VASCULAR BIOLOGY, ATHEROSCLEROSIS, AND ENDOTHELIUM BIOLOGY
Endothelial-to-Mesenchymal Transition in Bone Marrow and Spleen of Primary Myelo fibrosis
Benedetta Gaia Erba,* Cristian Gruppi,yMonica Corada,* Federica Pisati,*zVittorio Rosti,xNiccolo’ Bartalucci,{ Jean-Luc Villeval,kAlessandro Maria Vannucchi,{Giovanni Barosi,xAlessandra Balduini,yand Elisabetta Dejana*,**
From the Vascular Biology Laboratory,* The FIRC Institute of Molecular Oncology (IFOM) Foundation, Milan, Italy; the Department of Molecular Medicine,yUniversity of Pavia, Pavia, Italy; the Hystopatology Unit,zCogentech, Milan, Italy; the Centre for the Study and Treatment of Myelofibrosis,x Biotechnology Research Laboratories, Center for the Study of Myelofibrosis, Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo Foundation, Pavia, Italy; the Section of Haematology,{Department of Medical and Surgical Care, University of Florence, Florence, Italy; INSERM,kU1009, Institut Gustave Roussy and Université Paris XI, Villejuif, France; and the Department of Immunology, Genetics and Pathology,** University of Uppsala, Uppsala, Sweden
Accepted for publication April 4, 2017.
Address correspondence to Elisabetta Dejana, Ph.D., The FIRC Institute of Molecular Oncology (IFOM) Foundation, Via Adamello 16, 20139 Milan, Italy. E-mail:elisabetta.
dejana@ifom.eu.
Primary myelofibrosis is characterized by the development of fibrosis in the bone marrow that con- tributes to ineffective hematopoiesis. Bone marrowfibrosis is the result of a complex and not yet fully understood interaction among megakaryocytes, myeloid cells,fibroblasts, and endothelial cells. Here, we report that>30% of the endothelial cells in the small vessels of the bone marrow and spleen of patients with primary myelofibrosis have a mesenchymal phenotype, which is suggestive of the process known as endothelial-to-mesenchymal transition (EndMT). EndMT can be reproduced in vitro by incu- bation of cultured endothelial progenitor cells or spleen-derived endothelial cells with inflammatory cytokines. Megakaryocytes appear to be implicated in this process, because EndMT mainly occurs in the microvessels close to these cells, and because megakaryocyte-derived supernatantfluid can reproduce the EndMT switch in vitro. Furthermore, EndMT is an early event in a JAK2-V617F knock-in mouse model of primary myelofibrosis. Overall, these data show for the first time that microvascular endothelial cells in the bone marrow and spleen of patients with primary myelofibrosis show functional and morphologic changes that are associated to the mesenchymal phenotype. (Am J Pathol 2017, 187: 1879e1892;
http://dx.doi.org/10.1016/j.ajpath.2017.04.006)
Primary myelofibrosis (PMF) is one of the family of myeloproliferative neoplasms, which is a heterogeneous group of related diseases that also includes polycythemia vera and essential thrombocythemia.1 Myeloproliferative neoplasms are clonal malignant disorders that are charac- terized by oncogenic transformation of the hematopoietic progenitor cell compartment, which causes abnormal pro- liferation of the myeloid lineages driven by a hypersensi- tivity to regulatory growth factors.2 Fifty percent to 60%
of patients with myeloproliferative neoplasms harbor a so- matic gain-of-function mutation in the Janus kinase 2 (JAK2) gene, which results in the translation of the mutated JAK2-V617F protein. Because this mutation is in the pseudo-kinase JH2 domain, the autoinhibitory function of JAK2-V617F is impaired. This results in constitutive activation of the downstream JAK2 signaling pathways,
which include the STATs, mitogen-activated kinases, and phosphatidylinositol 3-kinaseeAKT.3Additional mutations in the hematopoietic compartment have also been identified, including for the thrombopoietin receptor and, as reported recently, the calreticulin genes, which lead to a similar phenotype.4,5
Supported by Associazione Italiana per la Ricerca sul Cancro Special Program Molecular Clinical Oncology 5 1000 grant to AIRC-Gruppo Italiano Malattie Mieloproliferative (AGIMM) and AIRC-IG investigator grant 14471; TELETHON grant GGP14149; the European Community WNT FOR BRAIN, European Research Council contract 268870; ITN BtRAIN grant 675619; Cariplo Foundation grant 2014-1038; the Swedish Science Council; Knut and Alice Wallenberg Foundation (E.D.); and Cariplo Foundation grant 2010-0807 (A.B).
B.G.E., C.G., and M.C. contributed equally to this work.
Disclosures: None declared.
ajp.amjpathol.org
PMF clonal myeloproliferation primarily induces bone marrow fibrosis and osteosclerosis, with extramedullary hematopoiesis in the spleen, which results in splenomegaly and abnormal cytokine overexpression.6,7 Debilitating symptoms include anemia, thrombosis, and general malaise.
Ineffective erythropoiesis and extramedullary hematopoiesis can be the cause of anemia and splenomegaly. No definitive therapy is currently available for patients with PMF. The only potentially curative treatment is allogeneic stem cell transplantation, although this is associated with high rates of mortality and morbidity.8
Despite being a hematopoietic disease, among the myeloproliferative neoplasms, involvement of the endothe- lial lineage has been reported only for PMF,9e11although this has been poorly characterized to date. In the bloodstream of patients with PMF, the numbers of endothelial progenitor cells (EPCs) are significantly higher than in healthy donors and for other myeloproliferative neoplasms.9 Moreover, a massive neoangiogenesis is observed in bone marrow12,13 and spleen14 of patients with PMF, which shows a high concentration of the proangiogenic factors vascular endo- thelial growth factor15andfibroblast growth factor-2.16
Althoughfibroblasts are believed to be directly implicated in fibrosis in PMF bone marrow, other cells might also contribute to this phenomenon, such as endothelial cells.
Under inflammatory conditions, endothelial cells can undergo a biological process known as endothelial-to-mesenchymal transition (EndMT), such as during arteriosclerosis17or in the tumor stroma.18 This phenotypic switch is characterized by coexpression of endothelial and mesenchymal markers and by changes to some of the phenotypic characteristics, such as acquisition of an elongatedfibroblastoid morphologic struc- ture and increased migratory properties.18,19Physiologically, EndMT takes place during the formation of the heart cushion in the embryo,20e22but it can be reactivated in adults under pathologic conditions.23e25 The EndMT program can be induced by transforming growth factor (TGF)-b and the bone morphogenetic proteins (BMPs), and it can significantly contribute tofibrotic progression in different diseases.23,26e28 In the present study, we show that endothelial cells from the microvasculature of bone marrow and spleen can un- dergo the EndMT switch during the development of PMF in both patients and a mouse model of PMF. This process occurs during the early stages of fibrotic degeneration and can be mediated by the release and activation of TGF-b and other inflammatory cytokines by megakaryocytes and platelets. Thus, EndMT can contribute to the development and maintenance of bone marrow fibrosis and spleen dysplasia in PMF.
Materials and Methods
Tissue Samples from PMF Patients
Bone marrow, spleen tissue, and EPCs from peripheral blood of patients with PMF were collected at the Istituto di
Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo (Pavia, Italy). The diagnosis of PMF was established according to the World Health Organization 2008 criteria29 and the Italian Consensus Conference criteria.30 The clinical features of patients with PMF are summarized in Supplemental Table S1. The healthy control subjects were staff members or donors for scientific research. All of the patients and donors approved and signed the informed consent. Bone marrow sections from healthy donors were purchased from Labospace (Milano, Italy).
Mice
The C57Bl/6J mice (8 to 12 weeks of age) for isolation of endothelial cells from the spleen (SECs) and of megakar- yocytes and for purification of platelets were purchased from Charles Rivers Laboratories (Calco, Italy).
The conditional JAK2FLEX/WT KI mice have been described previously,31 and they were crossed with trans- genic mice that expressed Cre-recombinase under the Vav promoter, to obtain mice expressing the JAK2-V617F mutation in heterozygosity, namely JAK2V617F/WTKI.32
The animal experimentation was approved by the FIRC Institute of Molecular Oncology Institutional Animal Care and Use Committee and was performed according to the guidelines for the regulation of animal experimentation of the Italian Ministry of Health.
Cells
Murine SECs were isolated and immortalized as described previously.33 Briefly, the mouse spleens were collected under sterile conditions and dissociated with 1.5 mg/mL collagenase type I (Roche, Mannehim, Germany) and DNase 25 mg/mL (Roche) for 3 hours at 37C. The resulting cell suspension was filtered through a nylon screen (70-mm mesh), centrifuged, and plated. Two days later, the heterogeneous population of SECs was infected with polyoma middle T antigen to specifically immortalize them.34
The SECs were cultured in 0.1% gelatin-coatedflasks in MCDB131 medium (Life Technologies, Paisley, UK), supplemented with 20% North American fetal bovine serum (HyClone, South Logan, UT), 50 mg/mL endothelial cell growth supplement (made from calf brain), 100 mg/mL heparin (Sigma-Aldrich, St. Louis, MO), 100 U/L penicillin/
streptomycin (Sigma-Aldrich), and 2 mmol/L L-glutamine (Sigma-Aldrich). The cell-starving medium consisted of MCDB131 medium with 1% bovine serum albumin (BSA;
EuroClone, Milano, Italy).
EPCs from healthy donors and patients with PMF were isolated according to protocols reported previously35 and were grown on collagen-Iecoated plates in Endothelial Cell Growth Medium-2 supplied with Endothelial Cell Growth Medium-2 MV SingleQuots kits (Lonza, Cologne, GmbH, Germany).
Cell Treatments
To induce the EndMT phenotype, the SECs and EPCs were stimulated with a cocktail of proinflammatory cytokines that included TGF-b, IL-1b, and tumor necrosis factor (TNF)-a (Peprotech, Rocky Hill, NJ), as described previously,36 or with BMP6 (R&D Systems, Minneapolis, MN). The SECs were also treated with the TGF-b inhibitors LY-2109761 (SelleckBio, Munich, Germany) and dorsomorphin homo- log 1 (DMH1; Tocris Bioscience, Bristol, UK).
Megakaryocyte Differentiation from Murine Fetal Liver
Murine fetal liver cells were collected at embryonic days 13 to 15, mechanically dissociated, and cultured for 3 days in Dulbecco’s modified Eagle’s medium (Life Technolo- gies) supplemented with 10% fetal calf serum (Life Technologies), 100 U/L penicillin/streptomycin, 2 mmol/L
L-glutamine, and 10 ng/mL recombinant mouse thrombo- poietin (PeproTech). The cells were then layered on a single-step gradient (1.5% to 3.0% BSA), and the mega- karyocytes were allowed to settle for 30 minutes at room temperature. The megakaryocyte pellet was washed and resuspended in complete medium, to a final concentration of 0.2 109cells/mL.37
Platelet Puri fication and in Vitro Induction of Platelet and Megakaryocyte Release Reactions
Platelet purification from peripheral blood of mice was performed as described previously.38 Briefly, the blood was collected from the inferior vena cava and pooled in buffered 3.2% citrate-dextrose solution, pH 5.2, and diluted with two parts of Ca2þ-free and Mg2þ-free Tyrode buffer, pH 6.5, followed by centrifugation at 300 g for 7 minutes, to obtain the platelet-rich plasma. This platelet- rich plasma was further diluted in two parts of Tyrode buffer, pH 6.5, supplemented with 0.6 U/mL apyrase a (New England BioLabs, Ipswich, MA) and centrifuged at 600 g for 15 minutes, to obtain the platelet pellet. The platelets were then resuspended in Tyrode buffer, pH 7.4, that contained 2 mmol/L CaCl2, to afinal concentration of 0.2 109platelets/mL.
The platelets and megakaryocytes were incubated with 2 U/mL recombinant thrombin (Sigma-Aldrich) for 15 minutes at room temperature. Then 2 U/mL recombinant hirudin (Sigma-Aldrich) was added for 15 minutes to inactivate the thrombin. An additional centrifugation at 600 g for 15 minutes led to the separation of the supernatantfluid from the pellet. The pellet was then resuspended in Tyrode buffer, pH 7.4, and both the supernatant fluid and the pellet were supplemented with 0.5% BSA and 70 mmol/L HCl. After 30 minutes, 70 mmol/L NaOH was added to neutralize the acid pH environment of the solutions.
For the TGF-b blocking experiments, the megakaryocyte supernatant fluids were preincubated with an antieTGF-b
neutralizing antibody (clone 1D11; R&D Systems) or the corresponding isotype control antibodies (R&D Systems).
IF Analysis
For immunofluorescence (IF), confluent cell monolayers werefixed with 4% formaldehyde that was freshly prepared from paraformaldehyde in phosphate-buffered saline (PBS).
After 15 minutes at room temperature, thefixed cells were permeabilized with 0.5% Triton X-100 for 5 minutes. The antibodies for the blocking (1 hour), primary (2 hours), and secondary (1 hour) incubations were diluted in PBS with 2%
BSA. The appropriate fluorophore-conjugated secondary antibodies were used (Alexa-488, Alexa-555, Alexa-647, respectively; Molecular Probes, Eugene, OR). The nuclei were visualized using DAPI.
Tissue Samples
Frozen bone marrow and spleen biopsies were sectioned (thickness, 5 mm) using a cryotome (Leica, Wetzlar, Germany), and the sections were mounted on Superfrost glass slides (Thermo Scientific, Waltham, MA) and dried.
The sections were then fixed for 15 minutes with 4%
formaldehyde that was freshly prepared from para- formaldehyde in PBS at room temperature, washed with PBS, permeabilized with 0.1% TritonX-100ePBS, and blocked with PBS containing 5% donkey serum and 2%
BSA for 3 hours at room temperature.
Murine paraffin-embedded bone marrow and spleen were sectioned (thickness, 5 mm) using a microtome (Leica) and deparaffinized through descending concentrations of ethanol. Antigen unmasking was performed in 0.25 mmol/L EDTA, pH 8.0, for 50 minutes at 95C. The sections were then blocked with Tris-buffered saline (TBS) containing 5%
donkey serum, 0.05% Triton X-100, and 2% BSA, for 1 hour at room temperature. Incubations with primary anti- bodies on both the cryosections and paraffin sections were performed in blocking solution overnight at 4C. Tissue sections were then washed and incubated with the appro- priatefluorophore-conjugated secondary antibodies (Molec- ular Probes) for 1 hour at room temperature. The nuclei were visualized with DAPI. The specimens were mounted with Vectashield (Vector Laboratories, Burlingame, CA).
Confocal microscopy was performed using a confocal mi- croscope (TCS SP2; Leica). Image acquisition was performed using a 63/1.4 NA oil-immersion objective (HCX PL APO 63 Lbd Bl; Leica), and with spectral detection bands and scanning modalities optimized for the removal of channel cross talk. The Leica confocal software version 2.61 and ImageJ software version 1.49 (NIH, Bethesda, MD) were used for the data analysis. Only adjustments of brightness and contrast were used in the preparation of the figures, using Photoshop software version 13.0.6 (Adobe, San Jose, CA).
Hematoxylin and eosin staining of the paraffin-embedded sections was performed according to standard protocols.
IP Analysis
For immunoprecipitation (IP) cells were solubilized in JS buffer (20 mmol/L HEPES, pH 7.5, 1.5 mmol/L MgCl2, 5 mmol/L EGTA, 150 mmol/L NaCl, 1% Triton X-100, 0.5%
glycerol), protease inhibitors (Roche), and phosphatase in- hibitors (Sigma-Aldrich) on ice for 20 minutes. Precleared cell extracts were subjected to antibody precipitation at 4C, and immunocomplexes were captured using protein G-Sepharose
beads (GE Healthcare Europe GmbH, Milano, Italy). Immu- noprecipitated material was separated using Tris-glycine sodium dodecyl sulfate-PAGE, blotted onto nitrocellulose membranes, and analyzed by standard methods.
Immunoblot Analysis
Total protein was extracted by solubilizing the cells in boiling Laemmli buffer. Lysates were incubated for 10 minutes at Figure 1 Endothelial cells of patients with primary myelofibrosis (PMF) show a transforming growth factor (TGF)-bemediated endothelial-to-mesenchymal transition (EndMT) phenotype. A and B: Representative images of bone marrow (A) and spleen (B) tissues derived from healthy donors and patients with PMF costained for vascular endothelial (VE)-cadherin (red) and the EndMT markersfibroblast-specific protein 1 (Fsp1; green), fibronectin (FN; green), and collagen I (green). DAPI staining (blue) shows bone marrow cell nuclei. Quantification of positive Fsp1 and FN vessels in spleen of healthy donors and patients with PMF (B). C: Representative images of immunofluorescence analysis for TGF-b downstream signaling molecules p-Smad2/3 (green) in splenic tissues from healthy donors and patients with PMF. Endothelial cells were stained for VE-cadherin (red). DAPI staining (blue) shows nuclei. Endothelial cell nuclei positive for p-Smad2/3 are highlighted by boxed areas. D: Quantification of endothelial cell expression of p-Smad2/3. Boxed areas are shown at higher magnification to the right. Data are expressed as means SD. n Z 3 healthy CTRL donors (AeC); n Z 8 patients with PMF (AeC). *P < 0.05, **P < 0.01. Scale bars: 30 mm (AeC, main images); 10 mm (AeC, higher magnification). CTRL, control; JAK2, Janus kinase 2; p-, phospho; wt, wild-type.
100C and then centrifuged at 10,000 g for 5 minutes to remove the cell debris. The supernatant fluids were collected, and their protein concentrations were determined using BCA Protein Assay kits (Pierce-ThermoFisher, Rockford, IL), according to the manufacturer’s instructions.
Equal amounts of protein were loaded onto the acrylamide gels at different concentrations and were separated using sodium dodecyl sulfate-PAGE, transferred to Protran nitro- cellulose hybridization transfer membranes (pore size, 0.2 mm; ThermoFisher, Rockford, IL), and blocked for 1 hour at room temperature in TBS with Tween-20 (TBST) and 5% BSA. The membranes were then incubated over- night at 4C, with the primary antibodies diluted in TBST containing 5% BSA. The membranes were rinsed at least three times with TBST and then incubated with horseradish
peroxidase-linked secondary antibodies (Cell Signaling Technology, Danvers, MA) for 1 hour at room temperature.
After three washes, the membranes were incubated with Amersham ECL Western blot (WB) detection reagents (Amersham Biosciences, Uppsala, Sweden) for 1 minute and exposed using a ChemiDoc XRS gel imaging system (Bio-Rad, Hercules, CA), for the required time.
RT-PCR
For the gene expression analysis, total RNA was isolated using RNaesy Micro kits (Qiagen, Valencia, CA), and 500 ng of total RNA was reverse transcribed with random hexamers (High-Capacity cDNA Archive kits; Applied Biosystems, Foster City, CA) according to the Figure 2 Endothelial progenitor cells (EPCs) undergo endothelial-to-mesenchymal transition (EndMT) under proinflammatory conditions. EPCs (assessed as endothelial colony-forming cells) from healthy donors and patients with PMF were treated with a cocktail of proinflammatory cytokines: 5 ng/mL transforming growth factor (TGF)-b1, 100 U/mL IL-1b, and 100 U/mL tumor necrosis factor (TNF)-a for 72 hours. A: Representative images of expression of endothelial specific markers vascular endothelial (VE)-cadherin (red) and claudin5 (green) in EPCs before and after treatment. DAPI staining (blue) shows nuclei.
B: Quantification of quantitative real-time RT-PCR analysis of EPCs for expression of mesenchymal markers Snail, CD44, N-cadherin, and fibronectin (FN). mRNA is expressed as 2DCt, as normalized to GAPDH. C: Representative images of phase contrast microscopy for morphologic structure of EPCs before (untreated) and after cytokine treatment. Data are expressed as means SD. n Z 3 healthy CTRL donors (A); n Z 6 patients with PMF (A); n Z 3 for each analysis (B).
**P< 0.01. Scale bars: 10 mm (A); 30 mm (C). CTRL, control; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; JAK2, Janus kinase 2; PMF, primary myelofibrosis; wt, wild-type.
manufacturer’s instructions. The cDNAs were amplified using the TaqMan gene expression assay (Applied Bio- systems) in a thermocycler (7900 HT; ABI/Prism, Applied Biosystems).
Antibodies
The following antibodies were used for IF, WB, and IP anal- ysis: antievascular endothelial (VE)-cadherin (IF, dilution 1:200; WB analysis, dilution 1:200; IP, dilution 1:100; Santa Cruz Biotechnology, Dallas, TX); anti-CD44 (IF, dilution 1:100; WB analysis, dilution 1:500; BD Biosciences, San Jose, CA); anti-CD41a (IF, dilution 1:100; eBioscience, San Diego, CA) antieN-cadherin (IF, dilution 1:200; WB analysis, dilu- tion 1:2500; IP, dilution 1:100; BD Biosciences); anti-collagen I (IF, dilution 1:200; Invitrogen, Waltham, MA); anti- efibroblast-specific protein 1 (FSP1; IF, dilution 1:200; WB analysis, dilution 1:500; Millipore-Merck, Milano, Italy), anti-claudin5 (IF, dilution 1:500; Abcam, Cambridge, UK), anti-fibronectin (WB analysis, dilution 1:1000; Abcam); anti- panTGF-b (IF, dilution 1:100; R&D Systems, Minneapolis, MN); antieTGF-b neutralizing (clone 1D11; R&D Systems);
anti-CD31 (IF, dilution 1:400; BD Biosciences); antiejunc- tional adhesion molecule-A (JAM-A); (IF, dilution 1:100;
BV12; Santa Cruz Biotechnology); antiea-catenin (IF, dilu- tion 1:200; WB analysis, dilution 1:500; Cell Signaling Technology, Danvers, MA); antieb-catenin (IF, dilution 1:200; WB analysis, dilution 1:500; Cell Signaling); antieg- catenin (IF, dilution 1:200; WB analysis, dilution 1:500; Cell Signaling); anti-p120 (IF, dilution 1:200; WB analysis, dilu- tion 1:500; Cell Signaling); anti-Smad1, anti-Smad3, antie phospho-Smad1 (WB analysis, dilution 1:1000; Cell Signaling); antiephospho-Smad3 (IF, dilution 1:100; Santa Cruz Biotechnology; WB analysis, dilution 1:500; Abcam);
anti-tubulin (WB analysis, dilution 1:2000; Sigma-Aldrich);
vinculin (WB analysis, dilution 1:5000; Sigma-Aldrich). The references on the specificity of the antibodies used are reported in the brochure of the respective cited supplier companies.
Statistical Analysis
Statistical significance was evaluated using two-tailed non- paired t-tests with the significance level at P < 0.05.
Results
Microvascular Endothelial Cells of Bone Marrow and Spleen of Patients with PMF Express Mesenchymal Markers
Analysis of the bone marrow sections of patients with PMF for expression of different markers demonstrated that a subpopulation of endothelial cells were costained with the specific marker VE-cadherin and Fsp1 (Figure 1A). The vessels were also surrounded by high levels offibronectin, which is usually low or absent in the bone marrow vasculature of healthy individuals. The spleen of the patients with PMF were then analyzed, where extramedullary hematopoiesis primarily takes place and the bone marrow hematopoietic niche can be recre- ated. Spleen sections showed that Fsp1 was coexpressed with VE-cadherin in the endothelial cells of PMF speci- mens but not in control spleens (Figure 1B). Furthermore, these vessels showed strong staining for fibronectin and collagen I (Figure 1B). Fsp1, fibronectin, and collagen I are not present in the normal vessels of the spleen, but they can be up-regulated when endothelial cells undergo EndMT.
Under other pathologic conditions, TGF-b is the main EndMT inducer.18,39e41 The spleen sections from the pa- tients with PMF showed strong staining with the antieTGF- b antibody (Supplemental Figure S1), and the number of endothelial cells that were positive for the TGF-b down- stream signaling molecule phospho-Smad2/3 was increased for the patients with PMF compared with controls (Figure 1, C and D). The presence of the JAK2-V617F mutation in the hematopoietic lineage did not significantly change the appearance of EndMT, TGF-b production or Smad phos- phorylation for the patients with PMF.
EPCs and SECs Develop EndMT in an In flammatory Environment
To investigate the molecular mechanisms through which endothelial cells undergo EndMT in the patients with PMF, circulating EPCs were isolated and cultured as endothelial colony-forming cells35; these cells have been reported to be
Figure 3 Inflammatory cytokines and tranforming growth factor (TGF)-b/bone morphogentic protein (BMP)-6 axis are responsible for the EndMT phenotype in spleen endothelial cells (SECs). SECs were treated with 5 ng/mL recombinant TGF-b1, 100 U/mL IL-1b, and 100 U/mL tumor necrosis factor (TNF)-a for 120 hours under starving condition. A: Representative images of phase contrast microscopy for morphologic structure of SECs before (untreated) and after cytokine treatment. B: Quantification of VE-cadherin and N-cadherin expression at transcriptional and protein levels, by RT-PCR and as representative images of Western blot analysis (right), respectively. C: Quantification of mRNA levels of claudin5, a-SMA, fibronectin (FN), fibroblast-specific protein 1 (Fsp1), collagen I, serpin1, Snail, Bmp6, and CD44. Western blot analysis (right) of CD44 and FN expression in SECs treated with proinflammatory cytokines. D:
Quantification of mRNA levels after 100 ng/mL BMP-6 treatment for 96 hours, for claudin5, a-SMA, FN, Fsp1, collagen I, serpin1, and Snail, by RT-PCR. For all RT-PCR analyses, mRNA is expressed as 2DCt, as normalized to 18 seconds. E: Representative Western blot analysis for expression of total levels of Smad1 and Smad3 and their phosphorylated forms (pSer463/465, pSer423/425, respectively) in SEC lysates after stimulation with 5 ng/mL TGF-b1, 100 U/mL IL-1b, and100 U/mL TNF-a, or with 100 ng/mL BMP-6, for 45 minutes. All Western blot analyses were representative of independent experiments with tubulin or vinculin used to monitor sample loading. Data are expressed as means SD. n Z 3 independent experiments for Western blot analysis. *P < 0.05,
**P< 0.01. Scale bar Z 30 mm. a-SMA, a-smooth muscle actin; EndMT, endothelial-to-mesenchymal transition; VE, vascular endothelial cadherin.
vascular endothelial growth factor receptor 2þ, VE- cadherinþ, von Willebrand factorþ, CD31þ, CD146þ, CD105þ, CD45, and CD14.35,42 EPCs circulate in high numbers in patients with PMF9 and can reach the spleen through the circulation. To reproduce the proinflammatory environment of the PMF disease, the EPCs were cultured and stimulated with a cocktail of cytokines, which included TGF-b, IL-1b, and TNF-a.36In a previous study,28we iden- tified several markers of EndMT that were very low or absent in the resting endothelium. Cytokine-stimulated EPCs from both healthy donors and the patients with PMF showed de- creases in the endothelial markers claudin-5 and VE-cadherin at cell-to-cell contacts (Figure 2A), with up-regulation of the mesenchymal markers Snail, N-cadherin,fibronectin, and CD44 (Figure 2B). Moreover, the switch toward a mesen- chymal phenotype induced cultured EPCs to acquire an elongated, fibroblast-like spindle shape (Figure 2C). No significant differences were found for the expression of the mesenchymal markers in the control EPCs or those from the patients with PMF, whether they had the JAK2-V617F mutation in the hematopoietic lineage (Figure 2B).
The spleen microenvironment of the patients with PMF has higher levels of TGF-b43 and increased microvessel density.14 A recent study showed that SECs of patients with PMF can express the JAK2-V617F mutation.11 This prompted us to also analyze the endothelial cells of splenic origin (ie, SECs), taking advantage of a murine SEC line.
Endothelial cells from the mouse spleen microvascula- ture were isolated, immortalized, and analyzed in terms of their morphologic structure and junctional marker expression. These murine SECs had the typical cobble- stone structure and the expected expression and distribu- tion of endothelial markers (ie, VE-cadherin, JAM-A, platelet endothelial cell adhesion molecule) (Figure 3 and Supplemental Figure S2). When these SECs were stimu- lated with TGF-b, IL-1b, or TNF-a, they acquired a spindle-shaped morphologic characteristic (Figure 3A) and showed partial down-regulation of VE-cadherin and up- regulation of N-cadherin, at both the transcriptional and protein levels (Figure 3B).
In the IF analysis, in endothelial cells activated by the cocktail of cytokines, N-cadherin was up-regulated and distributed at cell-to-cell contacts, but this was irregular and was not the case for all of the cells (Supplemental Figure S2). At this point, VE-cadherin was still expressed and present at the intercellular junctions. a-Catenin, b-catenin, g-catenin/plakoglobin, and p120 were also seen at the intercellular junctions. In the IP analysis followed by WB analysis (Supplemental Figure S3), N-cadherin was seen to associate with a-catenin, b-catenin, g-catenin/
plakoglobin, and p120, although the relative levels of the intracellular partners linked to N-cadherin was lower than to VE-cadherin.
Other junctional adhesion proteins, such as JAM-A and platelet endothelial cell adhesion molecule, were organized as expected at the intercellular junctions, in both the
control and cytokine-treated cells undergoing EndMT (Supplemental Figure S2).
The acquisition of the EndMT phenotype was further supported by down-regulation of endothelial markers, such as claudin5, and up-regulation of a-smooth muscle actin (a-SMA), fibronectin, Fsp1, collagen I, serpine1, Snail, BMP6, and CD44 (Figure 3C). The induction of CD44 and fibronectin was further validated by WB analysis (Figure 3C). Interestingly, among the genes that were up-regulated by the inflammatory cytokines, there was a significant increase in the BMP6 transcript (Figure 3C).
The BMPs are a group of heterogeneous ligands that belong to the TGF-b superfamily, and their aberrant expression is seen for vascular disorders.28BMP6 addition to SECs triggered the phosphorylation of Smad1 and Smad 3 (Figure 3E) and down-regulated the expression of endothelial-specific molecules, such as claudin5, with concurrent up-regulation of the EndMT mesenchymal markers a-SMA, fibronectin, Fsp11, collagen I, serpine1, and Snail (Figure 3D).
Overall, these data show that both human endothelial colony-forming cells of patients with PMF and murine SECs can undergo mesenchymal transition on activation with inflammatory cytokines. This effect is sustained through up-regulation of BMP6.
Megakaryocytes and Platelets Can Induce EndMT
To investigate the source of the inflammatory cytokines that can induce EndMT in vivo, the localization of megakaryo- cytes in the spleen of healthy donors and patients with PMF was determined. PMF megakaryocytes are characterized by dysplastic hyperplasia and by high levels of TGF-b.41,44,45 Only splenic tissue from the patients with PMF, and not from the healthy donors, showed cells that were positive for the megakaryocyte marker CD41a (Figure 4, A and B).
In the spleen sections from patients with PMF, endothelial cells that were double-positive for VE-cadherin and the mesenchymal marker Fsp1 were in close proximity to mega- karyocytes (Figure 4C). In contrast, in the healthy donors, megakaryocytes were low or absent, and the vessels did not stain for Fsp1 (Figure 4C). These data suggest that megakar- yocytes in spleens from patients with PMF can induce EndMT, as a close and potent source of TGF-b and other cytokines.
To determine whether inflammatory cytokines that are released from a-granules of both megakaryocytes and platelets, and in particular TGF-b, can indeed trigger EndMT, murine megakaryocytes and platelets were isolated from fetal liver and peripheral blood, respectively. An in vitro release reaction was then induced in the megakaryocytes (Figure 4, D and E) and in the platelet suspensions (Figure 4, F and G) to release the content of their granules, as described previ- ously.46 The resulting supernatant fluids and pellets were incubated with the SECs and were seen to activate the TGF-b signaling pathway, with induction of phosphorylation of both Smad1 and Smad3 (Figure 4, D and F) and
up-regulation of EndMT markers, such as a-SMA, BMP6, fibronectin, and serpine1 (Figure 4, E and G). To further understand the relevance of TGF-b in this system, SECs were incubated with the megakaryocyte supernatant fluid in the presence of an antieTGF-b blocking antibody, which strongly inhibited phosphorylation of the Smads and expression of the typical EndMT markers (Supplemental Figure S4).
Two available inhibitors were then tested: LY-2109761, as a paneTGF-b inhibitor of TbRI and TbRII kinases,47and DMH1, which acts on BMP signaling.28 On incubation of SECs with platelet-derived supernatantfluid and pellet, and in contrast to LY-2109761, DMH1 had little effect on phosphorylation of Smad1 and Smad3 (Figure 5A), but it significantly down-regulated the EndMT markers (Figure 5B). A possible explanation for this discrepancy is that during EndMT, the endothelial cells can up-regulate BMP6 (Figure 4D), which in turn will contribute to the
up-regulation of the EndMT markers (Figure 3D). DMH1 therefore might be more active on the slowly progressing up-regulation of the EndMT markers that is induced by BMP6 than on the early activation of the Smads.
It is therefore likely that megakaryocyte-derived and platelet-derived TGF-b in combination with endogenous endothelial BMP6 induces the EndMT switch observed in the bone marrow and spleen of patients with PMF.
EndMT Is an Early Event in the Development of Fibrosis in a Mouse Model of Myeloproliferative Disease
The JAK2V617F/WTKI murine model of disease31 was also studied, where the V617F mutation is expressed under heterozygosity in the hematopoietic compartment specif- ically, and where a myeloproliferative-like phenotype de- velops. To determine whether EndMT is an early event in fibrosis development, bone marrow and spleen sections
Figure 4 Megakaryocytes (MKs) are detected only in spleens of patients with primary myelofibrosis (PMF) and are in proximity of vessels undergoing endothelial-to-mesenchymal transition (EndMT) after the release of transforming growth factor (TGF)-b. A: Repre- sentative immunofluorescence images of spleen sections from healthy donors and patients with PMF for MKs stained with CD41a (red). DAPI staining (blue) shows nuclei. B: Quantification of total MKs perfield analyzed. C: Representative images for healthy donors and patients with PMF for expression of VE-cadherin (red), Fsp1 (green), and CD41a (purple). Arrowheads indicate vessels coex- pressing VE-cadherin and Fsp1; arrows, MKs. DAPI staining (blue) shows nuclei. DeG: Supernatant (SN) fluids and pellets obtained from MKs (D and E) or platelets (Plts; F and G) in vitro release re- action were added to SECs for 45 minutes. D and F: Representative Western blot analysis for phosphorylation of Smad1 (pSer463/465) and Smad3 (pSer423/425) and total Smad1 and Smad3 levels. Vin- culin was used to monitor sample loading. E and G: Quantification of RT-PCR analysis for SECs treated with SN or pellet from MKs (E) or Plts (G) for 40 hours to evaluate a-SMA, Bmp6,fibronectin (FN), and serpine1 expression. For all RT-PCR, mRNA is expressed as 2DCt, as normalized to 18 seconds. Data are expressed as means SD. n Z 2 healthy CTRL donors (A and C); nZ 6 patients with PMF (A and C);
nZ 3 independent experiments for Western blot analysis (E and G).
*P< 0.05, **P < 0.01. Scale bars: 50 mm (A); 30 mm (C). a-SMA, a-smooth muscle actin; Bmp6, bone morphogenetic protein 6; CTRL, control; Fsp1,fibroblast-specific protein 1; JAK2, Janus kinase 2;
p-, phospho-; SEC, spleen endothelial cell;VE, vascular endothelial cadherin; wt, wild-type.
from the JAK2V617F/WTKI mice were analyzed at the initial stage of the disease (ie, 2 months), to determine whether there were mesenchymal markers in endothelial cells before the apparentfibrosis.
Hematoxylin and eosin staining revealed increased numbers of megakaryocytes in bone marrow sections of the mutated mice (Figure 6A), which resembled the PMF pathologic process. Both the size and weight of the spleen of the mutated mice were increased, and there was clear loss of the architecture of the organ, with no further distinction between the red and white pulp (Figure 6A). Conversely, the morphologic characteristics of the liver and lungs were comparable between these mutated and wild-type mice.
Overall, this murine model reproduces the human myelo- proliferative neoplasm disease relatively accurately. IF analysis of the JAK2V617F/WT KI mutant showed that the bone marrowederived and spleen-derived endothelial cells coexpressed VE-cadherin and the mesenchymal markers Fsp1 and CD44 (Figure 6B), which was not the case for the JAK2FLEX/WT KI (wild-type) mice. This confirmed the ob- servations reported above for the spleen sections of the patients with PMF (Figure 1B).
Of note, up-regulation of Fsp1 and CD44 was already seen at the very early stages of disease development and before the detection of obvious organfibrosis. This suggests that EndMT transition might be an early event in, and a possible trigger of,fibrosis.
Overall, the data reported here show for thefirst time that there is a functional change in the microvascular endothelial cells toward a mesenchymal phenotype in the bone marrow and spleen of patients with PMF. The acquisition of these mesenchymal characteristics by endothelial cells is an early event that appears to be induced by the inflammatory
microenvironment where the vessels are located. The EndMT switch of the endothelial cells might contribute to the overall development of thefibrosis.
Discussion
PMF is a myeloproliferative disease that is characterized by abnormal proliferation of the myeloid lineages and exces- sive production of proinflammatory cytokines, which lead to the replacement of the hematopoietic environment with a fibrotic stroma. The fibrotic stroma is an important clinical problem, and intervention to limit fibrosis has so far been only palliative. Patients with PMF also experience neo- angiogenesis in their bone marrow and spleen tissues and show significant increases in circulating EPCs, which sug- gests that an alteration in endothelial cell homeostasis might contribute to the pathologic phenotype.
In the present study, we have introduced a novel patho- genetic mechanism that can act at the onset and during progression of bone marrow fibrosis in PMF. We have shown that endothelial cells of the microcirculation of the bone marrow and spleen of patients with PMF can undergo EndMT, thus acquiring fibroblastoid characteristics. This process of EndMT is important in embryogenesis20e22; however, in the adult it is associated tofibrotic degeneration in different organs, including the heart, lungs, and kidney.23e25
In both physiological and pathologic situations, EndMT is mediated by inflammatory cytokines, and in particular by TGF-b and the BMPs.40 It was recently reported that in PMF megakaryocytes constitutively produce and release high levels of bioactive TGF-b.41,45 We confirmed these Figure 5 Transforming growth factor (TGF)-b/bone morphogenetic protein (BMP) inhibitor treatment inhibits phosphorylation of Smads and down- regulates EndMT markers in spleen endothelial cells. A: Representative Western blot analysis for p-Smad1 (pSer463/465), p-Smad3 (pSer423/425), and total Smad1 and Smad3 levels. SECs were incubated for 45 minutes with supernatant (SN) and pellets derived from platelet release reaction in presence or absence of 10 mmol/L LY-2109761 or 2 mmol/L DMH1, added to cells 24 hours before the stimuli. Vinculin was used to verify sample loading. B: RT-PCR analysis for a-SMA,fibronectin (FN), serpine1, Snail, and N-cadherin expression in SECs treated for 40 hours with SN and pellets derived from platelet release reaction in the absence (vehicle) or presence of 10 mmol/L LY-2109761 or 2 mmol/L DMH1 added to the cells 24 hours before the stimuli. The mRNA is expressed as 2DCt, as normalized to 18 seconds. Data are expressed as means SD. n Z 4 independent experiments. *P < 0.05, **P < 0.01. a-SMA, a-smooth muscle actin; DMH1, dorsomorphin homolog 1; DMSO, dimethyl sulfoxide; EndMT, endothelial-to-mesenchymal transition; p-, phospho-; SEC, spleen endothelial cell.
observations in the spleen of these patients with PMF, where TGF-b localized in the stroma and Smad activation in the endothelial cells. We have also shown that the EndMT switch of endothelial cells is localized in regions where the vessels are contiguous to megakaryocytes. A direct incu- bation of SECs with megakaryocytes or platelet supernatant fluid and pellet induced EndMT, which was abrogated by TGF-b inhibitors. Other factors in the bone marrow stroma that can be released by megakaryocytes (such as IL-1a) can also contribute to EndMT and act synergistically with TGF-b.
Endogenous BMP6 is elevated in SECs undergoing EndMT, and in a previous study we demonstrated that this cytokine can significantly contribute to the EndMT switch.28 The involvement of the BMP subfamily in the
development of PMF is not new; high levels of BMP1, BMP6, and BMP7 and BMP-receptor 2 mRNAs have been reported for bone marrow biopsies from patients with PMF in the advanced stages of the disease,48 and BMPs have been suggested to have roles in sustaining the progression of fibrosis. These observations suggest that inhibitors of the Smad pathway that act downstream TGF-b or BMP6 might be effective in inhibiting or reverting EndMT and possibly in reducingfibrosis in vivo.
These data showed no significant differences in expres- sion of the EndMT markers in the vessels of patients car- rying the JAK2-V617F mutation in the hematopoietic lineage. Infection of endothelial cells in vitro with a vector expressing the JAK2 mutant also did not change their EndMT or their sensitivity to inflammatory stimuli (data not Figure 6 Bone marrow and splenic endothelial cells of JAK2V617F/WTKI mouse model expressfibroblast-specific protein 1 (Fsp1) and CD44 before showing the clearfibrotic phenotype. A: Representative H&E staining of bone marrow, spleen, liver, and lung tissue sections from JAK2FLEX/WTKI and JAK2V617F/WTKI mice at 2 months of age. Black arrows show megakaryocytes in bone marrow tissue. B: Representative immunofluorescence staining for vascular endothelial (VE)-cadherin (red) and Fsp1 (green) or CD44 (green) of bone marrow and spleen sections of JAK2FLEX/WTKI and JAK2V617F/WTKI mice at 2 months of age.
White arrows indicate costaining for VE-cadherin and Fsp1 or CD44. DAPI staining (blue) shows nuclei. nZ 4 JAK2FLEX/WTKI; nZ 3 JAK2V617F/WTKI. Scale bars: 50 mm (A, bone marrow); 200 mm (A, spleen); 500 mm (A, liver and lungs); 30 mm (B). H&E, hematoxylin and eosin.
shown). Thus, the proinflammatory microenvironment that is generated by deregulated cytokine production by he- matopoietic cells might have a more prominent role in EndMT development in patients with PMF, rather than the presence of the JAK2-V617F mutation in the endothelial cells.
EndMT is a relatively common reaction of different types of endothelial cells when in contact with an inflammatory microenvironment, as shown here and elsewhere.28,36 Although it is well known that megakaryocytes are pro- duced and differentiate in the bone marrow to originate platelets,49 we observed that they are present also in the spleen of patients with PMF and that they can contribute to EndMT also in this organ. Splenic fibrosis is not usually observed for patients with PMF. In the study of the mouse model of myeloproliferation here, we saw that they devel- oped splenicfibrosis quite early in the kinetics of evolution of the disease. This effect paralleled the expression of the EndMT markers also at very early stages of fibrotic degeneration. The apparent difference between the mouse model and patients with PMF might be due to the different histologic features of the splenomegaly. In mice, the evo- lution of the disease is particularly rapid, and splenomegaly is coupled to marked splenicfibrosis that can be seen from the early stages of the disease. In patients with PMF, the excessive splenomegaly is due to extramedullary hemato- poiesis,6,50and it develops before there are signs offibrosis, which might, however, have a later evolution. Indeed, we have also seen signs of EndMT in the spleen of some patients with prefibrotic PMF (eg, patients 3 and 5) (Supplemental Table S1 and Figure 1B), which further underlines the early appearance of this phenomenon in the progression of the disease.
In addition, the JAK2 mutation is not present in all of the patients with PMF, and other relevant mutations have been detected in association with this disease.5The animal model therefore reproduces only in part the most common features of PMF.
Although the role of EndMT might be to contribute to fibrosis, an alternative possibility is that the subset of endothelial cells that expresses the mesenchymal markers also contributes to the onset of the disease. The close as- sociation of endothelial cells with megakaryocytes suggests that there might be cross talk between these two cell types.
Endothelial cells that are undergoing EndMT might produce growth and differentiating factors that might contribute, in turn, to the altered functional behavior of megakaryocytes.
The gene forfibronectin was among those that were strongly up-regulated by the megakaryocyte and platelet release re- actions. Fibronectin regulates megakaryocyte interactions with the bone marrow niche to modulate expansion, matu- ration, and pro-platelet formation.51,52 Moreover, a recent study demonstrated that when bone marrow hematopoietic stem cells are cultured in an in vitro system in the presence of different extracellular matrix components, only fibro- nectin can increase cell survival and proliferation of the
bone marrow precursors, and in the presence of thrombo- poietin, fibronectin strongly increases the pool of mega- karyocyte progenitors.53 Therefore, it is tempting to speculate that in PMF the cross talk between endothelial cells and megakaryocytes gives rise to a pathologic autoefeed-forward process: the TGF-b produced by the megakaryocytes is sensed by the endothelial cells, which in response undergo EndMT and strongly up-regulate the expression offibronectin or other growth factors. This will continuously feed the megakaryocytes expansion and maturation, thus fostering myelofibrosis. Further and deeper analysis of the endothelial cellemegakaryocyte niche is needed to define this concept in more detail.
Acknowledgments
B.G.E., C.G., and M.C. performed the experiments and analyzed the data; B.G.E. and E.D. wrote the manuscript;
F.P. performed the immunohistochemical analysis of the human and mouse samples; V.R. and G.B. isolated and cultured the endothelial colony-forming cells from human samples and provided human bone marrow and spleen from patients with PMF; J.-L.V. developed the transgenic mice;
N.B. and A.M.V. provided transgenic mice and bone marrow and spleen samples; C.G. and A.B. designed and contributed to the experiments on megakaryocytes and to the overall strategy of the manuscript; E.D. designed the project and the experiments, discussed the data, and wrote the manuscript.
Supplemental Data
Supplemental material for this article can be found at http://dx.doi.org/10.1016/j.ajpath.2017.04.006.
References
1. Dameshek W: Some speculations on the myeloproliferative syn- dromes. Blood 1951, 6:372e375
2. Delhommeau F, Jeziorowska D, Marzac C, Casadevall N: Molecular aspects of myeloproliferative neoplasms. Int J Hematol 2010, 91:
165e173
3. Levine RL, Gilliland DG: JAK-2 mutations and their relevance to myeloproliferative disease. Curr Opin Hematol 2007, 14:43e47 4. Rumi E, Pietra D, Pascutto C, Guglielmelli P, Martinez-Trillos A,
Casetti I, Colomer D, Pieri L, Pratcorona M, Rotunno G, Sant’Antonio E, Bellini M, Cavalloni C, Mannarelli C, Milanesi C, Boveri E, Ferretti V, Astori C, Rosti V, Cervantes F, Barosi G, Vannucchi AM, Cazzola M; Associazione Italiana per la Ricerca sul Cancro Gruppo Italiano Malattie Mieloproliferative Investigators:
Clinical effect of driver mutations of JAK2, CALR, or MPL in pri- mary myelofibrosis. Blood 2014, 124:1062e1069
5. Vainchenker W, Kralovics R: Genetic basis and molecular patho- physiology of classical myeloproliferative neoplasms. Blood 2017, 129:667e679
6. Tefferi A: Myelofibrosis with myeloid metaplasia. N Engl J Med 2000, 342:1255e1265
7. Levine RL, Wadleigh M, Cools J, Ebert BL, Wernig G, Huntly BJ, Boggon TJ, Wlodarska I, Clark JJ, Moore S, Adelsperger J, Koo S, Lee JC, Gabriel S, Mercher T, D’Andrea A, Frohling S, Dohner K, Marynen P, Vandenberghe P, Mesa RA, Tefferi A, Griffin JD, Eck MJ, Sellers WR, Meyerson M, Golub TR, Lee SJ, Gilliland DG:
Activating mutation in the tyrosine kinase JAK2 in polycythemia vera, essential thrombocythemia, and myeloid metaplasia with myelofibrosis. Cancer Cell 2005, 7:387e397
8. Ballen KK, Shrestha S, Sobocinski KA, Zhang MJ, Bashey A, Bolwell BJ, Cervantes F, Devine SM, Gale RP, Gupta V, Hahn TE, Hogan WJ, Kroger N, Litzow MR, Marks DI, Maziarz RT, McCarthy PL, Schiller G, Schouten HC, Roy V, Wiernik PH, Horowitz MM, Giralt SA, Arora M: Outcome of transplantation for myelofibrosis. Biol Blood Marrow Transplant 2010, 16:358e367
9. Massa M, Rosti V, Ramajoli I, Campanelli R, Pecci A, Viarengo G, Meli V, Marchetti M, Hoffman R, Barosi G: Circulating CD34þ, CD133þ, and vascular endothelial growth factor receptor 2-positive endothelial progenitor cells in myelofibrosis with myeloid meta- plasia. J Clin Oncol 2005, 23:5688e5695
10. Rosti V, Bonetti E, Bergamaschi G, Campanelli R, Guglielmelli P, Maestri M, Magrini U, Massa M, Tinelli C, Viarengo G, Villani L, Primignani M, Vannucchi AM, Frassoni F, Barosi G; AGIMM In- vestigators: High frequency of endothelial colony forming cells marks a non-active myeloproliferative neoplasm with high risk of splanchnic vein thrombosis. PLoS One 2010, 5:e15277
11. Rosti V, Villani L, Riboni R, Poletto V, Bonetti E, Tozzi L, Bergamaschi G, Catarsi P, Dallera E, Novara F, Massa M, Campanelli R, Fois G, Peruzzi B, Lucioni M, Guglielmelli P, Pancrazzi A, Fiandrino G, Zuffardi O, Magrini U, Paulli M, Vannucchi AM, Barosi G; Associazione Italiana per la Ricerca sul Cancro Gruppo Italiano Malattie Mieloproliferative (AGIMM) in- vestigators: Spleen endothelial cells from patients with myelofibrosis harbor the JAK2V617F mutation. Blood 2013, 121:360e368 12. Mesa RA, Hanson CA, Rajkumar SV, Schroeder G, Tefferi A:
Evaluation and clinical correlations of bone marrow angiogenesis in myelofibrosis with myeloid metaplasia. Blood 2000, 96:3374e3380 13. Ni H, Barosi G, Hoffman R: Quantitative evaluation of bone marrow angiogenesis in idiopathic myelofibrosis. Am J Clin Pathol 2006, 126:241e247
14. Barosi G, Rosti V, Massa M, Viarengo GL, Pecci A, Necchi V, Ramaioli I, Campanelli R, Marchetti M, Bazzan M, Magrini U:
Spleen neoangiogenesis in patients with myelofibrosis with myeloid metaplasia. Br J Haematol 2004, 124:618e625
15. Di Raimondo F, Azzaro MP, Palumbo GA, Bagnato S, Stagno F, Giustolisi GM, Cacciola E, Sortino G, Guglielmo P, Giustolisi R:
Elevated vascular endothelial growth factor (VEGF) serum levels in idiopathic myelofibrosis. Leukemia 2001, 15:976e980
16. Martyre MC, Le Bousse-Kerdiles MC, Romquin N, Chevillard S, Praloran V, Demory JL, Dupriez B: Elevated levels of basic fibroblast growth factor in megakaryocytes and platelets from pa- tients with idiopathic myelofibrosis. Br J Haematol 1997, 97:
441e448
17. Chen PY, Qin L, Tellides G, Simons M: Fibroblast growth factor receptor 1 is a key inhibitor of TGF-b signaling in the endothelium.
Sci Signal 2014, 7:ra90
18. Kalluri R, Weinberg RA: The basics of epithelial-mesenchymal transition. J Clin Invest 2009, 119:1420e1428
19. Zeisberg EM, Potenta S, Xie L, Zeisberg M, Kalluri R: Discovery of endothelial to mesenchymal transition as a source for carcinoma- associatedfibroblasts. Cancer Res 2007, 67:10123e10128 20. Armstrong EJ, Bischoff J: Heart valve development: endothelial cell
signaling and differentiation. Circ Res 2004, 95:459e470
21. Liebner S, Cattelino A, Gallini R, Rudini N, Iurlaro M, Piccolo S, Dejana E: b-catenin is required for endothelial-mesenchymal trans- formation during heart cushion development in the mouse. J Cell Biol 2004, 166:359e367
22. MacGrogan D, Luna-Zurita L, de la Pompa JL: Notch signaling in cardiac valve development and disease. Birth Defects Res A Clin Mol Teratol 2011, 91:449e459
23. Zeisberg EM, Tarnavski O, Zeisberg M, Dorfman AL, McMullen JR, Gustafsson E, Chandraker A, Yuan X, Pu WT, Roberts AB, Neilson EG, Sayegh MH, Izumo S, Kalluri R: Endothelial-to- mesenchymal transition contributes to cardiacfibrosis. Nat Med 2007, 13:952e961
24. Zeisberg EM, Potenta SE, Sugimoto H, Zeisberg M, Kalluri R: Fi- broblasts in kidneyfibrosis emerge via endothelial-to-mesenchymal transition. J Am Soc Nephrol 2008, 19:2282e2287
25. Potenta S, Zeisberg E, Kalluri R: The role of endothelial-to- mesenchymal transition in cancer progression. Br J Cancer 2008, 99:1375e1379
26. Camenisch TD, Molin DG, Person A, Runyan RB, Gittenberger-de Groot AC, McDonald JA, Klewer SE: Temporal and distinct TGF-b ligand requirements during mouse and avian endocardial cushion morphogenesis. Dev Biol 2002, 248:170e181
27. Medici D, Potenta S, Kalluri R: Transforming growth factor-b2 promotes Snail-mediated endothelial-mesenchymal transition through convergence of Smad-dependent and Smad-independent signalling.
Biochem J 2011, 437:515e520
28. Maddaluno L, Rudini N, Cuttano R, Bravi L, Giampietro C, Corada M, Ferrarini L, Orsenigo F, Papa E, Boulday G, Tournier- Lasserve E, Chapon F, Richichi C, Retta SF, Lampugnani MG, Dejana E: EndMT contributes to the onset and progression of cerebral cavernous malformations. Nature 2013, 498:492e496
29. Thiele J, Kvasnicka HM: The 2008 WHO diagnostic criteria for polycythemia vera, essential thrombocythemia, and primary myelo- fibrosis. Curr Hematol Malig Rep 2009, 4:33e40
30. Barosi G, Ambrosetti A, Finelli C, Grossi A, Leoni P, Liberato NL, Petti MC, Pogliani E, Ricetti M, Rupoli S, Visani G, Tura S: The Italian consensus conference on diagnostic criteria for myelofibrosis with myeloid metaplasia. Br J Haematol 1999, 104:730e737 31. Hasan S, Lacout C, Marty C, Cuingnet M, Solary E, Vainchenker W,
Villeval JL: JAK2V617F expression in mice amplifies early he- matopoietic cells and gives them a competitive advantage that is hampered by IFNalpha. Blood 2013, 122:1464e1477
32. Bartalucci N, Tozzi L, Bogani C, Martinelli S, Rotunno G, Villeval JL, Vannucchi AM: Co-targeting the PI3K/mTOR and JAK2 signalling pathways produces synergistic activity against myelopro- liferative neoplasms. J Cell Mol Med 2013, 17:1385e1396 33. Balconi G, Spagnuolo R, Dejana E: Development of endothelial cell
lines from embryonic stem cells: a tool for studying genetically manipulated endothelial cells in vitro. Arterioscler Thromb Vasc Biol 2000, 20:1443e1451
34. Garlanda C, Parravicini C, Sironi M, De Rossi M, Wainstok de Calmanovici R, Carozzi F, Bussolino F, Colotta F, Mantovani A, Vecchi A: Progressive growth in immunodeficient mice and host cell recruitment by mouse endothelial cells transformed by polyoma middle-sized T antigen: implications for the pathogenesis of oppor- tunistic vascular tumors. Proc Natl Acad Sci U S A 1994, 91:
7291e7295
35. Ingram DA, Mead LE, Tanaka H, Meade V, Fenoglio A, Mortell K, Pollok K, Ferkowicz MJ, Gilley D, Yoder MC: Identification of a novel hierarchy of endothelial progenitor cells using human periph- eral and umbilical cord blood. Blood 2004, 104:2752e2760 36. Rieder F, Kessler SP, West GA, Bhilocha S, de la Motte C,
Sadler TM, Gopalan B, Stylianou E, Fiocchi C: Inflammation- induced endothelial-to-mesenchymal transition: a novel mechanism of intestinalfibrosis. Am J Pathol 2011, 179:2660e2673
37. Drachman JG, Sabath DF, Fox NE, Kaushansky K: Thrombopoietin signal transduction in purified murine megakaryocytes. Blood 1997, 89:483e492
38. Ruggeri ZM, Orje JN, Habermann R, Federici AB, Reininger AJ:
Activation-independent platelet adhesion and aggregation under elevated shear stress. Blood 2006, 108:1903e1910
39. Miettinen PJ, Ebner R, Lopez AR, Derynck R: TGF-b-induced transdifferentiation of mammary epithelial cells to mesenchymal cells:
involvement of type I receptors. J Cell Biol 1994, 127:2021e2036 40. Gonzalez DM, Medici D: Signaling mechanisms of the epithelial-
mesenchymal transition. Sci Signal 2014, 7:re8
41. Ciurea SO, Merchant D, Mahmud N, Ishii T, Zhao Y, Hu W, Bruno E, Barosi G, Xu M, Hoffman R: Pivotal contributions of megakaryocytes to the biology of idiopathic myelofibrosis. Blood 2007, 110:986e993
42. Piaggio G, Rosti V, Corselli M, Bertolotti F, Bergamaschi G, Pozzi S, Imperiale D, Chiavarina B, Bonetti E, Novara F, Sessarego M, Villani L, Garuti A, Massa M, Ghio R, Campanelli R, Bacigalupo A, Pecci A, Viarengo G, Zuffardi O, Frassoni F, Barosi G: Endothelial colony-forming cells from patients with chronic myeloproliferative disorders lack the disease-specific molecular clonality marker. Blood 2009, 114:3127e3130
43. Zingariello M, Martelli F, Ciaffoni F, Masiello F, Ghinassi B, D’Amore E, Massa M, Barosi G, Sancillo L, Li X, Goldberg JD, Rana RA, Migliaccio AR: Characterization of the TGF-b1 signaling abnormalities in the Gata1low mouse model of myelofibrosis. Blood 2013, 121:3345e3363
44. Balduini A, Badalucco S, Pugliano MT, Baev D, De Silvestri A, Cattaneo M, Rosti V, Barosi G: In vitro megakaryocyte differentia- tion and proplatelet formation in Ph-negative classical myeloprolif- erative neoplasms: distinct patterns in the different clinical phenotypes. PLoS One 2011, 6:e21015
45. Badalucco S, Di Buduo CA, Campanelli R, Pallotta I, Catarsi P, Rosti V, Kaplan DL, Barosi G, Massa M, Balduini A: Involvement of
TGF-b1 in autocrine regulation of proplatelet formation in healthy subjects and patients with primary myelofibrosis. Haematologica 2013, 98:514e517
46. Labelle M, Begum S, Hynes RO: Direct signaling between platelets and cancer cells induces an epithelial-mesenchymal-like transition and promotes metastasis. Cancer Cell 2011, 20:576e590
47. Chakrabarti L, Wang BD, Lee NH, Sandler AD: A mechanism linking Id2-TGF[beta] crosstalk to reversible adaptive plasticity in neuroblastoma. PLoS One 2013, 8:e83521
48. Bock O, Hoftmann J, Theophile K, Hussein K, Wiese B, Schlue J, Kreipe H: Bone morphogenetic proteins are overexpressed in the bone marrow of primary myelofibrosis and are apparently induced by fibrogenic cytokines. Am J Pathol 2008, 172:951e960
49. Patel SR, Hartwig JH, Italiano JE Jr: The biogenesis of platelets from megakaryocyte proplatelets. J Clin Invest 2005, 115:3348e3354 50. Barosi G: Myelofibrosis with myeloid metaplasia. Hematol Oncol
Clin North Am 2003, 17:1211e1226
51. Malara A, Gruppi C, Pallotta I, Spedden E, Tenni R, Raspanti M, Kaplan D, Tira ME, Staii C, Balduini A: Extracellular matrix struc- ture and nano-mechanics determine megakaryocyte function. Blood 2011, 118:4449e4453
52. Pallotta I, Lovett M, Rice W, Kaplan DL, Balduini A: Bone marrow osteoblastic niche: a new model to study physiological regulation of megakaryopoiesis. PLoS One 2009, 4:e8359
53. Malara A, Currao M, Gruppi C, Celesti G, Viarengo G, Buracchi C, Laghi L, Kaplan DL, Balduini A: Megakaryocytes contribute to the bone marrow-matrix environment by expressingfibronectin, type IV collagen, and laminin. Stem Cells 2014, 32:926e937
