© 1999 American Heart Association, Inc.
Original Contribution |
Thrombopoietin Stimulates Endothelial Cell Motility and Neoangiogenesis by a Platelet-Activating Factor–Dependent Mechanism
From the Dipartimento di Medicina Interna (M.F.B., P.D., G.G., L.P., G.C.) and Dipartimento di Fisiopatologia Clinica (E.B., L.D.S.), Università di Torino and Dipartimento di Scienze Cliniche e Biologiche, Torino, Italy.
Correspondence to Dr Giovanni Camussi, Cattedra di Nefrologia, Istituto di Nefro-Urologia, Ospedale Maggiore S. Giovanni Battista, Corso Dogliotti, 14, 10126 Torino, Italy. E-mail camussi{at}ddmc.unito.it
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Abstract—In this study, we demonstrate that human umbilical cord vein–derived endothelial cells (HUVECs) expressed c-Mpl, the thrombopoietin (TPO) receptor, and that TPO activates HUVECs in vitro, as indicated by directional migration, synthesis of 1-alkyl-/1-acyl-platelet-activating factor (PAF) and interleukin-8 (IL-8), and phosphorylation of the signal transducers and activators of transcription (STAT) STAT1 and STAT5B. The observation that WEB 2170 and CV3988, 2 structurally unrelated PAF receptor antagonists, prevented the motility of HUVECs induced by TPO suggests a role of PAF as secondary mediator. Moreover, kinetic analysis of TPO-induced tyrosine phosphorylation of STAT demonstrated that STAT5B activation temporally correlated with the synthesis of PAF. PAF, in turn, induced a rapid tyrosine phosphorylation of STAT5B and PAF receptor blockade, by WEB 2170, preventing both TPO- and PAF-mediated STAT5B activation. The in vivo angiogenic effect of TPO, studied in a mouse model of Matrigel implantation, demonstrated that TPO induced a dose-dependent angiogenic response that required the presence of heparin. Moreover, the in vivo angiogenic effect of TPO was inhibited by the PAF receptor antagonist WEB 2170 but not by the anti–basic fibroblast growth factor neutralizing antibody. These results indicate that the effects of TPO are not restricted to cells of hematopoietic lineages, because TPO is able to activate endothelial cells and to induce an angiogenic response in which the recruitment of endothelial cells is mediated by the synthesis of PAF. Moreover, biochemical analysis supports the hypothesis that STAT5B may be involved in the signaling pathway leading to PAF-dependent angiogenesis.
Key Words: hemopoietic growth factor • STAT • platelet-activating factor • signal transduction • neovascularization
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The c-mpl gene encodes a member of the hematopoietic receptor superfamily that shows, in the extracellular domain, high amino acid sequence homology with the erythropoietin receptor and with the common subunit (c) of the interleukin (IL)–3, IL-5, and granulocyte-macrophage colony-stimulating factor (GM-CSF) receptors.1 The identification and the cloning of the c-Mpl ligand demonstrated the involvement of c-mpl in megakaryopoiesis.2 3 Thus it was termed thrombopoietin (TPO).2 3 More recently it has been shown that TPO can directly modulate the response of mature platelets to several stimuli and thereby their homeostatic potential.4 The in vitro and in vivo effects of TPO are mediated by the presence, on the target cells, of c-Mpl. TPO, as the other member of the hematopoietic receptor superfamily, is devoid of intrinsic tyrosine kinase activity. However, biochemical events triggered by ligand-receptor binding are characterized by tyrosine phosphorylation of multiple intracellular molecules,5 including Janus kinase 2 (JAK2) and members of the signal transducers and activators of transcription (STAT).6 7 8
The JAKs are nontransmembrane protein tyrosine kinases that are rapidly tyrosine phosphorylated on ligand binding and play a critical signaling function for cytokine receptors.9 One such function is the activation of STAT proteins, which are latent cytoplasmic proteins that undergo rapid tyrosine phosphorylation and dimerization following cytokine stimulation.9 As a consequence they acquire DNA-binding activity, translocate into the nucleus, bind to specific promoter elements, and control the expression of target genes.9 Several lines of evidence indicate that activation of the JAK/STAT signaling pathway elicits a number of responses, including changes in cell function, growth, differentiation,9 and epithelial tubule formation.10 Moreover, the observation that the JAK/STAT pathway regulates multiple developmental processes in Drosophila, including the formation of ectopic adult wing veins,11 strongly suggests a role of this pathway in modulating new-vessel formation.
Although TPO has been initially defined as a lineage-specific late-acting growth factor, it has also been reported that this factor may exert effects on the erythroid and granulocytic population.2 Indeed, the hematopoietic progenitor cells express c-Mpl, and TPO has the ability to enhance proliferation of early erythroid progenitors and directly affects erythroid differentiation.3 Moreover, it has been recently demonstrated that TPO treatment also induces the functional activation of polymorphonuclear neutrophils.12 These observations suggest that TPO, beyond the maturation of megakaryocytic lineage, serves a spectrum of functions. These pleiotropic effects of TPO are shared by other growth factors such as IL-3, GM-CSF, and stem cell factor, which are known to be involved in regulation of the hematopoietic tissue homeostasis.13
It has been suggested that interaction between endothelial cells and stem cells plays a critical role in hematopoiesis during embryogenesis and possibly in adults.14 Mammalian embryogenesis hematopoiesis is thought to begin in the yolk sac in the blood islands lined by endothelial cells. These endothelial cells are in close proximity to the developing stem cells and may provide the microenvironment required for stem cell proliferation and differentiation.14 Several lines of evidence suggest that the endothelium of yolk sac plays a major role in the induction of hematopoiesis in the early embryo.15 It has also been suggested that stromal cells capable of mediating hematopoiesis in fetal liver and subsequently in bone marrow have endothelial characteristics.15 Therefore, fetal hematopoiesis parallels neoangiogenesis, which has a critical role in the development of the embryo. These observations suggest a bidirectional interaction between endothelium and hematopoietic cells, with the generation of a microenvironment favorable to the development of hematopoietic processes and, conversely, with the production of angiogenic factors by cells of hematopoietic lineage. The neoangiogenesis involves a concerted sequence of events, including directional migration and proliferation of endothelial cells and canalization of solid endothelial cords penetrating in the tissue.16 A number of diffusible angiogenic factors have been recently characterized, including cytokines and mediators produced from the myeloid cell line.16 Some of these angiogenic factors, such as vascular endothelial growth factor (VEGF),17 IL-8,18 and platelet-activating factor (PAF),19 can also be produced by endothelial cells and possess an autocrine modulatory role on neoangiogenesis.
The aim of the present study was to evaluate whether TPO is able to induce angiogenesis. We found that human umbilical cord vein–derived endothelial cells (HUVECs) express c-Mpl and that TPO activates HUVECs in vitro, as indicated by directional migration, tubulogenesis, synthesis of PAF and IL-8, and phosphorylation of STAT1 and STAT5B. In vivo, TPO induced an angiogenic effect in a mouse model in which Matrigel was used for the delivery of mediators. These in vitro and in vivo effects of TPO were inhibited by the PAF-receptor antagonist, WEB 2170, suggesting a role of PAF as secondary mediator in TPO-induced angiogenesis.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Reagents
Matrigel was obtained from Becton Dickinson Labware. Synthetic 1-alkyl-C16 PAF (1-hexadecyl-2-acetyl-sn-glyceryl-3-phosphorylcholine) (1-alkyl-PAF) was obtained from Bachem Feinchemikalien. 1-Acyl-C16 PAF (1-palmitoyl-2-acetyl-sn-glyceryl-3-phosphorylcholine) (1-acyl-PAF) was obtained by acetylation of 1-palmitoyl-sn-glyceryl-3-phosphorylcholine with acetic anhydride and dimethyl-aminopyridine as previously described.20 CV 3988, a structural analog of PAF that acts as specific PAF antagonist21 was from Takeda Chemical Industries. WEB 2170, a triazolodiazepine (hetrazepinoic) with potent and specific PAF–receptor antagonist activity,22 was obtained from Boehringer Ingelheim KG. Diazepam, a benzodiazepine without a significant PAF-inhibitory activity,23 was obtained from Roche S.p.A. Anti-mouse T-cell serum, anti-L3/T4 and anti-Ly2 monoclonal antibodies (mAbs), anti–MAC-1 FITC-conjugated mAb, and the correspondent irrelevant isotypic IgG controls were purchased from Cedarlane. FITC-conjugated anti-rabbit IgG and anti-mouse IgG were obtained from Cappel Laboratories. Rabbit anti-human von Willebrand factor (vWF), rabbit IgG, and anti-rabbit IgG mAb (RG-96) were obtained from Sigma. Rabbit polyclonal Ab anti–basic fibroblast growth factor (bFGF) was from Immunogenex. Silica gel 60F254 thin-layer chromatography (TLC) plates were obtained from Merck. µPorasil HPLC columns were provided from the Millipore chromatographic division (Waters). RPMI 1640 (GIBCO) and bovine calf serum (BCS) were from Hyclone Laboratories. Human TPO, a gift from Genzyme, was used throughout all experiments. Murine versus human TPO cDNA sequences revealed 71% sequence identity2 3 ; however, human TPO was active in an assay using murine megakaryocytic progenitors.2 3 The presence of endotoxin contamination of TPO preparation was tested by the Limulus amebocyte assay, and the concentration was<0.1 ng/mL. Sepharose-protein A and phospholipase C from Bacillus cereus were purchased from Sigma Chemical Co. Nitrocellulose filters, horseradish peroxidase–conjugated protein A, molecular-weight markers, [
-32P]dCTP, and the chemiluminescence
reagent were from Amersham. Poly(dI · dC):poly(dI · dC)
was obtained from Pharmacia.
Antisera
Polyclonal anti–c-Mpl antiserum raised in rabbit immunized with
a glutathione S-transferase (GST)–fusion protein containing
the amino acid sequence corresponding to residues 518 to 541 (GenBank
accession No. M90103) was obtained as previously
described.12 Anti–c-Mpl antiserum was affinity
purified on a column of Sepharose covalently cross-linked with the
GST-fusion protein. Specificity of anti–c-Mpl antiserum was assessed
by competition experiments. One hundred micrograms of affinity-purified
Igs were sequentially preadsorbed (twice) with 1.5 mg of GST-fusion
protein covalently cross-linked with the Sepharose-protein A for 1 hour
at 4°C. The recovered serum samples were used individually to
immunoprecipitate platelet lysates. Polyclonal antibodies to p91
and polyclonal rabbit antiserum to the common ß subunit of
IL-3/GM-CSF/IL-5 receptors were obtained and characterized as
previously described.24 The 4G10 and PY20
anti-phosphotyrosine antibodies were obtained from Upstate
Biotechnology and from Affinity Research Products Ltd,
respectively. Anti-STAT5A (L-20), anti-STAT5B (G-2), and anti-STAT5B
(C-17) antisera were purchased from Santa Cruz Biotechnology,
Inc.
Western Blot Analysis and Immunoprecipitation
Studies
HUVEC monolayers that had been tightly confluent for a minimum
of 24 hours were washed twice with PBS; serum starved for 4 hours at
37°C in endotoxin-free medium M199 containing 1% BSA, PBS (30%
vol/vol), sodium orthovanadate 0.2 mmol/L, and EDTA 1 mmol/L;
and incubated without or with TPO (20 ng/mL) or PAF (10 nmol/L) at
37°C for the time indicated. Protein concentration of HUVEC lysates
obtained as previously described24 was determined by the
Bradford technique, and the protein content of the samples was
normalized to 250 mg/sample by appropriate dilution with lysis buffer.
The samples were then adsorbed by antisera coupled to Sepharose-protein
A, and bound proteins were eluted and processed as previously
described.24
Preparation of Nuclear Extract and Gel Retardation Assay
Nuclear extracts from untreated and from TPO- and PAF-treated
HUVECs were prepared as described by Sadowski and
Gilman.25 The oligonucleotides used that
corresponded to the prolactin-inducible element of the ß-casein
promoter (PIE) were G GGG GGA CTT CTT GGA ATT AAG GGA and G GGG
TCC CTT AAT TCC AAG AAG TCC.26 Those used that
corresponded to the serum-inducible element (SIE) of c-fos were G GGG
CAT TTC CCG TAA ATC and G GGG GAT TTA CGG GAA ATG.24 The
annealed oligonucleotides were labeled by filling in
the overhanging ends with Klenow fragment in the presence of
[-32P]dCTP. Gel retardation reactions were
performed as previously described.24
In Vitro PAF Synthesis by Endothelial Cells
HUVECs were prepared, grown, and characterized as described
previously.27 In standard PAF synthesis assays, HUVECs at
the third or fourth passage, maintained for 24 hours in DMEM without
FCS, were stimulated in 1 mL of Iscove's medium containing 0.25% BSA
for various periods of time with TPO.
In Vitro Endothelial Cell Proliferation and
Migration Assay
Proliferative activity of HUVECs was assayed by direct cell
count as previously described.28 Migration of HUVECs or
CDC-HMEC-1 cells, kindly provided by Francisco J. Candal, was performed
in Boyden's chambers as previously described.29
Endothelial cells that passed across the polycarbonate
filter after addition in the lower compartment of the chamber of
vehicle alone (saline containing 0.25% BSA), PAF (10 nmol/L), or TPO
(1 to 10 ng/mL) were counted at 200 magnification. The involvement of
PAF-specific receptors in endothelial cell migration
was evaluated by endothelial-cell treatment with 3
µmol/L WEB 217022 or 5 µmol/L
CV3988,21 which are 2 chemically unrelated specific
PAF-receptor antagonists.
In vitro angiogenesis was studied on Matrigel-coated surface as described by Montesano and Orci.30 TPO was either incorporated in Matrigel or added to the medium. To evaluate cell viability, trypan blue exclusion staining was performed at the end of the experiments.
Murine Angiogenesis Assay
Female C57 mice were used at 6 to 8 weeks of age. Matrigel (8.13
mg/mL), in liquid form at 4°C, was mixed with the experimental
substances and injected (0.5 mL) into the abdominal subcutaneous tissue
of mice along the peritoneal midline.31 32 At various
times mice were killed and gels were recovered and processed for
histology and immunohistochemistry as previously
described.19 Vessel area and total Matrigel area were
planimetrically assessed from stained sections, as described by Kibbey
et al.33 We considered vessels only those structures
possessing a patent lumen and containing red blood cells. Results were
expressed as percentage±SEM of the vessel area relative to the total
Matrigel area. Animal procedures conformed to the Guide for the
Care and Use of Laboratory Animals (NIH publication No. 93-23,
revised 1985).
In Vivo Experimental Protocol
The angiogenic effect of various doses of TPO (1, 5, 10, 15, and
40 ng) in 0.5 mL of Matrigel was studied at different times and in the
absence or presence of 10 U/mL heparin. In selected experiments, the
effect of WEB 2170, a hetrazepinoic benzodiazepine with specific
PAF-receptor antagonist activity
(IC50=0.3 µmol/L on in vitro platelet
aggregation),22 on TPO-induced angiogenesis was evaluated.
WEB 2170 was included in the Matrigel plug (final concentration, 250
ng/mL) and injected intraperitoneally (10 mg/kg) 30
minutes before the subcutaneous injection and daily for 6 days.
Diazepam, a nonhetrazepinoic benzodiazepine with absent or minimal
activity as a PAF antagonist when tested on in vitro
platelet aggregation (IC50=260
µmol/L),23 was used instead of WEB 2170 as control. To
evaluate whether the angiogenic effect of TPO may depend on local
production of bFGF, 10 ng/mL of anti-bFGF neutralizing antibody
were added into Matrigel before injection. In preliminary experiments,
this dose of antibody was found to inhibit the angiogenesis induced by
50 ng bFGF.
The direct angiogenic effect of synthetic PAF and of PAF extracted and purified from Matrigel plugs of mice challenged with TPO was also studied.
Assay and Quantification of PAF
PAF extracted and purified by the Matrigel plugs obtained from
mice or from cultured HUVECs was measured by bioassay after extraction
and purification by TLC and HPLC.34 Characterization of
bioactive PAF was performed by comparison with synthetic PAF according
to previously described criteria.34
To study the incorporation of radioactive precursors, 5x105 HUVECs were incubated in 1 mL RPMI 1640 for 30 minutes with 30 mCi [3H]acetate (from Amity) before stimulation.34 Cell extracts were fractionated by TLC on aluminum-sheet silica-gel plates (silica gel 60, F254, 0.2-mm thickness; Merck) using a chloroform/methanol/acetic acid/water solvent (50:25:8:4) by volume. The plates were cut into 1-cm sections, and the radioactivity of each was measured. Radiolabeled PAF was used as a standard. To discriminate between the presence of an ester or ether group at the sn-1 position, the 1-radyl-2-[3H]acetylglycerols, obtained by phospholipase C treatment, were acetylated at position 3 by incubation for 16 hours at 37°C with 0.5 mL of acetic anhydride and 0.1 mL of pyridine.35
Immunofluorescence Study
For the study of TPO-induced STAT5B nuclear translocation,
coverslip-attached HUVECs were fixed for 5 minutes in 3%
paraformaldehyde in PBS (pH 7.4) containing 2% sucrose
and permeabilized with HEPES–Triton X-100 buffer
(20 mmol/L HEPES, pH 7.4, 300 mmol/L sucrose, 50 mmol/L
NaCl, 3 mmol/L MgCl2, and 0.5% Triton
X-100). STAT5B was detected by indirect
immunofluorescence with a mouse monoclonal
anti-STAT5B IgG1 (G-2; Santa Cruz) and a rabbit FITC-conjugated
anti-mouse IgG as secondary antibody (Cedarlane). As control, the
primary antibody was substituted by an irrelevant isotypic control
mouse IgG (Cedarlane).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Expression of TPO Receptor on HUVECs
To evaluate c-Mpl expression by HUVECs, cell lysates from HUVECs or platelets were immunoprecipitated with an anti–c-Mpl antiserum, raised against the intracellular domain of the receptor, or with a preimmune serum. The eluted proteins were subjected to 8% SDS-PAGE, and the filter was immunoblotted with the same antiserum. As shown in Figure 1A
, the
anti–c-Mpl antiserum was able to recognize TPO receptor both in
platelets and in HUVECs immunoprecipitated with the anti–c-Mpl
antiserum but not with the preimmune serum. The specificity of the
anti–c-Mpl antiserum was assessed by immunoabsorption experiments. As
shown in Figure 1B, after 2 steps of absorption the
immunoprecipitated band corresponding to the c-Mpl protein was barely
detectable.
|
To evaluate the specificity of TPO receptor engagement by the ligand,
we studied the ability of TPO to affect the signaling pathway of other
receptors similar to c-Mpl, such as the IL-3 receptor.1
Unstimulated and TPO- or IL-3–stimulated HUVECs were
immunoprecipitated with an antiserum to the ß signal–transducing
subunit of the IL-3 receptor that is also shared by the GM-CSF and the
IL-5 receptors.5 As indicated by the antiphosphotyrosine
immunoblot shown in Figure 2, TPO, unlike IL-3, was unable to induce phosphorylation
of both the IL-3 receptor ß subunit and the previously described
phosphotyrosine-associated proteins.24
|
In Vitro Growth and Migration of HUVECs
To evaluate cell proliferation, HUVECs were stimulated every 2
days with various concentrations of TPO or with bFGF as positive
control. As shown in Table 1, TPO did not
induce any significant increase in cell number, whereas bFGF stimulated
cell growth. In contrast, TPO, added to the lower well of the Boyden's
chamber, induced a dose-dependent migration of HUVECs across the 5-mm
(pore size) gelatin-coated polycarbonate filters (Figure 3A). The extent of migratory response
observed at 10 ng/mL concentration of TPO was similar to that induced
by 10 ng/mL bFGF or 10 nmol/L PAF. When TPO was also added to the upper
well of the Boyden's chamber, it failed to induce a significant
transmembrane migration (number of cell migrated: with 10 ng/mL TPO,
21.55±1.24; 10 nmol/L PAF, 18±3.2; and 10 ng/mL bFGF, 50.2±7.95; in
unstimulated control, 17.22±6.64). Similar results were obtained using
CDC-HMEC-1 microvascular human endothelial cells (data
not shown). As shown in Figure 3B, heparin and polymyxin B did
not affect HUVEC migration. These results indicate that the motogenic
effect of TPO is independent of the synthesis of a heparin-binding
growth factor and that it is not due to endotoxin contamination.
Pretreatment with WEB 2170 and CV 3988, 2 structurally unrelated
specific PAF-receptor antagonists,21 22
inhibited TPO-induced migration (Figure 3B).
Heat-inactivated TPO did not stimulate HUVEC migration.
Moreover, when HUVECs were plated on a Matrigel-coated surface, TPO
induced a rapid formation of cordlike structures of
endothelial cells (Figure 4). Cells move from their initial uniform
pattern of dispersal to form a series of cell clusters joined by long,
multicellular processes, previously defined as a "tube
formation."30 Cordlike structures of
endothelial cells appeared within 1 hour in
TPO-stimulated HUVECs (Figure 4B). A similar extent of cordlike
structures formation was observed both in experiments in which TPO was
incorporated into the Matrigel and in those in which TPO was added to
the incubation medium. In untreated controls spontaneous cordlike
structures of endothelial cells were absent at 1 hour
(Figure 4A) and started to be detectable after 3 hours (data not
shown). Cell viability of endothelial cells was >95%
as evaluated by trypan blue exclusion assay. No changes in cell
viability were observed when cells were stained before and at the end
of experiments. The formation of cordlike structures of
endothelial cells was an energy-requiring mechanism, as
it was abrogated by 10–2 mol/L
NaN3, an inhibitor of the oxidative
phosphorylation and glycolysis (data not shown). The
ability of the microvascular human endothelial cells,
CDC-HMEC-1, to form cordlike structures was also evaluated. As shown in
Figure 4, cordlike structure formation in Matrigel was
present after 1 hour of TPO stimulation (Figure 4D), whereas
it was absent in unstimulated CDC-HMEC-1 cells (Figure 4C).
|
|
|
Synthesis of PAF and IL-8 by TPO-Stimulated HUVECs
Since several cytokines were shown to induce PAF synthesis
by HUVECs, the ability of TPO to stimulate the synthesis of PAF was
studied. As shown in Table 2, HUVECs
synthesized PAF after stimulation with TPO in an amount comparable to
that induced by tumor necrosis factor- (TNF-) and thrombin, but
with a distinct time course. PAF synthesized after TPO stimulation
remained cell associated, being undetectable in the cell-free
supernatant. Using radioactive acetate as a substrate for PAF
synthesis, we found that PAF detected after stimulation with TPO was
newly synthesized. The TLC analysis of lipid fractions
extracted 30 minutes after addition of TPO to HUVECs preincubated with
[3H]acetate demonstrated the presence of 1 main
peak of radioactivity that comigrated with synthetic
[3H]C16-PAF. This peak was absent in the lipid
fractions extracted from unstimulated HUVECs. The
heterogeneity of acetylated
glycerophospholipids produced by HUVECs has been extensively
documented.36 Therefore, to evaluate the molecular species
of PAF produced by TPO-stimulated HUVECs, the
3H-labeled lipids extracted and purified by TLC
were modified into 1-radyl-2,3-diacetyl-glycerols as described
in Materials and Methods. When the labeled
1-radyl-2,3-diacetyl-glycerols (45 500 cpm) were separated by TLC, the
radioactive products appeared in 2 distinct regions of
chromatogram, corresponding to the retention front of
radiolabeled 1-palmitoyl-2,3-diacetylglycerol (peak I;
24 000±650 cpm) and 1-O-hexadecyl-2,3-diacetylglycerol
(peak II; 15 325±720 cpm). The major radiolabeled compound produced
by HUVECs was the 1-acyl derivative of PAF, which is 1000-fold less
active than the alkyl derivative in terms of platelet
activation.36 The compound migrating with the 1-alkyl
derivative of PAF accounted for 
37% of the radiolabeled
1-radyl-2,3-diacetyl-glycerol submitted to TLC.
|
Moreover, we evaluate the release of IL-8 by HUVECs to investigate
whether TPO was also able to stimulate the production of a
polypeptide mediator. TPO induced IL-8 production by HUVECs
with an early release detectable after 2 hours followed by a
production that was sustained up to 24 hours (Figure 5). Heat-inactivated TPO
failed to induce PAF synthesis and IL-8 production.
|
TPO Induces STAT1 and STAT5B but Not STAT5A Tyrosine
Phosphorylation in HUVECs
The involvement of the JAK/STAT pathway in TPO-induced HUVEC
activation was evaluated by the analysis of STAT1, STAT5A, and
STAT5B tyrosine phosphorylation. Unstimulated and
TPO-stimulated HUVECs were immunoprecipitated with an anti-STAT1
antiserum and immunoblotted with an antiphosphotyrosine
antibody. Kinetic analysis of STAT1 tyrosine
phosphorylation, reported in Figure 6A, demonstrated that both tyrosine
phosphorylation and dephosphorylation
of STAT1 occurred very rapidly in TPO-stimulated HUVECs (STAT1 tyrosine
phosphorylation peaked 5 minutes after TPO
stimulation), suggesting a direct effect of TPO in mediating the
activation of this STAT protein. When the same time-course experiments
were performed using a specific antiserum to STAT5A (L-20, Santa Cruz)
or STAT5B (G-2, Santa Cruz), we found that STAT5B, but not STAT5A,
became tyrosine phosphorylated on TPO stimulation
(Figure 6B). By contrast, in experiments not reported, when TPO
was added to the growth factor–dependent M-07e, cells either STAT5A or
STAT5B became tyrosine phosphorylated. Moreover,
kinetic analysis of STAT5B activation demonstrated a delayed
and persistent STAT5B tyrosine phosphorylation
detectable from 30 to 90 minutes (Figure 6B and 6C).
|
TPO-Mediated STAT5B but Not STAT1 Tyrosine
Phosphorylation Is Abrogated by the PAF-Receptor
Antagonist WEB 2170
The delayed tyrosine phosphorylation of STAT5B
observed after TPO stimulation suggested that this effect of TPO might
be indirect. The finding that HUVEC activation is associated with the
neosynthesis of several mediators, including PAF and IL-8, led us to
evaluate the ability of the specific PAF-receptor
antagonist, WEB 2170, or the IL-8 blocking antibody, to
prevent TPO-mediated STAT5B tyrosine phosphorylation.
Unstimulated HUVECs or HUVECs stimulated for 30 minutes with TPO,
pretreated or not pretreated with WEB 2170, were immunoprecipitated
with a specific antiserum to STAT5B (G-2, Santa Cruz) and
immunoblotted with an anti-phosphotyrosine antibody. The
results shown in Figure 7A demonstrate
that tyrosine phosphorylation of STAT5B is completely
abrogated by pretreatment of HUVECs with the PAF-receptor
antagonist WEB 2170. The effect of WEB 2170 was specific
for STAT5B, as it was unable to prevent STAT1 tyrosine
phosphorylation occurring 5 minutes after TPO addition
(Figure 7B). In contrast, the neutralizing anti–IL-8 antibody
did not affect tyrosine phosphorylation of STAT5B
induced by TPO (data not shown).
|
PAF Induces STAT5B but Not STAT5A Tyrosine Phosphorylation
The above results, demonstrating the effect of the PAF-receptor
antagonist, WEB 2170, in preventing TPO-mediated STAT5B
tyrosine phosphorylation, prompted us to investigate
the effect of PAF on STAT5A and STAT5B activation. Unstimulated and
1-alkyl-PAF–stimulated HUVECs were immunoprecipitated with an
antiserum to STAT5A (L-20, Santa Cruz) or STAT5B (G-2, Santa Cruz) and
immunoblotted with an anti-phosphotyrosine antibody. The
results reported in Figure 8 demonstrated
that PAF induced a rapid (peaking after 5 minutes) STAT5B, but not
STAT5A, tyrosine phosphorylation. We also compared the
effect on STAT5B activation of 1-alkyl-PAF with that of 1-acyl- PAF. As
shown in Figure 9, when 1-acyl-PAF was
used, at the same concentration of 1-alkyl-PAF, no tyrosine
phosphorylation of STAT5B was detected.
|
|
TPO- and PAF-Induced Nuclear Translocation of STAT
Proteins
The nuclear translocation of STAT1 and STAT5B was studied by gel
retardation assay, evaluating the formation of SIE or PIE
complexes in nuclear extract from untreated and TPO- or PAF-treated
HUVECs. The results reported in Figure 10A demonstrated the presence of a band
of SIE-binding activity that was specifically blocked by competition
with unlabeled SIE, but not with the unrelated
oligonucleotides corresponding to the PIE sequence.
That STAT1 was present in the TPO-induced DNA-protein complex was
evident from a supershift analysis in which an antibody to
STAT1, but not to STAT3, specifically caused a mobility shift of the
SIE-containing complex (Figure 10B). These data were
consistent with the inability of TPO to induce STAT3 activation
(data not shown). The results reported in Figure 11A demonstrated the presence of a band
of PIE-binding activity in the nuclear extracts from both TPO- and
PAF-treated HUVECs that were completely blocked by adding an excess of
unlabeled PIE. Supershifted experiments demonstrated that the
anti-STAT5B (C-17, Santa Cruz) but not the anti-STAT5A antiserum
altered the mobility shift of the PIE complex, indicating the presence
of STAT5B in this complex (Figure 11B). As control, nuclear
extracts from TPO-stimulated M-07e cells were preincubated with an
anti-STAT5A or an anti-STAT5B antiserum, and the results (not reported)
demonstrated the presence of both proteins in the TPO-induced PIE
complex.
|
|
Nuclear translocation of STAT5B was also evaluated by
immunofluorescence. As shown in Figure 12A, unstimulated HUVECs showed a
peripheral cytoplasmic staining for STAT5B. By contrast, a
nuclear staining of STAT5B was observed after incubation both with TPO
(Figure 12B and 12C) and PAF (Figure 12D), with a time
course that paralleled the corresponding
tyrosine-phosphorylation analysis. Controls
with irrelevant isotypic IgG1, used instead of the anti-STAT5B
monoclonal IgG1, were negative (data not shown).
|
In Vivo Angiogenic Effect of TPO
Figures 13 and 14 show the results of experiments
performed to evaluate the angiogenic effect of TPO in Matrigel
containing 64 U/mL heparin and various concentrations of TPO, TPO plus
WEB 2170, or sterile saline used as vehicle for TPO, which were
injected subcutaneously into mice. The
histological and morphometric analyses of
Matrigel plugs were performed upon euthanizing the mice 6 days later.
TPO induced a dose-dependent angiogenic response that was absent in
controls containing heparin plus saline (Figure 13A).
TPO-induced angiogenesis was significantly reduced in mice treated with
Matrigel containing TPO and WEB 2170 (250 ng/mL) followed by daily
intraperitoneal injection of WEB 2170 (10 mg/kg per
day) (Figure 13B). When diazepam was used instead of WEB
2170, no significant inhibition of TPO-induced angiogenesis was
observed (Figure 13B). In contrast, neutralizing antibody
against the bFGF did not significantly reduce TPO-induced angiogenesis.
In the absence of heparin, TPO induced cell migration within Matrigel
but not formation of canalized vessels (Figures 13B and 14C). The optimal angiogenic effect of TPO required a
concentration of 64 U/mL heparin. The same concentration was required
for the optimal angiogenic effect of bFGF31 or
VEGF.37 The requirement of heparin for the angiogenic
effect of TPO was also indicated by the effect of protamine, which
completely abrogated the angiogenesis induced by 10 ng/mL TPO (Figure 13B). These results suggest that the angiogenic effect of TPO
depends on the production of heparin-binding growth factors.
Indeed, it has been recently reported that TPO stimulates the
production of VEGF.38 The penetration into the
Matrigel of cords of vWF-positive endothelial cells was
observed after 6 and 24 hours of implantation of TPO-containing
Matrigel (data not shown). Canalization of vessels progressively
increased in the following days and reached its maximum at day 6, with
formation of microaneurismatic (Figure 14A) or linear (Figure 14B) structures containing red blood cells and leukocytes,
which were indicative of functional vessels. Sections of the gel were
stained with anti-vWF antibody to confirm the presence of
endothelial cells in association with the vessels
(Figure 14B, inset). Infiltra- tion into the Matrigel of T
lymphocytes (reactive with a pan-T mAb), CD4-positive T lymphocytes
(reactive with anti-L3/T4 mAb), CD8-positive T lymphocytes (reactive
with anti-Ly2 mAb), and polymorphonuclear neutrophils was
never observed. The staining of Matrigel sections for
aspecific esterase or for Mac-1 antigen showed that monocytes
were confined within the lumen of neoformed vessels but were absent
within the Matrigel. As shown in Figures 13B and 14D,
WEB 2170 markedly reduced the neovascularization of Matrigel induced by
TPO. In Matrigel containing as control either 64 U/mL heparin alone or
heparin plus heat-inactivated TPO, angiogenesis was never
observed (data not shown).
|
|
In some experiments Matrigel was submitted to PAF extraction at 6
hours, 24 hours, and 6 days after the beginning of the experiment. PAF
was detected at 6 and 24 hours in mice injected with Matrigel
containing TPO but not in controls (Figure 15). When the animals were injected
with Matrigel containing an amount of synthetic 1-alkyl-PAF comparable
with that extracted from TPO-treated mice (5 ng/mL), an angiogenic
response was observed (28±9.1 mean area percent; 5 experiments). In
contrast, 1-acyl-PAF (5 ng/mL) induced a slight migration of
endothelial cell into the Matrigel but not the
formation of canalized vessels (5.2±1.8 mean area percent; 5
experiments).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The aim of the present study was to evaluate whether endothelium can represent a target for TPO. The results obtained demonstrate that HUVECs express functional TPO receptors and that TPO is able to activate HUVECs in vitro and to stimulate angiogenesis in vivo via the synthesis of PAF.
The formation of new vessels involves a sequence of events including
directional migration and proliferation of endothelial
cells, as well as canalization of solid endothelial
cords penetrating into the tissue.16 A number of soluble
factors that may act with autocrine or paracrine mechanisms have been
found to modulate these events.16 In the present study
we demonstrate that TPO may engage in some of the
physiological events involved in the angiogenesis.
The in vitro study demonstrated that HUVECs express c-Mpl and that
receptor activation by TPO promotes directional migration of HUVECs. In
contrast, no proliferation was observed. Other angiogenic factors, such
as TNF, were shown to promote endothelial cell motility
without a direct stimulation of cell division. This suggests the
involvement of secondary mediators in the full development of the
angiogenic process. Indeed, it has been shown that the angiogenic
effect of TNF is mediated by the production of PAF and nitric
oxide.32 37 In the present study we observed that TPO
stimulated the synthesis of PAF and of IL-8. Despite the fact that both
of these mediators possess angiogenic properties, the experiments with
the PAF-receptor antagonists and with anti–IL-8 blocking
antibody indicate that the directional migration elicited in vitro by
TPO was mainly PAF dependent. Several studies demonstrated that HUVECs
can synthesize various molecular species of PAF with different
biological properties.36 PAF synthesized after TPO
stimulation contained 63% and 37% of 1-acyl-PAF and
1-alkyl-PAF, respectively. The study of the angiogenic potential of
these PAF analogues suggests that the 1-alkyl derivative of PAF mainly
accounts for the angiogenic activity observed in vivo.
The activation of HUVECs by TPO was also supported by biochemical studies indicating the involvement of STAT proteins. The STAT are cytosolic transcriptional factors that are tyrosine phosphorylated following ligand stimulation.9 They form homo- or heterodimers and translocate into the nucleus, where they bind and activate specific DNA sequences.9 Several lines of evidence indicate that TPO treatment both in proliferating cells and in terminally differentiated elements, such as platelets and neutrophils, leads to the activation of the STAT signaling pathway.6 8 Among the STAT proteins STAT1, STAT3, and STAT5 seem to be strictly related to TPO-mediated receptor activation.6 7 8 Our results demonstrate a rapid and transient tyrosine phosphorylation of STAT1 with a consequent formation of an SIE complex on TPO treatment. Unlike other STAT proteins, STAT5 consists of 2 proteins, which are referred to as STAT5A and STAT5B.9 39 These 2 proteins are encoded from 2 distinct yet highly homologous genes and can be differently activated by a wide range of cytokines.9 We found that TPO-mediated HUVEC activation is associated with STAT5B but not with STAT5A tyrosine phosphorylation. These data are consistent with the observed nuclear translocation of STAT5B and with the formation of a PIE complex containing STAT5B only. Similar patterns of STAT activation are triggered by a number of cytokine receptors sharing structural features with c-Mpl, including erythropoietin, GM-CSF, IL-3, and IL-5 receptors.1 Moreover, it has been recently reported that activation of c-Mpl in an erythroleukemia cell line may affect the signaling pathway of the GM-CSF/IL-3/IL-5 receptors via the phosphorylation of their common ß subunit.40 However, our finding that TPO was unable to activate the signaling through the common ß subunit suggests that STAT5 activation as well as the biological response of HUVECs to TPO is specifically triggered by TPO receptor engagement.
The activation of STAT5B was delayed and temporarily correlated with the synthesis of PAF induced by TPO. Moreover, the PAF-receptor inhibitor WEB 2170 was shown to prevent STAT5B activation but not STAT1 activation. These observations, together with the finding that 1-alkyl-PAF, but not 1-acyl-PAF, induced a rapid STAT5B tyrosine phosphorylation with subsequent translocation into the nucleus, suggest that TPO-induced STAT5B activation is secondary to the production of the alkyl derivative of PAF. Despite the absence of direct evidence for a role of STAT5B in mediating the angiogenesis induced by TPO, the observed correlation between the activation of STAT5B and the production of the angiogenic derivative of PAF suggests a potential role of the STAT pathway in the formation of new vessels. Indeed, it has been found that the JAK/STAT pathway is involved in the formation of ectopic adult wing veins in Drosophila.11
When TPO-containing Matrigel was injected subcutaneously in mice, a dose-dependent neoangiogenesis was observed. However, for the full expression of in vivo angiogenic properties, the bioactivity of PAF was required. Indeed, WEB 2170 significantly prevented TPO- but not bFGF-induced angiogenesis. Moreover, we observed that the TPO-induced angiogenesis was associated with the synthesis of PAF within the Matrigel. This study does not provide direct evidence of a cellular source of PAF produced within Matrigel after TPO stimulation; however, endothelial cells were the predominant cell population detectable at 6 and 24 hours, when PAF was synthesized within the Matrigel.
PAF is a phospholipid mediator with multiple biological
activities relevant for the development of several pathological and
physiological processes, such as inflammation,
shock, embryogenesis, and cell differentiation.36 41 It
has recently been shown that PAF directly stimulates in vitro migration
of endothelial cells and promotes in vivo angiogenesis
by a heparin-dependent mechanism.19 42 PAF acts, on
endothelial cells, through a specific
receptor,19 43 which belongs to the family of
"serpentine" receptors containing 7 -helical domains that weave
in and out of the plasma membrane.43 Specific
antagonists of PAF receptor have been synthesized and are
currently used in studies of the pathophysiology of PAF.41
Several lines of evidence indicate that PAF acts as secondary mediator
of angiogenesis induced by TNF and hepatocyte growth factor, as
PAF is actively synthesized in vivo during the angiogenic processes and
a PAF-receptor antagonist significantly reduced the
angiogenic response to these cytokines.32 44 The
present observation that PAF is synthesized by HUVECs stimulated in
vitro with TPO and in vivo within Matrigel during TPO-induced
neoangiogenesis, as well as the observation that a specific
PAF-receptor antagonist, WEB 2170, inhibits in vitro HUVEC
migration and, in vivo, the angiogenic process, suggests that this
phospholipid acts as a secondary mediator also for TPO. PAF may act as
a mediator of cell-to-cell communication involved in the amplification
of the signal triggered by TPO by determining the production of
other endothelial-derived heparin-dependent angiogenic
factors that may account for the in vivo endothelial
cell proliferation. Indeed, PAF has been shown to induce the expression
within Matrigel of several angiogenic factors, including VEGF and its
specific receptor flk-1.45 On the other hand, it has been
recently shown that TPO triggers the production of VEGF in
c-Mpl-expressing cells.38 Therefore, one can speculate
that PAF may stimulate migration of endothelial cells
while VEGF may provide the signal for cell proliferation. Further
studies are needed to investigate the
pathophysiological role of TPO in angiogenesis. The
observation that fetal hematopoiesis parallels neoangiogenesis suggests
a bidirectional interaction between endothelial
and hematopoietic cells. In this context, TPO may contribute to
generate a microenvironment favorable to a coordinated development of
hematopoiesis and angiogenesis.
In conclusion, the results of the present study indicate that TPO is able to activate endothelial cells and to induce an angiogenic response that is mediated by the synthesis of PAF. Moreover, biochemical analysis supports the hypothesis that STAT5B may be involved in the signaling pathway leading to PAF-dependent angiogenesis. The failure to demonstrate a defect in vasculogenesis/angiogenesis by targeted inactivation of either c-mpl3 or STAT5B46 genes may be explained by the redundancy of angiogenic mechanisms acting in vivo. To further investigate the role of STAT5B-dependent signaling pathway in vasculogenesis and neoangiogenesis, generation of transgenic mice with a specific endothelial promoter containing a dominant negative STAT5B construct, is under development in our laboratory.
|
Acknowledgments |
|---|
This work was supported by a grant from the Associazione Italiana per la Ricerca sul Cancro (to G.C. and L.P.) and by the Consiglio Nazionale delle Ricerche (CNR; targeted project, "Biotechnology" to G.C.). We thank Dr Sylvie Gisselbrecht (INSERM, Hopital Cochin, Paris, France) for the generous gift of the anti-STAT5A and anti-STAT5B antisera, Prof Yosef Yarden (Weizmann Institute for Technology, Rehovot, Israel) for the generous gift of the anti-STAT1 antiserum, and Dr L. Silengo for discussion and helpful suggestions. We thank Dr Edwin Ades and Francisco J. Candal of Centers for Disease Control and Prevention (Atlanta, Ga) and Dr Thomas Lawley of Emory University (Atlanta, Ga) for providing the CDC.HMEC-1 cell line.
|
Footnotes |
|---|
1 Both authors contributed equally to this study.
Received January 19, 1998; accepted November 23, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-
Vigon I, Mornon JP, Cocault L, Mitjavila MT,
Tambourin P, Gisselbrecht S, Souyri M. Molecular cloning and
characterization of MPL, the human homologue of v-mpl oncogene:
identification of a member of the hemopoietic growth factor
superfamily. Proc Natl Acad Sci U S A. 1992;89:5640–5644.
[Abstract/Free Full Text] - Wendling F, Maraskovsky E, Debili N, Florindo C, Teepe M, Titeux M, Methia N, Breton-Gorius J, Cosman D, Vainchenker W. c-Mpl ligand is a humoral regulator of megakaryocytopoiesis. Nature. 1994;369:571–574.[Medline] [Order article via Infotrieve]
-
Gurney AL, Carver-Moore K, de Sauvage FJ, Moore MW.
Thrombocytopenia in c-mpl-deficient mice. Science. 1994;265:1445–1447.
[Abstract/Free Full Text] -
Montrucchio G, Brizzi MF, Calosso G, Marengo S,
Pegoraro L, Camussi G. Effect of recombinant human megakaryocyte growth
and development factor on platelet activation. Blood. 1996;87:2762–2768.
[Abstract/Free Full Text] - Kishimoto T, Taga T, Akira S. Cytokine signal transduction. Cell. 1994;76:253–262.[Medline] [Order article via Infotrieve]
-
Gurney AL, Wang SC, Henzel WJ, de Sauvage F. Distinct
regions of c-MPL cytoplasmic domain are coupled to the JAK-STAT signal
transduction pathway and Shc phosphorylation.
Proc Natl Acad Sci U S A. 1995;92:5292–5296.
[Abstract/Free Full Text] -
Tortolani PJ, Johnston JA, Bacon C. M, McVicar DW,
Shimosaka A, Linnekin D, Longo DL, O'Shea JJ. Thrombopoietin induces
tyrosine phosphorylation and activation of the Janus
kinase, JAK2. Blood. 1995;85:3444–3451.
[Abstract/Free Full Text] - Pallard C, Gouilleux F, Benit L, Cocault L, Souyri M, Levy D, Groner B, Gisselbrecht S, Dusanter-Fourt I. Thrombopoietin activates a STAT5-like factor in hematopoietic cells. EMBO J. 1995;14:2847–2856.[Medline] [Order article via Infotrieve]
-
Darnell JE Jr. STATs and gene regulation.
Science. 1997;277:1630–1635.
[Abstract/Free Full Text] - Boccaccio C, Andò M, Tamagnone L, Bardelli A, Michieli P, Battistini C, Comoglio PM. Induction of epithelial tubules by growth factor HGF depends on the STAT pathway. Nature. 1998;391:285–288.[Medline] [Order article via Infotrieve]
-
Yan R, Luo H, Darnell JE Jr, Dearolf CR. A JAK-STAT
pathway regulates wing vein formation in Drosophila.
Proc Natl Acad Sci U S A. 1996;93:5842–5847.
[Abstract/Free Full Text] - Brizzi MF, Battaglia E, Rosso A, Strippoli P, Montrucchio G, Camussi G, Pegoraro L. Regulation of polymorphonuclear cell activation by thrombopoietin. J Clin Invest. 1997;99:1576–1584.[Medline] [Order article via Infotrieve]
- Ketley NJ, Newland AC. Hemopoietic growth factors. Postgrad Med J. 1997;73:215–221.[Abstract]
- Risau W. Differentiation of endothelium. FASEB J. 1995;9:926–933.[Abstract]
- Risau W. Vasculogenesis angiogenesis and endothelial cell differentiation during embryonic development. In: Feinberg RN, Sherer GK, Auerbach R, eds. The Development of the Vascular System. Basel, Switzerland: Karger; 1991:1–24.
- Folkman J. Angiogenesis and angiogenesis inhibition: an overview. EXS. 1997;79:1–8.[Medline] [Order article via Infotrieve]
- Koch AE, Harlow LA, Haines GK, Amento EP, Unemori EN, Wong WL, Pope RM, Ferrara N. Vascular endothelial growth factor: a cytokine-modulating endothelial function in rheumatoid arthritis. J Immunol. 1994;152:4149–4156.[Abstract]
-
Koch AE, Polverini PJ, Kunkel SL, Harlow LA, Di Pietro
LA, Elner VM, Elner SG, Strieter RM. Interleukin-8 as a
macrophage-derived mediator of angiogenesis.
Science. 1992;258:1798–1801.
[Abstract/Free Full Text] - Camussi G, Montrucchio G, Lupia E, DeMartino A, Perona L, Arese M, Vercellone A, Toniolo A, Bussolino F. Platelet-activating factor directly stimulates in vitro migration of endothelial cells and promotes in vivo angiogenesis by a heparin-dependent mechanism. J Immunol. 1995;154:6492–6501.[Abstract]
-
Montrucchio G, Lupia E, De Martino A, Silvestro L,
Rizea Savu S, Cacace G, De Filippi PG, Emanuelli G, Camussi G. Plasmin
promotes an endothelium-dependent adhesion of
neutrophils, involvement of platelet activating factor and
P-selectin. Circulation. 1996;93:2152–2160.
[Abstract/Free Full Text] - Ter