Published online before print April 1, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
and IFN-
* Experimental Transplantation and Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, Maryland;
Department of Biomedical Sciences and Biotechnology, Division of Human Anatomy, University of Brescia, Italy; and
Translational Genomics Research Institute, Laboratory of Cancer Drug Development, Phoenix, Arizona
1Correspondence: Experimental Transplantation and Immunology Branch, Center for Cancer Research, National Cancer Institute, Building 10, Room 12N226, MSC 1907, Bethesda, MD 20892. E-mail: tosatog{at}mail.nih.gov
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) and interferon-
(IFN-) selectively promote expression of the antiangiogenic molecules, IFN-inducible protein-10, monokine induced by IFN-, tryptophanyl-tRNA synthetase, and tissue inhibitor of metalmetalloproteinase-1, and inhibit expression of the proangiogenic molecules, platelet-endothelial cell adhesion molecule-1, vascular endothelial growth factor receptor-2, stromal cell-derived factor-1 (SDF-1), collagen type IV, endothelial cell growth factor-1, and carcinoembryonic antigen-related cell adhesion molecule-1. Reduced endothelial cell expression of SDF-1 protein by TNF- and IFN- disrupts extracellular matrix-dependent endothelial cell tube formation, an in vitro morphogenic process that recapitulates critical steps in angiogenesis. Replacement of SDF-1 onto the endothelial cell surface reconstitutes this morphogenic process. In vivo, TNF- and IFN- inhibit growth factor-induced angiogenesis and SDF-1 expression in endothelial cells. These results demonstrate that SDF-1/CXC chemokine receptor-4 constitutes a TNF-- and IFN--regulated signaling system that plays a critical role in mediating angiogenesis inhibition by these inflammatory cytokines.
Key Words: chemokine • cytokine • endothelium • inflammation
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The vascular endothelium is an important participant in the local inflammatory response [6 , 7 ]. Within minutes of tissue injury, pronounced, hemodynamic changes take place, including dilation of blood vessels and increased vascular permeability. The vasculature undergoes a number of modifications that are critical to the recruitment of inflammatory cells from the bloodstream to the sites of inflammation [1 , 3 , 8 ]. Inflammatory monocytes, macrophages, platelets, mast cells, and other leukocytes produce vascular endothelial growth factor (VEGF)-A, platelet-derived growth factor (PDGF), basic fibroblast growth factor (bFGF), angiopoietin-1, hepatocyte growth factor, insulin-like growth factor, and other proangiogenic factors, which promote neovascularization of the granulation tissue [6 , 7 , 9 ]. In spite of the early presence of proangiogenic factors, vascular repair occurs toward the end of the inflammatory response when the damaged tissue is replaced by fibroblast proliferation [10 ]. Thus, it is reasonable to speculate that negative regulators of angiogenesis are likely important in reducing or delaying the angiogenic response, but their characterization and mode of action are currently incomplete.
The proinflammatory cytokines tumor necrosis factor (TNF-) and interferon- (IFN-), which are often coexpressed at inflammatory sites and can act synergistically in vivo and in vitro, have been shown to variously modulate the survival, proliferation, migration, and other functions of endothelial cells that express functional receptors for these mediators [11
, 12
]. Both cytokines, together or individually, can inhibit endothelial cell proliferation and attachment in vitro, regulate vß3 integrin function in endothelial cells, and promote expression of the intercellular adhesion molecule-1 and endothelial leukocyte adhesion molecule-1 adhesion molecules and the chemokine regulated on activation, normal T expressed and secreted [13
14
15
]. TNF- can diminish endothelial cell migration but promote endothelial cell expression of the proangiogenic chemokine interleukin (IL)-8 [16
, 17
]. IFN- can induce endothelial cell expression of IFN-inducible protein-10 (IP-10) and monokine induced by IFN- (Mig), chemokines that exert angiotoxic and antiangiogenic activities in vivo [18
, 19
].
To gain a better understanding of the diverse effects TNF- and IFN- exert on endothelial cells, we broadly evaluated patterns of angiogenesis-related gene regulation and function in endothelial cells exposed to these inflammatory mediators at concentrations likely present at sites of severe inflammation. We identify a selected set of TNF-- and IFN--regulated genes, which exert pro- or antiangiogenic function and find that TNF- and IFN- inhibit endothelial cell morphogenesis into vascular structures and angiogenesis even in the presence of proangiogenic stimuli. We further show that diminished expression of stromal cell-derived factor-1 (SDF-1) and CXC chemokine receptor 4 (CXCR4) plays a critical role in mediating inhibition of endothelial cell function by TNF- and IFN-.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(hSDF-1), monoclonal anti-human mouse SDF-1 antibody [immunoglobulin G (IgG)1, clone 79014.111], and mouse IgG1 isotype control (hybridoma clone 11711.11) were purchased from R&D Systems (Minneapolis, MN). Rabbit anti-hSDF-1 antigen affinity-purified polyclonal antibody was from Peprotech (Rocky Hill, NJ). Mouse IgG2A antihuman CXCR4–phycoerythrin (PE) monoclonal antibody (mAb; IgG2A, clone 12G5) was purchased from BD PharMingen (San Diego, CA). Affinity-purified goat antiactin antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Fluorescein isothiocyanate (FITC)-conjugated, affinity-purified F(ab')2 fragment goat anti-mouse IgG was from Jackson ImmunoResearch (West Grove, PA). Peroxidase-linked donkey anti-rabbit IgG antibody was from Amersham Pharmacia Biotech (Piscataway, NJ). Human and murine TNF- and IFN- were from R&D Systems. PKH2 green fluorescence cell-linker, propidium iodide (PI), and RNase-A were from Sigma Chemical Co. (St Louis, MO).
Cells and cell cultures
Human umbilical vein endothelial cells (HUVECs), prepared from umbilical cord by 0.1% collagenase II (Worthington Biochemical, Freeehold, NJ) digestion, were propagated through passage 5 in M199 (Gibco-BRL, Grand Island, NY) culture medium with 20% newborn calf serum (Sigma Chemical Co.), 5% human AB serum, 1.6 mM L-glutamine (Gibco-BRL), 50 mg/ml porcine heparin (Sigma Chemical Co.), 50 µg/ml ascorbate (Fisher Scientific, Fair Lawn, NJ), 15 mM HEPES buffer (Calbiochem-Behring, La Jolla, CA), and 15 µg/ml endothelial cell growth supplement (a crude extract of bovine neural tissue containing bFGF and acidic FGF; Sigma Chemical Co.). Human stromal cells (HS-5, American Type Culture Collection, Manassas, VA) were cultured in 10% RPMI-1640 medium (Gibco-BRL) with 10% fetal bovine serum (FBS; Biofluids, Rockville, MD).
Microarray hybridization
Total RNA (20 µg), extracted from HUVECs, untreated and treated for 6 h with TNF- (10 ng/ml) and IFN- (100 ng/ml), were reverse-transcribed using oligo-(dT) with Superscript II (Stratagene, La Jolla, CA) and were then labeled with Cy5–deoxycytidine 5'-triphosphate (dCTP) and Cy3–dCTP, respectively (Amersham Pharmacia Biotech). The labeled cDNAs were hybridized overnight at 60°C onto a cDNA microarray synthesized at the National Cancer Institute (NCI; Bethesda, MD) microarray core facility (Hs-UniGEM2-vB6.0) containing 9128 spotted Incyte clones. Fluorescence intensities of the hybridized spots were measured by a laser confocal scanner (Agilent Technologies, Palo Alto, CA). The data were analyzed with the DEARRAY software [20
]. After background subtraction, average intensities at each clone in the cytokine-treated hybridization were divided by the average intensity of the corresponding clone in the untreated hybridization. Ratios were normalized based on the distribution of all targets on the array. Low-quality measurements (e.g., mean fluorescence intensities <100 for both channels) were excluded from further analysis and were treated as missing values. Overall quality of the microarray experiment was evaluated based on the number of high-quality measurements. Statistical significance of the normalized ratios was calculated based on 99% confidence intervals of the distribution of all targets before exclusion. Only high-quality values were taken into consideration for further analysis. For each clone, the calibrated treated/untreated ratio was calculated and normalized for each experiment. If all arrays delivered high-quality results for each clone, the calibrated ratios were averaged. If only one array was acceptable for that clone, only that value was accepted. Based on the normalization, a 99% confidence interval was generated for each array for the entire distribution of calibrated ratios for each experiment. The results shown include only high-quality results. Results below or above the values marking the 99% confidence interval were considered significantly different from the mean distribution at a P< 0.01 level of significance. The actual higher and lower limit values of this confidence interval for all arrays were 0.58 and 1.75. Calibrated ratio values within the confidence interval (0.58–1.75) were regarded as not statistically significantly different from the mean.
RNA isolation and reverse transcriptase-polymerase chain reaction (RT-PCR)
Total RNA was extracted using TRI Reagent (Molecular Research Center, Cincinnati, OH). Semiquantitative RT-PCR was performed as described [21
]. cDNA was synthesized from 2 µg total RNA using the SuperScript preamplification system (Gibco-BRL). The number of amplification cycles was determined experimentally for each primer pair to fit the linear part of the sigmoid curve, reflecting the relationship between the number of amplification cycles and the amount of PCR product. RT-PCR detection of quantitative differences in mRNA for each gene product was established by serial dilutions of input cDNA used in PCR assays. Amplification was performed in a 50-µl reaction mixture using 5 µl cDNA, platinum Taq DNA polymerase (Gibco-BRL), 1 µl dNTP mixture (10 mM, Gibco-BRL), and specific SDF-1 primers at appropriate annealing temperatures. SDF-1 and glyceraldehyde 3-phosphate dehydrogenase (G3PDH) were amplified for 30 and 28 cycles, respectively. PCR products were loaded onto a 1.8% agarose gel (NuSieve agarose, FMC, Rockland, ME) and were visualized by staining with ethidium bromide. G3PDH cDNA was used as an internal standard to quantify cDNA for other genes. Primer sequences for hSDF-1and G3PDH were performed as described previously [22
].
Western immunoblotting
Cell lysates from 1 x 106 cells were solubilized in tricine sodium dodecyl sulfate sample buffer (Novex, San Diego, CA), boiled, and run through 10–20% tricine gel (Novex). After transfer, Immobilon-P membranes (Millipore, Bedford, MA) were incubated overnight with rabbit anti-hSDF-1 antibody (Peprotech; 0.5 µg/ml); bound antibody was decteted with affinity-purified, peroxidase-linked donkey anti-rabbit IgG antibody (Amersham Pharmacia Biotech). Relative protein expression levels were estimated by membrane rehybridization with goat antiactin antibody.
Flow cytometry
HUVECs were detached with 2 mM EDTA and 1% FBS in phosphate-buffered saline (PBS), washed twice with ice-cold binding buffer [RPMI 1640, 20 mM HEPES, 1% bovine serum albumin (BSA)], blocked with mouse IgG isotype (IgG2A or IgG1) control for 30 min at 1 µg/ml, and incubated (5x105/ml in 100 ml PBS–0.1% BSA) with an anti-human CXCR4–PE antibody (12G5 clone, BD PharMingen). Cells were also stained with murine monoclonal anti-SDF-1 antibody (clone 11711.11, R&D Systems). After washing, cells were incubated with FITC-labeled goat anti-mouse F(ab')2 fragment antibodies for 30 min at 4°C. To detect dead cells, HUVECs were stained for 30 min at room temperature with PI in the presence of RNase 100 µ/ml in PBS. Intracytoplasmatic expression of SDF-1 was tested after cell permeabilization with 0.1% saponin. Data were collected from 5 x 103 viable cells using a FACSCalibur cytofluorometer (Becton Dickinson, Franklin Lakes, NJ) and analyzed using CellQuest software (Becton Dickinson). Background fluorescence was assessed through staining with isotype-matched antibodies.
Cell-cycle analysis
Cell-cycle analysis was carried out by two-color immunofluorescence as described [23
] on HUVECs, stained first for surface SDF-1 with anti-SDF-1 mAb (R&D Systems) and then after treatment with paraformaldehyde and ethanol, with PI (Sigma Chemical Co). One-hundred thousand events were acquired, and an analysis of cell cycle on HUVEC cells was performed with the aid of CellQuest software (Becton Dickinson).
PKH2 staining and Matrigel tube formation assay
HUVECs were stained with the green fluorescent dye PKH2 (Sigma Chemical Co.), according to the manufacturer’s instructions. Briefly, 106 HUVECs were centrifuged, and 1 ml diluent (provided in the kit) was added to each tube. Cells were thoroughly resuspended in diluent and then transferred to the tube containing 1 ml PKH2 green fluorescent dye. Cells were then incubated in the dye solution for 5 min, followed by a 1-min incubation with 2 ml FBS (Gibco-BRL) to stop the reaction. Subsequently, 4 ml RPMI 1640 with L-glutamine (Mediatech, Huntingford, VA) was added, and the cells centrifuged at 150 g for 10 min. The supernatant was decanted, and the cells were washed three times in RPMI-1640 culture medium. Some of the PKH2-stained cells were treated with TNF- (10 ng/ml) plus IFN- (100 ng/ml) for 6 h at 37°C, and subsequently, some of these cells were loaded with SDF-1 by incubation at 4°C with SDF-1 (1 µg) for 30 min. Unstained cells were left untreated or were treated with pertussis toxin (PTX; 100 ng/ml) or with anti-human CXCR4 antibody (clone 12G5, R&D Systems; 10 µg/ml) for 1 h at 37°C and 4°C, respectively. The cells were plated on Matrigel-coated vessels in complete culture medium and were incubated for 16 h. After incubation, cells were photographed under phase-contrast microscopy with fluorescence, and images were imported into Adobe Photoshop®. Tubes were examined at low-power magnification (5x). At least 10 fields were examined per well; each experimental condition was tested in triplicate.
In vivo Matrigel angiogenesis assay
The in vivo Matrigel angiogenesis assay was performed as described previously [24
]. An aliquot (0.5 ml) of Matrigel (Becton Dickinson Labware), alone or with the desired additives, was injected subcutaneously (s.c.) into the mid-abdominal region of female BALB/c athymic mice 6–8 weeks old. Additives included murine bFGF (150 ng/ml, R&D Systems), murine TNF- (30 ng/ml), plus IFN- (300 ng/ml). There were six to eight mice per group. After 7 days, the animals were killed, and Matrigel plugs were removed, fixed in 10% neutral-buffered formalin solution (Sigma Chemical Co.), and embedded in paraffin. Tissues were sectioned (5-mm thick), and slides were stained with Masson trichrome (American Histolabs, Gaithersburg, MD). Quantitative analysis of angiogenesis was performed using IPLab software (BioVision Technologies, Exton, PA). The results are expressed as the mean (SEM) area (expressed in µm2), occupied by cells within a Matrigel field measuring 1.0 x 106 µm2.
Immunohistochemistry
Matrigel plugs were fixed in 10% neutral-buffered formalin solution (Sigma Chemical Co.), embedded in paraffin, sectioned at 4 µm, and stained for SDF-1 as described [22
]. Antigen retrieval was performed by steaming the sections in citrate buffer (10 mmol/L), pH 6.0, and 0.01% Tween for 30 min. After blocking with 3% goat serum for 15 min, the sections were incubated with mouse monoclonal anti-SDF-1 antibody (clone 79018.111, dilution 1:50, R&D Systems) overnight at room temperature. Bound antibody was detected with a biotin-conjugated secondary antibody formulation for recognition of rabbit and mouse Igs (Ventana Medical System, Tucson, AZ). After addition of an avidin–horseradish peroxidase conjugate, the enzyme complex was visualized with 3,3'-diaminobenzidine tetrachloride and copper sulfate.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and IFN-, we performed cDNA microarray experiments in which we compared gene expression profiles of HUVECs cultured in medium alone to that of HUVECs stimulated for 6 h with TNF- (10 ng/ml) plus IFN- (100 ng/ml). These concentrations were selected to reflect cytokine levels presumably detected at sites of severe inflammation [4
, 5
, 14
]. Duplicate microarray hybridizations were performed on cDNA microarray slides (NCI core microarray facility). The results were analyzed using National Institutes of Health image processing software [20
] and were then censored for quality and fluorescence intensity for each microarray spot. From a total of 9128 clones, 3787–4786 clones were accepted for further study. Table 1 lists results from these experiments, which were selected on the basis of: reflecting genes that have been previously reported to be induced in endothelial cells by TNF- and/or IFN-, thus serving as an internal validation of the microarray results; genes linked to cell-cycle regulation; and genes previously linked to various aspects of endothelial cell function. Within the category of genes known to be induced by IFN- in endothelial cells, we confirmed that expression of the IFN--inducible chemokines, IP-10, CXCL11 (previously known as IFN--inducible protein-9), and others, was significantly enhanced in HUVECs after exposure to IFN- and TNF-. It is noteworthy that the IP-10 and CXCL11 genes were the most highly induced genes. We also confirmed that expression of the TNF--inducible TNFAIP2 and TNFAIP3 proteins and others was significantly enhanced in HUVECs after exposure to TNF- and IFN-. Within the category of genes linked to cell-cycle regulation, we found that p16 and p21 are induced by and TNF- and IFN-, suggesting that these cytokines may induce cell-cycle arrest in HUVECs. Among genes that have been previously linked to regulation of vascular endothelial cell function and angiogenesis, we found expression of VEGFR1, angiopoietin-1, Tie-1, PF4, PDGF-A, thrombospondin-1, integrin ß1, endostatin, PDGF-A, MMP9, and others to be similar in endothelial cells cultured with or without TNF- plus IFN-. By contrast, we found expression of a selected set of angiogenesis-related genes to be significantly regulated by TNF- and IFN- in endothelial cells. The IL-8, IP-10, Mig, WARS, VEGF-C, and TIMP-1 genes were significantly induced, whereas the PECAM (CD31), angiopoietin-2, VEGF-R2, COL4A3, ECGF-1, CEACAM1, and SDF-1 genes were significantly inhibited (Table 1) . |
View this table: [in a new window] |
Table 1. Gene Expression Profiling of Endothelial Cells after Treatment with TNF- Plus IFN-
|
and IFN- reduce endothelial cell expression of SDF-1, but this chemokine is known to critically regulate endothelial cell morphogenesis [22
], we wished to confirm that TNF- (10 ng/ml) and IFN- (100 ng/ml) can reduce SDF-1 expression in HUVECs. Using RT-PCR, we found SDF-1 expression to be reduced in endothelial cells exposed to TNF- (10 ng/ml) and IFN- (100 ng/ml) for 6 h, and IP-10 expression was enhanced (Fig. 1A
), validating the results of microarray analysis. In addition, we found that TNF- alone (10 ng/ml) and IFN- alone (100 ng/ml) did not reduce SDF-1 expression in HUVECs after 6 h (Fig. 1A)
or 24 h (not shown). By immunoblotting, we determined that TNF- (10 ng/ml) and IFN- (100 ng/ml) together reduce levels of SDF-1 protein expression in HUVECs after 48 h incubation (Fig. 1B)
. Using fluorescein-activated cell sorter (FACS) analysis for detection of intracellular SDF-1 in permeabilized HUVECs, we confirmed that TNF- and IFN- together can reduce levels of intracellular SDF-1 and established that this effect is observed at TNF- concentrations ranging between 1 and 10 ng/ml in the presence of IFN- concentrations of 100 ng/ml (Fig. 2A
). After 48 h culture with TNF- (10 ng/ml) plus IFN- (100 ng/ml), the proportion of HUVECs expressing surface SDF-1 was also reduced (18.7% as opposed to 35.2%, Fig. 2B
). Using PI to mark cell DNA and study cell-cycle distribution by FACS analysis, we established that reduction of cell-surface SDF-1 expression induced by TNF- (10 ng/ml) plus IFN- (100 ng/ml) occurred in cells at all stages of the cell cycle (Fig. 2B)
. As cDNA for CXCR4, the specific SDF-1 receptor, was not present on the microarray used, we examined the effects of TNF- plus IFN- on surface CXCR4 expression in HUVECs. By FACS, we found that after 48 h culture with TNF- (10 ng/ml) plus IFN- (100 ng/ml), only 1.45% of HUVECs display surface CXCR4 (Fig. 2C)
.
 View larger version (69K): [in a new window] |
Figure 1. Regulation of SDF-1 and IP-10 gene expression in endothelial cells. (A) Semiquantitative RT-PCR analysis of SDF-1, IP-10, and G3PDH expression in HUVECs incubated for 6 h in medium alone or medium supplemented with TNF- (10 ng/ml) alone, IFN- (100 ng/ml) alone, or TNF- (10 ng/ml) plus IFN- (100 ng/ml). Prior to stimulation, HUVECs were starved by 24 h incubation in M199 culture medium containing 2% AB serum. (B) Western blot analysis of SDF-1 expression in HUVECs cultured for 24 h in medium alone or medium supplemented with TNF- (10 ng/ml) and IFN- (100 ng/ml). Loading accuracy was tested by membrane reprobing with actin-specific antibodies. Recombinant SDF-1 (rSDF-1; 10 ng) was included as a control.
|
 View larger version (22K): [in a new window] |
Figure 2. TNF- and IFN- reduce SDF-1 and CXCR4 expression in HUVECs. (A) Flow cytometric analysis of intracellular SDF-1 in HUVECs incubated for 24 h in medium alone or medium supplemented with TNF- (1 or 10 ng/ml) plus IFN- (10 or 100 ng/ml). After permeabilization, SDF-1 was detected by staining with mouse monoclonal anti-SDF-1 antibody (R&D Systems Clone 79018.111) or isotype-matched control (R&D Systems Hybridoma Clone 44716.14) followed by FITC-labeled goat anti-mouse IgG antibodies. (B) Flow cytometric analysis of surface SDF-1 expression in HUVECs cultured for 48 h in medium alone or in medium supplemented with TNF- (10 ng/ml) plus IFN- (100 ng/ml). After staining with anti-SDF-1 or control antibodies, the cells were treated with paraformaldehyde and ethanol and were double-stained with PI for cell-cycle distribution. The results reflect two-color analysis of SDF-1 and PI fluorescence intensities. (C) Surface CXCR4 expression in HUVECs cultured for 24 h in medium alone or medium supplemented with TNF- (10 ng/ml) and IFN- (100 ng/ml). Cells were stained with anti-CXCR4 antibodies (R&D Systems Clone 44716) or control IgG1 followed by FITC-labeled goat anti-mouse IgG antibodies. FSC, Forward scatter.
|
and IFN- on the ability of HUVECs to form Matrigel-dependent tubular structures. Compared with untreated cells (Fig. 3A
, left panel), HUVECs exposed for 6 h to TNF- (10 ng/ml) plus IFN- (100 ng/ml) did not form the characteristic network of tubular structures on Matrigel (Fig. 3A , right panel), which was visualized by fluorescence microscopy of PKH2-stained HUVECs (upper panels) and phase-contrast microscopy (lower panels). Instead of forming tubes, TNF-- plus IFN--treated HUVECs clustered in groups of rounded cells that were mostly viable (Fig. 3A
, right panel), as judged by trypan blue staining (not shown). To examine the relative role of reduced cell-surface SDF-1 expression to the disruption of tube formation by TNF- plus IFN-, we first tested whether SDF-1 loaded on the surface of HUVECs pretreated (6 h) with TNF- (10 ng/ml) plus IFN- (100 ng/ml) reconstituted tube formation. After treatment with TNF- plus IFN- HUVECs could be coated with recombinant SDF-1 (achieved by 1 h incubation of HUVECs with 1 µg SDF-1 and verified by FACS) but did not reacquire an ability to form tubes on Matrigel (not shown). As we had determined (Fig. 2C)
that levels of CXCR4 expression on HUVECs are also markedly reduced by treatment with TNF- (10 ng/ml) plus IFN- (100 ng/ml), we reasoned that the disruption of tube formation by TNF- and IFN- could be a result of the combined reduction of surface SDF-1 and CXCR4. To test for this possibility, we performed mixing experiments in which we fluorescein-labeled with PKH2 one of the two populations. When we mixed together untreated HUVECs, 50% of which were fluorescein-labeled with PKH2, tubes were formed in which labeled and unlabeled cells composed mosaic tubular structures (Fig. 3B
, left panels). This result indicated that the fluorescent label did not disrupt tube formation by HUVECs. We then tested the effect of treatment with TNF- (10 ng/ml) plus IFN- (100 ng/ml). We found that HUVECs treated with TNF- (10 ng/ml) plus IFN- (100 ng/ml), and subsequently, fluorescein-labeled with PKH2, could not form tubular structures when mixed with untreated HUVECs not labeled with PKH2 (not shown). By contrast, we found that HUVECs treated with TNF- (10 ng/ml) plus IFN- (100 ng/ml), coated with SDF-1, and fluorescein-labeled with KPH2 could form tubular structures when mixed with untreated HUVECs (not labeled with PKH2), as evidenced by the appearance of mosaic tubular structures derived from fluorescein-labeled (cells treated with TNF- plus IFN- and coated with SDF-1) and unlabeled, untreated cells (Fig. 3B
, right panels). These structures appeared indistinguishable from those derived from mixing together untreated HUVECs, 50% of which were fluorescein-labeled with PKH2 (Fig. 3B
, left panels), providing evidence that reduction of surface SDF-1 is critical to the disruption of Matrigel-induced tube formation by TNF- and IFN-. To test for the contribution of surface CXCR4, we examined separately the effects of PTX, an inhibitor of G1 receptor signaling, and neutralizing antibodies to CXCR4. When mixed with nonfluorescein-labeled HUVECs pretreated with PTX (Fig. 3C
, left panel) or with anti-CXCR4 antibodies (Fig. 3C
, right panel), fluorescein-labeled HUVECs treated with TNF- (10 ng/ml) plus IFN- (100 ng/ml), and subsequently, coated with SDF-1, did not form tubular structures, providing evidence that a functional CXCR4 is required for tube formation by HUVECs expressing surface SDF-1. Together, these experiments demonstrate that disruption of ECM-dependent tube formation by HUVEC treatment with TNF- plus IFN- is dependent on reduction of SDF-1 and CXCR4 expression on HUVEC.
 View larger version (37K): [in a new window] |
Figure 3. TNF- and IFN- regulate ECM-dependent tube formation by endothelial cells. (A) Control HUVECs (left panel) and HUVECs exposed for 6 h to 10 ng/ml TNF- plus 100 ng/ml IFN- (right panel) were labeled with the green fluorescence dye PKH2 before plating on Matrigel-coated wells. (B) Control HUVECs (left panel), representing a mixture of 50% green fluorescent cells after staining with PKH2 and 50% unstained cells, and test HUVECs (right panel), representing a mixture of 50% cells treated for 6 h with 10 ng/ml TNF- plus 100 ng/ml IFN- and then loaded with SDF-1 (1 µg) and fluorescein-labeled with PKH2 and 50% control, unstained HUVEC, were incubated on Matrigel for 18 h. (C) HUVECs treated for 6 h with 10 ng/ml TNF- plus 100 ng/ml IFN-, loaded with SDF-1 and fluorescein-labeled with PKH2 were mixed (50/50) with unstained HUVECs treated with PTX (left panel) or with antibodies against CXCR4 (right panel). (A–C) Representative images obtained with an epifluorescence microscope (upper panels) and phase-contrast microscope (lower panels).
|
and IFN- can inhibit neovascularization in vivo. We used a previously described angiogenesis assay, in which bFGF mixed in with Matrigel and injected s.c. in mice promotes endothelial cell sprouting from adjacent vessels, invasion into the plug, and formation of capillaries containing red cells [24
, 25
]. Microscopic evaluation (representative images in Fig. 4A
and 4B
) and quantification (Table 2
) of the Matrigel plugs removed from the mice after 7 days revealed considerable cell infiltration in the plugs containing bFGF compared with plugs without bFGF. TNF- (30 ng/ml) and IFN- (300 ng/ml) mixed in with the Matrigel markedly inhibited bFGF-induced cell infiltration into the plug and new vessel formation (representative image in Fig. 4D
and Table 2
, Exps. 1 and 2). By immunohistochemistry, a proportion of cells infiltrating the Matrigel plugs impregnated with bFGF expressed cytoplasmic SDF-1 (representative image in Fig. 4B
, brown-staining cells indicated by the arrows). By contrast, the cells within plugs containing TNF- and IFN- together with bFGF were SDF-1-negative (representative image in Fig. 4D
). Only a few cells lining capillaries outside the plug were SDF-1-positive (representative image in Fig. 4D
, brown-staining cells indicated by arrowhead).
 View larger version (142K): [in a new window] |
Figure 4. Neovascularization and SDF-1 expression in vivo are inhibited by TNF- and IFN-. Groups of mice were injected s.c. with (A) Matrigel alone, (B) Matrigel plus bFGF, (C) Matrigel plus TNF- (30 ng/ml) and IFN- (300 ng/ml), or (D) Matrigel plus bFGF plus TNF- (30 ng/ml) and IFN- (300 ng/ml), and plugs were removed after 7 days. Histological sections from Matrigel plugs embedded in paraffin were immunostained with anti-SDF-1 antibodies and counterstained with hematoxylin and eosin. The arrows point to cells staining for SDF-1 (brown) within the Matrigel plug, and the arrowheads point to cells staining for SDF-1 (brown) outside the Matrigel plug.
|
|
View this table: [in a new window] |
Table 2. TNF- and IFN- Inhibit Growth Factor-Induced Angiogenesis In Vivo
|
and IFN-, we examined whether excess SDF-1 could reverse this inhibition and whether PTX itself could inhibit angiogenesis induced by bFGF. SDF-1, even at high concentrations (1 µg/ml), failed to reverse angiogenesis inhibition induced by TNF- and IFN- (Table 2
, Exp. 2). This result is in agreement with the observation that SDF-1 alone did not reverse inhibition of tube formation-induced TNF- and IFN- and is explained by the concomitant reduction of CXCR4 induced by TNF- and IFN- (Fig. 2)
. Indeed, we found that PTX markedly reduced neovascularization induced by bFGF (Table 2
, Exp. 3), providing evidence that Gi signaling is critically required in this Matrigel-based angiogenesis system. These experiments demonstrate that TNF- and IFN- can inhibit neovascularization and suggest a causative role for impaired SDF-1/CXCR4 signaling in this process. |
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and IFN-, used at concentrations one would expect to find in inflammatory tissues with pronounced tissue injury [4
, 5
, 14
], markedly inhibit neovascularization in vivo in the presence of optimal concentrations of the potent proangiogenic growth factor bFGF. Based on these results, TNF- and IFN- would be expected to inhibit neovascularization at sites of inflammation even within a strongly proangiogenic milieu generated by hypoxia and inflammatory cells.
We systematically examined endothelial cell expression of a large panel of genes that have been linked to regulation of endothelial cell growth and angiogenesis. We find expression of selected angiogenesis-related genes to be significantly regulated by TNF- and IFN-. Some of these genes, including IP-10, Mig, and IL-8, are known to be TNF-- and IFN--induced genes. However, others, including WARS, VEGF-C, TIMP-1, PECAM-1, angiopoietin-2, VEGFR-2, COL4A3, and CECAM1, are not known to be TNF-- and IFN--regulated genes in endothelial cells. Based on previous functional results, increased endothelial cell expression of the angiogenic factors IL-8 [17
] and VEGF-C [26
] and reduced expression of COL4A-3, which constitutes a source of natural inhibitors of angiogenesis [27
, 28
], would be expected to stimulate endothelial cell growth [29
]. The effect of reduced expression of angiopoietin-2 would be variable and context-dependent [29
30
31
32
33
]. By contrast, the transcriptional regulation of other genes in endothelial cells, induced (IP-10, Mig, WARS, and TIMP-1) or inhibited [kinase insert domain-containing receptor (KDR), SDF-1, ECGF-1, and CEACAM1] by TNF- and IFN-, would be expected to reduce or delay vascular repair at sites of inflammation. IP-10 and Mig, acting directly on endothelial cells or through the action of natural killer and T lymphocytes they attract, can be toxic for endothelial cells and reduce their ability to form new vessels [19
, 24
, 25
]. WARS is an enzyme that catalyzes the acetylation of tRNA with tryptophan, which exists in two forms, long and short, generated by alternative splicing [34
]. The short form of WARS is angiostatic in a variety of assays [34
, 35
], and a recombinant fragment of WARS, similar in length to the natural short form of the protein, inhibited retinal angiogenesis in vivo [36
]. Increased expression of TIMP-1 may compromise ECM remodeling and release of angiogenic factors. Like other TIMPs, TIMP-1 can bind and inactivate MMPs [37
], a family of endopeptidases that can degrade components of the ECM [38
], and facilitate endothelial cell assembly into vascular structures [6
]. In addition, MMPs can serve to liberate matrix-bound angiogenic factors, including VEGF-A and bFGF [7
]. KDR/VEGFR-2 is the signaling receptor for VEGF-A, which is recognized as a key endogenous regulator of angiogenesis [39
]. PECAM/CD31 is a major constituent of endothelial cell intercellular junctions, which has been implicated in integrin activation, transendothelial migration, and regulation of angiogenesis [40
]. ECGF-1 and CEACAM-1 are previously identified stimulators of endothelial cell growth and angiogenesis, at least in some experimental conditions [30
31
32
].
In addition to regulating IP-10, Mig, WARS, TIMP-1, KDR, PECAM, ECGF-1, and CEACAM-1 in a manner that would be expected to limit angiogenesis, our experiments show that TNF- and IFN- substantially reduce the expression of SDF-1 and CXCR4 in endothelial cells in vitro and in vivo. It was previously reported that TNF- and IL-1ß can reduce SDF-1 expression in fibroblasts [41
] and that IFN- can reduce CXCR4 expression in vascular endothelial cells [42
], but it was not previously known that inflammatory cytokines can reduce SDF-1 in endothelial cells. The present experiments show that reduction of SDF-1 and CXCR4 expression in endothelial cells is critical to inhibition of angiogenesis by TNF- and IFN-. Previously, we reported that surface expression of SDF-1 and its unique receptor, CXCR4, is required for endothelial cell morphogenesis into primordial vessels [22
]. Endothelial cells constitutively express SDF-1 bound to cell-surface proteoglycans [43
44
45
46
47
48
], which triggers CXCR4 signaling on adjacent endothelial-initiating cell motility [22
]. We now show that replacement of surface SDF-1, lost during exposure to TNF- and IFN-, allows the endothelial cells to regain an ability to form normal tubular structures as long as they encounter endothelial cells that express a functional CXCR4 receptor. These results confirm a critical role of SDF-1/CXCR4 signaling in mediating early events in endothelial cell alignment into vascular structures and demonstrate that reduced SDF-1/CXCR4 expression plays a critical role in mediating angiogenesis inhibition by TNF- and IFN-.
Thus, we demonstrate that TNF- and IFN- transcriptionally regulate a set of endothelial cell genes in a manner that would be expected to result in impairment of different steps in angiogenesis. Among these genes, we find reduced expression of SDF-1/CXCR4 to play a critical, early role in mediating the antiangiogenic effects of TNF- and IFN-.
Received December 2, 2003; revised January 29, 2004; accepted February 16, 2004.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
: biology and role in pathogenesis Adv. Immunol. 62,61-130[Medline]Vß3 in the disruption of the tumor vasculature induced by TNF and IFN- Nat. Med. 4,408-414[CrossRef][Medline] plus TNF- and inhibition by IL-4 and IL-13 J. Immunol. 154,1870-1878[Abstract], a potent inhibitor of endothelial cell growth in vitro, is angiogenic in vivo Proc. Natl. Acad. Sci. USA 84,5277-5281, IP-10, and Mig expression in Epstein-Barr virus-induced infectious mononucleosis and posttransplant lymphoproliferative disease Am. J. Pathol. 155,257-265-inducible tryptophanyl-tRNA synthetase includes alternative splicing J. Biol. Chem. 270,397-403} and {ß} explains functional diversity Blood 103,2452-2459 associates with heparan sulfates through the first ß-strand of the chemokine J. Biol. Chem. 274,23916-23925This article has been cited by other articles:
 |
|
 |
|
  B-C Lee, H-C Hsu, W-Y I Tseng, M-Y M Su, S-Y Chen, Y-W Wu, K-L Chien, and M-F Chen Effect of cardiac rehabilitation on angiogenic cytokines in postinfarction patients Heart, June 15, 2009; 95(12): 1012 - 1018. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Hohensinner, C. Kaun, K. Rychli, A. Niessner, S. Pfaffenberger, G. Rega, A. Furnkranz, P. Uhrin, J. Zaujec, T. Afonyushkin, et al. The inflammatory mediator oncostatin M induces stromal derived factor-1 in human adult cardiac cells FASEB J, March 1, 2009; 23(3): 774 - 782. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Sameermahmood, M. Balasubramanyam, T. Saravanan, and M. Rema Curcumin Modulates SDF-1{alpha}/CXCR4-Induced Migration of Human Retinal Endothelial Cells (HRECs) Invest. Ophthalmol. Vis. Sci., August 1, 2008; 49(8): 3305 - 3311. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Yano, Z. Liu, J. Donovan, M. K. Thomas, and J. F. Habener Stromal Cell Derived Factor-1 (SDF-1)/CXCL12 Attenuates Diabetes in Mice and Promotes Pancreatic {beta}-Cell Survival by Activation of the Prosurvival Kinase Akt Diabetes, December 1, 2007; 56(12): 2946 - 2957. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Rajashekhar, M. Grow, A. Willuweit, C. E. Patterson, and M. Clauss Divergent and convergent effects on gene expression and function in acute versus chronic endothelial activation Physiol Genomics, September 11, 2007; 31(1): 104 - 113. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Chen, H. Bai, Y. Shao, M. Arzigian, V. Janzen, E. Attar, Y. Xie, D. T. Scadden, and Z. Z. Wang Stromal Cell-Derived Factor-1/CXCR4 Signaling Modifies the Capillary-Like Organization of Human Embryonic Stem Cell-Derived Endothelium In Vitro Stem Cells, February 1, 2007; 25(2): 392 - 401. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ohira, M. Sasaki, K. Harada, Y. Sato, Y. Zen, K. Isse, K. Kozaka, A. Ishikawa, K. Oda, Y. Nimura, et al. Possible Regulation of Migration of Intrahepatic Cholangiocarcinoma Cells by Interaction of CXCR4 Expressed in Carcinoma Cells with Tumor Necrosis Factor-{alpha} and Stromal-Derived Factor-1 Released in Stroma Am. J. Pathol., April 1, 2006; 168(4): 1155 - 1168. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Tzima, J. S. Reader, M. Irani-Tehrani, K. L. Ewalt, M. A. Schwartz, and P. Schimmel VE-cadherin Links tRNA Synthetase Cytokine to Anti-angiogenic Function J. Biol. Chem., January 28, 2005; 280(4): 2405 - 2408. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Frantz, K. A. Vincent, O. Feron, and R. A. Kelly Innate Immunity and Angiogenesis Circ. Res., January 7, 2005; 96(1): 15 - 26. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
