Effects of interferons and their interactions with other ligands in human aortic valve cells

PROGRAMA DE DOCTORADO EN INVESTIGACIÓN BIOMÉDICA TESIS DOCTORAL: EFFECTS OF INTERFERONS AND THEIR INTERACTIONS WITH OTHER LIGANDS IN HUMAN AORTIC VALVE CELLS Presentada por Iván Parra Izquierdo para optar al grado de doctor por la Universidad de Valladolid Dirigida por: Dra. María del Carmen García Rodríguez Secretaría Administrativa. Escuela de Doctorado. Casa del Estudiante. C/ Real de Burgos s/n. 47011-Valladolid. ESPAÑA Tfno.: + 34 983 184343; + 34 983 423908; + 34 983 186471 - Fax 983 186397 - E-mail: negociado.escuela.d[email protected] Form 1T DEFENCE AUTHORIZATION BY THE PhD THESIS SUPERVISOR (As required by section 7.2. of the Regulation concerning doctoral thesis defense at UVa) TO THE CHAIRMAN OF THE PhD BOARD OF THE UNIVERSIDAD DE VALLADOLID Full name: Mª del Carmen García Rodríguez ID: 12323936F Department: Instituto de Biología y Genética Molecular University or Research Institution: Instituto de Biología y Genética Molecular Postal address: (IBGM, C/ Sanz y Forés 3, laboratorio E10, 47003 Valladolid E-mail: [email protected] I hereby authorize the defense of the PhD thesis entitled “EFFECTS OF INTERFERONS AND THEIR INTERACTIONS WITH OTHER LIGANDS IN HUMAN AORTIC VALVE CELLS” Written by Mr. Iván Parra Izquierdo under my supervision, within the PhD program in Investigación Biomédica de la Universidad de Valladolid. After taking into account the following considerations, the thesis includes novel contributions to the field and meets the requirements for the Degree of Doctor of Philosophy. Thesis supervisor, Place and date: Valladolid, June 28th, 2019 Signature: INDEX INDEX ABBREVIATIONS .............................................................................................................................. 1 ABSTRACT .......................................................................................................................................... 7 INTRODUCTION.............................................................................................................................. 11 I.1-Mammalian heart valves ........................................................................................................ 11 I.1.1-Heart valve physiology .................................................................................................... 13 I.1.2-Valvular heart diseases .................................................................................................... 15 I.2-Calcific aortic valve disease .................................................................................................... 19 I.2.1-CAVD risk factors ........................................................................................................... 19 I.2.2-CAVD diagnosis .............................................................................................................. 21 I.2.3-CAVD management and treatment ................................................................................. 22 I.3-Molecular mechanisms of CAVD ........................................................................................... 24 I.3.1-Initial triggers of CAVD: haemodynamic and genetic factors ........................................ 24 I.3.2-Initiation phase: endothelial dysfunction and inflammation ........................................... 26 I.3.3-Propagation phase: fibrosis, angiogenesis, and osteogenesis .......................................... 32 I.3.4-Last phase ........................................................................................................................ 35 I.4-Interferons and JAK/STAT pathways .................................................................................. 37 I.4.1-IFN types and their receptors .......................................................................................... 37 I.4.2-JAK/STAT pathways ...................................................................................................... 39 I.4.3-JAK/STAT pathways, IFN and cardiovascular disease .................................................. 43 I.5-Toll-like receptors ................................................................................................................... 45 I.5.1-TLR types and ligands ..................................................................................................... 46 I.5.2-TLR signalling ................................................................................................................. 48 I.5.3-TLR and cardiovascular disease ...................................................................................... 51 HYPOTHESIS AND OBJECTIVES ............................................................................................... 57 MATERIALS AND METHODS ...................................................................................................... 61 M.1-Human valve samples ........................................................................................................... 61 M.2-Human VIC and VEC isolation and culture ...................................................................... 62 M.2.1-Cell isolation ................................................................................................................. 62 M.2.2-Cell culture .................................................................................................................... 62 M.2.3-Human valve endothelial cell purification by cell sorting ............................................ 63 M.3-Valve cell characterization by immunofluorescence .......................................................... 63 M.3.1-α-SMA detection in VIC ............................................................................................... 63 M.3.2-CD31 and VWF detection in VEC ................................................................................ 64 M.4-Protocol for cell activation ................................................................................................... 65 M.5-Quantitative reverse transcription polymerase chain reaction (RT-qPCR) ................... 67 M.5.1-RNA purification ........................................................................................................... 67 M.5.2-Retrotranscriptase reaction ............................................................................................ 68 M.5.3-Quantitative PCR .......................................................................................................... 68 M.6-Protein analysis by immunoblot (Western Blot) ................................................................ 70 M.6.1-Protein extraction and quantitation by the bicinchoninic acid (BCA) assay................. 70 M.6.2-Western blot procedure: SDS-PAGE and transfer ........................................................ 71 M.6.3-Protein immunodetection and visualization by chemiluminescence ............................ 71 M.7-Enzyme-linked immunosorbent assay (ELISA) ................................................................. 73 M.8-Proliferation assay ................................................................................................................. 74 M.9-In vitro calcification assays ................................................................................................... 74 M.9.1-Staining of calcium-phosphate crystals with Alizarin Red dye .................................... 74 M.9.2-Calcium deposits quantification .................................................................................... 75 M.10-Apoptosis/necrosis assay by flow cytometry ..................................................................... 75 M.11-Ectopic phosphatase activity .............................................................................................. 76 M.12-HIF-1α detection by immunofluorescence ........................................................................ 76 M.13-Dynamic adhesion assays in side-specific VEC ................................................................ 76 M.14-Migration (wound healing) assay in side-specific VEC ................................................... 77 M.15-RNA interference assays ..................................................................................................... 78 M.16-Conditioned medium experiments ..................................................................................... 78 M.17-Statistical analysis ............................................................................................................... 79 RESULTS ........................................................................................................................................... 83 R.1-Common effects of type I and II IFN in human VIC .......................................................... 84 R.1.1-Human aortic valve tissue and explanted cells express IFN receptors .......................... 85 R.1.2-IFN activate several signalling pathways in control VIC .............................................. 85 R.1.3-IFN cooperate with LPS to induce a pro-inflammatory phenotype in VIC ................... 90 R.1.4-IFN drive VIC differentiation towards a pro-osteogenic phenotype ............................. 96 R.1.5-IFN and LPS induce VIC calcification to a higher extent in male cells ........................ 98 R.2-Specific mechanisms of IFN-α in VIC ................................................................................ 106 R.2.1-IFN-α and TLR cooperation is specific for TLR2-4 ligands ....................................... 107 R.2.2-IFN-α drives VIC differentiation towards an osteoblast-like phenotype ..................... 107 R.2.3-BMP-2 signalling plays a role on male preferential calcification ................................ 110 R.2.4-Female-specific Akt activation is a protective mechanism for calcification ............... 113 R.3-Specific mechanisms of IFN-γ in VIC ................................................................................ 115 R.3.1-IFN-γ and LPS interplay promotes HIF-1α induction in VIC ..................................... 116 R.3.2-IFN-γ treatment induces a pro-angiogenic phenotype in male VIC ............................ 122 R.3.3-Chemical stabilization of HIF-1α increases VIC calcification .................................... 123 R.3.4-Signalling pathways involved in sex-differences upon IFN-γ+LPS treatment ............ 124 R.3.5-Mechanistic differences of IFN in osteogenic differentiation...................................... 126 R.4-Correlation of IFN findings in intact valve tissue ............................................................. 129 R.4.1-Expression of IFN receptors, HIF-1α and related molecules in control and calcified valves from males .................................................................................................................. 130 R.4.2-Intrinsic sex-differences in calcified valve tissue ........................................................ 131 R.5-Type I IFN signalling mediates Poly(I:C) effects in VIC ................................................. 134 R.5.1-dsRNA treatment activates IFN signalling in VIC ...................................................... 135 R.5.2-JAK/STAT signalling blockade reduced Poly(I:C)-induced effects ............................ 138 R.6-Relevance of IFN-γ+LPS effects in the AV context and VEC side-specific effects ........ 143 R.6.1-Effects of secreted factors by VIC activated with IFN-γ+LPS in VEC ....................... 143 R.6.2-Effects of IFN-γ and TNF-α in side-specific VEC ...................................................... 146 DISCUSSION ................................................................................................................................... 157 D.1-IFN as pro-inflammatory and osteogenic cytokines in VIC ............................................. 157 D.1.1-IFN-α effects in VIC .................................................................................................... 157 D.1.2-IFN-γ effects in VIC .................................................................................................... 161 D.1.3-Mechanistic differences between type I and II IFN ..................................................... 163 D.2-Novel insights into TLR4/3 ligands in VIC ........................................................................ 164 D.2.1-Novel insights into LPS effects in VIC ....................................................................... 164 D.2.2-Type I IFN signalling mediates Poly(I:C) effects in VIC ............................................ 165 D.3-JAK/STAT and TLR interplay in VIC .............................................................................. 167 D.3.1-IFN-α and LPS interplay on inflammation and osteogenesis ...................................... 168 D.3.2-IFN-γ and LPS interplay on inflammation, angiogenesis and osteogenesis ................ 169 D.4-Sex-differences in VIC responses ....................................................................................... 171 D.5-VIC-VEC communication and its potential role for CAVD ............................................ 175 D.6-IFN-γ and TNF-α effects in VEC ........................................................................................ 176 D.7-JAK/STAT as potential therapeutic targets for CAVD ................................................... 179 ABSTRACT 7 ABSTRACT Introduction: Calcific aortic valve disease (CAVD) is the most common aetiology of acquired aortic valve disease. Initially considered a passive and degenerative process, the current view has redefined CAVD as an athero-inflammatory process in early stages that then progresses into a more complex condition. Supporting the role of inflammation-induced valve calcification are growing evidences on the induction of inflammation and subsequent pro-calcifying responses in valve interstitial (VIC) and endothelial cells (VEC) by immune mediators such as Toll-like receptor (TLR) ligands and tumor necrosis factor-α (TNF-α). Objectives: Prompted by evidences of a constitutive interferon (IFN) activity associated to ectopic calcification in a rare disease, the Singleton-Merten syndrome, and the infiltration of immune cells secreting IFN in diseased valves, we aimed to explore the role of these cytokines and Janus kinases (JAK)/Signal transducers and activators of transcription (STAT) pathways on CAVD pathogenesis. Materials and methods: Human aortic valve cells explanted from patients with no valve disease were used as a model for studying CAVD underlying processes. Cells were exposed to recombinant IFN-α or IFN-γ combined or not with TLR ligands (in VIC) or TNF-α (in VEC). Inflammation, calcification and angiogenic responses as well as cell proliferation and apoptosis were studied using different cellular and molecular biology techniques such as Western Blot, ELISA, qPCR, flow cytometry, immunofluorescence and calcification assays. In addition, total protein and RNA were extracted from both non-mineralized and calcified valves for gene and protein expression analysis. Results: IFN activate several signalling pathways and promote inflammatory responses in VIC, characterized by adhesion molecule expression and nuclear factor-κB activation. In addition, IFN trigger VIC differentiation towards a pro-osteogenic phenotype and promote calcific nodule formation in high-phosphate conditions. Strikingly, we found an IFN-Lipopolysaccharide (a TLR4 ligand) interplay that further potentiates the pro-inflammatory, pro-angiogenic and pro-osteogenic responses. Significant findings include the blockade of the responses JAK inhibitors currently used in clinics named Jakinibs. Additionally, this study provides several new insights about sexdifferences in CAVD, with unexpected greater responses to IFN and LPS in male cells and identified some of the underlying molecular mechanisms such as larger protein kinase B/Akt activation in cells 8 from females, and larger osteogenic signalling, hypoxia-inducible factor (HIF)-1α and extracellular signal-regulated kinases activation in male cells. We also found sex-differences in gene expression profile in calcified valve tissue that correlate with lower valve calcification in female patients. Furthermore, our work demonstrates a novel role for type I IFN signalling on TLR3-mediated effects in VIC. Finally, data unveiled IFN-γ as a pro-inflammatory cytokine in VEC as well as differences on monocyte adhesion to side-specific VEC monolayers in response to IFN-γ and TNF-α with potential relevance in CAVD pathogenesis. Conclusions: Our findings highlight that IFN act as pro-inflammatory and pro-osteogenic cytokines in valve cells and support the model of inflammation-induced calcification, the concept of additional pro-angiogenic mechanisms beyond valve thickening and hypoxia, and the notion of CAVD as a sex-divergent disease from the early inflammatory stages. Clinically relevant findings include the blockade of IFN-induced responses by Jakinibs, and by a HIF-1α inhibitor. JAK/STAT and HIF-1α pathways emerge as potential therapeutic targets for CAVD. INTRODUCTION 11 INTRODUCTION I.1-Mammalian heart valves In animals, the heart is composed of red muscle and a variable number of chambers. Mammalian hearts are composed of four different cavities, two atria and two ventricles (Figure I) that function as a pump to simultaneously deliver blood to the lungs and the rest of the body as follows: a) The deoxygenated blood from the different tissues is delivered to the right atrium through the superior and inferior vena cava. This blood is then transferred to the right ventricle, which pumps it to the lungs through the pulmonary artery. b) The oxygenated blood returns to the left atrium of the heart through the pulmonary veins. After reaching the left ventricle, it gets pumped through the aorta artery to the rest of the body. During this process, the oxygen is delivered to the different tissues and the unoxygenated blood finally returns to the right atrium, starting a new cycle. The human heart has different valves whose function is to regulate blood flow direction: the semilunar valves, aortic (AV) and pulmonary, and the atrioventricular valves, mitral and tricuspid. The pulmonary valve regulates the flow from the right ventricle to the pulmonary artery, whereas the AV regulates the flow from the left ventricle to the aorta. The mitral valve separates the left ventricle and atrium, whereas the tricuspid valve separates the right ventricle and atrium (Figure I). The cardiac valves open and close over 3 x 109 times in a lifespan of 70 years. The left-side valves, mitral and aortic, regulate the highest-pressure process, which reaches a transvalvular pressure of about 120 and 80 mmHg respectively (systemic circulation). In contrast, the transvalvular pressure on the right-side valves, tricuspid and pulmonary, is approximately of 25 and 10 mmHg, respectively (pulmonary circulation).1 To note, the AV, which is the main focus of this thesis, is exposed to an unique mechanical environment within the cardiovascular system, with the opposing sides of the valve experiencing markedly different hemodynamic shear stresses from the surrounding blood flow. 12 Figure I. Representation of the human heart anatomy. Taken from: http://nursingmedic.blogspot.com/2010/11/anatomy-of-heart.html. The cardiac cycle is characterized by a mixture of electric and mechanical events. In the most simplified model, the cardiac cycle can be divided into two phases named diastole and systole. During the former one, the atrioventricular valves are open and allow the blood to flow into the heart and to reach the ventricles, which are in a relaxed state. When the ventricles are full of blood, the atrioventricular valves close and the semilunar valves open. At the same time, the electrical stimulation triggers the contraction of the ventricles, which subsequently pumps the blood outside the heart through the pulmonary and aorta arteries. It is therefore crucial for the correct performance of the heart that the four valves preserve their physiology and function. However, some conditions such as infections, degeneration and congenital disorders can lead to valve malfunctioning, which in general affects the whole cardiac operation.1 13 I.1.1-Heart valve physiology I.1.1.1-Structure of the valves The physiology of the heart valves differs mainly in their structure. The atrioventricular valves are composed of two (mitral valve) or three leaflets (tricuspid valve) that have the same structure. These valves, but not the semilunar valves, also have a specialized supportive structure connecting the valve to the ventricles, which is composed of three chordae tendineae that insert into two papillary muscles.2 In contrast, the semilunar valves are normally composed of three cusps attached to a cylindrical structure named root that connect the ventricles to the major arteries. The importance of the AV is remarked by its connection to the right and left coronary arteries; in fact, the AV cusps are classified as the right, left, and non-coronary.3 The functionality of the heart valves is accomplished by a complex and specialized organization of cells and extracellular matrix layers. In general, the same composition and cell types can be found in the four heart valves. The outer surfaces of the valves are covered by a continuous endothelium and the valve interstice is arranged in three main layers named fibrosa, spongiosa and ventricularis, each of them showing specific structures and functions. The fibrosa layer is oriented to the major arteries or to the atrium depending on the valve. It is composed of a dense connective tissue containing collagen fibres oriented circumferentially that supports most of the haemodynamic challenges of the valves. The spongiosa is the middle layer and contains mainly glycosaminoglycans whose major function is to support and facilitate the movements of the valve cups. Finally, the ventricularis layer is oriented to the ventricle and composed of elastin fibres oriented radially, thereby allowing valve flexibility in each cycle. In humans, the AV is an avascular and usually tricuspid structure of ≤ 1 mm in thickness that is attached to the aorta via a fibrous annulus named aortic root. A schematic representation of the AV structure is shown in Figure II. I.1.1.2-Cellular components of the valves Under normal conditions, the resident valve cells can be grouped in valvular endothelial cells (VEC) and valvular interstitial cells (VIC) (Figure II). VEC: These cells populate the outer layers of the valve, the endothelium. VEC are specialized cells with homeostatic functions such as the regulation of nutrient transmission, extracellular matrix synthesis, inflammation and VIC phenotype, and also have anti-thrombotic 20 Diabetes mellitus: This epidemic disease exhibits a high rate of prevalence and strong correlation with the development of cardiovascular diseases. The association of diabetes with CAVD has been demonstrated by numerous reports, including the Multi-Ethnic Study of Atherosclerosis, which demonstrated an increase in AV calcium content in patients with diabetes mellitus.27 However, little is known about the effects of this disease on the progression of AS. Obesity and renal failure: The correlation of obesity with CAVD progression is still not clear, although some studies support obesity as a risk factor for the disease, indicating that an elevated body mass index (≥ 30 kg/m2) correlates with CAVD development.28 Regarding renal failure, chronic kidney disease has been shown to have a powerful impact on the presence and severity of AS.29 Lipids and oxidized phospholipid transporters: Dyslipidaemia is a central and very-well known contributor to the development of atherosclerosis, in which low density lipoprotein (LDL)- cholesterol levels play a major role. In fact, a meta-analysis of 27 different clinical trials revealed that lowering LDL-cholesterol levels with statins reduced up to 25% the risk for cardiovascular mortality and non-fatal myocardial infarction.30 LDL-cholesterol levels have also been associated with a greater prevalence of CAVD.14 The similarities between atherosclerosis and CAVD early stages, as well as the promising in vitro effects of statins in VIC, prompted to develop a series of randomized clinical trials using different statins for the treatment of AV sclerosis or AS, i.e. the Scottish Aortic Stenosis and Lipid Lowering Trial, Impact on Regression (SALTIRE),31 the Simvastatin and Ezetimibe in Aortic Stenosis (SEAS),32 and the Aortic Stenosis Progression Observation: Measuring Effects of Rosuvastatin (ASTRONOMER).33 Unfortunately, all these trials failed to show any relationship between statin treatment and a significant reduction in the progression of CAVD. A potential reason of this failure could be a late intervention when the disease has progressed and exhibit more complex mechanisms in which lipids may not play a key role. Lipoprotein A (Lp(a)), a transporter of oxidized phospholipids considered a well-studied risk factor for cardiovascular disease, has been pointed as an important predictor of AV disease.34 Increasing evidences indicate its potential role as a therapeutic target for the disease. A recent multimodality imaging analysis of a clinically representative cohort of patients with AS revealed the association of Lp(a) and oxidized phospholipids-apoB levels with increased AV calcification and faster disease progression, suggesting that lowering Lp(a) or inactivating oxidized phospholipids may slow AS progression.35 21 Bicuspid aortic valve: Bicuspid AV is the most common congenital cardiac abnormality and has been associated with mutations in the NOTCH1 gene.36 The different physiology of the valve strongly affects the haemodynamic regulation and the wall shear stress supported by the tissue, which promotes a more frequent and accelerated development of CAVD as compared to tricuspid AV.37 I.2.2-CAVD diagnosis The detection of cardiac calcifications is an emerging predictive value for future cardiovascular events. A recent review outlined the importance of the early detection by innovative techniques of both atherosclerosis and cardiac calcifications in order to improve the current prediction based on classical cardiovascular risk factors and the use of validated algorithms, i.e., the Framingham Risk Score, the Pooled Cohort Equations, and the European SCORE Risk Charts.38 Therefore, it is mandatory to develop more accurate techniques allowing the clinicians not only to detect the presence of cardiac calcifications, but also their severity. The conventional technique for the diagnosis of CAVD is two-dimensional echocardiogram, however the limitations of this technique to distinguish the degree of stenosis are prompting clinicians to develop novel and complementary techniques such as computed tomography calcium scoring, and positron emission tomography. Doppler-Echocardiography: This relatively inexpensive and harmless technique is based on the use of sound waves to produce video images of the heart (Figure IVA). The echocardiogram is a key diagnostic tool providing information about the condition of the AV and its causes and severity, and it can also be used to detect other valvular diseases, such as aortic regurgitation, mitral stenosis and mitral regurgitation. The main parameter indicating AS is the valve area, although since it shows some limitations, additional measurements such as flow rate, mean pressure gradient, ventricular function, size and wall thickness, degree of valve calcification, and blood pressure help to define more accurately the condition of the valve.16 Depending on the indicated parameters, four different grades of severity of AS can be defined:16 (i) High-gradient aortic stenosis: valve area ≤ 1 cm2 and mean gradient > 40 mmHg. (ii) Low-flow, low-gradient aortic stenosis with reduced ejection fraction: valve area ≤ 1 cm2, mean gradient < 40 mmHg, ejection fraction < 50%, and stroke volume index (SVi) ≤ 35 mL/m2. In this case, further investigation with low dose of dobutamin stress during echocardiography is necessary to truly distinguish AS from other conditions generating similar cardiovascular outcomes. (iii) Low- 22 flow, low-gradient aortic stenosis with preserved ejection fraction: valve area ≤ 1 cm2 and mean gradient < 40 mmHg, ejection fraction ≥ 50% and SVi ≤35 mL/m2. (iv) Normal-flow, low-gradient aortic stenosis with preserved ejection fraction: valve area ≤ 1 cm2 and mean gradient < 40 mmHg, ejection fraction ≥ 50% and SVi > 35 mL/m2. Computed tomography (CT) calcium Scoring: This technique, based on the use of X-rays, can provide a more detailed, reproducible, and accurate assessment of the calcification burden in the AV than echocardiography, demonstrating a strong association and diagnostic value for severe AS (Figure IVB), and it could be a future clinical routine technique.39 Positron emission tomography (PET): It is a non-invasive technique that allows to monitor the activity of specific biological processes within tissues, including the AV, by the use of different radiotracers binding to regions where a biological process is occurring. In the case of the AV, PET allows not only the detection of calcifications but also the evaluation of the extent of inflammation. The positron-emitting radiotracer 18F-fluoride binds to regions of newly developing microcalcification with a good predictive value as demonstrated in a recent study.40 Therefore, optimized 18F-fluoride PET-CT holds a promise as a powerful research technique that could improve our understanding of the disease and be used as a biomarker of calcification activity in clinical trials of novel therapies.40 A comparison of calcification imaging obtained with different techniques is shown in Figure IV. Figure IV. Comparison of different methods for imaging calcification in the same patient with AS. (A) 2-dimensional echocardiography. (B) CT calcium scoring. (C) 18F-fluoride PET-CT. Reproduced with permission from41. I.2.3-CAVD management and treatment The disease is usually detected by echocardiography performed for routine examinations or after the presence of systolic murmurs. Its management depends on a wide variety of factors, especially the severity of the disease and whether this is worsening in subsequent follow-up 23 appointments. Initial symptoms of AS usually include fatigue, easy tiring, loss of energy, swelling of the ankles, palpitations, shortness of breath, chest pain, and dizziness or loss of consciousness. If AV sclerosis is detected, the intervention is usually preventive, with changes in the lifestyle whose main objective is to reduce the influence of risk factors on the progression of the disease. If the disease progresses to moderate AS, the patient requires treatment by a cardiologist as well as a follow-up annual echocardiogram. These recommendations are sometimes accompanied by preventive treatments, such as angiotensin-converting enzyme inhibitors, β-blockers, and/or diuretics. However, no medical treatment has been approved or recommended for directly addressing CAVD,16 and when severe AS is detected, even if there are no symptoms, surgery is the only recommended option. Four main surgical options arise for CAVD patients:16 Aortic valve repair: It is a common procedure for the case of leaking bicuspid aortic valves consisting on reshaping the aortic valve cusps, thus allowing the valve to open and close more completely. It has some advantages as compared to valve replacement, such as a lower tendency towards clot formation, and the lower probability of infection. Balloon valvuloplasty: It is not a commonly used procedure. The technique consists in the introduction of a flexible and thin tube tipped with a deflated balloon that is directed to the valve, where it is inflated to stretch the narrowed valve. Then the balloon deflates and is removed. This method is used for either children or adults who cannot undergo open-heart surgery. Aortic valve replacement: This procedure consists in the replacement of the dysfunctional heart valve for either a biological or a synthetic valve in an open-heart surgery. Mechanical valves are commonly composed of carbon, metal, or plastic. Their main advantage is the improved durability compared to biological valves, however, they present higher chances of blood clot formation. Biological valves from animal tissue, mostly porcine, although sometimes human donors, show less durability but lower chances towards blot clot formation. A third option is called “The Ross Procedure”, which consists in the substitution of the damaged aortic valve for a pulmonary allograft. A valve taken from a cadaver is then used to replace the pulmonary valve of the patient. The Ross procedure is used in patients younger than 40 to 50 years to avoid the long-term use of anticoagulant medications after surgery.16 Transcatheter aortic valve replacement: Also called transcatheter aortic valve implantation or TAVI, it is a newly developed technique that has improved the surgery conditions and lowered the chances for an open-heart intervention. Transcatheter AV replacement is the equivalent process of an arterial 24 stent in the AV context and provides a functional replacement to the valve site through a catheter. The new valve is expanded pushing the old cusps out of the way and manages the regulation of the blood flow.16 The election of the adequate procedure depends on the disease severity as well as the physical condition and age of the patient. Over the last years, the development of these less invasive techniques has improved the quality of life after the procedure. However, the risk of surgery in elderly patients and the problems that sometimes arise after valve surgery reinforce the necessity of the identification of novel drug targets to delay or reverse the disease without a surgical intervention. The understanding of the molecular mechanisms leading to valve leaflet thickening and calcification is a matter of urgency for the scientific community regarding the increasing prevalence of aortic valvular disease in developed and undeveloped countries. I.3-Molecular mechanisms of CAVD CAVD is currently considered an active and multifaceted pathobiological process. Its molecular mechanisms are not fully elucidated although several processes are known to be relevant to CAVD pathogenesis, i.e. endothelial dysfunction, inflammation, fibrosis, angiogenesis, and calcification.41 The causes involved in the initiation of AV sclerosis are still unknown, but emerging evidences point to endothelial damage and dysfunction as the first step of the disease.41 Once the endothelium is disrupted, it is essential for the maintenance of the valve performance that the homeostatic mechanisms repair the damage so the valve can return to its normal function. However, in some cases this process does not work properly, thus initiating the development of CAVD. The disease progression can be divided into three main stages.42 The first stage is an atherosclerosis-like process in which immune cell infiltration and lipid deposition play a key role on the activation of resident valve cells. During this phase, the initial microcalcifications are also developed within the valve cusps. The second and third phases are characterized by the development of macrocalcifications as well as other pathological processes such as neoangiogenesis. In this section we summarize the current knowledge on CAVD initiation and progression. I.3.1-Initial triggers of CAVD: haemodynamic and genetic factors The aetiology of the disease can be infective, degenerative, or only congenital. Rheumatic CAVD seems to be caused by an unresolved throat infection with Streptococcus that triggers the 25 development of rheumatic heart disease, which is a chronic condition resulting from acute rheumatic fever after the infection with the former pathogen. Both rheumatic fever and rheumatic heart disease may cause damage to the heart valves that can result in stenosis and regurgitation.43 For non-rheumatic tricuspid CAVD, the focus of the current study, the initial causes of the disease remain still unidentified, although a wide amount of evidences points to a role of the hostile haemodynamic environment to which the AV is exposed. The interplay between fluid haemodynamic and different challenges including present risk factors, infections, or increased serum levels of inflammatory mediators, is thought to play a key role on CAVD initiation. As noted earlier, the fluid dynamic context of the AV differs from the rest of the valves, and the mechanical environment has been related to the preferential disease progression as compared to another semilunar valve with physiologically similarities, the pulmonary valve.44 The high peripheral resistance developed during each cardiac cycle causes a substantial diastolic back pressure on the valve in the closed position, which affects the endothelial layer. This pressure creates an oscillatory shear stress that has been commonly linked to atheroprone regions,45 only in the aortic side of the valve (Figure V). Therefore, as mentioned earlier, aVEC are exposed to a different haemodynamic environment than vVEC and vascular endothelial cells,46 the latter exposed to a laminar blood flow with high and constant pressure in the straight regions (> 15 dynes/cm2), and a disturbed and much lower flow in curved regions (< 4 dynes/cm2).45 Hostile shear stress affects not only to the outside layer and aVEC, but also influences the behaviour of VIC from the fibrosa layer, which suffer from high mechanical stress especially during diastole (Figure V). The physiological relevance of the complex haemodynamic environment in the AV could trigger different outcomes not only by its own effects, but also by potentiating the responsiveness of VEC and VIC to different insults, i.e., inflammatory stimuli. The notion of a key role for the fluid dynamics in CAVD development is further supported by the higher rate of disease found in bicuspid aortic valves, which are exposed to an even more hostile environment than tricuspid valves, characterized by a strong eccentric jet coming out of the valve and the associated vortex formation. In bicuspid AV, the most common congenital cardiac abnormality, the presence of only two abnormal cusps drastically alters the vortex evolution both temporally and spatially,44 consequently most bicuspid AV patients suffer cardiac events and need surgery for valve replacement at earlier ages. 26 Figure V. Representation of the different wall shear stress underwent by the aortic and the ventricular side of the valve during a cardiac cycle. (A) Under systolic mechanical forces, the ventricularis VEC experience straight shear while the VIC feel bending forces. (B) Diastolic mechanical forces include compression of the fibrosa VEC (AVEC) and high stress tensile strain on the aortic VIC (AVIC). Reproduced with permission from44. The complex environment within the AV is thought to be accurately regulated, and the disruption of valve homeostasis by external factors or conditions is actively controlled to maintain valve function. However, when the endothelial damage is not properly repaired, homeostasis is disrupted, and the diseased state can progress to the first stage of CAVD. Three stages for CAVD pathogenesis have been proposed based on clinical data and innovative optical molecular imaging techniques; the initiation and propagation phases, which are asymptomatic stages, and a late-stage phase with clinical manifestations. A scheme describing the three phases of CAVD is shown in Figure VI. I.3.2-Initiation phase: endothelial dysfunction and inflammation The first stage of the disease seems to be initiated after an unresolved endothelial damage. During this phase, inflammation has been pointed as the driving mechanism, with a good amount of evidences suggesting that CAVD is an athero-inflammatory disease at least in its initial phases.21 At the later stages of this phase, the first microcalcification nodules are formed, as disclosed by innovative optical molecular imaging that allowed the detection of early calcifications.47,41 27 I.3.2.1-Endothelial dysfunction The pathologic environment created by disturbed blood flow can in turn activate resident valve cells (both VIC and VEC). The contribution of these cells to the progression of the first stages of the disease gained attention over the last decades by the demonstration of their active roles in the complex regulation of valve mineralization. The first report showing a role for the endothelium in CAVD, published in 2003, demonstrated the association between AV sclerosis and systemic endothelial dysfunction.48 In this line, an early event in CAVD is the production of reactive oxidative stress that co-localizes with calcified regions, as demonstrated in explanted human calcified AV.49 Additionally, the treatment with the endothelial nitric oxide synthetase (eNOS) inhibitor L-NAME, reduced the oxidative stress in calcified valves, pointing to the endothelium as the major source of oxidative stress within the valve.49 In addition, endothelial dysfunction and oxidative stress can be further promoted by pro-inflammatory cytokines, mainly tumor necrosis factor-α (TNF-α), which are secreted by infiltrated immune cells in the early stages of the disease.50 The relevance of the endothelium in CAVD initiation is also supported by the differential transcriptional profile of porcine aVEC and vVEC, the former showing lower expression levels of protective and anticalcifying genes such as osteoprotegerin (OPG),51 which correlates with the fibrosa layer-preferential calcification. Finally, pro-inflammatory cytokines52 and the altered extracellular matrix53 can trigger endothelial-to-mesenchymal transition (EndMT), a source of osteogenic cells within the valve. Collectively, these evidences stress out the importance of the endothelium in maintaining AV homeostasis, and demonstrate that the potential disruption of VEC homeostasis further contributes to CAVD initiation. I.3.2.2-Inflammation: Immune cell infiltration, lipid deposition, and NF-  B activation Inflammation is a well-known inductor of vascular calcification54 and also plays a role on CAVD initial stages (Figure VI). The inflammatory response is a demonstrated hallmark of CAVD, supported by increasing evidences in both clinical and cellular studies as described in the next paragraphs:55,56,57 Immune cell infiltration: After endothelial dysfunction occurs, the expression of adhesion molecules, i.e., intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), facilitates the infiltration of immune cells such as monocytes, mast cells and T lymphocytes.58 Infiltrated monocytes are then differentiated into macrophages, which secrete a wide range of inflammatory cytokines such as TNF-α and extracellular matrix remodelling molecules such as 28 MMP. Additionally, not only macrophages but also T lymphocytes secrete active cytokines that affect resident valve cells behaviour, i.e., interleukin (IL)-6 and IL-8. Strikingly, high serum levels of IL-6 have been linked to the development of cardiovascular events, and this cytokine is overexpressed in CAVD and is strong inducer of in vitro calcification of VIC.59 Furthermore, another important cytokine with a reported role in CAVD is IL-1β, which is known to induce a proinflammatory phenotype60 and the secretion of matrix remodelling molecules in VIC.61 Finally, the activation of innate immunity also leads to the expression of several factors involved in tissue remodelling such as MMP and transforming growth factor-β1 (TGF-β1), which is overexpressed in CAVD and promotes VIC calcification mainly via apoptosis-related mechanisms.62 In general, the signalling triggered by all these cytokines converges on the activation of the transcription factor nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), whose chronic activation promotes VIC calcification (reviewed in56). Collectively, these evidences point to VIC as the master and most important sensors of the inflammatory milieu during early CAVD and as the major promoters of calcification in the context of the AV. Figure VI. Inflammation as a key mechanism of CAVD. Molecular imaging detects 3 stages of calcification in CAVD: initiation, propagation, and late-stage calcification, in mice co-injected with nanoparticles to visualize macrophages (green) and near-infrared fluorescence imaging agent to detect calcification (red). Reproduced with permission from42. 29 Lipoprotein deposition and lipids metabolism: In addition to immune cell infiltration, the initiation of the inflammatory process in the AV might be driven by lipid accumulation within the valve cusps. Circulating Lp(a) and LDL are also deposited within the valve, the latter being oxidized during the process (oxLDL). Strikingly, oxLDL, inflammatory cell infiltrates and the expression of TNF-α colocalize in calcified valves.63 Other important mediator of the initial stages of CAVD is Lp(a), which carries a high content in lysophosphatidylcholine. The enzyme autotaxin, which is transported to the valve by Lp(a) and secreted by VIC, can transform lysophosphatidylcholine into lysophosphatidic acid that subsequently promotes inflammation and mineralization of the AV.64 Supporting this notion, recent reports revealed that Lp(a) and oxidized phospholipids promote valve calcification in patients with AS,35 and the treatment of VIC with Lp(a) strongly promotes cell mineralization.65 Another important pathway in the regulation of inflammation is the arachidonic acid route, the precursor of lipid mediators like leukotrienes and prostaglandins. The first enzyme of this route with a reported role in CAVD is 5-lipoxygenase, which is overexpressed in CAVD and triggers leukotriene synthesis and the subsequent inflammation of VIC.66 Regarding prostaglandins, the inducible enzyme cyclooxygenase-2 is expressed by VIC isolated from AS valves and it has been shown to mediate calcification in a mouse Klotho model of CAVD.67 Pathogen-derived molecules: Other subsets of immunostimulatory molecules can be derived from pathogens such as virus and bacteria. In fact, a significant pathogen cargo has been found together with immune cells in inflamed areas of calcified valves, as demonstrated in a study detecting Chlamydia pneumoniae, Helicobacter pylori, Cytomegalovirus, Epstein-Barr virus, and Herpes simplex virus in the fibrosa layer of stenotic valves.68 Moreover, bacteria associated with chronic periodontal infection have also been detected in the fibrosa layer of the AV,69 and inoculation of oral bacteria is reported to cause CAVD in a rabbit model of low-grade endocarditis.70 Moreover, pathogen-associated molecular patterns (PAMP) can be sensed by innate immune receptors, including Toll-like receptors (TLR). These receptors, known to mediate inflammation and to play a crucial role in the control of infection and the maintenance of tissue homeostasis, have been associated to CAVD pathogenesis,71 as explained at length below. Renin-angiotensin system activation: The pro-pathogenic milieu of the initial stages of CAVD is not restricted to cytokines and lipid mediators; in fact, angiotensin converting enzyme and its product angiotensin II are expressed and colocalize with LDL in calcified aortic valves.72 Moreover, angiotensin II type I receptor is expressed by fibroblasts only in valve lesions.72 Furthermore, data from a genetically hyperlipidaemic animal model associated angiotensin II to leaflet thickening and endothelial derangements, a phenomena linked to the early phase of AS. In ApoE-deficient and 36 initial atherosclerotic process has lost importance as the main mechanism in favour of a positive feedback loop of abnormal extracellular matrix and calcific nodules effects in VIC.42 With time, the degeneration of the valve can trigger additional cardiac problems such as left ventricular hypertrophy. When CAVD reaches its final stages, valve area is significantly reduced, and aortic regurgitation occurs, and when this is severe, surgical intervention is usually the selected therapy for patients. Altogether, current knowledge from clinical and experimental evidences points to CAVD as a complex and active disease, with a wide variety of pathologic processes implicated and the participation of several cytokines and cellular types (Figure VII). Figure VII. Simplified scheme of the three stages of the disease, indicating the different cellular types and processes taking part. ECM indicates extracellular matrix. Modified with permission from71. 37 I.4-Interferons and JAK/STAT pathways Cytokines such as interferons (IFN) and their signalling pathways have been associated to cardiovascular diseases such as atherogenesis.109 Here, we summarize the current knowledge on IFN function, signalling routes and their reported role on cardiovascular diseases. I.4.1-IFN types and their receptors In 1957, Alick Isaacs and Jean Lindenmann reported the phenomenon of “virus interference”, setting the basis for the discovery of the IFN, which are a group of soluble glycoproteins involved in a wide variety of cellular responses, especially the response to viral infections.110 IFN allow for communication between cells by triggering the protective defences of the immune system to eradicate pathogens or tumours. There are 3 different families of IFN, named type I, II and III IFN. Whereas type I IFN are predominantly expressed by innate immune cells and activated fibroblasts, the functionally similar type III IFN, (IFN-λ1-4), are more restricted and primarily act on epithelial surfaces.111 Finally, type II IFN (IFN-γ), is synthesized by natural killer and T cells and exerts antiviral functions mainly by the activation of macrophages.112 I.4.1.1-Type I IFN Members of type I IFN include IFN-α, β, κ, δ, ε, τ, ω, and ζ. These pleiotropic cytokines are associated not only to the response to viral, but also to bacterial infection. Most cell types secrete type I IFN in response to a viral challenge. All subtypes can signal through a receptor composed of the IFN-α/β receptor subunits 1 (IFNAR1) and 2 (IFNAR2).113 Type I IFN binding to IFNAR1/2 complex triggers the activation of Janus kinases (JAK)/Signal transducers and activators of transcription (STAT) signalling pathways and IFN-sensitive response elements that regulate the expression of several genes (Figure VIII). The genes activated by type I IFN regulate a large number of cellular functions such as: interferon regulatory factor 3 (IRF3) and IRF7 expression, dendritic cell activation, T cell survival, NK cell activation, chemokine expression, lymphatic node retention, as well as antiproliferative and antiviral effects.111,114 Remarkably, classical and important antiviral actions of type I IFN include the induction of the double-stranded RNA (dsRNA)-activated protein kinase R, which inhibits the cellular translational machinery, and 2′-5′-oligoadenylate synthetase/RNAseL, which degrades foreign RNA.115 38 I.4.1.2-Type II IFN The only member of type II IFN family is IFN-γ. This cytokine is sensed by a heterodimeric receptor composed of the IFN-γ receptor subunits 1 (IFNGR1) and 2 (IFNGR2) (Figure VIII). IFN- γ activates the so-called gamma-activated sequences (GAS), which have important roles in tissue homeostasis, immune and inflammatory responses, and tumour immunosurveillance.112 Specific immune functions of GAS genes are related to IRF1 expression, upregulation of MHC pathways, chemokine expression, Treg cell inhibition, Th1 cell differentiation and antiproliferative actions. IFN-γ also mediates the polarization of macrophages to an ‘M1-like’ state, which is characterized by increased pro-inflammatory activity and macrophage resistance to tolerogenic and anti-inflammatory factors.114 Moreover it should be noted that, in some cases, type I IFN induce a cross-signalling that activates the canonical IFN-γ-mediated formation of STAT1 dimers and the activation of GAS (Figure VIII).111 The importance of type I and II IFN relies not only on the effects triggered by their signalling, but also on their role as master cross-regulators of other immune pathways. In this regard, a substantial cooperation between the signalling pathways induced by type I IFN and the TNF-α pathway has been described.111 The role of IFN-γ as a master regulator of inflammatory pathways is widely known. Remarkably, the regulation of TLR responses by IFN-γ and STAT1 has been reported in different systems, thus reinforcing the role of this cytokine as a key connector of the innate and adaptive immunity responses.116 I.4.1.3-Type III IFN IFN-λ 1–4 proteins, also termed IL-28 and IL-29, are structurally more closely related to the IL-10 family.117 This subtype is poorly understood, although is known to share some functionalities with type I IFN by also activating JAK signalling pathways and inducing the expression of IFN- stimulated genes. Recent information demonstrates its importance in some types of virus or fungal infections by exhibiting antiviral effects at barrier surfaces, i.e. respiratory and gastrointestinal tracts, as well as the blood-brain barrier.118 I.4.1.4-Interferon regulatory factors The expression of IFN is closely related to the IRF transcription factors. In mammalians, nine different IRF family members have been described.119 Their structure comprises a conserved aminoterminal DNA binding domain with a helix-loop-helix shape and a motif containing five tryptophan 39 residues. The IRF-association domains 1 and 2 are located in the carboxyterminal region and mediate homodimeric and heterodimeric interactions with other IRF family members, transcription factors, and co-factors.119 IRF are important molecules mediating the activation of immune cells, and their functions include virus-mediated activation of IFN, and modulation of cell growth, differentiation, apoptosis, and immune system activity.120 Figure VIII. Type I and II IFN signalling. Schematic representation of canonical type I and II IFN signalling. ISRE indicates IFN-sensitive response elements; PIAS, protein inhibitor of activated STAT; SHP, short heterodimer partners; SOCS, suppressors of cytokine signalling. Taken with permission from114. I.4.2-JAK/STAT pathways The JAK/STAT pathways are pleiotropic cascades used to transduce a multitude of signals for development and homeostasis in animals, from human to flies. JAK/STAT are master regulatory routes involved in the sensing of more than 50 different cytokines and growth factors121 that coordinate the innate and adaptive immunity as well as the regulation of cell fate, including 40 proliferation, differentiation, migration, apoptosis, and cell survival. The JAK/STAT response strongly depends on the signal, tissue, and cellular context.122,123 Three major groups of proteins participate in JAK/STAT signalling upon ligand binding: (i) the receptors, which sense a large range of different cytokines and growth factors; (ii) the JAK tyrosine kinases, which initiate the cascade upon ligand-receptor binding; (iii) the STAT transcription factors, which dimerize and translocate to the nucleus to activate the transcription of target genes. The cytokine receptors, including IFN receptors, consist of two or more transmembrane polypeptide chains, and are associated with one or more JAK.115,124 Ligand binding causes the dimerization of the receptors that brings the associated JAK close in proximity, thus promoting its autoactivation through trans-phosphorylation of tyrosine residues. The activated JAK then phosphorylate signature tyrosine residues in the cytoplasmic region of the receptors, thus creating docking sites for the binding of the Src-homology2 (SH2) domains of STAT transcription factors. Upon recruitment to the receptor complex, STAT proteins are phosphorylated by JAK at tyrosine residues from the SH2 and transactivation domains, causing its separation from the receptor and the formation of STAT homo or heterodimers, which is followed by translocation to the nucleus, where they function as transcription factors by binding to conserved DNA recognition sites.115,124 I.4.2.1-Janus kinases or JAK The JAK were first named “just another kinases”. However, upon the discovery of their key implication on the sensing of IFN and other cytokines, they were called Janus kinases in honour to the two-faced roman god Janus, because they contain two similar phosphate-transferring domains. In mammalians, the JAK family comprises four members: JAK1, JAK2, JAK3, and the tyrosine kinase TYK2. They are typically located in the endosomes and the plasma membrane, along with their cognate receptors. These proteins are large size kinases (120-140 kDa) that contain a tyrosine kinase domain, which is essential for the enzymatic activity of JAK and contains conserved tyrosine residues necessary for JAK activation, and a catalytically inactive pseudo-kinase domain. Both regions bind the receptors through amino-terminal band-4.1, ezrin, radixin, moesin domains. JAK1, JAK2 and TYK2 are ubiquitously expressed, whereas the expression of JAK3 seems to be restricted to cells of the hematopoietic lineage and vascular smooth muscle cells (VSMC).125,123,124 JAK1 is involved in the response to many cytokines such as type I and II IFN and several IL (2, 4, 7, 9, 10, 13, 15, 20, 21, 22, 28). JAK2 mediates the response to hormone-like cytokines such as 41 the growth hormone, prolactin, erythropoietin and thrombopoietin, and to cytokines involved in hematopoietic cell development, such as IL-3 and granulocyte macrophage colony-stimulating factor and IFN-γ. JAK3 exclusively mediates signals through cytokines that use the common γ chain, which are IL-2, 4, 7, 9, 15, 21. TYK2 associates with cytokine receptors that signal through JAK1 or JAK2, but not with JAK3, and plays an essential role on the transduction of several cytokines such as type I IFN, IL-6, IL-10, and IL-12/IL-23. The function of TYK2 also seems to be associated with the cellular response to lipopolysaccharide (LPS), although it is not clear whether this is a direct or indirect mechanism.125,124 I.4.2.2-STAT The activation of JAK triggers signalling through STAT transcription factors. Mammalians have 7 types of STAT transcription factors: STAT1, 2, 3, 4, 5a, 5b, and 6. They are composed of 750 to 900 amino acids and each STAT consists of seven conserved functional domains: The N-terminal domain, a coiled-coil domain, a central DNA-binding domain, a linker region, a SH2 domain followed by a single conserved tyrosine residue, and a C-terminal transcriptional activation domain. As noted earlier, upon ligand binding, STAT are recruited to the receptor and activated by tyrosine phosphorylation by JAK, which in turn promotes their dissociation and the formation of STAT dimers. Then, STAT dimers migrate to the nucleus through the nuclear pore complex by a mechanism dependent on importin α-5, where they activate different subset of genes. STAT1, 3, 4, 5, and 6 form homodimer complexes and some of them form heterodimeric or heterotrimeric complexes, an aspect that will be detailed below. In addition, all the STAT proteins, except STAT2, can be phosphorylated on at least one serine residue by various serine kinases such as p38, c-Jun N terminal kinases (JNK), and phosphatidylinositol 3 kinase (PI3K), which further enhances the STAT- mediated gene transcription.124,122 STAT family members include: STAT1: The founding member of the STAT family is widely expressed, reaching high levels of expression in heart, thymus and spleen. It is involved in type I and II IFN signalling (Figure VIII). Upon type I IFN binding, the receptor forms a complex composed of STAT1, STAT2 and IRF9. Then, this complex undergoes nuclear translocation and subsequent binding to IFN-sensitive response elements in target genes. Upon IFN-γ binding to its receptor, STAT1 forms homodimers that translocate to the nucleus and bind GAS involved in the response to viral and bacterial infections (Figure VIII). In general, STAT1 signalling is pro-inflammatory and anti-proliferative, although some other context-dependent effects have been reported.124 In addition, the regulation of STAT1 responses is tightly linked to the regulation of STAT3. In fact, STAT3 is usually related with an 42 enhanced proliferation and anti-inflammatory activity. Thus, the balance between the activation of both STAT proteins could determine the extent of the inflammation as well as the cell fate. In addition to IFN, other cytokines such as IL (2, 3, 6, 10, 11, 12, 15, 17, 22) and growth factors such as fibroblast growth factor-1, granulocyte macrophage colony-stimulating factor or VEGF amongst others signal through STAT1.121 STAT2: It is expressed in most tissues. Initially described as a component of the complex formed upon type I IFN challenge, it is known to be critically involved in the regulation of type I IFN autocrine loop. It is also involved in the responses to IL-17 epithelial growth factors, angiotensin, and urokinase-type plasminogen activator.121 STAT3: It has been classically related to IL-6 signalling, promoting acute phase gene expression and being expressed in most tissues. In addition, it transduces signals from a wide range of cytokines, i.e, IFN, IL (6, 11, 10, 11, 19, 20, 21, 22, 24, 26, 27, 31), granulocyte colony-stimulating factor, and leptin, among others, as well as from several growth factors and oncogenes. In this regard, the JAK/STAT3 axis has been proposed as a potential therapeutic target for the treatment of solid tumours since it is constitutively activated in many cancers and promote cell proliferation.126 Moreover, it plays an essential role on Th17 differentiation in mice and humans. Strikingly, regarding the inflammatory response, STAT3 paradoxically promotes inflammation in some settings while inhibits it in others.122 STAT4: Its expression is restricted to myeloid cells, thymus and testis. Its main activator is IL-12, being therefore critical for the Th1 differentiation process as well as for the IL-12-dependent activation of NK cells. Other cytokines or growth factors that activate STAT4 are IFN, IL-13, IL-17, IL-23, and urokinase-type plasminogen activator.124,127 STAT5: It consists of two isoforms, STAT5a and STAT5b, which exhibit differential expression in muscle, brain, mammary gland, and secretary organs. Together with STAT3, they exhibit the highest degree of homology to invertebrate STAT and are functionally pleiotropic. They play a role in the biological response to IFN, IL (2, 3, 5, 7, 9, 15), and granulocyte macrophage colony-stimulating factor among others.121,127 They play an essential role in erythropoiesis and lymphopoiesis.124 STAT6: It plays a critical role for the IL-4/IL-13-dependent polarization of naïve CD4+ lymphocytes into Th2 effectors, as well as on the proliferation and maturation of B cells. Intriguingly, STAT6 also mediates IFN responses and its homodimers recognize GAS elements.124,127 43 I.4.2.3-Mechanisms of JAK/STAT regulation JAK/STAT signalling has a wide variety of positive and negative regulators that control and prevent inappropriate cytokine responses. Within minutes after ligand binding, the STAT are already activated and located in the nucleus, whereas after some hours, they are exported back to the cytoplasm. The most important negative feedback loops regulating STAT activation include the dephosphorylation by protein tyrosine phosphatases, the direct inhibition by protein inhibitors of activated STAT (PIAS) and suppressor of cytokine signalling (SOCS), and the nuclear export of STAT.124,111 Several phosphatases have been related to the attenuation of JAK/STAT signalling, such as short heterodimer partners 1 and 2 or protein tyrosine phosphatases1B. These phosphatases target either the receptor complex at the cell membrane or the phosphorylated STAT molecules in the nucleus. Only short heterodimer partners-2 and TC-protein tyrosine phosphatase have been implicated in nuclear STAT dephosphorylation, which is critical for STAT nuclear export. Regarding STAT translocation, several evidences point to a continuous balance between nuclear import and export processes even under inactivated conditions. The underlying mechanisms are still poorly understood, but translocation seems to be regulated by multiple nuclear export and nuclear localization sequence elements. The PIAS family of proteins (PIAS1, PIAS3, PIASx and PIASy) negatively regulate STAT signalling by mechanisms dependent either on their binding to STAT or by acting as E3 SUMO-protein ligases. The SOCS proteins antagonize STAT activation in a classically known feedback loop and play an essential role in the inhibition of both IFN-γ and IL-6 responses. Mechanisms of SOCS-induced STAT downregulation include the physical obstruction of STAT recruitment to the receptor complex, as well as the blockade of JAK kinase activity. In addition, SOCS proteins facilitate the ubiquitination of JAK, thus targeting them for proteasomal degradation. Additional mechanisms for the regulation of STAT include covalent modifications, such as serine phosphorylation, acetylation, and O-glycosylation.124,128 I.4.3-JAK/STAT pathways, IFN and cardiovascular disease JAK/STAT signalling plays a role on several disorders, including cardiovascular diseases. Human JAK mutations have been related to numerous diseases, including severe combined immune deficiency, certain leukemias, polycythaemia vera, and other myeloproliferative disorders. In this base, JAK have become attractive targets for the development of therapeutics for a variety of hematopoietic and immune system disorders, and there are several JAK inhibitors approved for the 44 treatment of autoimmune diseases. For example, tofacitinib is used for the treatment of rheumatoid arthritis,129 and the JAK1/2 inhibitor ruxolitinib for polycitemia vera130 and myelofibrosis.131 Dysregulation of JAK/STAT signalling is also associated with various cardiovascular diseases, exhibiting a complex and diverse signalling with differential effects depending on the context. Indeed, STAT3 has been shown as a protective transcription factor for compensatory hypertrophy by promoting a reduction in apoptosis.132 Besides, STAT1 activation in the heart has been associated with cell death and ischemic disease progression.133 Regarding plaque formation, IFN are thought to be pro-atherogenic cytokines, since they exacerbate the inflammatory response and are master regulators of VSMC fate and apoptosis.109 In addition, the JAK/STAT pathways can also regulate another inflammatory routes. In this line, IFN-α acts as an inflammatory amplifier by increasing TLR4 expression and cytokine production, which turns into plaque destabilization.134 In addition, STAT1 acts as the key regulator of the interplay between TLR4 ligands and IFN-γ, which potentiates a pro-atherogenic state in human atherosclerosis.135 In CAVD, not straightforward associations of IFN with the disease have been reported to date, although other cytokines like IL-6, which activates JAK/STAT among other pathways,136 have been reported to promote VIC mineralization.59 Interestingly, gene profiling analysis of severe calcified stenotic human AV identified the overexpression of IFN-related cytokines, i.e., monokine induced by IFN-γ, now renamed to CXCL9.137 Importantly, IFN-γ is secreted by T lymphocytes in calcified valves in an active form, and impairs the calcium resorption potential of osteoclasts.138 The most intriguing evidence associating enhanced type I IFN activity and ectopic calcification in non-skeletal tissues, i.e aorta and AV, derives from a rare autosomal dominant disease named Singleton-Merten syndrome (SMS), first reported in 1973.139 These patients showed abnormalities in dentition and hands, as well as severe calcification and intimal weakening of the aortic arch and valve.139 In some cases of SMS, the early onset of calcification in the aorta and AV can be observed in children or young adults.140 The causes of this rare condition are still not completely understood, but emerging evidences point to mutations on genes encoding to cytosolic nucleic acid sensors as the potential cause. Strikingly, this disease is currently considered a type I interferonopathy whose pathobiology links calcification with constitutive type I IFN activity.141,142 Particularly, mutations affecting two retinoic acid-inducible gene I (RIG-I)-like receptors family members143,142 named melanoma differentiation-associated gene 5144 and RIG-I140 have been identified as the plausible cause of the disease. DDX58 mutations cause the constitutive activation of 45 RIG-I, which in turn triggers type I interferon production.140 Collectively, these evidences directly link a constitutive type I IFN activity with ectopic calcification in non-skeletal tissues, including aorta and heart valves. In the cardiovascular system, as mentioned earlier, some effects of IFN-JAK/STAT on atherogenesis are potentiated by cooperation with TLR signalling pathways,134,135 which has been associated to inflammation and CAVD. A putative interplay of IFN and TLR signalling in the context of CAVD, which shares some pathobiological similarities with atherogenesis, has not been explored. I.5-Toll-like receptors The TLR are part of a family of germline-encoded innate immunity receptors called pattern recognition receptors (PRR). In addition to TLR, the PRR family also includes the RIG-I, retinoic acid-inducible gene-I like receptors or RLR, the Nod-like receptors or NLR, and C-type lectin receptors.145 The PRR are located in cellular membranes and cytosolic compartments of a wide range of cellular types within the human body. Furthermore, in addition to their role on the regulation of innate immunity, PRR are also important for the activation of the acquired immunity. These receptors are sensors not only of PAMP from a myriad of microorganisms, but also of damageassociated molecular patterns (DAMP) released upon tissue injury. First, PAMP are usually very conserved pathogen motifs with low mutation rates that are not present in mammalians.145 Second, DAMP are molecules derived from damaged tissue, blood vessels or necrotic cells, i.e., nucleic acids, which activate TLR receptors in the absence of pathogens, a process known as sterile inflammation.146 The role of TLR in the immune response was first described in 1996 when Hoffman and colleagues demonstrated a role for these receptors in the host response to pathogens, specifically upon challenge with Aspergillus fumigatus.147 One important hit regarding the TLR family was the discovery of TLR4 as the sensor of the Gram-negative bacteria pattern LPS.148 These evidences, among others, were a major step forward in the innate immunity field, in which it was previously thought that the pathogens were not recognized by specific receptors. Regarding their structure, these receptors are type I transmembrane glycoproteins containing an extracellular leucine-rich repeat domain, a transmembrane domain, and an intracellular Toll-IL-1 receptor (TIR) domain.146 The extracellular domains possess segments of 24-29 amino acids containing the conserved sequence LXXLXLXX, being L leucine and X any other amino acid. The intracellular domains are homologs 52 In the context of cardiovascular diseases, TLR have been described to be involved in the pathogenesis of atherosclerosis, acute coronary syndromes, stroke, viral myocarditis, sepsis, ischemia/reperfusion injury, and heart failure.159 Several lines of evidence implicate TLR signalling in the inflammatory process underlying atherosclerosis. For example, TLR 1, 2, 4 and 5 are expressed in atherosclerotic plaques by both resident vascular cells and leukocytes. Moreover, TLR4 is up-regulated in atheroprone areas of vascular vessels. Finally, TLR engagement triggers the expression of adhesion molecules, inflammatory cytokines, and matrix remodelling molecules in vascular endothelial and smooth muscle cells.159 The role of TLR in CAVD has been addressed by our group and others.71 Although there is not a straightforward association of increased TLR signalling with CAVD development in in vivo model, increasing evidences demonstrate that these pathways are up-regulated in calcified valves.160 In vitro evidences also support a role for TLR in the initiation of CAVD. In this regard, VIC, which are candidates to suffer calcification, express high levels of TLR2 and TLR4, and respond to its ligands by triggering inflammation and subsequent osteogenic responses (Figure X).71 Emerging work from our group and others stressed out the importance of these receptors as master sensors of a wide variety of molecules in the AV context,71 demonstrating that TLR activation leads to osteogenesis and calcification of VIC.71 In addition, in vivo administration of LPS in C57BL/6 mice induced AV lesions and thickening.161 As previously indicated, bacterial and viral cargo has been detected in calcified valves.68 Meng and colleagues first showed that the Gram-negative bacteria motif LPS promoted an inflammatory and osteogenic phenotype in VIC.162 A further study revealed that TLR2 and TLR4 were the most expressed TLR in VIC, and demonstrated a link between its increased expression/signalling and the induction of responses relevant to CAVD pathogenesis.163 Work from our group further disclosed that the nucleic acid sensor TLR3 also triggered inflammatory and osteogenic responses in VIC, suggesting a role of viral patterns in CAVD initiation.164 Extensive work has been performed since then to identify potential TLR ligands, PAMP and DAMP, in the context of the AV (Figure X). 53 Figure X. TLR ligands and its effects in VIC. TLR engagement by several PAMP and DAMP triggers inflammatory and osteogenic responses in VIC. BGN indicates, biglycan; HMGB1, high-mobility group box- 1; PGN, peptidoglycan; S1P, sphingosine 1-phosphate. Reproduced with permission from71. In addition to sensing PAMP, TLR activation can also occur in the AV via sensing different extracellular matrix proteins and molecules that are up-regulated in the disease. In this line, soluble biglycan, a small proteoglycan that is overexpressed in CAVD,165 activates TLR signalling in VIC promoting inflammation and osteogenesis.166 Matrilin-2 is another extracellular matrix accumulated in calcified human nodules that induces inflammation and calcification of VIC via TLR2 and TLR4 signaling.167 Other example of PAMP is high mobility group box 1, a nuclear regulatory protein that is secreted under certain conditions, and act as a pro-inflammatory cytokine. It has been detected overexpressed in tissue and plasma levels of CAVD patients,168 and it has also been demonstrated that, once secreted, high mobility group box 1 induces pro-osteogenic responses in VIC via activation of TLR4.169 Finally, other important molecules with marked implications in CAVD progression are oxLDL, which have been shown as modulators of the LPS-induced TLR responses by promoting a synergistic increase of osteogenic markers expression in human VIC.170 More importantly, a later study focused on the protective effects of IL-37 showed that TLR2 and TLR4 blockade with neutralizing antibodies prevented the oxLDL-induced activation of NF-κB and ERK in VIC.161 Additional research is needed to elucidate the role of TLR in vivo and to evaluate whether TLR signalling could be a potential therapeutic target for CAVD. HYPOTHESIS AND OBJECTIVES 57 HYPOTHESIS AND OBJECTIVES Based on recent findings linking type I IFN activity with the early onset of calcification in the atypical SMS,142 the presence of T lymphocytes in aortic valves,171 and the release of type IFN upon TLR activation of VIC,164 we hypothesized that IFN could play a role on CAVD pathogenesis via activation of inflammation and the subsequent osteogenesis in aortic valve cells. The specific objectives of the study were: 1-To identify the different signalling pathways activated by recombinant IFN-α and IFN-γ in human VIC and to check a potential cooperation with the TLR4 ligand LPS. 2-To determine whether recombinant IFN-α and IFN-γ, alone or combined with LPS, upregulate CAVD underlying processes such as inflammation, osteoblastic differentiation, calcific nodule formation, and angiogenesis, in human VIC. 3-To elucidate the molecular mechanisms regulating IFN responses, as well as to analyse potential differences between IFN-α and IFN-γ in VIC. 4-To confirm the in vitro findings of VIC with the gene and protein expression profile of valve tissue obtained from control and calcified human aortic valves. 5-To analyse whether IFN secretion accounts for the TLR3 ligand Poly(I:C)-induced effects in VIC. 6-To check potential valve-side effects of IFN-γ on inflammation and monocyte adhesion in VEC isolated from the aortic and the ventricular side of the valve. MATERIALS AND METHODS 61 MATERIALS AND METHODS The composition of buffers, the references and source of reagents, kits, and materials used in the study are listed on the supplementary Annex 1. M.1-Human valve samples Human tricuspid valve samples were obtained from patients who underwent surgery in the Hospital Clínico Universitario of Valladolid. Diagnosis protocols and surgical procedures for heart transplantation and valve replacement followed the current European guidelines. Research complies with good scientific practice and with the Helsinki Declaration. The study design had been approved by the local ethical committee of the Hospital Clínico Universitario of Valladolid (IRB protocol number PI 15-263). All the patients were previously informed of the procedure and signed the consent for valve donation. Control non-mineralized valves were from recipients of heart transplantation with valve disease excluded by echocardiography. Calcified valves included explanted valves from patients who underwent valve replacement after non-rheumatic CAVD diagnosis. The criteria for the inclusion of the calcified valves in the study were an aortic valve area of 0.7 ± 0.2 cm2, a haemodynamic peak gradient of 88 ± 22 mmHg, and a mean gradient of 56 ± 15 mmHg. Representative pictures of control and calcified valves are shown in Figure XI. Collected specimens were kept on ice until they were either frozen or processed as indicated below. The patient demographics are shown at the beginning of each part in the Results section. Figure XI. Photographs of a control (left) and a calcified (right) aortic valve included in the study. 68 by mechanical disintegration with an Omni tissue homogenizer (Omni International). For cells, total RNA was extracted using 1 mL of reagent per approximately 300,000 cells. Then, chloroform extraction was performed using 200 μL of chloroform per mL of reagent. After mixing thoroughly, samples were centrifuged at 12,000 rpm for 12 min. The aqueous phase was extracted by gently pipetting and mixed with 500 μL of isopropanol, followed by centrifugation at 12,000 rpm at 4 ºC for 10 min. The supernatant was discarded, and the pellet was washed twice with 75% ethanol and then air-dried. RNA was resuspended into 20 μL of diethyl pyrocarbonate (DEPC)-treated water, previously prepared by treating water with 0.1% DEPC for 2 h at 37 ºC and subsequently autoclaved to remove the DEPC. Total RNA was quantified by spectrophotometry (Nanodrop 1000, Thermo Scientific, Waltham, MA, USA). M.5.2-Retrotranscriptase reaction 1.5 μg of RNA were used to synthesize first-strand complementary DNA by the reverse transcription reaction. First, the RNA was mixed with 300 ng of random primers, 1μL of 10 nM triphosphate deoxyribonucleotides mix (dNTPs) and sterile distilled water up to 12 μL. The mixture was heated for 5 min at 65 ºC followed by a quick chill on ice to allow primer annealing. Then, 4 μL of 5X First-Strand Buffer, 1 μL of ribonucleases inhibitor (RNasin) and 2 μL of 2 mM dithiothreitol (DTT) were added. The mixture was gently homogenized and incubated at 37 ºC for 2 min. Finally, 1 μL of 200 U/µL Moloney Murine Leukemia Virus retro-transcriptase (M-MLV) was added to a final volume of 20µL and incubated as follows: 10 min at room temperature; 50 min at 37 ºC; 15 min at 70 ºC. Generated cDNA was stored at -20 ºC until use. M.5.3-Quantitative PCR cDNA was amplified by real-time PCR using a KAPA SYBR® FAST qPCR master mix (2X) kit. A total of 20 ng of cDNA generated in the previous step were added in a final volume of 20 μL. For each target gene, 0.4 μL of 10 μM of the corresponding primers (Sigma-Aldrich, St.Louis, MO) were used. Primers previously reported and showing a 100% of complementarity with the target gene in a bioinformatics tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) were used. All the primers used in the study are listed in Table 3. The qPCR reaction was carried out using a LightCycler480® (Roche Diagnostics, Rotkreuz, Switzerland) under the following conditions: (i) 5 min of denaturalization at 95 ºC; (ii) 45 cycles of 15 seconds at 95 ºC, 20 seconds at 60 ºC, and 5 seconds at 72 ºC; and (iii) a final cycle of 5 seconds at 95 ºC and 1 minute at 55 ºC, followed by 10 69 seconds at 4 ºC. Two technical replicates were performed for each condition. Basal transcript levels were referred to the housekeeping gene glyceraldehyde-3 phosphate dehydrogenase (GAPDH) value as (2-Ct, Ct = cycle threshold value). Relative transcript levels were calculated based on the 2-Ct method (relative to GAPDH and to basal conditions). Table 3: PCR primers used in the study. Gene name Forward primer sequence (5´-3´) Reverse primer sequence (5´-3´) Size (bp) ACAN ACTCTGGGTTTTCGTGACTCT ACACTCAGCGAGTTGTCATGG 81 AKT1 AGCGACGTGGCTATTGTGAAG GCCATCAATTCTTGAGGAGGAAGT 97 BCL-2 GGTGGGGTCATGTGTGTGG CGGTTCAGGTACTCAGTCATCC 89 BMP2 ACCCGCTGTCTTCTAGCGT CTCAGGACCTCGTCAGAGGG 140 CNMD CTGGATCACGAAGGAATCTGT ACCATGCCCAAGATACGGG 176 GAPDH TGCCAAATATGATGACATCAAGAA GGAGTGGGTGTCGCTGTTG 121 HIF1A AGTGTACCCTAACTAGCCGA GTGCAGTGCAATACCTTCC 70 IBSP CCCCACCTTTTGGGAAAACCA TCCCCGTTCTCACTTTCATAGAT 109 IFNAR1 TGCCATGCCAGAAGATAGTG TTAGGTGCTCAGGCTTCCAG 156 IFNβ CAACTTGCTTGGATTCCTACAAAG TATTCAAGCCTCCCATTCAATTG 81 IFNGR1 AAAGTCAGAAGAATTTGCTGTAT ACTGAAGGGTGAAATATGTC 108 IL6 CACCTCTTCAGAACGAATTG CTAGGTATACCTCAAACTCC 239 IL8 ATGACTTCCAAGCTGGCCGT TCCTTGGCAAAACTGCACCT 82 IRF1 TTCCCTCTTCCACTCGGAGT GATATCTGGCAGGGAGTTCA 378 IRF3 TCTGCCCTCAACCGCAAAGAAG TACTGCCTCCACCATTGGTGTC 151 IRF8 GCTGATCAAGGAGCCTTCTG ACCAGTCTGGAAGGAGCTGA 98 MSX2 TGCAGAGCGTGCAGAGTTC GGCAGCATAGGTTTTGCAGC 144 MGP CTTAGCGGTAGTAACTTTGTG AGAGCCTTCTCGGATCCTCTC 151 NOS3 CCAGCTAGCCAAAGTCACCAT GTCTCGGAGCCATACAGGATT 354 OSX TGCTTGAGGAGGAAGTTCAC AGGTCACTGCCCACAGAGTA 148 RUNX2 TGGTTACTGTCATGGCGGGTA TCTCAGATCGTTGAACCTTGCTA 101 SMURF1 TGTGAAAAACACATTGGACCCA ACGCTAATGGTTATCGAATCCG 81 SOST CCGGAGCTGGAGAACAACAAG GCACTGGCCGGAGCACACC 186 SOX9 AGCGAACGCACATCAAGAC CTGTAGGCGATCTGTTGGGG 85 TNAP CAAAGGCTTCTTCTTGCTGGT AAGGGCTTCTTGTCTGTGTC 258 VEGFA ATCTGCATGGTGATGTTGGA GGGCAGAATCATCACGAAGT 218 70 M.6-Protein analysis by immunoblot (Western Blot) This technique, based on the detection of target proteins with specific antibodies, consists on the separation of protein by size using electrophoresis, and their transfer to a membrane for the subsequent immunologic detection. In this study we followed the original sodium dodecyl sulphate (SDS)-PAGE method described by Laemmli in 1970,173 used in previous reports from our laboratory.164,174 Buffers composition is indicated in Annex 1. M.6.1-Protein extraction and quantitation by the bicinchoninic acid (BCA) assay Cells were grown to confluence in 6- or 12-well plates and treated as previously indicated. Before the lysis, supernatants were removed and stored at -80 ºC until use and cells were washed with cold PBS. Whole cell extracts: Cells were lysed with TNE buffer plus phosphatase inhibitors: (1 mM Na3VO4; 5 mM NaF) and protease inhibitors (10 μg/mL aprotinin, 10 μg/mL leupeptin; 1 mM phenylmetilsulfonyl fluoride (PMSF)). Typically, 75 μL of lysis buffer were used for 6-well plates and 50 μL for 12-well plates, and a cell scraper was used to facilitate cellular detachment and lysis. Cell extracts were centrifuged at 13,200 rpm for 12 min at 4 ºC and the supernatant containing solubilized proteins was collected to assay protein concentration. Cytoplasmic and nuclear extracts: Cells were lysed using the Active Motif Nuclear and cytoplasmic extract kit following manufacturer`s protocol. Protein extracts from valve leaflet tissue: Specimens were first powdered using a mortar and a pestle with liquid nitrogen, and then resuspended in RIPA lysis buffer supplemented with protease/phosphatase inhibitors as previously indicated in a proportion of 5 uL of buffer per mg of tissue. BCA protein assay: Total protein content of cell extracts was calculated using the BCA assay, which is based on the capacity of proteins to reduce Cu2+ to Cu+ in an alkaline environment. For the measurement, a Pierce BCA protein assay kit was used with some modifications of the manufacturer’s procedures for microplate applications. Briefly, 5 μL of each sample were incubated with 100 μL of the working BCA solution (mixing 50 parts of reagent A containing bicinchoninic acid in alkaline buffer and 1 part of reagent B containing cupric sulphate) for 30 min at 37 ºC. The absorbance at 570 nm was determined using a Versamax microplate reader (Molecular Devices, 71 Sunnyvale, CA). Protein concentration was determined by interpolation in a standard curve with known concentrations of BSA. M.6.2-Western blot procedure: SDS-PAGE and transfer For Western blot analysis, usually 25 μg of protein were mixed with Laemli buffer, then heated at 100 ºC for 5 min and stored at -20 ºC until use. Protein separation was carried out in denaturing conditions in polyacrylamide gels, typically 8%-10% acrylamide depending on the proteins of interest. Electrophoretic separation was carried out applying a constant power of 25 mA for each gel using a power source (Bio-Rad. Hercules, CA). A molecular weight protein ladder was included for use as a size standard. Proteins were then transferred to hybond polyvinylidene difluoride membranes that had been previously activated in methanol for 1 min followed by rewetting in distilled water for 2 min. A wet-transfer system (Bio- Rad, Hercules, CA) was used at 400 mA for 2 h. M.6.3-Protein immunodetection and visualization by chemiluminescence Non-specific binding was prevented by incubating the membrane with blocking buffer containing TBS-0.5% Tween buffer (TTBS) plus either 5% of non-fat milk or BSA, the latter for phospho-antibodies, for 1 h at room temperature on a rocker platform. Primary antibodies were prepared as indicated by the manufacturer, typically in 5% milk or BSA in TTBS supplemented with 0.02% sodium azide. The membranes were incubated overnight with the primary antibody and later washed 3 times for 10 min each in TTBS, followed by 1 h incubation at room temperature with the appropriate horseradish peroxidase-conjugated secondary antibody in 3% milk in TTBS. A compilation of primary and secondary antibodies used in the study is shown in Table 4. Antibodylabelled proteins were then detected by chemiluminescence reaction by using horseradish peroxidase blotting substrate ECL. Blots were exposed to autoradiography films and developed in a dark chamber. Signals were scanned with a densitometer GS-800 (Bio-Rad, Hercules, CA) and analysed by Quantity One software (Bio-Rad, Hercules, CA). Results were expressed as arbitrary units (a.u) normalized to the corresponding loading control (β-tubulin or actin) and to basal conditions. If necessary, a harsh stripping buffer solution containing β-mercaptoethanol and SDS was used to eliminate previous signals from overlapping proteins following a protocol available online (https://www.abcam.com/protocols/western-blot-membrane-stripping-for-restaining-protocol). Briefly, membranes were incubated in harsh stripping buffer for 45 min at 50 ºC under shaking 72 conditions, then rinsed in tap water, and later washed extensively in TTBS. Before reusing the blot, it was incubated with blocking buffer. Table 4: Primary and secondary antibodies used for Western Blot. References and vendors in Annex 1. Immunoblot human primary antibodies Target protein (human) Molecular weight (kDa) Working dilution Actin 42 1:1000 eNOS 140 1:1000 HIF-1α 120 / 110 1:1000 Histone-H3 17 1:500 ICAM-1 100 1:100 NF-κB-p65 65 1:500 pAkt (Ser473) 65 1:500 pp44/42 MAPK (Erk1/2) (Thr202/Tyr204) 44 / 42 1:1000 pNF-κB-p65 (Ser536) 65 1:1000 pp38 (Tyr182/Thr180) 38 1:1000 pSAPK/JNK (Thr183/Tyr185) 54 / 46 1:1000 pSTAT1 (Ser727) 95 1:1000 pSTAT1 (Tyr701) 95 1:1000 pSTAT3 (Tyr705) 90 1:1000 RUNX2 65 1:1000 STAT1 95 1:1000 STAT3 95 1:1000 β-tubulin 55 1:20000 VCAM-1 110 1:200 Immunoblot secondary antibodies (Horseradish peroxidase-conjugated) Target primary Working dilution Goat-anti-mouse IgG 1:3000 Goat anti-rabbit IgG 1:2000 73 M.7-Enzyme-linked immunosorbent assay (ELISA) One application of this antibody-based technique is the detection and quantification of secreted proteins. The antigen of interest is usually immobilized on a solid surface and then complexed with an antibody conjugated to an enzyme, typically HRP. In the assay, detection is accomplished by assessing the conjugated enzyme activity by oxidation of chromogenic reagents by a peroxidase like 3,3’,5,5’-tetramethylbenzidine producing a detectable signal, most commonly a colour change. Quantification of the target protein is done based on a standard curve made of known concentrations of the protein of interest. Protein secretion was evaluated in the supernatants of cells activated with the corresponding stimuli for 24-48 h, typically 1 mL/well (6-well plate) of cell activation medium. Collected supernatants were analysed by ELISA following manufacturer’s protocol. A Versamax microplate reader (Molecular Devices, Sunnyvale, CA) was used to measure absorbance at 450 nm. Sample concentration was calculated by extrapolation from the standard curve, and values were normalized to total cell protein content. For prostaglandin E2 (PGE2) analysis, an online free software (https://www.myassays.com/arbor-assays-pge2-enzyme-immunoassay-kit-k051-h.assay). was used to perform the calculations. All the commercial ELISA kits are listed in Table 5. Table 5: Commercial ELISA kits used in the study. Target protein Supernatant dilution BMP2 No dilution IFN-β Untreated and stimulated:1/10 IL-6 Untreated: 1/20 Stimulated: 1/100 IL-8 Untreated: No dilution Stimulated: 1/40 IP-10 Untreated: No dilution Stimulated: 1/40 MMP-1 No dilution PGE2 No dilution VEGF-A Untreated: 1/2 Stimulated: 1/5 74 M.8-Proliferation assay Cell proliferation was assayed by an indirect method based on the conversion of 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) into a purple and insoluble product named formazan by the activity of mitochondrial enzymes of living cells.175 For the assay, 2,500 cells were first seeded in 96-well plates. The following day, cells were stimulated in 200 μL of activation media for 12 days, changing medium and stimuli every 3 days. At day 12, cells were visually examined, washed with PBS, and incubated with 100 μL of phenol-red- free M199 medium supplemented with MTT at a final concentration of 1 mg/mL for 4 h at 37 ºC /5 % CO2. Formazan crystal extraction was performed by adding 100 μL of an SDS-HCl solution (10% SDS in 0.01 N HCl) followed by a 4 h incubation at 37 ºC. Results were evaluated by measuring the absorbance at 570 nm and 630 nm (background) in a Versamax Microplate reader (Molecular Devices, Sunnyvale, CA). Data were expressed as A570-A630 at day 12. M.9-In vitro calcification assays In vitro calcification assays were performed in high-phosphate and low serum conditions following previously reported protocols.176 VIC grown until reaching approximately 90% of confluence in 6- or 12-well plates, were incubated in calcification medium (M199 containing 1% FBS, 3 mmol/L Pi and 1% antibiotic-antimycotic solution) for a period ranging from 7 to 21 days, depending on the stimuli and the experiment design. Medium and treatments were replaced every 3 days. Evaluation of calcification was later performed by using two different and independent protocols and technical replicates, as indicated below. M.9.1-Staining of calcium-phosphate crystals with Alizarin Red dye An anthraquinone derivative, alizarin red (AR), was used for visualization and semiquantitation of calcification nodules in VIC. Cells, previously washed with PBS, were incubated with a dye solution (1.4% AR in water, pH 4.2 adjusted with NH4OH) for 2 min unless otherwise indicated. Then, cells were washed at least 3 times with PBS and visualized under a phase-contrast microscope (Nikon Eclipse TS100) connected to a digital camera. AR dye was then extracted following the acetic acid extraction protocol described by Gregory et al.177 Briefly, cells were incubated in 10% acetic acid for 2 h at room temperature and then harvested by scraping. Later, the solution was heated at 85 ºC for 15 min, followed by cooling on ice for 10 min. After centrifugation, 75 the aqueous phase was extracted and neutralized with 1M NH4OH. Absorbance at 450 nm was measured with a Versamax Microplate reader (Molecular Devices, Sunnyvale, CA). Data were expressed as a fold increase of absorbance referred to basal conditions. M.9.2-Calcium deposits quantification A quantitative colorimetric calcium QuantichromTM kit, based on the binding of a phenolsulphonephthalein dye to free calcium forming a very stable blue coloured complex, was used to quantify calcium deposits. Briefly, cells were first washed once with PBS and calcium deposits were then extracted with 0.6 N HCl for 24 h at 4 ºC. Supernatants were collected and 5 µL were used to quantitate calcium deposits following manufacturer´s protocol. The intensity of the colour at 612 nm, which is directly proportional to the calcium content of the sample, was evaluated. Cells were lysed in 0.1 N NaOH with 0.1% SDS and total protein concentration was determined using the BCA assay. Data were expressed as mg/dL of calcium and normalized to total protein content. It should be noted that calcium quantitation cannot be performed in AR-stained plates; therefore, cells in separate plates were processed in parallel and assayed for each method. M.10-Apoptosis/necrosis assay by flow cytometry Annexin V and propidium iodide staining followed by flow cytometry analysis was used to evaluate cell apoptosis and necrosis. The principle of annexin V assay is based on the capacity of this protein to bind with high affinity to phosphatidylserine residues exposed to the cell surface in the presence of calcium, a phenomenon occurring at the early stages of canonical apoptosis.178 Propidium iodide is a cell membrane-impermeant dye binding to double-stranded DNA that is generally excluded from viable cells and can be excited at 488 nm thus emitting at a maximal wavelength of 617 nm. For the assay, approximately 100,000 cells were grown to confluence and then treated with the indicated ligands for 7 days in calcification medium. Medium was replaced every 2 days. To avoid cell death caused by trypsinization, cells were detached using StemPro Accutase. Then, cells were spun down and resuspended in Annexin V binding buffer. Cell staining was performed following manufacturer`s protocol by incubating with FITC ApoScreen® annexin V for 15 min on ice followed by two washes and staining with ApoScreen® propidium iodide. Immediately, stained cells were analysed by a GalliosTM Flow Cytometer (Beckman Coulter Inc, US). Data analysis was 76 performed using Kaluza® Flow Analysis Software (Beckman Coulter Inc, US). The percentage of apoptotic cells was calculated based on the number of annexin-V positive cells (B1+B2 quadrants). M.11-Ectopic phosphatase activity Ectopic phosphatase activity assessment was based on the conversion of p-nitrophenyl phosphate to p-nitrophenol by extracellular phosphatases and the detection of the generated chromogenic product by measuring the absorbance at 405 nm.179 For the assay, cells cultured in 12-well plates were activated in M199 medium supplemented with 1% FBSi and 1% antibiotic/antimycotic solution for 21 days in an incubator at 37 ºC with 5% CO2 and humidified atmosphere. Medium and stimuli were replaced every 3 days. For the enzymatic assay, cells were incubated with 10 mmol/L of p-nitrophenyl phosphate in phenol-red-free M199 medium for 12 h at 37 ºC. Then, absorbance at 405 nm was measured and phosphatase activity was calculated using the Beer-Lambert law and a specific molar extinction coefficient of 18.75 mM-1 x cm-1 for p-nitrophenyl phosphate. Data were expressed as nmol/min and normalized to μg of total protein content determined by the BCA assay. M.12-HIF-1α detection by immunofluorescence Immunofluorescence staining was used to analyse the cellular location of HIF-1α protein and to confirm its nuclear translocation upon cell activation. A total of 10,000 VIC seeded on uncoated coverslips overnight were then activated with the corresponding stimuli for 24 h. The staining protocol and fluorescence quantitation was as the used for α-SMA detection (Section M.3.1). Primary antibody was rabbit anti-human HIF-1α antibody (1:100 in PBS-1% BSA); Secondary antibody was an Alexa Fluor 488 goat anti-rabbit IgG antibody (1:1000 in PBS-1% BSA). M.13-Dynamic adhesion assays in side-specific VEC Dynamic VEC-monocyte adhesion assays were based on the simulation of different flow patterns in the edge compared to the centre of a 6-well plate.172 This system simulated a flow pattern comparable to the one on the ventricular side of the valve (unidirectional and higher-magnitude shear stress, edge of the well) or the aortic side of the valve (multidirectional and lower-magnitude shear stress, centre of the well), although the flow magnitude was lower than the physiological in each side. A total of 5 independent pairs of aVEC and vVEC monolayers were treated with IFN-γ and/or 77 TNF-α, and then analysed for adhesion to the monocytic cell line THP-1, which were previously stained with calcein. For the assay, approximately 150,000 side-specific VEC were seeded in 6-well plates precoated with 1% gelatin. The next day, medium was replaced with 1.9 mL of EGM-2 growth medium and sheared on an orbital shaker set at the angular velocity of 150 rpm for 48 h. Stimuli were added to VEC monolayers, which were sheared for additional 24 h before performing the THP-1 adhesion assay. Before the assay, THP-1 cells (ATCC®; Middlesex, UK; Ref. TIB-202™) were cultured in RPMI-1640 medium supplemented with 1% L-glutamine and 10% FBSi. For the adhesion assay, THP-1 cells resuspended in pre-warmed RPMI medium at 106 cells/mL of density were stained with calcein-AM (1:1000 dilution) for 30 min at 37 ºC / 5% CO2. Then, THP-1 cells were washed 3 times with pre-warmed EGM-2 medium and added to VEC monolayers at a density of 106 cells in 1.5 mL of EGM-2 medium without supplements. Cells were then incubated for 1 h at 37 ºC without swirling. After 3 washes with pre-warmed medium, cells were fixed for 10 min with 4% p-formaldehyde. Non-adhered THP-1 cells were removed by washing twice with PBS. Fluorescence microscope images were taken using a Nikon DMx1200 camera (Nikon Instruments, Melville, NY) coupled on a Zeiss Axioscope Microscope (Carl Zeiss, Oberkochen, Germany). Images, 115X magnification, were captured using NIS elements software (Nikon Instruments, Melville, NY). Ten pictures were taken using a custom guide to discern the edge and the centre of the wells. In parallel, an additional 6-well plate with VEC monolayer in a was used to evaluate cell number by nuclei staining with DAPI for 10 min. Calcein stained THP-1 counting and VEC nuclei counting were performed using a custom-made macro in ImageJ (U.S., NIH). Data were expressed as total number of THP-1 cells adhered, total number of VEC, and normalized adhesion, THP-1/VEC ratio. M.14-Migration (wound healing) assay in side-specific VEC For migration assays, cells were seeded in 6-well plates for 36 h and then serum-starved overnight in EGM-2 supplemented with 2% FBSi. The following day a scratch was generated using a P200 pipette tip from the top to the bottom of each well. Cells were washed twice with pre-warmed medium to eliminate the dead ones and stimulated in 1.9 mL of starvation medium for 24 h. Next, cells were washed twice with pre-warmed medium, and microphotographs of the scratch area were 84 R.1-Common effects of type I and II IFN in human VIC The overall effects of IFN on the pathophysiology of CAVD were investigated in a cellular model of VIC isolated from control non-calcified aortic valves from males and females. In addition, calcified valve tissue from male patients was also used. Clinical features of both groups were compared as indicated in Methods. Characteristics and comorbidities were not significantly different between male and female control groups (Table 6). All patients were Caucasian, and all female donors were post-menopausal. Table 6: Clinical features of the patients whose valves were included in section 1 of the study. Control valves Calcified valves Patient characteristics Male n=16 Female n=7 p-value Male n=9 Age 61 ± 2 58 ± 3 0.48 77 ± 3 Hypertension 3 (19%) 0 0.53 3 (33%) Hypercholesterolemia 4 (25%) 1 (14%) 0.56 3 (33%) Diabetes mellitus 5 (31%) 3 (43%) 0.66 2 (22%) Smoking 4 (25%) 1 (14%) 0.56 3 (33%) Renal failure 2 (13%) 0 0.32 1 (11%) Aetiology of heart failure 7 idiopathic, 7 ischemic 1 alcoholic 1 sarcoidosis 4 idiopathic, 2 ischemic, 1 post-cardiotomy shock 0.31 - Aortic valve area (cm2) - - - 0.75 ± 0.11 Peak gradient (mmHg) - - - 68.75 ± 18.77 Mean gradient (mmHg) - - - 42.44 ± 14.88 Statins - - - 6 (66%) 85 R.1.1-Human aortic valve tissue and explanted cells express IFN receptors First, we analysed whether control human valve tissue and isolated VIC express IFN receptors. qPCR analysis revealed that non-calcified human aortic valve tissue expresses IFNAR1 and IFNGR1 genes, the former one showing higher transcripts levels (Figure 1A, left panel). However, the expression of both genes was similar in isolated VIC (Figure 1A, right panel). When comparing control and calcified valve tissue, similar transcript levels of IFNAR1 gene were observed (Figure 1B, left graph), while IFNGR1 expression was significantly higher in calcified valves (Figure 1B, right panel). Figure 1. Interferon receptors expression in valve tissue and VIC. Total RNA was extracted from valve tissue and VIC and analysed by qPCR for type I and type II IFN receptor subunit 1 mRNA expression, IFNAR1 and IFNGR1, respectively. (A) Receptor expression for control valve tissue (left panel) and isolated VIC (right panel); n≥6 for each group. (B) IFNAR1 (left panel) and IFNGR1 (right panel) gene expression in control and calcified valve tissue; n≥6 for each group. Grey dots, control male VIC or valve tissue; grey triangles, calcified male tissue. * indicates statistical significance; *p<0.05; **p<0.01. R.1.2-IFN activate several signalling pathways in control VIC Next, potential IFN-activated signalling pathways in VIC were explored. Given that inflammation plays a major role at the early stages of CAVD and that IFN are pro-inflammatory agents in some contexts, we examined the phosphorylation of the master transcription factor for inflammation, NF-κB. Data revealed that both recombinant type I (IFN-α) and type II (IFN-γ) IFN promoted NF-B-p65 phosphorylation at 24 h, being the effect enhanced upon co-stimulation with the TLR4 ligand LPS (Figure 2A-B). This effect was dose-dependent, reaching a maximum at a concentration of 100 ng/mL, a dose used throughout the study unless otherwise indicated. In addition, both IFN also promoted NF-κB-p65 phosphorylation at early times (0-60 min), although no significant potentiation in combination with LPS was observed (Figure 2C-D). 86 Figure 2. Interferons promote NF-κB-p65 phosphorylation with further potentiation by LPS. Control VIC from male patients were stimulated with the indicated ligands for either 24 h (A-B) or 0-60 min (C-D), and cell lysates analysed by Western Blot for NF-κB-p65 phosphorylation. Representative blots and densitometric analysis corresponding to the treatment with IFN-α (A, C) or IFN-γ (B, D); n≥4 for each treatment. Unless otherwise indicated, IFNα refers to 100 ng/mL IFN-α; IFNγ, 100 ng/mL IFN-γ; LPS, 1 μg/mL LPS. In the study, n indicates the number of VIC isolates from independent valve donors. * indicates statistical significance as compared with basal conditions or as indicated by the line below; *p<0.05; **p<0.01; ***p<0.001. Canonical IFN signalling pathways were then investigated first focusing on STAT1, which is known to play a major role on type I and II IFN signalling.112,111 Western Blot analysis showed that both IFN-α and IFN-γ promoted STAT1 phosphorylation at different sites at early time points (0-60 min) (Figure 3A-B). In addition, IFN-γ strongly induced STAT1 activation at 24 h, with further potentiation in combination with LPS (Figure 3C). 87 Figure 3. Interferons promote STAT1 phosphorylation with further potentiation by LPS. Control VIC from male patients were stimulated with the indicated ligands for 0-60 min (A-B) or 24h (C). Cell lysates were analysed by Western Blot for STAT1 phosphorylation using antibodies specific for the serine 727 and tyrosine 701 phosphorylation sites. Representative blots and corresponding densitometric analysis; n≥4. Doses, n, statistics, and symbols as in Figure 2. To further examine the activation of NF-κB-p65 and STAT1 transcription factors in stimulated VIC, we explored their nuclear translocation by analysing nuclear and cytoplasmic extracts. These experiments revealed that both transcription factors were present in the nucleus upon combined treatment of either IFN-α (Figure 4A) or IFN-γ (Figure 4B) combined with LPS. Remarkably, pre-treatment with ruxolitinib, a JAK inhibitor,180 abrogated the IFN+LPS-induced translocation of both transcription factors (Figure 4A-B). 88 Figure 4. Interferons combined with LPS promote STAT1 and NF-κB translocation. (A-B) Control VIC from males were stimulated with the indicated ligands and nuclear and cytoplasmic extracts were analysed by Western Blot. Representative blots after IFN-α+LPS (A) or IFN-γ+LPS (B) treatment; n=3. Doses and n as in Figure 2. Ruxo indicates 6 µM ruxolitinib; HisH3, histone-H3, used as nuclear loading control. Finally, we explored additional pathways related to IFN and TLR signalling. Western blot analysis showed that IFN-α also promoted the early activation of STAT3 (Figure 5), a transcription factor with intriguing context-dependent effects given its reported role as promotor of tumour growth,126 whereas it has cardioprotective effects.181 In addition, IFN-α was also able to activate protein kinase B (Akt), a survival-promoting kinase whose activation triggers extracellular matrix deposition and whose overexpression prevents calcification in the valve context.182,100 Moreover, results disclosed the early activation of the MAPK p38, ERK, and JNK (Figure 5), which have been associated to VIC calcification.183,166 Likewise, IFN-γ induced the early activation of STAT3, Akt and MAPK p38, ERK, and JNK (Figure 6). It is noteworthy that LPS did not activate IFN-related transcription factors such as STAT1 (Figure 3) and STAT3 (Figures 5-6). Collectively, data demonstrate the IFN-mediated activation of several intracellular signalling routes in VIC. 89 Figure 5. Additional signalling pathways activated by IFN-α and LPS in VIC. Control male VIC were stimulated with the indicated ligands and whole cell extracts analysed by Western Blot. (A) Representative blots for the indicated proteins; n=4. (B) Densitometric analysis corresponding to the phosphorylation of STAT3, Akt, ERK, JNK and p38. Doses, statistics, n, and symbols as in Figure 2. 90 Figure 6. Additional signalling pathways activated by IFN-γ and LPS in VIC. Control male VIC were stimulated with the indicated ligands and whole cell extracts were analysed by Western Blot. (A) Representative blots for the indicated proteins; n=4. (B) Densitometric analysis corresponding to the phosphorylation of STAT3, Akt, ERK, JNK and p38. Doses, statistics, n, and symbols as in Figure 2. R.1.3-IFN cooperate with LPS to induce a pro-inflammatory phenotype in VIC Next, we sought to determine the effects of the IFN-mediated activation of NF-B and STAT on downstream inflammatory molecules with a reported role on CAVD.56 Our group previously described that VIC secrete considerable amounts of IL-6 and IL-8 upon TLR4 activation with LPS.164 In the present study, ELISA analysis revealed that IFN-α and IFN-γ slightly induced IL-6 (Figure 7A) but not IL-8 secretion in VIC (Figure 7B). Conversely, both IFN potentiated the LPS- induced secretion of IL-6 and IL-8 (Figure 7A-B). Strikingly, statistical analysis revealed unexpected sex-differences in IL secretion in response to LPS alone or combined with IFN-α (Figure 7A-B, left panels). 91 Figure 7. Interferons enhance the LPS-induced secretion of IL-6 and IL-8 in VIC. Control male and female VIC were treated with the indicated ligands for 24 h and supernatants collected to assay cytokine secretion by ELISA. (A) IL-6 secretion upon treatment; n=6. (B) IL-8 secretion upon treatment; n=5. Data were normalized to total protein content in each sample. Grey dots and bars, control male AVIC; white dots and bars, control female VIC. IFNα refers to 100 ng/mL IFN-α; IFNγ, 100 ng/mL IFN-γ, LPS, 1 μg/mL LPS. * indicates statistical significance compared to the corresponding untreated conditions or as indicated by the line below. # indicates statistical significance compared to the same treatment of the other group. *,#P<0.05, **,##P<0.01, ***,###P<0.001. Two-way ANOVA statistics are shown on the box above each plot: I indicates interaction; T, treatment; S, sex. n indicates the number of independent pairs of male/female VIC isolates from different donors. Additionally, we evaluated adhesion molecule expression because the serum levels of ICAM- 1 and VCAM-1 are increased in patients with non-rheumatic valve disease.184 Immunoblot experiments revealed that VIC stimulation with IFN-α and IFN-γ triggered ICAM-1 and VCAM-1 protein expression at 24 h. Strikingly, co-treatment with LPS resulted in a greater induction of adhesion molecule expression (Figure 8A-C). Moreover, statistical analysis revealed that, in contrast to IL secretion, no significant differences on adhesion molecule expression were found when comparing cells from males and females (Figure 8B-C). 92 Figure 8. Interferons induce adhesion molecule expression with further potentiation by LPS in VIC. Control male and female VIC were treated with the indicated ligands for 24 h and whole cell lysates were analysed by Western Blot. (A) Representative blots of ICAM-1 and VCAM-1 protein levels in male cells. (B) Densitometric analysis of ICAM-1 expression for IFN-α (left panel, n=6) and IFN-γ (right panel, n=5) in cells from males and females. (C) Densitometric analysis of VCAM-1 expression; n=5. Each pair of male and female VIC were analysed in the same blot. Doses, symbols, n, and statistics were as in Figure 7. To further examine the role of IFN and the interplay with LPS on inflammation, we explored the secretion of inflammatory lipid mediators such as prostaglandin E2, which is induced upon TLR3/4 activation of human VIC.164 ELISA analysis confirmed previous data demonstrating LPS- mediated induction of PGE2 secretion and further disclosed that IFN did not promote PGE2 secretion 93 but strongly potentiated the LPS-induced effect (Figure 9). No statistical differences among male and female cells in PGE2 secretion were observed. Figure 9. Interferons enhance the LPS-induced secretion of PGE2. Control male and female VIC were treated with the indicated ligands for 24 h and supernatants collected to assay PGE2 secretion; n=6. Doses, symbols, n, and statistics were as in Figure 7. Additionally, because fibrosis is another major mechanism of CAVD,43 we explored whether IFN induce the secretion of matrix remodelling molecules such as MMP-1. The analysis of supernatants from activated VIC showed that neither IFN nor LPS alone induced MMP-1 secretion. Conversely, MMP-1 secretion was detected upon co-stimulation with both agents (Figure 10), suggesting an IFN-LPS interplay on matrix remodelling molecules. Moreover, the IFN-α and LPS interplay exhibited a sex-differential response, since the secretion of MMP-1 was only significant in male cells (Figure 10, left panel). In contrast, the combination of IFN-γ-LPS did not show significant sex-differences (Figure 10, right panel). 100 Figure 16. IFN-γ and LPS induce VIC calcification to a higher extent in cells from males. Control male and female VIC were treated with the indicated ligands for 14 days in calcification medium and analysed as in Figure 15. (A-B) Representative microphotographs (A) and staining level quantitation of ARS (B); n=5. Black line indicates 50 µm. (C) Calcium deposition quantification was normalized to total protein content; n=5. Doses, statistics, n and symbols were as in Figure 7. To investigate whether IFN-induced calcification was receptor-mediated and could be pharmacologically targeted, we used Jakinibs, which reduced the IFN+LPS-induced inflammation (Figures 11-12). VIC were pre-treated with tofacitinib and ruxolitinib before stimulation in calcifying medium. Notably, the JAK1/TYK2 inhibitor tofacitinib dramatically reduced calcification in cells treated with either IFN-α alone or combined with LPS (Figure 17A-C). 101 Figure 17. Tofacitinib dramatically decreases IFN-α-induced calcification. (A-C) Control male VIC were pre-treated or not with tofacitinib, stimulated for 21 days in calcifying medium and analysed as in Figure 15. (A-B) Representative microphotographs of ARS (A) and ARS quantitation (B). Black line indicates 50 µm. (C) Calcium deposition; n=4 VIC. Tofa indicates 6 µM tofacitinib. Doses, statistics, n, and symbols as in Figure 2. Likewise, the pre-treatment with the JAK1/2 inhibitor ruxolitinib strongly decreased both nodule formation and calcium deposition induced by IFN-γ alone and in combination with LPS (Figure 18A-C). Together, data revealed that IFN-mediated calcification of VIC can be pharmacologically blocked by Jakinibs currently used in therapy. 102 Figure 18. Ruxolitinib markedly decreases IFN-γ-induced calcification. (A-C) Control male VIC were pre-treated or not with ruxolitinib and stimulated for 14 days in calcifying medium and analysed as in Figure 15. (A-B) Representative microphotographs (A) and quantitation of ARS (B); n=4. Black line indicates 50 µm. (C) Calcium deposition quantification; n=4. Ruxo indicates 6 µM ruxolitinib. Doses, statistics, n, and symbols as in Figure 2. The next aim was to explore the mechanisms leading to IFN-mediated calcification under high-phosphate conditions. It is known that both osteogenic and dystrophic calcification play a role on phosphate-induced calcification of vascular cells.186,187 A potential role for dystrophic osteogenesis in VIC calcification was explored by assessing apoptosis and necrosis by flow cytometry. Data showed that not IFN-α neither LPS alone induced VIC apoptosis. However, both stimuli together significantly increased cell death. Strikingly, sex-differences were observed for IFN- α+LPS treatment, given that it increased apoptosis only in male VIC (Figure 19A-B). 103 Figure 19. IFN-α and LPS combination induces apoptosis in male VIC. Male and female VIC were treated for 7 days in calcifying medium and then analysed by flow cytometry for Annexin V-FITC and propidium iodide staining. (A) Representative flow cytometry plot. (B) Statistical analysis of apoptotic cells counts (B1+B2 quadrants); n=5. Doses, statistics, n, and symbols as in Figure 7. Conversely, under the same conditions, IFN-γ alone and together with LPS statistically increased apoptosis in male and female VIC, although a higher effect was found for IFN-γ in male cells (Figure 20A-B). Finally, Jakinibs significantly decreased IFN-induced apoptosis in male cells (Figure 21A-B). Collectively, data suggest a role for apoptosis in IFN-mediated calcification. 104 Figure 20. IFN-γ alone and together with LPS induces apoptosis in male and female VIC. Male and female VIC were cultured and treated for 7 days in calcifying medium and then analysed by flow cytometry for Annexin V-FITC and propidium iodide staining. (A) Representative flow cytometry plot. (B) Statistical analysis of apoptotic cells counts (B1+B2 quadrants); n=5. Doses, statistics, n, and symbols as in Figure 7. To note, calcification assays (Figures 15-16) and flow cytometry analysis (Figures 19-20) revealed similar responses in cells from both sexes in the absence of stimulus, thus suggesting a major role for IFN and LPS, but not for Pi, in the induction of the sex-differential responses. 105 Figure 21. Jakinibs reduce the IFN+LPS-mediated apoptosis of VIC. (A-B) Male VIC were pre-treated with Jakinibs before stimulation and apoptosis analysis was performed as in Figure 19. Representative flow cytometry plot and statistical analysis of apoptotic cells counts (B1+B2 quadrants) for IFN-α+LPS (A) treatment (n=4) and for IFN-γ+LPS treatment (n=4). Ruxo indicates 6 µM ruxolitinib, Tofa, 6µM tofacitinib. Doses, statistics, n, and symbols and as in Figure 2. Taken together, results from this section demonstrate that type I and II IFN promoted common effects on CAVD relevant pathologic processes in a cellular model of valve cells. The major findings include the induction of pro-inflammatory and pro-osteogenic responses on VIC that exhibited sex differences and can be abrogated by FDA-approved Jakinibs. Additionally, a marked positive interplay with the TLR4 ligand LPS was disclosed. Moreover, IFN alone and combined with LPS lead to an increased Pi-induced calcification that was significantly higher in cells obtained from males, thus correlating with clinical findings in calcified aortic valves.188 106 R.2-Specific mechanisms of IFN-α in VIC In this part of the study, the specific mechanisms by which IFN-α exerts its effects on VIC and the interplay with TLR ligands were investigated. The analysis was performed in VIC isolated from control valves from male and female patients. As shown in Table 7, patient characteristics and comorbidities were not significantly different between sex groups. To note, all patients were Caucasian, and female donors were post-menopausal. Table 7: Clinical features of the patients included in the section 2 of the study. Control valves Patient characteristics Male n=14 Female n=6 p-value Age 60 ± 7 57 ± 9 0.49 Hypertension 2 (14%) 1 (16%) 0.89 Hypercholesterolemia 4 (28%) 1 (16%) 0.57 Diabetes mellitus 4 (28%) 2 (33%) 0.8 Smoking 4 (28%) 0 0.14 Renal failure 2 (14%) 0 0.32 Aetiology of heart failure 7 idiopathic, 5 ischemic 1 alcoholic 1 sarcoidosis 4 idiopathic 2 ischemic 0.41 107 R.2.1-IFN-α and TLR cooperation is specific for TLR2-4 ligands Considering the cooperation between IFN-α and LPS on VIC inflammation (Figures 2, 7-9), we explored potential interplay with other TLR ligands. Interestingly, NF-κB was activated upon combination of IFN-α with several TLR agonists, namely Pam2CSK4 (a TLR2/6 ligand) and the dsRNA analogous Poly(I:C) (a TLR3 ligand) (Figure 22). Conversely, Flagellin (a TLR5 ligand) and Imiquimod (a TLR7 ligand) did not show significant cooperation with IFN-α (Figure 22), although Flagellin+IFN-α treatment increased NF-κB phosphorylation in almost all experiments. Together, data suggest that IFN-α-TLR cooperation is more prominent for TLR2-4 ligands in VIC. Figure 22. IFN-α cooperates with TLR2/6 and TLR3 ligands to phosphorylate NF-κB in VIC. Cells from control male valves were stimulated with the indicated TLR ligands for 24 h and whole cell lysates were analysed by Western Blot; n=5. Pam2 (TLR2/6) refers to 100 ng/mL of Pam2CSK4; Flag (TLR5), 1 µg/mL flagellin; IFNα, to 100 ng/mL IFN-α; Imiq (TLR7), 5 µg/mL Imiquimod; PolyIC (TLR3), 1 µg/mL polyinosinic:polycytidylic acid. Statistics, n, and symbols as in Figure 2. R.2.2-IFN-α drives VIC differentiation towards an osteoblast-like phenotype As shown in Figure 13, the exposure of VIC to IFN-α resulted on marked changes on cell morphology. To demonstrate the IFN-α-mediated cell differentiation, and because VIC resemble an activated myofibroblast phenotype in our culture conditions,189 we analysed the expression of the myofibroblast marker α-SMA. Immunofluorescence analysis revealed that the protein levels of α- SMA decreased after 48 h of treatment, with no apparent differences between male and female cells 108 (Figure 23A-B). Together, these findings demonstrate the IFN-α-mediated loss of myofibroblast phenotype, thus supporting the hypothesized VIC differentiation upon activation. Figure 23. IFN-α and LPS promote α-SMA downregulation in VIC. Control male and female VIC were treated with the indicated ligands for 48 h and α-SMA protein levels were analysed by immunofluorescence. (A) Images are representative of male VIC. Green, FITC; blue, DNA staining (DAPI); white line indicates 50 µm. (B) Total corrected cellular fluorescence (TCCF) was calculated as indicated in Methods; n=5. Doses, statistics, n, and symbols as in Figure 7. Because osteoblastic differentiation is a common finding in calcified human valves,190 we analysed the gene expression profile of different bone-related transcription factors and osteocyte markers. qPCR analysis revealed that the co-stimulation with IFN-α and LPS resulted in a malespecific upregulation of Sclerostin (SOST), an osteocyte marker increased in CAVD patients` serum191 (Figure 24A). In the attempt to better understand the VIC phenotype, we investigated the most important osteoblastic transcription factors reported to play a role in CAVD. Consistent with SOST induction, only in male VIC IFN-α+LPS treatment upregulated the expression of RUNX2/CBFA1 gene (Figure 24B), which is the master transcription factor for osteoblastic differentiation that has been associated to VIC differentiation.91 Then, we explored additional osteoblastic transcription factors and found that the mRNA levels of MSX2, a gene associated to valve calcification,192 were not increased by IFN-α but by LPS treatment in both male and female VIC at 24 h (Figure 24C). Otherwise, the expression of OSX gene, which has also been related associated to CAVD,89 was only upregulated in male cells specifically after IFN-α treatment (Figure 24D). Together, these results are consistent with a differentiation process accelerated by the combination of IFN-α with LPS and exhibiting sex-differences. 109 Figure 24. IFN-α and LPS interplay potentiates VIC differentiation. Control male and female VIC were treated with the indicated ligands for 24 h and mRNA extracted for qPCR experiments. (A) Relative SOST expression. (B-D) Relative expression of the osteoblastic transcription factors RUNX2 (B), MSX2 (C) and OSX (D); n=5. Doses, statistics, n, and symbols as in Figure 7. Based on the fact that ectopic phosphatase activity is necessary for osteoblastic extracellular matrix formation,101 the effects of IFN-α on TNAP expression were investigated. Herein, qPCR analysis revealed the upregulation of TNAP mRNA levels upon exposure of male cells to IFN-α and LPS for 24 h (Figure 25A). Conversely, no effect was observed in female VIC, thus confirming sex differences at early times (Figure 25A). Based on this result, we performed a long-term functional assay for ectopic phosphatase activity. After 21 days of treatment in activation medium, IFN-α or LPS alone and the combination of both stimuli significantly increased phosphatase activity in both male and female cells (Figure 25B). Collectively, these findings and previous data suggest that female cells also suffered an osteoblast-like differentiation process that seems to be delayed at early times. 116 R.3.1-IFN-γ and LPS interplay promotes HIF-1α induction in VIC As noted in the introduction, the master transcription factor for angiogenesis, HIF-1α, has been found in calcified valves co-localizing with calcifying nodules,107 but little is known about the underlying mechanisms of its expression. As expected, Western Blot analysis of VIC revealed no expression of HIF-1α in basal conditions (Figure 30A). However, we found that exposing VIC to IFN-γ slightly induced HIF-1α protein levels under normoxic conditions. Remarkably, this effect was strongly potentiated by LPS (Figure 30A). Moreover, the interplay with LPS on HIF-1α induction was specific for IFN-γ, since it was not observed with IFN-α (Figure 30B). Together, data disclose an immune, non-hypoxic, and IFN-γ-specific mechanism of HIF-1α induction. Figure 30. IFN-γ and LPS co-stimulation promotes HIF-1α protein stabilization. (A-B) Control male VIC were treated with the indicated ligands for 24 h and whole cell lysates analysed for HIF-1α protein levels. Representative blots and densitometric analysis of n=4-7. Doses, symbols, n, and statistics as in Figure 2. We next analysed HIF-1α nuclear translocation as an indicator of its activation. Two independent methods of cellular localization, cell immunostaining and Western blot analysis of nuclear and cytoplasmic extracts, demonstrated that HIF-1α was present in the nucleus upon costimulation of VIC with IFN-γ and LPS (Figure 31A-C). Remarkably, ruxolitinib pre-treatment abrogated HIF-1α nuclear translocation (Figure 31C). These data are consistent with a JAK/STAT- mediated activation of HIF-1α. 117 Figure 31. HIF-1α is translocated to the nucleus upon IFN-γ and LPS co-treatment of VIC under normoxic conditions. (A-B) Male VIC seeded on coverslips were treated for 24 h with IFN-γ+LPS and HIF- 1α protein was analysed by immunostaining as described in Methods. Representative microphotographs of HIF-1α (green) and DAPI (blue) staining, either separated or merged (A), and statistics of n=4 (B). Total corrected cellular fluorescence (TCCF) was calculated as indicated in Methods. White line indicates 50 M. (C) Male VIC were treated for the indicated times and nuclear (Nuc) and cytoplasmic (Cyt) extracts analysed by Western Blot. HisH3 indicates histone-H3; Ruxo,6 µM ruxolitinib. Doses, statistics, n, and symbols as in Figure 2. Once confirmed the immune-mediated activation of HIF-1α, we explored its regulation. RT- qPCR data showed that the treatment with IFN-γ and/or LPS did not alter HIF1A gene expression in VIC (Figure 32A), thus correlating with its well-known post-transcriptional regulation by stabilization via prolyl hydroxylase inhibition.198 Next, we used a pharmacological approach of several signalling routes that revealed a complex regulation of HIF-1α induction in VIC. First, the effect was receptor-dependent since HIF-1α protein induction was abrogated by the inhibition of both JAK and TLR activation (Figure 32B). Additionally, several downstream pathways, such as NF-κB, Akt and MAPK, played a role on HIF-1α induction (Figure 32B). Among HIF-1α selective inhibitors available, we tested a drug named PX-478199 that is currently used in a clinical trial for cancer treatment. Western Blot analysis revealed that this drug inhibited the IFN-γ+LPS-mediated expression of HIF-1α in a dose-dependent manner (Figure 32C). Given that PX-478 exerts some 118 potential off-target effects in transcription at high doses,199 the minimal dose exhibiting inhibitory effects in VIC, 40 µM, was chosen for the subsequent experiments. Figure 32. Immune induction of HIF-1α protein is controlled by several signalling pathways and abrogated by PX-478. (A) Male VIC were treated for 24 h and HIF1A mRNA levels analysed by qPCR; n=5. (B-C) Cells were pre-treated with the indicated inhibitors and activated for 24 h with IFN-γ+LPS for the analysis of HIF-1α protein levels. Representative blots and densitometric analysis of n=3-4. Data were normalized to IFN-γ+LPS induction (100%). Inhibitors and antagonist abbreviations and doses as in figure 11. PX indicates PX-478. Statistics, n and symbols as in Figure 2. To accurately elucidate the pathway controlling HIF-1α expression, gain and loss of function experiments were performed. First, we used cobalt chloride (CoCl2) a well-known chemical stabilizer of HIF-1α.200 Western Blot analysis revealed that treatment of male VIC with CoCl2 induced HIF-1α protein expression in a dose-dependent manner (Figure 33A-B). Conversely, chemical stabilization of HIF-1α did not induce either STAT1 or STAT3 phosphorylation (Figure 33A-B). These findings, together with HIF-1α induction blockade by Jakinibs (Figure 32B), suggest that this transcription factor is downstream of JAK/STAT signalling. Given the pro-apoptotic effects of CoCl2 reported in mitral valve interstitial cells at doses above 100 µM,201 the dose of 100 µM was selected for the subsequent experiments. 119 Figure 33. CoCl2 induces HIF-1α expression but not STAT1/3 activation. Male VIC were treated for 24 h with the indicated ligands and whole cell lysates analysed by Western Blot. (A) Representative blot for HIF- 1α, pSTAT1 (Tyr701), and pSTAT3 (Tyr705) levels. (B) Densitometric analysis and statistics; n=4. L indicates LPS; Ut, untreated. Doses, n, statistics and symbols as in Figure 2. We then performed loss of function experiments using a siRNA Silencer® select predesigned and validated siRNA duplexes for STAT1, STAT3 and HIF1A genes that were used to clearly define the signalling pathways involved in IFN-γ+LPS effects. Western blot analysis confirmed gene knockdown, since siRNA duplexes for STAT1 and STAT3 genes markedly reduced their protein levels, to an 18% and 5% respectively (Figure 34A). Based on these set-up experiments, the time of 72 h post-transfection was chosen for the silencing experiments. The efficiency of HIF1A silencing was demonstrated by analysing HIF-1α protein expression upon VIC stimulation with both immune and chemical inductors for 24 h (Figure 34B). HIF1A knockdown completely depleted both the CoCl2 and IFN-γ+LPS-mediated induction of HIF-1α protein (Figure 34B). Then, the effects of silencing STAT1/3 and HIF1A genes were analysed. Remarkably, data disclosed that STAT1 gene silencing completely abrogated HIF-1α protein induction upon cell activation with IFN-γ+LPS and CoCl2 (Figure 34B). In contrast, HIF1A 120 knockdown did not statistically affect either STAT1 protein levels or its phosphorylation (Tyr701) upon cell stimulation (Figure 34B). Figure 34. STAT1, but not STAT3, gene knockdown depletes HIF-1α induction by IFN-γ+LPS. Gene silencing experiments were performed as described in Methods. siRNA duplexes for STAT1 (si-STAT1), STAT3 (si-STAT3), and HIF1A (si-HIF1A) genes, and siRNA negative control (Ctrl siRNA) were used. (A) Representative blot and densitometric analysis confirming STAT1/3 gene knockdown at 72 h; n=3 male VIC. (B) Male VIC transfected as indicated for 72 h were activated with IFNγ+LPS or 100 µM CoCl2 for 24 h; n=4. Representative blot and densitometric analysis of n=4. Doses, statistics, n, and symbols as in Figure 2. Next, STAT3 knockdown showed a trend to reduce the stimuli-mediated HIF-1α induction, but data did not reach statistical significance (Figure 34B). Finally, HIF-1α knockdown did not significantly affect either STAT3 protein levels or its activation upon cell stimulation (Figure 35). 121 Altogether these data demonstrate that HIF-1α is downstream of STAT signalling. Moreover, STAT1 is the main transcription factor controlling the induction of HIF-1α expression by JAK/STAT and TLR pathways interplay. Figure 35. HIF1A and STAT1 knockdown does not affect STAT3 signalling upon stimulation. Gene silencing experiments were performed as described in Methods; siRNA duplexes and activation were as in Figure 34. Representative blots and the corresponding densitometric analysis of n=4. Doses, n, statistics, and symbols as in Figure 2. Densitometric analysis of knockdown experiments showed that STAT1 and STAT3 protein levels were statistically increased after 24 h of treatment with IFN-γ+LPS and negative control siRNA (Figures 34B and 35). These unexpected findings lead us to speculate whether these stimuli would increase STAT1 protein levels in non-transfected cells as an additional mechanism of regulation beyond its phosphorylation. The data revealed that cell activation for 24 h also significantly increased STAT1 total protein levels (Figure 36). This finding parallels previous reports, pointing to total STAT1 total protein levels increase as a mechanism of prolongation of STAT1 signalling and the expression of IFN-induced genes, apart from its phosphorylation.202 122 Figure 36. IFN-γ and LPS co-stimulation increases STAT1 total protein levels. Male VIC were treated with IFN-γ+LPS for 24 h and whole cell lysates analysed by Western Blot. Representative blot and densitometric analysis of n=4. Doses, n, statistics, and symbols as in Figure 2. R.3.2-IFN-γ treatment induces a pro-angiogenic phenotype in male VIC Because HIF-1α is the master transcription factor for angiogenesis, we next explored its downstream molecules and their involvement on VIC responses. Among the molecules downstream of HIF-1α with a crucial role on angiogenesis is VEGF-A, whose expression is increased in calcified valves.107 We found that VEGF-A was secreted by male VIC in response to IFN-γ, LPS, and the combination of both stimuli (Figure 37A). Additionally, significant sex-differences were observed since IFN-γ did not induce VEGF-A secretion in female cells, which also displayed lower secretion upon IFN-γ+LPS treatment (Figure 37A). In keeping with JAK and HIF-1α-mediated effects, pharmacological intervention with ruxolitinib and PX-478 blocked VEGF-A secretion induced by IFN-γ+LPS in male VIC (Figure 37B). To confirm the pro-angiogenic state induced by stimuli in VIC, we also explored the expression of the anti-angiogenic gene CNMD, which is known to maintain the valvular function by preventing angiogenesis.105 As shown in Figure 37C, CNMD expression was significantly reduced upon treatment with IFN-γ alone or combined with LPS in male VIC. Additionally, female VIC showed a high variation in the response but not statistically significant differences between activated and untreated conditions (Figure 37C). Together, data suggest that IFN-γ+LPS treatment triggered pro-angiogenic effects in VIC with the involvement of JAK and HIF-1α pathways and sexdifferential responses. 123 Figure 37. IFN-γ induces a pro-angiogenic phenotype potentiated by LPS via HIF-1 in male VIC. (A) Male and female VIC were treated with the indicated ligands for 48 h and supernatants collected for assaying VEGF-A secretion; n=5. (B) Male cells were pre-treated or not with ruxolitinib or PX-478 before activation and VEGF-A secretion was quantitated as in A; n=5-7. Data were normalized to IFN-γ+LPS value (100%). (C) Male and female VIC were treated for 24 h and CNMD gene expression was assessed by qPCR; n= 5 for each group. Doses, colour code, symbols, n, and statistics as in Figure 7. R.3.3-Chemical stabilization of HIF-1α increases VIC calcification To elucidate the role for HIF-1α on calcification, we performed calcification assays exposing cells to CoCl2. The chemical induction of HIF-1α significantly increased calcification levels in VIC, although it was less potent that the immune stimulation with IFN-γ+LPS (Figure 38A-C). These data disclosed a role for HIF-1α on the induction of VIC calcification. 124 Figure 38. Chemical induction of HIF-1α increases VIC calcification levels. Male VIC were treated with either IFN-γ+LPS or CoCl2 for 7 days in calcifying medium. (A-B) Representative microphotographs (A) and quantification of the relative ARS levels (B); n=4. (C) Calcium deposits measurements; n=4. Doses were as in Figure 34. Statistics, n and symbols as in Figure 2. R.3.4-Signalling pathways involved in sex-differences upon IFN-γ+LPS treatment Next, we explored the signalling pathways that could account for sex differences in VIC responses to IFN-γ+LPS. We found that the ERK was differentially activated in male and female cells, the former ones having a statistically greater phosphorylation of these kinases (Figure 39A), which have been previously described as pro-calcifying pathways in the valve context.91,166 Given the differences on the pro-angiogenic profile between male and female VIC (Figure 37), we also explored potential sex differences on HIF-1α induction as the upstream explanation. Strikingly, after 48 h, IFN-γ+LPS treatment resulted in a greater induction of HIF-1α in male cells (Figure 39B), thus confirming sex differences at this level. 125 Figure 39. HIF-1α and ERK are preferentially activated in male VIC upon IFN-γ and LPS treatment. Male and female VIC were treated with the indicated ligands for 24 h (A) or 48 h (B) and whole cell lysates analysed by Western blot. (A) Representative blot of ERK phosphorylation and densitometric analysis of the bands normalized to total ERK content; n=5. (B) Representative blot of HIF-1α protein induction and densitometric analysis of n=5. Doses, statistics, n, and symbols as in Figure 7. Based on the differential activation of ERK and HIF-1α, we aimed to determine their role on VIC calcification. A pharmacological approach disclosed that inhibition of HIF-1α by PX-478 and ERK activation blockade with PD98059 decreased the calcification levels induced by IFN-γ+LPS (Figure 40A-C). Data obtained with the HIF-1α inhibitor are in line with the increased calcification showed upon chemical stabilization of HIF-1α with CoCl2 (Figure 38) and suggest a role of HIF-1α on the immune-induced calcification. Altogether, these data suggest that the sex-biased activation of ERK and HIF-1α may be mechanisms accounting for the differential calcification levels in male and female VIC. 132 Figure 46. Sex differences in IFNGR and angiogenesis-related genes in calcified tissue. (A-C) Tissue homogenates from male and female calcified valves were used for mRNA analysis of the indicated genes; n=9 for each group. Symbols and statistics as in Figure 45. We also studied the protein levels of two relevant transcription factors for the IFN-mediated effects in VIC, namely STAT1 and HIF-1α. Western Blot analysis of tissue homogenates showed that the expression of HIF-1α was variable between samples from males and females, and its basal expression showed no apparent significant differences between sexes (Figure 47A-B). Although, it should be noted the low number of female available for this analysis and we cannot rule out possible technical issues that may affect HIF-1α protein stability. Remarkably, total STAT1 protein levels were significantly higher in valve tissue homogenates from males (Figure 47B), consistent with the higher IFN-mediated effects in male VIC. Figure 47. Sex differences in STAT1 protein levels and variable expression of HIF-1α among calcified valves. (A-B) Tissue homogenates from male and female calcified valves were analysed for protein expression. Representative blot (A) and densitometric analysis of the indicated proteins; n≥4 in each group. Symbols and statistics as in Figure 45. We finally explored the expression of inhibitors of calcification (MGP) and angiogenesis (CNMD) downregulated in VIC upon IFN treatment (Figures 14B and 37C). We found that the 133 expression of both inhibitors was higher in female calcified valves (Figure 48A-B). These data support the notion of the maintenance of protective mechanisms in female valves, which correlates with the increased downregulation of these genes in male VIC upon stimulation with inflammatory molecules. Figure 48. Sex differences in anti-calcification and anti-angiogenic genes among calcified valves. (A-B) Tissue homogenates from male and female calcified valves were used for mRNA analysis of the indicated genes; n=9 for each group. Symbols and statistics as in Figure 45. Altogether, the data on this chapter show basal differences between male control and calcified valves that correlate to findings on cultured VIC activated with IFN. Moreover, data argue for some intrinsic differences between sexes in valves. Sex differences in the response of control VIC to IFN also correlate with some findings in calcified valves such as lower MGP and CNMD in females and higher VEGFA in males. In addition, STAT1 protein levels are higher in calcified valves from males, suggesting a potential role for JAK/STAT on sex-differences in CAVD. Collectively, these intrinsic differences suggest protective or delaying mechanisms in females that could explain the sexdifferential clinical outcomes. However, these findings warrant further investigation with more accurately techniques such as immunohistochemistry or transcriptomic/proteomic analysis. . 134 R.5-Type I IFN signalling mediates Poly(I:C) effects in VIC In this section of the study, the evaluation of TLR3 signalling and interplay with IFN pathways was performed in VIC from control non-mineralized male valves. Table 10 includes the characteristics of all the patients, which were all Caucasian. Table 10: Clinical features of the patients whose valves were included in this section of the study. Control valves Patient characteristics Male n=9 Age (Range) 57 ± 9 Hypertension 1 (11%) Hypercholesterolemia 2 (22%) Diabetes mellitus 0 (0%) Smoking 0 (0%) Renal failure 0 (0%) Aetiology of heart failure 5 idiopathic, 3 ischemic, 1 hypertrophic 135 R.5.1-dsRNA treatment activates IFN signalling in VIC Emerging evidences point to an important role of viral or cell-derived dsRNA in cardiovascular diseases.203 Previous data from our group disclosed that the dsRNA analogous poly(I:C) promotes the secretion of inflammatory and anti-viral molecules by VIC, including IFN- and IFN-related proteins.164 Additionally, Meng and colleagues have recently demonstrated that poly(I:C)-induced inflammatory and osteogenic effects are mainly mediated by the TLR3-IRF3 axis on VIC,204,205 although the underlying mechanisms are not well understood. Our data from sections R.1 and R.2 demonstrated that type I IFN promote inflammatory and osteogenic responses in VIC. Based on these evidences, we hypothesized that TLR3 effects are mediated by type I interferon secretion and the ensuing activation of JAK/STAT pathways. First, we tested the overall effects of dsRNA in VIC using Poly(I:C) as a potent inducer of TLR3 activation. In addition, we also explored a potential TLR3-IFN-γ interplay. The MTT assay showed that Poly(I:C) significantly reduced VIC proliferation after long-term treatment (Figure 49A), in contrast to the TLR4 ligand LPS, which showed no significant effects in VIC growth (Figure 13B). In addition, qPCR analysis disclosed that Poly(I:C) treatment strongly increased TLR3 expression, (Figure 49B), thus suggesting a potential autocrine loop promoting a prolonged activation. Because Poly(I:C) is known to induce VIC calcification,164 we explored pro-osteogenic and myofibroblast markers as indicators of cell differentiation. Poly(I:C) markedly increased BMP2 expression, whereas it triggered the downregulation of the calcification inhibitor MGP and the myofibroblast marker ACTA2, although no interplay with IFN-γ was observed (Figure 49C). These findings are consistent with a dsRNA-mediated induction of VIC differentiation. Strikingly, data resemble the IFN-α-induced effects, which were characterized by BMP-2 upregulation (Figure 26), proliferation inhibition (Figure 13), and MGP gene and α-SMA protein expression downregulation (Figures 14 and 23). 136 Figure 49. Poly(I:C) treatment reduces cell proliferation and promotes cell differentiation and TLR3 expression. (A) Non-confluent cells were grown for 12 days, replacing medium and stimuli every 3 days; n=6 male VIC. (B-C) Male VIC were treated for 24 h with the indicated ligands and harvested for qPCR analysis of TLR3 (B) and the indicated differentiation markers expression (C); n=4-5. Poly indicates 1 µg/mL of polyinosinic:polycytidylic acid. IFNγ, 100 ng/mL IFN-γ. Doses, n, color code, and * as in Figure 2. Next, we sought to investigate whether Poly(I:C) could activate IFN-related signalling. Gene expression analysis revealed that VIC exposure to this ligand induced a strong upregulation on IFNB1 gene expression (Figure 50A). The secretion of this cytokine was then assayed in supernatants from activated cells. Strikingly, Poly(I:C), as well as IFN-γ, significantly induced IFN- β1 secretion, although no interplay between stimuli was observed (Figure 50B). Moreover, upregulation of IFN signalling was further demonstrated by analysing the expression of downstream transcription factors from the IRF family. As shown in Figure 50C, IRF1, 3 and 8 mRNA levels were significantly increased after Poly(I:C) treatment, although no interplay with IFN-γ was found. Together, data disclose type I IFN secretion and the upregulation of IFN signalling molecules upon VIC exposure to synthetic dsRNA. 137 Figure 50. Poly(I:C) treatment promotes IFN-β1 expression and secretion and IRF upregulation in VIC. (A-C) Male VIC were treated with the indicated ligands for 24 h. mRNA was analysed for IFN-related genes by qPCR (A, C), and IFN- was assayed in cell supernatants by ELISA (B); n=5. Doses as in Figure 49; n, symbols, and statistics as in Figure 2. Finally, we investigated IFN signalling cascade activation. As shown by Western Blot, TLR3 activation resulted in a marked phosphorylation of STAT1 (Figure 51). Strikingly, STAT1 activation was blocked by ruxolitinib, indicating a role of JAK signalling on the Poly(I:C)-mediated activation of STAT1. Overall, data demonstrate that a synthetic dsRNA triggers IFN-β secretion and the subsequent activation of IFN-related signalling pathways in VIC. 138 Figure 51. Poly(I:C) treatment promotes STAT1 activation. Male VIC were treated with the indicated ligands for 24 h and whole cell lysates analysed by Western Blot for STAT1 phosphorylation. Representative blot and densitometric analysis of n=4. Ruxo indicates 6 µM ruxolitinib. Doses as in Figure 49; n, symbols, and statistics as in Figure 2. R.5.2-JAK/STAT signalling blockade reduced Poly(I:C)-induced effects Given the previous findings, we sought to demonstrate a role for JAK/STAT signalling on the reported Poly(I:C)-induced responses in VIC, namely inflammation and calcification.164 Initially, we studied the master transcription factor for inflammation and found that Poly(I:C) induced a strong activation of NF-κB (Figure 52A). Then, we explored adhesion molecules downstream this factor. Our data confirmed that Poly(I:C) is a strong inducer of ICAM-1 (Figure 52B), as previously reported by our group,164 and uncovered the upregulation of VCAM-1 protein levels (Figure 52C). Remarkably, the pre-treatment with ruxolitinib completely depleted the marked Poly(I:C)-mediated activation of NF-κB (Figure 52A), and strongly reduced adhesion molecule expression (Figure 52B-C). Together, data are consistent with JAK/STAT involvement in dsRNA-mediated pro-inflammatory responses in VIC. 139 Figure 52. Poly(I:C) treatment induces inflammatory molecules, being its effects reduced by ruxolitinib pre-treatment. (A-C) Male VIC were treated with the indicated ligands for 24 h and whole cell lysates analysed by Western Blot for NF-κB activation (A) and adhesion molecule expression (B-C); n=4. Ruxo indicates 6 µM ruxolitinib. Doses as in Figure 49; n, symbols, and statistics as in Figure 2. Our group previously described the cytokine profile induced by Poly(I:C), which is charazterized by IL-6 and 8 secretion as well as the secretion of an IFN-related cytokine, interferon gamma-induced protein 10 (IP-10), also known as C-X-C motif chemokine 10.164 In this basis, we analysed whether JAK/STAT signalling activation mediates Poly(I:C)-induced cytokine profile. As shown in Figure 53, ruxolitinib pretreatment markedly abrogated IL secretion (Figure 53A) and completely depleted IP-10 secretion in VIC (Figure 53B), further demonstrating a role for JAK/STAT pathways in the inflammatory responses upon TLR3 activation. 140 Figure 53. Ruxolitinib pre-treatment blunts the Poly(I:C)-induced cytokine profile in VIC (A-B) Male VIC were treated with the indicated ligands for 24 h and supernatants were collected to asses cytokine secretion; n=4. Ruxo indicates 6 µM ruxolitinib. Doses, symbols, n, and statistics as in Figure 2. Our next step was to investigate the effects of Poly(I:C) on calcification under high phosphate conditions. Calcification assays revealed that Poly(I:C) is a strong inducer of VIC calcification, being its effects more robust that the induced by IFN-γ (Figure 54A-C). To note, nodule formation and calcium deposition were observed at relatively short times (7-9 days) as compared to type I and type II IFN-mediated calcification (Figures 15-16). Strikingly, VIC calcification was strongly potentiated upon combination of Poly(I:C) with IFN-γ (Figure 54A-C), thus indicating an interplay between Poly(I:C) and IFN-γ on VIC calcification in high phosphate conditions. 141 Figure 54. Poly(I:C) treatment induces VIC calcification with further potentiation by IFN-γ. (A-C) VIC were treated with the indicated ligands for 9 days in calcification medium and mineralization was evaluated by ARS (A-B) and calcium deposits quantitation (C); n=4. Doses, symbols, n, and statistics as in Figure 2. Finally, to elucidate the underlying mechanism, we investigated whether type I IFN signalling blockade blunts Poly(I:C)-induced calcification. Remarkably, a type I IFN receptor neutralising antibody (IFNAR Ab) markedly reduced both nodule formation and calcium deposition upon Poly(I:C) treatment, while an isotype control had no significant effects (Figure 55). These findings support a major role for IFNAR signalling on TLR3-induced calcification under high phosphate conditions and unravel a mechanism accounting for the TLR3 agonist Poly(I:C)-mediated effects in VIC via Type I IFN/JAK/STAT pathways. Data in this section showed that Poly(I:C) exerts similar anti-proliferative, pro-inflammatory and pro-osteogenic effects to those observed with IFN-α (sections R.1 and R.2) and exhibited differences in cell proliferation as compared to the TLR4 ligand LPS. The major findings are that Poly(I:C) mediates its marked effects, at least in part, by triggering type I IFN secretion and signalling in VIC, subsequently promoting inflammation and calcification that can be blocked by type I IFN signalling inhibition. Moreover, data disclosed a positive cooperation with IFN-γ on VIC calcification.