was used as assay buffer, and all steps of the staining process were performed at 4 °C. HUVECs were analyzed for expression of ADAM15 by incubation with a mouse monoclonal antibody against ADAM15 (5 μg/ml) followed by incubation with APC-conjugated anti-mouse antibody (1:200). Isotype controls were used in parallel. The fluorescence signal was detected by flow cytometry (LRS Fortessa, BD Biosciences) and analyzed with FlowJo 8.7.3 software (Tree Star, Inc., Ashland, USA).

Gene expression in cells cultured under flow conditions or static conditions as control were analyzed in three independent experiments using the GeneChip® Human Transcriptome Array 2.0 (Affymetrix, Santa Clara, CA, USA) as described [25]. To assess differential expression, CEL files were normalized by robust multiarray average (RMA) to facilitate comparisons across arrays using the AltAnalyze 2.0.9 software suite (Gladstone Institutes, San Francisco, CA, USA). Default analysis options (analysis of gene expression via rawp) were used.

2.11. Apoptosis assay

2.8. RT-qPCR

Apoptosis was induced by treatment with 10 ng/ml TNF or by growth factor depletion with serum free medium for 24 h. Cell apoptosis was then determined using three different approaches. For end point analysis, cells were stained with the Annexin V Apoptosis Detection Kit eFluor 450 (FITC-Annexin V and 7-AAD, eBioscience, San Diego, USA) according to the manufacturer's protocol and the fluorescence signal was detected by flow cytometry (see above). For live-cell imaging, cells were incubated with FITC-labelled annexin V in Hanks Buffered Salt Solution (HBSS) for 30 min. After washing with HBSS, the relative apoptosis was determined as ratio of the green fluorescent area and the total cell area in at least 8 images for each condition using the IncuCyte ZOOM microscope with software 2016B (Essen Bioscience, Ann Arbor, USA). Additionally, caspase-3/7 activity assays were performed. For this, cells were incubated with medium containing the IncuCyte Caspase-3/7 Green Reagent (1:1000 dilution, Essen Bioscience, Ann Arbor, USA) and the IncuCyte CytoTox Red Reagent (1:4000 dilution, Essen Bioscience, Ann Arbor, USA) for live/dead staining. Apoptosis was induced as mentioned above by treatment with TNF or growth factor depletion. In addition also the effect of TAPI (10 μM), simvastatin (1 μM) and an anti-ADAM15 IgG (10 μg/ml, clone 23G9) on apoptosis was tested. Green and red fluorescence were imaged every hour over a period of 72 h with the IncuCyte ZOOM and cells positive for green or red fluorescence were counted with the IncuCyte ZOOM software 2016B.

The mRNA levels were quantified by RT-qPCR analysis and normalized to the mRNA level of different reference genes. The most stable reference genes were determined with the geNorm algorithm included in the qbase+ software (biogazelle, Gent, Belgium). Based on these results, glyceraldehyde 3-phosphate dehydrogenase (GAPDH, TATAbinding protein (TBP) and cytochrome C1 (CYC1) were chosen as most stable reference genes. In the first set of experiments mRNA levels were normalized against all three reference genes. For ADAM15 knockdown experiments the reference genes GAPDH and TBP were used to exclude an effect of the knockdown on the expression of the reference genes. All other RT-qPCRs were normalized to the most stable reference gene GAPDH. In case of normalization with multiple reference genes the term reference gene index was used and the composition was explained in the figure legend. RNA was extracted using RNeasy Kit (Qiagen, Hilden, Germany) and quantified photometrically. For each set of experiments equal amounts of RNA were reverse transcribed using PrimeScript™ RT Reagent Kit (Takara Bio Europe, St-Germain-en-Laye, France) and PCR reactions were performed using SYBR Premix Ex Taq II (Takara Bio Europe) according to the manufacturer's protocol. The specific primers and annealing temperatures are listed in Supplemental Table 1. All PCR reactions were run on a LightCycler® 480 system (Roche, Basel, Switzerland) with the following protocol: 45 cycles of 10 s denaturation at 95 °C, followed by 10 s annealing at the indicated temperature and 15 s amplification at 72 °C. Standard curves were determined by a serial dilution of a defined cDNA standard within each data set. Relative quantification was performed with the E-Method from Roche Applied Bioscience using the LightCycler®480 software 1.5 (Roche).

2.12. Statistics Quantitative data are shown as mean and SD calculated from at least three independent experiments. Data were analyzed by general 53

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

mixed model analysis (PROC GLIMMIX, SAS 9.4, SAS Institute Inc., Cary, USA) and assumed to be derived from either log normal or negative binomial (count data) distributions; residual plots and the Shapiro-Wilk test were used as diagnostics. In case of heteroscedasticity (according to the covtest statement), the degrees of freedom were adjusted by the Kenward-Roger approximation. Differences between the parameters were tested with Student's t-test or in case of heteroscedasticity with the Mann-Whitney U test. Multiple comparisons were corrected by false discovery rate (FDR). A p-value < .05 was considered significant. 3. Results 3.1. Shear stress upregulates endothelial ADAM 15 expression Human umbilical vein endothelial cells (HUVECs) were cultured under static or different flow conditions. As expected, endothelial cells under flow conditions acquired an elongated shape and aligned in the direction of flow (Suppl. Fig. 1A). To further validate our experimental setup, we analyzed the mRNA expression of NOS3 and EDN1 that are known to be regulated by shear stress. Indeed, expression of NOS3 was induced with increased levels of shear stress while the expression of EDN1 was reduced (Suppl. Fig. 1B-C). We next confirmed the described effects of flow culture on proliferation and cell survival [2,3]. As expected, cell survival was increased upon flow culture as shown by reduced annexin V-binding (Suppl. Fig. 2A). Effects on cell proliferation were not observed in this setup, since cells were seeded with a high density to rapidly reach confluence (Suppl. Fig. 2B). This system was then used to study the influence of shear stress on the expression of ADAM protease family members. HUVECs were cultured under static (0 dyn/cm2) or flow (30 dyn/cm2) conditions, and the transcriptome was investigated by genome-wide analysis with the GeneChip HTA 2.0 (Affymetrix, Santa Clara, CA, USA). Only three ADAMs showed significant upregulation (ADAM15 and ADAM19) or downregulation (ADAM23) by a factor of more than two (Fig. 1A). Endothelial ADAM15 has been implicated in several pathologies [18,27–30] while the role of ADAM19 and ADAM23 still remains unclear. These reports suggest that the observed regulation of ADAM15 could impact relevant endothelial cell functions and we therefore focused on the further investigation of ADAM15. The induction of ADAM15 mRNA expression by increasing levels of shear stress was first verified by quantitative PCR (Fig. 1B). Of note, the kinetic of ADAM15 mRNA expression was very similar to that of NOS3. Since divergent reports exist whether proinflammatory mediators such as TNF can upregulate ADAM15 mRNA expression in endothelial cells [18,31], we additionally exposed HUVECs to TNF. This, however, did not further affect the induction of ADAM15 by shear stress (Fig. 1C). Moreover, a similar upregulation of ADAM15 was seen in HUVECs, human umbilical artery endothelial cells (HUAECs) and human pulmonary endothelial cells (HPMECs) indicating that this regulation by shear stress is independent of the vascular type of endothelial cells (Fig. 1D). In subsequent experiments HUVECs were used to further study the mechanism and function of shear stress induced ADAM15 regulation. Upregulation of ADAM15 protein in response to shear stress was confirmed by Western blotting with an antibody against mature ADAM15 (Fig. 2A). Under static conditions ADAM15 was detectable as a single band in the cell lysates. The molecular weight of approximately 90 kDa corresponded well to the reported molecular weight of mature ADAM15 [26]. This band was clearly more prominent when cells were exposed to shear stress. Additionally, ADAM15 surface expression was analyzed by flow cytometry, showing basal expression of the protease at the surface of endothelial cells under static culture conditions but clear upregulation in response to shear stress (Fig. 2B, C).

Fig. 1. Transcriptional expression of ADAM family members by cultured endothelial cells in response to flow. A: HUVECs were cultured with and without flow for 48 h and genome-wide transcriptome analysis was performed by microarray analysis (HTA 2.0 chip, Affymetrix, Santa Clara, CA, USA). Data for all known ADAMs expressed in humans are shown as fold change of endothelial mRNA expression under flow culture versus static culture determined in three independent experiments each with HUVEC from a different donor. Data are shown as mean fold change in mRNA expression of three independent experiments. B: HUVECs were cultured for 24 h under different conditions of flow resulting in the indicated laminar shear stress. Cells were then studied for ADAM15 mRNA expression in relation to a reference gene index consisting of GAPDH, TBP and CYC1. C: HUVECs were stimulated with 10 ng/ml TNF or left unstimulated under static or flow conditions and studied for ADAM15 mRNA expression in relation to that of GAPDH. D: Endothelial cells from different vascular beds were cultured as described above and compared for mRNA upregulation of ADAM15 in relation to that of GAPDH. Data shown in B-D are presented as mean + SD from of least three independent experiments. Statistical differences to the static control are indicated by asterisks (* p < .05, ** p < .01, *** p < .001).

3.2. ADAM15 upregulation by shear stress involves KLF2 It is well known that the transcription factor KLF2 is induced by shear stress and is responsible for the regulation of the majority of shear stress sensitive genes [32–34]. Interestingly, three characteristic CCGCCC elements for binding of KLF2 can be found in the ADAM15 promoter region (Suppl. Fig. 3). We therefore questioned whether ADAM15 is also regulated by KLF2. HUVECs were cultured under static or different flow conditions and the mRNA expression of KLF2 was analyzed. We confirmed that KLF2 mRNA expression is strongly 54

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 2. Endothelial shear stress enhances expression of ADAM15 mRNA and protein. A-C: Upregulation of ADAM15 protein expression in response to shear stress was studied in cell lysates by Western blotting (A) and on the cell surface by flow cytometry (B, C). Results are shown as representative blot and densitometric analysis (A), as summary of mean fluorescence intensities or representative histogram B, C). All quantitative data are presented as mean + SD from at least four (A) or three (B) independent experiments. Statistical differences to the static control are indicated by asterisks (* p < .05, ** p < .01). Fig. 3. Regulation of ADAM15 correlates with expression of KLF2. A: HUVECs were cultured under static and flow conditions for 24 h and KLF2 mRNA expression was analyzed. B-D: HUVECs were treated for 48 h with 1 μM simvastatin or control vehicle (DMSO) in the presence or absence of 10 μM GGPP. Cells were then analyzed for mRNA expression of KLF2 (B), NOS3 (C) and ADAM15 (D). E: HUVECs were treated for 48 h with or without simvastatin, and ADAM15 protein expression was analyzed by Western blotting Results are shown as representative blot F: HUVECs were treated for 48 h with 1 μM simvastatin or control vehicle (DMSO) in the presence or absence of 10 μM GGPP. Subsequently, ADAM15 surface expression was analyzed by flow cytometry. Results are shown as summary of mean fluorescence intensities. Quantitative data are shown as mean + SD of at least three independent experiments. Statistical differences to the control are indicated by asterisks (* p < .05, *** p < .001) and in B – D and F differences between the treatments are indicated by rhombs (# p < .05, ## p < .01, ### p < .001).

induced by shear stress in our experimental setup (Fig. 3A). Under static conditions KLF2 is only weakly expressed. However, it is reported that KLF2 expression can also be induced under these conditions by inhibition of the mevalonate pathway with simvastatin and that this effect is abrogated by geranylgeranyl pyrophosphate (GGPP) which is a downstream product of the mevalonate pathway [16,24]. To study the effects of this KLF2 regulation on ADAM15, we treated HUVECs with simvastatin prodrug alone or together with GGPP for 24 h. The treated cells were then analyzed for KLF2 mRNA expression to confirm its regulation by simvastatin exposure. Additionally, we analyzed mRNA expression of NOS3 and ADAM15 as known or potential KLF2 target genes, respectively (Fig. 3B-D). The expression of KLF2 was clearly induced by simvastatin, which could be prevented by GGPP treatment (Fig. 3B). The known KLF2 target gene NOS3 was regulated in the same manner by simvastatin and GGPP (Fig. 3C). ADAM15 expression was also induced by simvastatin, which could be completely suppressed by GGPP (Fig. 3D). The induction of ADAM15 by simvastatin could also be observed on protein level by Western blot analysis (Fig. 3E) and by flow cytometry (Fig. 3F). To exclude KLF2-independent effects of simvastatin and GGPP, we also tested the effect of lentiviral KLF2 overexpression in HUVECs. HUVECs were transduced with lentivirus carrying an empty vector or a vector coding for KLF2 mRNA. Subsequently, transduced cells were analyzed for KLF2, NOS3 and ADAM15 mRNA expression (Fig. 4A-C). Cells transduced with the lentivirus for KLF2 overexpression showed a clear upregulation in KLF2 mRNA expression, which correlated with mRNA induction of NOS3 and ADAM15. As final proof for our hypothesis that KLF2 is required for

upregulation of ADAM15 expression, we established a KLF2 knockdown in HUVECs via siRNA. After 24 h cell culture at static or flow conditions, KLF2, NOS3 and ADAM15 mRNA expression was analyzed (Fig. 4D-F). Due to the fact that KLF2 is unstable under static conditions [35], the degradation of KLF2 mRNA could not be further increased by siRNA transfection. However, flow conditions lead to a profound upregulation of KLF2 mRNA in cells treated with control siRNA and this upregulation was completely suppressed by siRNA against KLF2 (Fig. 4D). As predicted, we could also observe a shear stress-mediated induction of ADAM15 and NOS3 mRNA expression in cells transfected with control siRNA, but there was no induction in KLF2 knockdown cells (Fig. 4E, F). These findings confirm that ADAM15 mRNA upregulation by shear stress is mediated by the transcription factor KLF2. 55

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 4. Shear stress induced upregulation of ADAM15 is suppressed by KLF2 knockdown. A-C: HUVECs were transduced with lentivirus for overexpression of ADAM15 or control vector. After 24 h, cells were cultured under static or flow conditions for 24 h and analyzed for their mRNA expression of KLF2 (A) NOS3 (B) and ADAM15 (C) D-F: HUVECs were transfected with non-targeting control siRNA or siRNA against KLF2. After 24 h, cells were cultured under static or flow conditions for 24 h and analyzed for their mRNA expression of KLF2 (D) NOS3 (E) and ADAM15 (F). All data are shown as mean + SD and of at least five independent experiments. Statistical differences to the control are indicated by (** p < .01, *** p < .001) and in D - F differences due to KLF2 knockdown are indicated by rhombs (# p < .05, ## p < .01).

3.3. Inflammatory gene induction and apoptosis are enhanced in the absence of ADAM15

apoptosis in synovial fibroblasts [36], we also investigated endothelial cell survival. In the first experiments, cell death of HUVECs with and without ADAM15 knockdown was induced by growth factor depletion for 24 h and the number of apoptotic cells was measured by annexin V staining and flow cytometry. ADAM15 knockdown caused only a slight non-significant increase in cell death under basal conditions, but clearly increased the induction of cell death by growth factor depletion (Fig. 6A). In a second set of experiments endothelial cell death was induced by stimulation with TNF for 24 h. As a more direct measure of induced cell death the number of annexin V positive apoptotic cells in the intact cell layer were analyzed via microscopic live cell imaging. ADAM15 knockdown had a weak effect on unstimulated cells, but again strongly enhanced induction of cell death by TNF (Fig. 6B, C). As a different indicator of cell death, we investigated degradation of poly [ADP-ribose] polymerase 1 (PARP-1) by Western blotting. Both tested ADAM15 shRNA sequences reduced the presence of intact full length PARP-1 in untreated and TNF treated cells (Fig. 6D). The TNF-induced reduction was significant for only one sequence but the second sequence showed a similar trend. Nonetheless, all experiments suggest that ADAM15 can promote endothelial cell survival and resistance against induced cell death. This was further corroborated by a caspase 3/7 activity assay with life/dead staining demonstrating increased caspase 3/7 activity in unstimulated, TNF stimulated and growth factor depleted cells when expression of ADAM15 is suppressed (Fig. 7 A-F). This was not associated with increased necrosis of unstimulated cells (Suppl. Fig. 7A, B) or TNF stimulated cells (Suppl. Fig. 7C, D) and in growth factor depleted cells (Suppl. Fig. 7E, F) only an early transient increase in necrosis was observed that was not significant. Of note, treatment with the metalloproteinase inhibitor TAPI which is known to block several ADAM- and MMP-family members led to an

To study which function could be mediated by ADAM15 in endothelial cells under flow, we transduced endothelial cells with a lentivirus encoding shRNA for knockdown of ADAM15. The successful knockdown was confirmed by Western blotting and flow cytometry (Fig. 5A-C). We could not observe any differences in cell proliferation, migration and angiogenesis between the control and the ADAM15 knockdown cells under static conditions (Suppl. Fig. 4A-C). Proliferation under flow condition was also not affected by ADAM15 knockdown (Suppl. Fig. 5). We next studied whether ADAM15 would affect the induction of inflammatory genes in endothelial cells (Fig. 5D-I). The mRNA levels of ICAM1 (Intercellular Adhesion Molecule 1), VCAM1 (Vascular Cell Adhesion Molecule 1), CX3CL1 (C-X3-C Motif Chemokine Ligand1/ Fractalkine), MCP1 (Monocyte Chemoattractant Protein 1) and IL8 (Interleukin 8) were studied in HUVECs transduced with control shRNA or shRNA against ADAM15. The investigated genes were expressed at very low basal levels in unstimulated cells and no relevant difference upon ADAM15 knockdown was observed. However, when the cells were treated with TNF to induce an inflammatory response the expression of the inflammatory markers was stronger in cells with knockdown of ADAM15. For the chemokine CX3CL1 this difference was significant and for the other inflammatory markers a similar trend was observed. By contrast, no difference was seen for the protease ADAM10 (Suppl. Fig. 6) excluding the possibility that the results are caused by more general effects on gene transcription. These data indicate that the inflammatory response in endothelial cells under static conditions can be attenuated in the presence of ADAM15. Since ADAM15 has been reported to mediate resistance against 56

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 5. ADAM15 knockdown increases inflammatory gene induction by TNF. HUVECs were transduced with a lentivirus encoding non-targeting control shRNA or shRNA against ADAM15. A-C: The cells were controlled for knockdown of ADAM15 protein expression in the cell lysate (A) and on the cell surface (B, C). Results are shown as representative Western blot with a summary of the densitometric analysis (A), as summary of the mean fluorescence intensities measured by flow cytometry (B) and as representative histogram (C). D-I: HUVECs with or without ADAM15 knockdown were left unstimulated or stimulated with 10 ng/ml TNF for 24 h. Subsequently, cells were harvested and analyzed for mRNA expression of ADAM15 (D), IL8 (E), CX3CL1 (F), MCP1 (G), ICAM1 (H) and VCAM1 (I) in relation to a reference gene index consisting of GAPDH and TBP. All quantitative data are shown as mean + SD of at least three independent experiments. Statistical differences to the control are indicated by asterisks (* p < .05, ** p < .01, *** p < .001) and in D-I differences between control shRNA and ADAM15 shRNA are indicated by rhombs (# p < .05, ### p < .001).

induction of endothelial apoptosis to a similar extend as ADAM15 knockdown (Fig. 8A-B). To more specifically target ADAM15 we treated the cells with an anti-ADAM15 antibody. Also this antibody induced endothelial apoptosis to a similar extend as the inhibitor (Fig. 8A-B). These results indicate a protective role of a metalloproteinase activity as well as ADAM15 in endothelial apoptosis. Since ADAM15 is upregulated in endothelial cells by shear stress, the contribution of the protease to endothelial cell survival may become more prominent under flow conditions. We already showed that that the amount of annexin V positive cells is profoundly reduced by shear stress (Suppl. Fig. 2A). To investigate whether ADAM15 would contribute to this increased endothelial cell survival under flow, HUVECs with and without knockdown of ADAM15 were cultured under flow conditions for 24 h and stained with annexin V. HUVECs with ADAM15 knockdown showed a clearly increased number of annexin V positive cells (Fig. 8C). Of note, the annexin V-positive cells are localized slightly above the monolayer indicating that they round up and become displaced from the monolayer but remain attached to neighboring cells when they undergo cell death. These data indicate that the upregulation of ADAM15 in response to shear stress contributes to the survival of endothelial cells under these conditions.

4. Discussion Our study demonstrates that exposure of primary cultured endothelial cells to conditions of high shear stress induces ADAM15 expression. We here provide pharmacological and genetic evidence that this upregulation is mediated by the transcription factor KLF2. As a consequence of this, upregulation of ADAM15 contributes to enhanced survival of endothelial cells exposed to laminar shear stress. We found that exposure of endothelial cells to flow conditions significantly affects the expression profile of ADAM family proteases. The well-studied members of this family, ADAM10 and ADAM17, are constitutively expressed by endothelial cells under static conditions. While ADAM10 is not further regulated slight induction of ADAM17 was observed. In contrast, ADAM15 was expressed to some degree under static conditions but considerably upregulated by flow conditions. This was confirmed in venous, arterial and microvascular endothelial cells and therefore does not seem to be limited by functional specialization of the endothelial cells. The ADAM expression profile can be modulated by proinflammatory mediators such as TNF, which has been shown to enhance ADAM8 expression [37]. However, in our setup TNF per se did not affect mRNA expression of ADAM15 under static or flow conditions. This is consistent with earlier observations showing that neither cytokine activation nor endothelial damage can enhance ADAM15 57

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 6. ADAM15 knockdown decreases endothelial cell survival. HUVECs were transduced with a lentivirus encoding non-targeting control shRNA or shRNA against ADAM15. A: HUVECs with or without ADAM15 knockdown were left untreated or challenged by depletion of growth factors for 24 h. Subsequently, cells were harvested, stained for annexin V positive cells and investigated by flow cytometry. B-C: HUVECs with or without ADAM15 knockdown were stimulated with 10 ng/ml TNF or left unstimulated for 24 h. The intact cell layer was stained with FITCannexin V and analyzed by microscopy. The area of cells with green fluorescence was expressed in relation to the total area of cells (B) and representative images for each condition are shown (C). D: Cell lysates from HUVECs with or without ADAM15 knockdown were analyzed for presence of unfragmented PARP-1 by Western blotting and quantified signals were normalized to the signal of cells transduced with control shRNA. For this experiment two different shRNA sequences (#1 and #2) were tested. Both are complementary to the coding region of all ADAM15 transcript variants. There binding to positions are 1021–1043 (#1) and 1447–1467 (#2). All quantitative data are shown as mean + SD of at least three independent experiments. Statistical differences to the control are indicated by asterisks (* p < .05, ** p < .01, *** p < .001) and differences between control shRNA and ADAM15 shRNA are indicated by rhombs (# p < .05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

for endothelial cells. It is not yet clear to what extent ADAM15 is responsible for KLF-mediated effects in endothelial cells. In contrast to ADAM15, KLF2 may even enhance apoptosis. Taniguchi et al. have reported that transient overexpression of KLF2 leads to increased annexin V binding indicating induced apoptosis [40]. It can be envisaged that under the tested conditions KLF2 induces other genes with proapoptotic activity that override the antiapoptotic activity of induced ADAM15. Moreover, it is not clear how ADAM15 exerts this antiapoptotic function. We observed that the metalloproteinase inhibitor TAPI can promote endothelial apoptosis to a similar extent as ADAM15 knockdown or treatment with an antibody against ADAM15. However, the results do not yet allow a final conclusion whether ADAM15 activity is required, since the antibody may have blocked other non enzymatic functions of ADAM15 as well or the inhibitor may have affected antiapototic activity of relevant proteases other than ADAM15. In principle, ADAM15 acts as an adhesion molecule and as a protease, and both activities could be implicated in mediating cell survival. The disintegrin domain of ADAM15 is known to bind to different integrins [30]. Such binding could initiate integrin mediated inside-out signaling which can play a crucial role in cell migration, differentiation and cell survival. For example, shear stress mediated cell signaling via complex formation of platelet endothelial cell adhesion molecule 1 (PECAM-1), vascular endothelial cadherin (VE-cadherin) and vascular endothelial growth factor receptors (VEGFRs) [41] is mediated by integrins leading to downstream signaling through Src kinase and focal adhesion kinase (FAK) [42]. Interestingly, ADAM15 has been described to colocalize with VE-cadherin in adherence junctions and to directly interact with FAK via its cytoplasmic domain [43,44]. The activation of integrin and Src/FAK signaling can contribute to increased cell survival as observed with osteocytes exposed to mechanical strain [45]. Moreover, in chondrocytes overexpression of ADAM15 leads to increased cell survival due to increased Src/FAK depended expression of X-linked inhibitor of apoptosis (XIAP) [38]. These reports indicate that ADAM15

expression [18]. Thus, upregulation of ADAM15 appears to be restricted to very special conditions and pathways such as those occurring under flow. Interestingly, under these conditions also ADAM19 is upregulated while ADAM23 is downregulated but it cannot be speculated on the consequences of this regulation since only little is known about the function of these ADAMs. Multiple evidence is provided in the present study that the flow induced regulation of ADAM15 in endothelial cells is brought about by the induction of the transcription factor KLF2. It is well known that KLF2 expression is induced in endothelial cells by flow conditions [32]. It is thought that this induction is mediated by Rho GTPase. This explains why inhibition of the mevalonate pathway by statins, which suppresses functionally relevant geranylation of Rho, induces KLF2 expression. Geranylation of Rho by GGPP abrogates the statin effect on KLF2 expression [16]. Alongside with KLF2 regulation we observed the same regulation of NOS3 and ADAM15 by flow application, by statin treatment or by GGPP. By knockdown we show that KLF2 is required for induction of ADAM15 and NOS3 by flow. Moreover, by overexpression of KLF2 in static cultures we show that the increased presence of KLF2 is sufficient to induce enhanced ADAM15 and NOS3 expression. The data show that ADAM15 is a KLF2 target gene similar to NOS3. In fact, possible KLF2 binding sites could be located in the ADAM15 promoter. Although this suggests that KLF2 exerts its activity on ADAM15 by direct binding to the ADAM15 promoter, the involvement of other indirect mechanisms involving other transcription factors cannot be excluded at this stage. Our study indicates that silencing of ADAM15 increases TNF-induced gene transcription as well as the susceptibility of endothelial cells to induced cell death. Conversely, it can be assumed that the upregulation of ADAM15 contributes to enhanced endothelial survival under flow conditions. Other ADAMs have been linked to cell survival, and there are several reports that also ADAM15 promotes fibroblast and chondrocyte survival [36,38,39]. In fact, we found a similar function 58

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 7. ADAM15 knockdown increases endothelial caspase-3/7 activity. A-F: HUVECs were transduced with a lentivirus encoding non-targeting control shRNA or shRNA against ADAM15. Subsequently, transduced HUVECs were left untreated (A,B), stimulated with 10 ng/ml TNF (C,D) or challenged by depletion of growth factors (E,F). A fluorescence based activity assay using the IncuCyte Caspase-3/7 Green Reagent allowed visualization of caspase-3/7 activity over a period of 72 h via life microscopy. Fluorescence images were taken every hour and caspase-3/7 positive cells were counted with the IncuCyte ZOOM software 2016B. Data are presented as time course of the caspase-3/7 positive cell count per image (A, C and E) or as area under the curve of these time courses (B, D and F). All quantitative data are shown as mean + SD of at least four independent experiments. Statistical differences to the control are indicated by asterisks (** p < .01, *** p < .001). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

enhances signaling pathways improving cell survival in different cell types. It is also possible that ADAM15 protects from cell death via its proteolytic activity. ADAM15 has been reported to cleave growth factors including epidermal growth factor receptor (EGFR) ligands [46]. This may allow autocrine signaling by the cleaved growth factor which will then confer to resistance against apoptosis [47]. The further study of the involved mechanism with targeted mutations of the disintegrin, the metalloproteinase domain or the intracellular C-terminus is challenging since overexpression of active exogenous ADAM15 does not per se increase cell survival in primary endothelial cells (unpublished observations). For this reason, knock-in mutations of endogenous ADAM15 in endothelial cells and subsequent induction by flow experiments need to be performed for an in-depth study of the mechanism of ADAM15 function in these cells. ADAM15 has been implicated in angiogenesis and pathological neovascularization in mice [19]. In our assays with cultured primary endothelial cells knockdown of ADAM15 neither affected tube formation nor proliferation. Since flow conditions provide a differentiation signal for endothelial cells it seems unlikely that the observed induction

of endothelial ADAM15 under flow conditions will lead to enhanced angiogenesis. In contrast to endogenous ADAM15, antiangiogenetic activities have been attributed to exogenously applied recombinant ADAM15 consisting of only the disintegrin domain [48]. This may become further complicated by the circumstance that cell expressed ADAM15 can be released in exosomes. The increased presence of released ADAM15 may potentially counteract the angiogenetic activities of the cell expressed protease. Furthermore, anti- or proangiogenetic functions may be balanced differently under conditions of endothelial culture or in the in vivo situation in microvessels [49,50]. Although our study provides evidence that ADAM15 may have an anti-inflammatory effect in endothelial cells also pro-inflammatory activities of ADAM15 have to be considered. ADAM15 on the surface of endothelial cells has been described as adhesion receptor for platelet GPIIb/IIIa and by this ADAM15 could promote thrombus formation in cardiovascular pathologies [31]. In addition, ADAM15 is known to be expressed in endothelial junctions and the protease has been implicated in VE-cadherin shedding [21,43]. This may explain the reported finding that ADAM15 can promote endothelial hyperpermeability and edema 59

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

Fig. 8. ADAM15 knockdown decreases endothelial cell survival under flow. A-B: HUVECs were left untreated or treated with 10 μM TAPI or 10 μg/ml anti-hADAM15 IgG. Caspase-3/7 activity was visualized for 72 h by a fluorescence based activity assay using the IncuCyte Caspase-3/7 Green Reagent. Fluorescence images were taken every hour and caspase-3/7 positive cells were counted with the IncuCyte ZOOM software 2016B. Data are presented as time course of the caspase-3/7 positive cell count per image (A) or as area under the curve of these time courses (B). C: HUVECs transduced with control shRNA or shRNA against ADAM15 were cultured under flow conditions and stained with FITC-annexin V. The area of green fluorescent cells was determined in relation to the total area of cells. Representative images of FITCannexin V staining are shown for every condition. Quantitative data are shown as mean + SD for at least four (A-B) or six (C) independent experiments. Statistical differences are indicated by asterisks (* p < .05). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Appendix A. Supplementary data

formation [20–22]. The absence of ADAM15 at regions of low flow may therefore also be regarded as a protective response that reduces the risk of immediate inflammatory events such as thrombus or edema formation. However, this short-term protection may be gained to the disadvantage of decreased endothelial survival. As we show in our study, loss of ADAM15 reduces endothelial survival and this may be of particular relevance at sites of low flow. In this case, the endothelial cells cannot cope with stress factors and this may lead to increased TNFinduced inflammatory gene transcription, endothelial damage and vascular dysfunction. This would then in the long run promote inflammatory lesion development and potentially also edema formation and platelet activation. Thus, the reported proinflammatory activity of ADAM15 may be limited to immediate events while in a long-term perspective the presence of ADAM15 could be beneficial by promoting endothelial survival and functional integrity.

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.yjmcc.2019.06.017. References [1] K.A. Barbee, P.F. Davies, R. Lal, Shear stress-induced reorganization of the surface topography of living endothelial cells imaged by atomic force microscopy, Circ. Res. 74 (1994) 163–171. [2] S. Dimmeler, B. Assmus, C. Hermann, J. Haendeler, A.M. Zeiher, Fluid shear stress stimulates phosphorylation of Akt in human endothelial cells: involvement in suppression of apoptosis, Circ. Res. 83 (1998) 334–341. [3] S. Akimoto, M. Mitsumata, T. Sasaguri, Y. Yoshida, Laminar shear stress inhibits vascular endothelial cell proliferation by inducing cyclin-dependent kinase inhibitor p21(Sdi1/Cip1/Waf1), Circ. Res. 86 (2000) 185–190. [4] B.S. Conklin, R.P. Vito, C. Chen, Effect of low shear stress on permeability and occludin expression in porcine artery endothelial cells, World J. Surg. 31 (2007) 733–743. [5] N.G. dela Paz, T.E. Walshe, L.L. Leach, M. Saint-Geniez, P.A. D'Amore, Role of shear-stress-induced VEGF expression in endothelial cell survival, J. Cell Sci. 125 (2012) 831–843. [6] T.G. Papaioannou, C. Stefanadis, Vascular wall shear stress: basic principles and methods, Hell. J. Cardiol. 46 (2005) 9–15. [7] A.G. Koutsiaris, S.V. Tachmitzi, N. Batis, Wall shear stress quantification in the human conjunctival pre-capillary arterioles in vivo, Microvasc. Res. 85 (2013) 34–39. [8] G. Dai, M.R. Kaazempur-Mofrad, S. Natarajan, Y. Zhang, S. Vaughn, B.R. Blackman, et al., Distinct endothelial phenotypes evoked by arterial waveforms derived from atherosclerosis-susceptible and -resistant regions of human vasculature, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 14871–14876. [9] J. Suo, D.E. Ferrara, D. Sorescu, R.E. Guldberg, W.R. Taylor, D.P. Giddens, Hemodynamic shear stresses in mouse aortas: implications for atherogenesis, Arterioscler. Thromb. Vasc. Biol. 27 (2007) 346–351. [10] A.G. Passerini, D.C. Polacek, C. Shi, N.M. Francesco, E. Manduchi, G.R. Grant, et al., Coexisting proinflammatory and antioxidative endothelial transcription profiles in a disturbed flow region of the adult porcine aorta, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 2482–2487. [11] S. Mohan, N. Mohan, A.J. Valente, E.A. Sprague, Regulation of low shear flow-

Acknowledgement We thank Tanja Woopen and Melanie Esser for expert technical assistance. Sources of funding This work was supported in part by the IZKF (project IZKF-T11-5 of AL and DD and project T11-3 of HJ) and the ERS boost fund (project OPBF071 of AL and US) of the RWTH Aachen University and by the Deutsche Forschungsgemeinschaft (project LU869/7-1 of AL and project SCHN 587/14-1 of US). Disclosures The authors have no conflicts of interest to declare. 60

Journal of Molecular and Cellular Cardiology 134 (2019) 51–61

A. Babendreyer, et al.

[12] [13]

[14] [15] [16] [17] [18] [19] [20] [21] [22]

[23] [24] [25]

[26] [27] [28]

[29] [30]

mediates inflammation, FASEB J. 22 (2007) 641–653. [31] H. Langer, A.E. May, A. Bültmann, M. Gawaz, ADAM 15 is an adhesion receptor for platelet GPIIb-IIIa and induces platelet activation, Thromb. Haemost. 94 (2005) 555–561. [32] R.J. Dekker, S. Van Soest, R.D. Fontijn, S. Salamanca, P.G. De Groot, E. VanBavel, et al., Prolonged fluid shear stress induces a distinct set of endothelial cell genes, most specifically lung Krüppel-like factor (KLF2), Blood. 100 (2002) 1689–1698. [33] J.O. Fledderus, J.V. Van Thienen, R.A. Boon, R.J. Dekker, J. Rohlena, O.L. Volger, et al., Prolonged shear stress and KLF2 suppress constitutive proinflammatory transcription through inhibition of ATF2, Blood. 109 (2007) 4249–4257. [34] R.A. Boon, J.G. Horrevoets, Key transcriptional regulators of the vasoprotective effects of shear stress Shear stress, Mol. Cell (2009) 39–43. [35] J.V. van Thienen, J.O. Fledderus, R.J. Dekker, J. Rohlena, G.A. van IJzendoorn, N.A. Kootstra, et al., Shear stress sustains atheroprotective endothelial KLF2 expression more potently than statins through mRNA stabilization, Cardiovasc. Res. 72 (2006) 231–240. [36] B.B. Böhm, I. Freund, K. Krause, R.W. Kinne, H. Burkhardt, ADAM15 adds to apoptosis resistance of synovial fibroblasts by modulating focal adhesion kinase signaling, Arthritis Rheum. 65 (2013) 2826–2834. [37] D. Dreymueller, J. Pruessmeyer, J. Schumacher, S. Fellendorf, F.M. Hess, A. Seifert, et al., The metalloproteinase ADAM8 promotes leukocyte recruitment in vitro and in acute lung inflammation, Am. J. Phys. Lung Cell. Mol. Phys. 313 (2017) L602–L614. [38] B. Böhm, S. Hess, K. Krause, A. Schirner, W. Ewald, T. Aigner, et al., ADAM15 exerts an antiapoptotic effect on osteoarthritic chondrocytes via up-regulation of the Xlinked inhibitor of apoptosis, Arthritis Rheum. 62 (2010) 1372–1382. [39] B.B. Böhm, T. Aigner, B. Roy, T.A. Brodie, C.P. Blobel, H. Burkhardt, Homeostatic effects of the metalloproteinase disintegrin ADAM15 in degenerative cartilage remodeling, Arthritis Rheum. 52 (2005) 1100–1109. [40] H. Taniguchi, F.V. Jacinto, A. Villanueva, A.F. Fernandez, H. Yamamoto, F.J. Carmona, et al., Silencing of Kruppel-like factor 2 by the histone methyltransferase EZH2 in human cancer, Oncogene. 31 (2012) 1988–1994. [41] B.G. Coon, N. Baeyens, J. Han, M. Budatha, T.D. Ross, J.S. Fang, et al., Intramembrane binding of VE-cadherin to VEGFR2 and VEGFR3 assembles the endothelial mechanosensory complex, J. Cell Biol. 208 (2015) 975–986. [42] D.E. Conway, M.T. Breckenridge, E. Hinde, E. Gratton, C.S. Chen, M.A. Schwartz, Fluid shear stress on endothelial cells modulates mechanical tension across VEcadherin and PECAM-1, Curr. Biol. 23 (2013) 1024–1030. [43] C. Ham, ADAM15 is an Adherens junction molecule whose surface expression can be driven by VE-cadherin, Exp. Cell Res. 279 (2002) 239–247. [44] D. Fried, B.B. Böhm, K. Krause, H. Burkhardt, ADAM15 protein amplifies focal adhesion kinase phosphorylation under genotoxic stress conditions, J. Biol. Chem. 287 (2012) 21214–21223. [45] T. Bellido, Osteocyte apoptosis induces bone resorption and impairs the skeletal response to weightlessness, BoneKEy-Osteovision. 4 (2007) 252–256. [46] S. Hart, O.M. Fischer, N. Prenzel, E. Zwick-Wallasch, M. Schneider, L. Hennighausen, et al., GPCR-induced migration of breast carcinoma cells depends on both EGFR signal transactivation and EGFR-independent pathways, Biol. Chem. 386 (2005) 845–855. [47] B. Schäfer, B. Marg, A. Gschwind, A. Ullrich, Distinct ADAM metalloproteinases regulate G protein-coupled receptor-induced cell proliferation and survival, J. Biol. Chem. 279 (2004) 47929–47938. [48] V. Trochon-Joseph, D. Martel-Renoir, L.M. Mir, A. Thomaïdis, P. Opolon, E. Connault, et al., Evidence of antiangiogenic and antimetastatic activities of the recombinant disintegrin domain of metargidin, Cancer Res. 64 (2004) 2062–2069. [49] H.D. Lee, Y.H. Kim, B.H. Koo, D.S. Kim, The ADAM15 ectodomain is shed from secretory exosomes, BMB Rep. 48 (2015) 277–282. [50] H.D. Lee, B.-H. Koo, Y.H. Kim, O.-H. Jeon, D.-S. Kim, Exosome release of ADAM15 and the functional implications of human macrophage-derived ADAM15 exosomes, FASEB J. 26 (2012) 3084–3095.

induced HAEC VCAM-1 expression and monocyte adhesion, Am. J. Phys. 276 (1999) C1100–C1107. C. Cheng, D. Tempel, R. Van Haperen, H.C. De Boer, D. Segers, M. Huisman, et al., Shear stress – induced changes in atherosclerotic plaque composition are modulated by chemokines, J. Clin. Invest. 117 (2007) 616–626. J.S. DeVerse, A.S. Sandhu, N. Mendoza, C.M. Edwards, C. Sun, S.I. Simon, et al., Shear stress modulates VCAM-1 expression in response to TNF-α and dietary lipids via interferon regulatory factor-1 in cultured endothelium, Am. J. Physiol. Heart Circ. Physiol. 305 (2013) H1149–H1157. M.E. Davis, H. Cai, G.R. Drummond, D.G. Harrison, Shear stress regulates endothelial nitric oxide synthase expression through c-Src by divergent signaling pathways, Circ. Res. 89 (2001) 1073–1080. R.J. Dekker, J.V. van Thienen, J. Rohlena, S.C. de Jager, Y.W. Elderkamp, J. Seppen, et al., Endothelial KLF2 links local arterial shear stress levels to the expression of vascular tone-regulating genes, Am. J. Pathol. 167 (2005) 609–618. S. SenBanerjee, Z. Lin, G.B. Atkins, D.M. Greif, R.M. Rao, A. Kumar, et al., KLF2 is a novel transcriptional regulator of endothelial proinflammatory activation, J. Exp. Med. 199 (2004) 1305–1315. D. Dreymueller, J. Pruessmeyer, E. Groth, A. Ludwig, The role of ADAM-mediated shedding in vascular biology, Eur. J. Cell Biol. 91 (2012) 472–485. B. Herren, E.W. Raines, R. Ross, Expression of a disintegrin-like protein in cultured human vascular cells and in vivo, FASEB J. 11 (1997) 173–180. K. Horiuchi, G. Weskamp, L. Lum, H.-P. Hammes, H. Cai, T.A. Brodie, et al., Potential role for ADAM15 in pathological neovascularization in mice, Mol. Cell. Biol. 23 (2003) 5614–5624. C. Sun, M.H. Wu, M. Guo, M.L. Day, E.S. Lee, S.Y. Yuan, ADAM15 regulates endothelial permeability and neutrophil migration via Src/ERK1/2 signalling, Cardiovasc. Res. 87 (2010) 348–355. C. Sun, M.H. Wu, E.S. Lee, S.Y. Yuan, ADAM15 contributes to atherosclerosis by mediating endothelial barrier dysfunction via Src family kinase activity, Arterioscler. Thromb. Vasc. Biol. 32 (2012) 2444–2451. C. Sun, R.S. Beard, D.L. McLean, R.R. Rigor, T. Konia, M.H. Wu, et al., ADAM15 deficiency attenuates pulmonary hyperpermeability and acute lung injury in lipopolysaccharide-treated mice, Am. J. Phys. Lung Cell. Mol. Phys. 304 (2013) L135–L142. N. Schwarz, J. Pruessmeyer, F.M. Hess, D. Dreymueller, E. Pantaler, A. Koelsch, et al., Requirements for leukocyte transmigration via the transmembrane chemokine CX3CL1, Cell. Mol. Life Sci. 67 (2010) 4233–4248. K.M. Parmar, V. Nambudiri, G. Dai, H.B. Larman, M.A. Gimbrone, G. GarcíaCardeña, Statins exert endothelial atheroprotective effects via the KLF2 transcription factor, J. Biol. Chem. 280 (2005) 26714–26719. T. Pasqualon, J. Pruessmeyer, S. Weidenfeld, A. Babendreyer, E. Groth, J. Schumacher, et al., A transmembrane C-terminal fragment of syndecan-1 is generated by the metalloproteinase ADAM17 and promotes lung epithelial tumor cell migration and lung metastasis formation, Cell. Mol. Life Sci. 72 (2015) 3783–3801. T. Maretzky, G. Yang, O. Ouerfelli, C.M. Overall, S. Worpenberg, U. Hassiepen, et al., Characterization of the catalytic activity of the membrane-anchored metalloproteinase ADAM15 in cell-based assays, Biochem. J. 420 (2009) 105–113. A. Bültmann, Z. Li, S. Wagner, M. Gawaz, M. Ungerer, H. Langer, et al., Loss of protease activity of ADAM15 abolishes protective effects on plaque progression in atherosclerosis, Int. J. Cardiol. 152 (2011) 382–385. K. Komiya, H. Enomoto, I. Inoki, S. Okazaki, Y. Fujita, E. Ikeda, et al., Expression of ADAM15 in rheumatoid synovium: up-regulation by vascular endothelial growth factor and possible implications for angiogenesis, Arthritis Res. Ther. 7 (2005) R1158–R1173. J.-F. Mosnier, A. Jarry, C. Bou-Hanna, M.G. Denis, D. Merlin, C.L. Laboisse, ADAM15 upregulation and interaction with multiple binding partners in inflammatory bowel disease, Lab. Investig. 86 (2006) 1064–1073. L. Charrier-Hisamuddin, C.L. Laboisse, D. Merlin, ADAM-15: a metalloprotease that

61