Background
The human genome harbors 104 genes encoding for cysteine-based (Cys-based) phosphatases classified into three classes on the basis of their amino acid sequences and catalytic domains [
1]. Dual-specificity phosphatases (DUSPs) or
Vaccinia-H1-like (VH1-Like) enzymes represent the largest group of class I of Cys-based motif phosphatases and is represented by 61 members with diverse substrates specificity ranging from mRNA to inositol phospholipids, p-Ser/p-Thr and p-Tyr. Among these 61 phosphatases, 11 are specific for the MAPKs ERK, JNK and p38 and are known as the typical DUSPs or MAPK specific phosphatases (MKPs). The second group of the VH1-like phosphatases is known as the atypical DUSPs (A-DUSPs) represented by 19 small enzymes (with less than 250aa) and are poorly characterized (reviewed in [
2]). Considering the important role of MAPKs in the regulation of different cellular functions and their involvement in different human diseases including cancer [
3‐
5], the activation as well as inhibition processes of these serine/threonine kinase family has been well characterized. Therefore, among all DUSPs, the MKPs have been the most characterized
in vitro and
in vivo. The expression of several phosphatases belonging to this group is altered in human cancer (reviewed in [
6]).
DUSP3, also called
V accinia H 1-
R elated (VHR), is the founding member of the dual-specificity protein phosphatases group. It consists of a 185 amino acids (Mr 21 kDa) catalytic domain but no apparent targeting domain or docking site and is encoded by
DUSP3/Dusp3 gene [
7]. The crystal structure of DUSP3 has been solved and shows a shallow active site allowing DUSP3 to act on both pTyr and pThr in its substrates [
8]. DUSP3 has been reported to dephosphorylate the MAPKs ERK and JNK, but not p38 [
7‐
9]. More recently, EGFR and ErbB2 were reported as direct new substrates for this phosphatase in a non-small cell lung cancer cell line NSCLC [
10]. Unlike many other MKPs, DUSP3 expression is not induced in response to activation of MAPKs, but is regulated during cell cycle progression [
11,
12]. In a previous study, we have shown that in HeLa cells, the knockdown of endogenous DUSP3 using RNA interference induces cell cycle arrest at G1/S and G2/M phases and is accompanied by the hyperactivation of ERK1/2 and JNK1/2 [
11,
12]. In line with this finding, DUSP3 was found up-regulated in human cancers and in several cancer cell lines. Indeed, we reported that DUSP3 is highly expressed in cervical carcinomas and in several cervix cancer cell lines [
13]. This phosphatase is also highly expressed in human prostate cancer and in the LNCaP human prostate adenocarcinoma cell line [
14]. On the other hand, recent reports showed that DUSP3 is downregulated in NSCLC and when overexpressed in these cells, it leads to decreased cell proliferation and reduced tumor growth in a xenograft mouse model [
10]. In line with these findings, Min Gyu Lee’s group reported recently that DUSP3 downregulation in NSCLC tumors, when correlated with high levels of the histone H3 lysine 36 (H3K36) demethylase, KDM2A, is associated with poor prognosis for the patients [
15]. In the same study, the authors demonstrated that KDM2A activates ERK1/2 through epigenetic repression of
DUSP3 expression via demethylation by H3K36 at the
DUSP3 locus. DUSP3 has also been found downregulated in breast carcinomas [
16]. These studies clearly suggest that DUSP3 plays complex and contradictory roles in tumorigenesis that could be cell type-dependent. However, most of these studies were performed either
in vitro, using recombinant proteins, or in cell lines, using transient overexpression or siRNA knockdown. Furthermore, all these studies were focused on tumor cells without taking into account the host cells. Therefore, the physiological function of DUSP3 is unknown.
We report herein that DUSP3 is highly expressed in endothelial cells (EC), depletion of which causes an inhibition of EC in vitro tubulogenesis. To investigate the physiological functions of DUSP3, we generated a new mutant mouse strain deficient for Dusp3 gene. The obtained DUSP3-deficient mice were viable and had no apparent phenotype or spontaneous pathology, suggesting that these mice could be useful to study DUSP3’s role in different pathological conditions. Indeed, by applying different in vivo, ex vivo and in vitro models, we provide evidence that DUSP3 plays an important and non-redundant role in angiogenesis.
Discussion
The physiological function and possible involvement of the A-DUSP family members in cancer is largely unknown. The lack of knockout mice for A-DUSPs is probably one of the major limitations in the determination of the physiological function of these phosphatases. So far, out of 19 A-DUSPs, only 3 were disrupted in mice, STYX [
25], DUSP14 [
26] and laforin [
27]. However, the role of these phosphatases in cancer and angiogenesis were not investigated in these mutant mice. Thus, the physiological function of A-DUSPs in cancer and angiogenesis is still unknown.
We report here the generation of a new mouse strain lacking Dusp3 gene, encoding for the atypical dual specificity phosphatase DUSP3. The mutant DUSP3−/− mice develop normally and do not have any spontaneous evident pathology, making them a good in vivo tool to investigate the role of DUSP3 in different diseases. By applying different in vivo and ex vivo models to these knockout mice, we provide evidence for a new physiological role of DUSP3 in neovascularization. We also report that DUSP3 is highly expressed in human endothelial cells and demonstrate its essential role for in vitro primary human endothelial cell angiogenic sprouting function.
The fact that DUSP3-deficient mice are born displaying no vascular defects under normal conditions could be explained by a redundant function of DUSP3 shared with other DUSPs. Indeed, several DUSPs have overlapping substrates specificity, especially among MAPKs. This makes it difficult to assign a specific physiological role for a specific DUSP in a specific tissue. It is conceivable that conditional knockout mice lacking DUSP3 only in the endothelial cells may display a vascular phenotype during embryonic vascular development than the full knockout mice. However, we found that DUSP3-deficiency prevented neo-vascularisation of Matrigel plugs and LLC xenograft tumors suggesting that DUSP3 plays an important and non-redundant function in tumor-induced angiogenesis.
Using microarray analysis, we evaluated the expression levels of all DUSPs transcripts in the Matrigel plugs extracts retrieved form DUSP3
−/− and WT mice. We found that among all DUSPs (typical and atypical), DUSP1/MKP1 and DUSP23/VHZ were significantly downregulated in Matrigel plugs retrieved RNAs (Table
2) and we did not observe an increase in any DUSP in the absence of DUSP3. Altered expression of DUSP1/MKP1 has been reported in different human cancer (reviewed in [
28]). In angiogenesis, DUSP1/MKP1 expression is associated with increased invasiveness of NSCLC due to an increased expression of VEGFC, suggesting that DUSP1 inhibition could be a good strategy to inhibit tumor invasion and angiogenesis [
29]. Therefore, the observed decrease of neo-angiogenesis in our model could also be due to the decreased DUSP1/MKP1 expression. The other possibility could be that the observed decrease of DUSP1 reflects the decreased number of endothelial and smooth muscle cells in the Matrigels infiltrates (Table
1). As for DUSP23/VHZ, little is known about this phosphatase function. However, DUSP23/VHZ is highly expressed in several human cancers and could play a role in cell cycle regulation [
30]. The cellular distribution of DUSP23 is not known. Therefore, it is difficult to conclude if, in our case, the observed decrease of this phosphatase transcript in DUSP3
−/− retrieved Matrigels is due to DUSP3 deficiency or reflects the decrease of EC and smooth muscle cells infiltration in Matrigels.
Table 2
Dusps genes expression profile from Dusp3-knockout versus WT Matrigel plugs
Typical Dusps
|
Dusp1
| MKP1 | 0.60 | 0.041* |
Dusp2
| PAC-1 | 1.0 | 0.96 |
Dusp4
| MKP2 | 0.48 | 0.18 |
Dusp5
| DUSP5; hVH3; B23 | NP | |
Dusp6
| MKP3 | 1 | 0.9 |
Dusp7
| MKPX | 0.9 | 0.6 |
Dusp8
| DUSP8 | 0.97 | 0.88 |
Dusp9
| MKP4 | ND | |
Dusp10
| MKP5 | 1.5 | 0.23 |
Dusp16
| MKP7 | 1.2 | 0.33 |
Styxl1
| MK-STYX | ND | |
Atypical Dusps
|
Epm2a
| Laforin | ND | |
Dusp3
| VHR | ND | |
Dusp11
| PIR1 | 1 | 0.75 |
Dusp13
| MDSP; TMDP; DUSP13 | 1.8 | 0.38 |
Dusp14
| MKP6 | 0.84 | 0.51 |
Dusp15
| VHY | 1.1 | 0.56 |
Dusp18
| LMW-DSP20 | 0.54 | 0.11 |
Dusp319
| SKRP1; LMW-DSP3 | 1.3 | 0.057 |
Dusp21
| LMW-DSP21 | ND | |
Dusp22
| JSP-1; JKAP; MKPX; VHX | 1.4 | 0.078 |
Dusp23
| LDP-3; VHZ | 0.85 | 0.04* |
Dusp26
| MKP8, NEAP | 1.3 | 0.71 |
Dupd1
| DUPD1 | 2.5 | 0.18 |
Dusp28
| VHP | 1.1 | 0.31 |
Styx
| STYX | 1.1 | 0.43 |
In the aortic ring assay, we found that DUSP3 deficiency prevented the sprouting in response to the angiogenic growth factor b-FGF. This finding was further supported by the significant decrease of angiogenic sprouting in the HUVECs spheroid model after
in vitro downregulation of DUSP3 using RNA interference. Although the underlying mechanism is not clear, these findings suggest that DUSP3 plays an important role in the b-FGF receptor signaling pathways. FGF signalings are involved in a plethora of biological processes leading to: activation of cell proliferation, inhibition of apoptotic signals, activation of cell migration in different cell types and promotion of angiogenesis. FGF activates cell proliferation mainly through the Raf-MEK-ERK MAPK pathway (reviewed in [
19]). The fact that DUSP3 dowregulation in HUVECs did not affect cell proliferation and ERK1/2 activation suggests that DUSP3 is dispensable for the b-FGF-induced cell proliferation. We can also exclude the involvement of DUSP3 in FGF PI3K/Akt pro-survival/anti-apoptotic pathway as DUSP3 depletion did not impact HUVECs apoptosis. On the other hand, the phosphorylation of Akt was normally induced in DUSP3-depleted HUVECs. FGF plays also a crucial role in cell migration and angiogenesis. This effect is mediated through the PI3K and PLCγ/PKC activation pathways [
19]. We have indeed demonstrated that DUSP3 depletion in HUVECs affected PKC basal and b-FGF induced phosphorylation. PKCs represent a large family of enzyme activated by two secondary messengers, calcium (Ca2+) and diacylglycerol (DAG). Ca2+ increases the affinity of PKC for lipids and DAG induces a high affinity interaction with the membrane leading to its activation [
31,
32]. To be ready for activation by Ca2+ and DAG, PKC is first phosphorylated by both phosphoinositide-dependent kinase 1 [
33] and by autophosphorylation [
34]. The autophosphorylation of PKC on serine 660 residue is important for the stability of the enzyme conformation and downstream signal transduction [
35,
36]. In absence of DUSP3, we found that this autophosphorylation site (Ser660) is hyperphosphorylated, suggesting that PKC is in a ready state to be activated. However, the anti-phospho-PKC Ser660 antibody used detects endogenous levels of several PKC isoforms. To investigate which PKC isoform is affected by DUSP3 depletion, immunoprecipitation of all the isoforms, followed by immunoblotting with phospho-PKC bII Ser660 is required. What is clear so far is that DUSP3 is involved in FGF-induced PKC activation in MAPKs-independent manner. Upon activation with FGF, DUSP3-depleted cells showed a very slight increase in the phosphorylation of the autophosphorylation site of the PKC family proteins compared to the control. This could be due to the fact the hyperactivated status of PKC at basal levels leads to an unresponsive signaling pathway.
We have also investigated if the most recently identified DUSP3 substrate, EGFR in H1299 cells [
10], could be affected by DUSP3 depletion in HUVECs. EGFR is an important player in diverse biological processes and is actually targeted by different approaches in various human malignancies [
37]. Tyrosine phosphorylation is an important post-translational modification for EGFR-induced signaling after ligand binding. We found that EGFR tyrosine phosphorylation was not affected by DUSP3 deficiency in HUVEC cells neither at basal levels, nor after EGF activation suggesting that DUSP3 is not targeting EGFR in endothelial cells. These results were compatible with recent study where Wagner
et al. showed that EGFR was not regulated by DUSP3 in the primary NSCLC tumor cells and in the NSCLC cell line H460 [
15].
In the DUSP3
−/− mice, we also found that the activity of ERK1/2 and JNK1/2 were not affected by DUSP3 deficiency in B cells, T cells, macrophages and platelets (unpublished observations). However, we failed in testing this in mice primary endothelial cells as the purification of sufficient number of these cells without affecting the basal activity of MAPKs was challenging. This is not the first time that previously characterized DUSP substrate specificity is not confirmed in a knockout mice model. Indeed, deficiency of DUSP2/PAC1, a known phosphatase for ERK and p38, does not lead to enhanced ERK and p38 phosphorylation but rather causes an enhanced JNK phosphorylation, suggesting a crosstalk between the different MAPKs that contribute to the observed changes in DUSP2
−/− mice [
38]. Similarly, knockout of DUSP10/MKP5, a phosphatase known to target p38, does not cause p38 hyperphosphorylation [
39]. These inconsistencies are probably due to the use of
in vitro overexpression/downregulation systems during previous characterizations of DUPS’s substrate specificity, which may not faithfully reflect the outcomes from DUSP-deficient primary cells. Alternatively, the lack of a particular DUSP may be compensated by other DUSPs.
Methods
Generation of DUSP3 knockout mice by disruption of Dusp3 locus
The DUSP3 knockout (KO) mouse was generated by replacing the Exon II with the Neo gene by homologous recombination. A 2.3 Kb fragment containing Exon I and a 4.5 Kb fragment containing the 5’ region of Intron II of the
Dusp3 gene were cloned inside the plasmid pPNT and the plasmid was transfected into the 129/SvJ embryonic stem (ES) cell line by electroporation. G418 and Ganciclovir resistant ES clones were screened by PCR using a forward primer located in the
Dusp3 gene, outside the 2.3 kb fragment cloned in the plasmid, and a reverse primer located in the Neo gene. The proper homologous recombination was verified by Southern hybridization analysis, detecting an additional 4.5 Kb fragment after
XbaI digestion and hybridization with a probe located in the 5’ region of the
Dusp3 gene. Two recombinant ES cell lines were injected into blastocysts of C57BL/6 mice producing chimeras that were mated with C57BL/6 mice to generate heterozygous founders. ES transfection and blastocyst injection were performed at the Moores Cancer Center/Transgenic and Gene Targeting core facility at UCSD.
http://cancer.ucsd.edu/Research/Shared/tgm/default2.asp. Heterozygous mice were mated to generate +/+ and −/− littermates to be used for experimentation. Mice were weaned and ear-marked at day 21. At week 4, 2 mm of tail was cut for genotyping using a surgical blade. Total DNA was extracted from tail tip using High Pure PCR template preparation kit (Roche, Vilvoorde, Belgium) and 0.1 μg was used as a template in 50 μl of a final reaction mixture which contained the
Dusp3 primers 5′GTGTGAGCTGCACTTTCCAA3′ and 5′GGTGACTGGGTGAAGAATGG3′, together with the Neo primer 5′TTGCCAAGTTCTAATTCCATCAGA3′. The reaction generates a 456 bp fragment from the
Dusp3 gene and a 365 bp fragment from the recombinant construct.
Ethical statement
All mice experiments and procedures were carried out following the guidelines and in agreement with the animal ethics committee of the University of Liège. All the work was covered by the ethical licence: 858 “understanding the role of DUSP3 in angiogenesis”.
Antibodies and reagents
Anti-Von Willebrand Factor (vWF) antibody was from Dako (Dako, Heverlee, Belgium). Anti-DUSP3 antibody used for immunohistochemistry, basic-Fibroblast growth factor (b-FGF) and heparin were from R&D (R&D Systems, Minneapolis, MN). Anti-DUSP3 antibody used for western blots, as well as the anti-CD31, Matrigel and Dispase were from BD Biosciences (BD Biosciences, San Jose, CA). Anti-phosphotyrosine antibody (4G10) was from Millipore (Millipore, Overijse, Belgium). Anti-phospho-ERK1/2 (Thr202/Tyr204), anti-ERK1/2, anti-cJun, anti-pospho-Akt (Ser473), anti-Akt, anti-EGFR, anti-phospho-PKC pan (Ser660) antibodies and SAPK/JNK kinase assay kit were all from Cell Signaling (Cell Signaling, Danvers, MA). HRP conjugated anti-mouse and anti-rabbit secondary antibodies and enhanced chemiluminescence kit (ECL) were from GE Healthcare (GE Healthcare Europe GmbH, Diegem, Belgium). Double stranded siRNA used as a non-targeting control (siCTL) was from Dharmacon (Thermo Scientific-Dharmacon, Erembodegem-Aalst, Belgium). Double stranded siRNAs used for DUSP3 silencing were from Eurogentec (Eurogentec, Seraing, Belgium) and sequences were siDUSP3-1 (GGCAGAAGAUGGACGUCAA), siDUSP3-2 (GGUCCUUCAUGCACGUCAA). Anti-rat Alexa 594 secondary antibody were from Life Technology (Life Technology, Gent, Belgium). GeneTrans II was from Mo Bi Tec (Mo Bi Tec, Gottingen, Germany). [3H] thymidine was from Perkin Elmer (Perkin Elmer, Zaventem, Belgium). Anti-GAPDH antibody and Fluorescein isothiocyanate-dextran (FITC-Dextran) were from Sigma (Sigma-Aldrich, Diegem, Belgium). Collagen R was from Serva (Serva, Heidelberg, Germany).
Cell culture and siRNA transfection and cellular proliferation
Human Umbilical Vein Endothelial Cells (HUVEC), EBM medium and EGM Singlequot were purchased from Lonza (Lonza, Basel, Switzerland). HUVECs were maintained in EGM (EBM + EGM Singlequot). Early cell passages (2 to 6) were transfected with non-targeting siCTL (150nM) or with 2 different DUSP3 targeting siRNA (150nM) using Gene
Trans II transfection reagent as a vehicle (3.5 μL/1 mL). HUVEC were used for experiments 72 hours after transfection. Cellular proliferation was measured as previously reported [
11,
17]. The Lewis Lung Carcinoma (LLC) cells were cultured in DMEM (Lonza, Basel, Switzerland) supplemented with 10% heat-inactivated fetal bovine serum in a humidified atmosphere of 5% CO2 at 37°C.
Immunohistochemistry
Human cervix carcinoma paraffin embedded serial sections (4 μm) were incubated during one hour in the oven at 60°C, deparaffinized and rehydrated using successive baths as follow: 2 × 5 min in xylol, 2 × 2 min in 100% ethanol, 1 × 1 min in 95% ethanol, 1 × 2 min in 70% ethanol and 2 × 2 min in dH2O. Antigen retrieval was performed using Target retrieval solution (Dako) for 40 min at 99°C. After 20 min at room temperature (RT), endogenous peroxydases were inhibited using Peroxydase blocking solution (Dako) during 10 min at RT. Background staining was reduced by incubating the slides in 10% normal goat serum/PBS for 30 min at RT. Sections were then subsequently incubated with the primary anti-DUSP3 (dilution: 1/50) or anti-vWF (dilution: 1/200) antibodies for 1 h at RT then with the HRP conjugated anti-mouse or anti-rabbit secondary antibody at 1/200 dilution during 1 h at RT. Staining was revealed using 3,3’-Diaminobenzidine (DAB) chromogen and slides were counterstained with haematoxylin.
Tubulogenesis Matrigel assay
To perform tube formation assay, 200 μL of Matrigel were put in 24 well culture plates and incubated for 2 hour at 37°C to allow gelling. Dissociated (3 × 103) HUVECs were diluted in the appropriate medium and added onto the Matrigel layer. After 24 h, tube formation was visualized using phase-contrast microscopy. Total tube length and number of intersections were quantified using Image J software (National Institutes of Health, Bethesda, MD).
Time-lapse video microscopy
siCTL and siDUSP3 transfected HUVECs (105 cells) were seeded on gellified Matrigel layer in 2 wells Lab-Tek chamber slides (Thermo Fisher Scientific, Waltham, MA, USA) and transferred to the stage of a Nikon A1R microscope (Nikon, Wavre, Belgium) equipped with x, y and z axes and maintained at 37°C and 12 hours. Images were acquired every 10 minutes using Nis Elements software (Nikon, Wavre, Belgium) and saved as ND2 files. Individual files were then combined and processed into AVI Movies using Nis Elements software. Representative snap shots were taken from siCTL and siDUSP3 conditions at different time intervals.
In vivo Matrigel angiogenesis assay and LLC cells injection
DUSP3+/+ and DUSP3−/− mice were subcutaneously injected in the two flanks with 500 μl of Matrigel supplemented with b-FGF (250 ng/ml) and Heparin (0.0138 mg/ml). Ten days later, Matrigel plugs were carefully harvested, weighted and digested with Dispase for 1 h at 37°C. The hemoglobin content was determined by a colorimetric assay using Drabkin’s reagent (Sigma-Aldrich). In separated experiments, 5 min prior mice sacrifice, freshly prepared FITC-Dextran (100 mg/kg) was injected in the tail vein. Matrigel plugs were frozen in Tissue-tek for subsequent immuno-fluorescence analysis.
For LLC tumor cells injection, mice were subcutaneously injected in the flanks with 106 LLC cells. Seven days later, tumors were carefully harvested, weighted and mechanically grounded using a homogenizer. The hemoglobin content was determined using Drabkin’s reagent colorimetric assay.
Immunofluorescence staining
For immunofluorescent staining of frozen Matrigel plugs, sections of 7 μm were fixed in ice-cold acetone for 2 min then in methanol (4°C) for 5 min. After blocking in PBS containing 10% normal goat serum for 30 min at RT, slides were incubated for one hour with anti-CD31. Slides were then washed. Immunoreactivity was revealed using anti-rat Alexa 594 secondary antibody. CD31 staining and injected FITC-dextran fluorescence were visualized under Olympus Vanox AHBT3 epifluorescent microscope (Olympus, Aartselaar, Belgium). The number of CD31+ blood vessels sections and total FITC-Dextran fluorescence intensity were quantified using Imaris software (Imaris, Bitplane, Zurich, Switzerland).
Mouse aortic ring assay
Mouse aortic ring assay was performed as previously described [
40]. Briefly, 1 mm long mice aortic rings explants were cultured in collagen gel (1,5 mg/ml). The aortic rings were either non-stimulated, stimulated with autologous serum or stimulated with 20 ng/ml of b-FGF. The explants were cultured for 9 Days at 37°C and 5% C02 and photographed using Zeiss Axiovert 25 (Zeiss, Zaventem, Belgium). Microvessel intersections number and maximal length of vessels outgrowth were quantified with the Aphelion 3.2 software from Adsis (Meythet, France).
Spheroid sprouting assay
To generate the spheroids, we proceeded as previously reported [
41]. Briefly, HUVECs resuspended in EBM containing 0.24% high viscosity methyl cellulose (Sigma-Aldrich) were seeded in 96 well round bottom non-adherent plates and cultured overnight at 37°C. Each spheroid contained 10
3 cells. Single spheroids were collected, embedded in rat tail collagen type 1 gel (Corning, Seneffe, Belgium) and cultured for 48 hours at 37°C in 2% FBS supplemented EBM with 75 ng/ml phorbol-12 myristate 13-acetate (PMA) or 10 ng/ml b-FGF. To quantify the sprouting, the mean number of sprout in each condition was counted.
Cell lysates, immunoprecipitation, western blot and SAPK/JNK Kinase assay
For western blot experiments, cells were stimulated for the indicated time points and lysed using RIPA buffer (50 mMTris-HCl (pH = 8.0), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM orthovanadate, complete protease inhibitor cocktail tablets EDTA free and 1 mM phenylmethylsulfonyl fluoride) on ice during 20 minutes. Lysates were next clarified by centrifugation at 21000 g during 20 min at 4°C. The resulting supernatants were collected and protein concentrations were determined using the colorimetric Bradford reagent (Bio-Rad, Nazareth, Belgium). Samples were next denaturated at 95°C in Laemmli buffer. To investigate the SAPK/JNK activity, we used the SAPK/JNK kinase assay kit following the instructions of the manufacturer. Briefly, cells were stimulated for the indicated time and lysed with the cell lysis buffer provided. Cell lysates were incubated overnight at 4°C with Phospho-SAPK/JNK Rabbit mAb sepharose beads with constant agitation. Kinase assay was performed by adding c-Jun recombinant protein and ATP to the beads with 1× Kinase buffer and incubated for 30 min at 30°C. The reaction was stopped by adding SDS Laemmli buffer and boiled at 95°C for 5 min. Samples were then run on SDS-PAGE gel and transferred to Hybond-nitrocellulose membranes. To block the non-specific binding sites, membranes were incubated for one hour in Tris-buffered saline-Tween 20 containing 5% of non fat milk or 3% BSA. Membranes were next incubated with anti-Phospho-c-Jun and anti-c-Jun. Immunoprecipitations of EGFR were carried out following previously reported protocols [
42]. To evaluate the efficiency of siRNA transfection, the phosphorylation of ERK1/2 and Akt, cell lysates from transfected endothelial cells were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were next immunoblotted with anti-DUSP3, anti-phospho-ERK1/2 and anti-phospho-Akt antibodies. Membranes were next stripped, blocked and immunoblotted with anti-GAPDH, anti-Akt and anti-ERK1/2 antibodies for normalization. Immunoreactivity was then revealed using HRP conjugated secondary antibodies. The blots were developed by enhanced chemiluminescence (Amersham, Gent, Belgium) according to the manufacturer’s instructions.
Microarray analysis and gene expression profiles
Total RNA was isolated from b-FGF containing Matrigel plugs retrieved from DUSP3+/+ and DUSP3−/− mice 10 days after sub-cutaneous injection. RNA was prepared using Trizol reagent (Roche). The yield of the extracted RNA was determined using spectrophotometer by measuring the optical density at 260 nm. The purity and quality of the extracted RNA were evaluated using the Experion RNA StdSens Analysis kit (Bio-Rad Laboratories, Hercules, CA). High quality RNA with RNA Quality Indicator (RQI) score greater than 8 was used for microarray experiment. Gene expression profiling was performed using Illumina’s multi-sample format Mouse WG-6 V2 BeadChip containing 45281 transcripts and profiles six samples simultaneously on a single chip (Illumina Inc., San Diego, CA). For each sample, 250 ng of total RNA was labeled using Illumina Total Prep RNA Amplification kit (Ambion, Austin, TX) according to the manufacturer’s instructions. Briefly, double stranded cDNA was synthesized using T7-oligo (dT) primers and followed by an in vitro transcription reaction to amplify antisense RNA (aRNA), while biotin was incorporated into the synthesized aRNA probe. The aRNA probe was then purified and quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA).
Biotinylated cRNA probe was hybridized to the Mouse WG-6 V2 BeadChip Array (Illumina). Labeled aRNA (1500 ng) was used for hybridization to each array. The hybridization, washing and scanning, were performed according to the manufacturer’s instructions. The arrays were scanned using a BeadArray Reader (Illumina). The microarray images were registered and extracted automatically during the scan according to the manufacturer’s default settings. Raw microarray intensity data were analysed with the Genome Studio software normalized using the quantile normalization method according to the manufacturer’s recommendation. The probes were considered as expressed by filtering data on Detection p-value lower than 0.05. Data are presented as the ratio of the average values obtained from 2 separate pools of Matrigels retrieved from 3 DUSP3−/− mice on 2 separate pools of Matrigels retrieved from 3 DUSP3+/+ mice and the corresponding p-value was determined using unpaired student’s t test. A value of p < 0.05 was considered as statistically significant.
Statistical analysis
The student t-test was used to assess statistical differences between different groups. Results were considered as significant if p-value < 0.05. Results are presented as mean ± SEM. Prism software (GraphPad, San Diego, CA) was used to perform statistical analysis. * = p < 0.05, ** = p < 0.01, *** = p < 0.001.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
AM performed, western blotting, proliferation assays, immunohistochemistry, siRNA transfection, in vitro tubulogenesis, in vivo Matrigel plug and aortic ring assays. CE helped for the aortic ring and the spheroid assays and for the in vivo Matrigel assay. BK helped for the in vivo Matrigel plug assays and edited the manuscript. FC made the cloning to generate DUSP3 knockout mouse. BS helped for the quantification of tubulogenesis, spheroid and the aortic rings sprouting. MM and FD helped for live imaging assay. SP and VM helped for the mice handling, cell transfection and western blotting. TZ performed the LLC xenograft tumor assay in mice. PD helped for mice breeding and handling. TM supervised and provided the grant support for the generation of the DUSP3 knockout mouse. MM and LM participated in discussion of the results and edited the manuscript. NA and GC participated in design of the experiments and discussion and edited the manuscript. SR designed and supervised the study, drafted the manuscript and provided grant support for this study. All authors read and approved the final version of this manuscript.