| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Experimental Therapeutics, Preclinical Pharmacology |
1 Department of Neurosurgery and 2 Laboratory of Pathology, Erasmus University Hospital; 3 Laboratory of Toxicology, Institute of Pharmacy; and 4 Department of Logical and Numerical Systems, Faculty of Applied Sciences, Université Libre de Bruxelles, Brussels, Belgium
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Experimental Design: The influences of G17 and G17 receptor antagonists were characterized in vitro in terms of angiogenesis on human umbilical vein endothelial cell (HUVEC) tubulogenesis processes on Matrigel and in vivo with respect to U373 orthotopic glioma xenografts. The influence of phosphatidylinositol 3'-kinase, protein kinase C, and nuclear factor-
B inhibitors was characterized in vitro on G17-mediated HUVEC tubulogenesis. G17-mediated release of interleukin (IL)-8 from HUVECs and G17-induced modifications in nuclear factor-B DNA binding activity were characterized by means of specific enzyme-linked immunosorbent assays. The influence of G17 on E- and P-selectin expression was determined by means of computer-assisted microscopy, whereas the influence of E- and P-selectin on HUVEC migration was approached by means of antisense oligonucleotides. The chemotactic influence of G17 and IL-8 on HUVEC migration was characterized by means of computer-assisted videomicroscopy with Dunn chambers.
Results: Messenger RNAs for cholecystokinin (CCK)A, CCKB, and CCKC receptors were present in HUVECs and microvessels dissected from a human glioblastoma. Whereas G17 significantly increased the levels of angiogenesis in vivo in the U373 experimental glioma model and in vitro in the HUVECs, the CCKB receptor antagonist L365,260 significantly counteracted the G17-mediated proangiogenic effects. G17 chemoattracted HUVECs, whereas IL-8 failed to do so. IL-8 receptor 
(CXCR1) and IL-8 receptor ß (CXCR2) mRNAs were not detected in these endothelial cells. Gastrin significantly (but only transiently) decreased the level of expression of E-selectin, but not P-selectin, whereas IL-8 increased the expression of E-selectin. Specific antisense oligonucleotides against E- and P-selectin significantly decreased HUVEC tubulogenesis processes in vitro on Matrigel.
Conclusions: The present study shows that gastrin has marked proangiogenic effects in vivo on experimental gliomas and in vitro on HUVECs. This effect depends in part on the level of E-selectin activation, but not on IL-8 expression/release by HUVECs.
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cholecystokinin (CCK; refs. 19 and 20 ) and gastrin (21, 22, 23) were identified in the brain and cerebrospinal fluid about three decades ago. In fact, CCK is the most abundant peptide system in the brain, and the brain is the main production site of CCK (24) . In contrast, gastrin production in the brain is limited to oxytocinergic hypothalamo-pituitary neurons and a few cerebellar and vagal neurons (24) . In normal mammals, the antral G cells in the stomach and proximal duodenum are the main sites of gastrin synthesis (24) . Of the various forms of gastrin found in blood or tissue, 90% are gastrin-17 (G17; ref. 24 ). The plasma concentrations of gastrin are more than 10 times greater than those of CCK (25) . Gastrin can be produced locally as an autocrine or paracrine growth factor in various tumor types (24 , 26, 27, 28) , including brain tumors (29) . Gastrin- and CCK-related peptides have the same biologically active COOH-terminal pentapeptide amide sequence, and gastrin can signal through several receptors and binding sites. Three receptors to which gastrin can bind have already been cloned. The CCKA (also named CCK1) receptor displays a high level of affinity in binding carboxy-amidated and tyrosyl-sulfated CCK peptides and a low level of affinity in binding nonsulfated CCK peptides and gastrin peptides (24 , 30) . The CCKB (also named CCK2) receptor is less selective than the CCKA receptor because it binds tyrosyl-sulfated and nonsulfated CCK peptides, gastrin, short COOH-terminal CCK, and gastrin fragments with an almost similar degree of affinity (24 , 30) . CCKA and CCKB receptors can homo- or heterodimerize (31) . Gastrin can also bind to the 78-kDa gastrin-binding protein, i.e., the so-called "CCKC gastrin receptor" (32) . Whereas CCKA receptors are not expressed by glioma cells (33 , 34) , CCKB receptors are expressed in low-grade glioma cells only (33) . CCKC receptors could be expressed by a large proportion of glioma cells (34) . In addition to the three cloned receptors, several gastrin-binding proteins have also been described, but they have not yet been cloned. A variant of the CCKB receptor, also labeled CCK-C (for "CCK cancer") but different from the 78-kDa CCKC receptor, has been described in human pancreatic cancers (35) . A selective receptor for glycine-extended gastrin has been described on the rat pancreatic carcinoma cell line AR4-2J (36) . Singh et al. (37) describe a "novel" gastrin receptor, and Rehfeld et al. (38) show that specific binding sites for the COOH-terminal tetrapeptide of CCK or G17 exist on hog pancreatic islets. We have also identified specific binding sites for the COOH-terminal heptapeptide of gastrin on U373 human glioblastoma cells (16) , which do not express CCKA or CCKB receptors (14 , 16 , 17 , 34) but do express CCKC receptor (34) . We used this U373 model to investigate whether CCKA and CCKB receptor antagonists are able to significantly increase the survival periods of U373 orthotopic xenograft-bearing nude mice by specifically reducing neoangiogenesis in these experimental gliomas. We performed polymerase chain reaction (PCR) analyses to investigate whether CCKA, CCKB, and CCKC receptor mRNAs are present in HUVECs and GBM microvessels.
IL-8 is a potent proangiogenic factor (39, 40, 41)
and is stored in the Weibel-Palade body of endothelial cells, from which it can be rapidly released on stimulation by histamine or thrombin (42)
. Gastrin not only induces the expression of IL-8 mRNAs but also stimulates its release in human AGS gastric epithelial cancer cells (43
, 44)
. We analyzed the influence of gastrin on IL-8 release in HUVECs. In addition, gastrin induces IL-8 expression in gastric cancer epithelial cells through the activation of nuclear factor (NF)-B (43)
. We thus analyzed (a) whether gastrin can activate NF-B in HUVECs (using a colorimetric-based assay for measuring NF-B DNA binding capacity; ref. 45
) and (b) whether specific NF-B inhibitors can prevent gastrin-mediated tubulogenesis in HUVECs. Pagliocca et al. (46)
show that gastrin induces branching morphogenesis in human gastric cancer AGS cells by activation of protein kinase C (PKC) and phosphatidylinositol 3'-kinase (PI3K). We investigated whether PI3K and PKC inhibitors can prevent gastrin-induced tubulogenesis in HUVECs. We then made use of computer-assisted videomicroscopy of HUVECs cultured in Dunn chambers (47)
to analyze whether gastrin and IL-8 have any chemotactic influence on HUVECs.
The level of expression of E-selectin is also under the control of NF-B (48
, 49)
, and soluble P-selectin can induce endothelial cell migration (50)
. Both E- and P-selectins are induced in endothelial cells by proangiogenic cytokines such as tumor necrosis factor (TNF)- or IL-1ß (49)
. We thus analyzed the patterns of expression of E- and P-selectins in HUVECs both with and without stimulation by gastrin, IL-8, and IL-1ß. We also used antisense oligonucleotides directed against E- and P-selectins to analyze whether E- and P-selectins are directly involved in HUVEC tubulogenesis processes.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Human HUVECs were established to grow as in vitro monolayers by means of a method (53) adapted from the procedure described by Gimbrone et al. (54) . Three distinct batches of HUVEC primocultures (Huvec 1, Huvec 2, and Huvec 3) have been used in the present work.
Compounds
Gastrin (pGlu-Gly-Pro-Trp-Leu-Glu-Glu-Glu-Glu-Glu-Ala-Tyr-Gly-Trp-Met-Asp-Phe-NH2) and CCK [H-Asp-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2] were purchased from Sigma (Bornem, Belgium). The L365,260 (3R-3-[N'-(3-methylphenyl)ureido]-1,3-dihydro-1-methyl-5-phenyl-2H-1,4-benzodiazepine-2-one) and L364,718 (3S(-)-N-(2,3-dihydro-1-methyl-2-oxo-5-phenyl-1H-1,4-benzodiazepine-3-yl-1H-indole-2-carboxamide) compounds were provided by ML Laboratories PLC (London, United Kingdom). Human IL-8 and IL-1ß were obtained from R&D Systems (Oxon, United Kingdom). Ro320,432 (bisindolmaleimide XI; 2-(8-[(dimethylamino)methyl]-6,7,8,9-tetrahydropyrido[1,2-a]indol-3-yl)-3-(1-methylindol-3-yl)maleimide) and LY-294,002 (C19H17NO3·HCl) were obtained from Sigma-Aldrich (Bornem, Belgium). caffeic acid phenethyl ester (CAPE), BAY-11-7085 [(E)-3-[(4-t-butylphenylsulfonyl]-2-propenenitrile] and Parthenolide (C15H20O3) were obtained from Biomol Research Laboratories Inc. (Plymouth Meeting, PA).
In vivo Stereotactic Procedures and Treatments
Nude Rats with Surgery.
The brains of 8-week–old female nude rats (150 g; Hsd:RH-nu; Harland, Horst, the Netherlands) were stereotactically grafted with 106 human U373 cells. The U373 glioma-bearing rats were split into three groups: one group was left untreated (control), whereas the other two groups underwent surgery to remove the bulk of their tumors (the number of grafted animals in each group is indicated in the Fig. 1
legend). The location of each tumor was established by means of nuclear magnetic resonance imaging. During the surgery carried out on the U373 glioma-bearing rats, a micropump (Alzet micro-osmotic pump; model 1002; Alza Corp., Palo Alto, CA) was installed subcutaneously on each rat’s back immediately after the removal of the U373 tumor bulk. The end of the catheter was implanted directly into the brain cavity resulting from the tumor resection and maintained on the skull with cement. Illustrations of the full procedure are given elsewhere (14)
.
|
). Three H&E-stained slides were analyzed for each brain, and the number of vessels like the one illustrated in Fig. 1C
were recorded at x400 magnification using a 10 x 10 grid fitted into the microscope eyepiece. Five fields were analyzed per slide, totaling 15 fields per U373 tumor.
Nude Mice without Surgery for Short-Term Angiogenesis-Related Analyses.
The brains of thirty-five 8-week–old female nu/nu mice (21–23 g; Iffa Credo, Charles Rivers, Arbresle, France) were stereotactically implanted with 106 U373 cells, as detailed elsewhere (51)
. The 35 mice were split into seven groups of equal size 10 days after tumor graft. The first group (control) received three intracranial (through the hole of tumor cell stereotactic implantation) administrations of 20 µL saline per week (on Mondays, Wednesdays, and Fridays), for 4 consecutive weeks. The second group received twelve 20-µL intracranial administrations containing 10 mg/kg gastrin instead of saline, and the third and fourth groups received 12 administrations of 10 mg/kg L365,260 (a CCKB receptor antagonist) and L364,718 (a CCKA receptor antagonist), respectively. The fifth and sixth groups were the recipients of 12 concomitant administrations of G17 + L365,260 and G17 + L364,718, respectively, whereas the final group underwent twelve 20-µL intracranial administrations containing 10 mg/kg CCK-8. The administrations started at the 14th day after tumor graft, and all of the animals were euthanized on the 43rd day after tumor graft, i.e., 24 hours after the final intracranial administration. The brains were processed for histology to determine the levels of angiogenesis as detailed above for the nude rats.
Unlike Schmidt et al. (7) , we did not use Alzet osmotic minipumps because gastrin was not stable over the 28 days required for the experiments (data not shown).
Nude Mice without Surgery for Long-Term Survival– and Angiogenesis–Related Analyses.
The brains of thirty 8-week–old female nu/nu mice (21–23 g; Iffa Credo, Charles Rivers) were stereotactically implanted with 106 U373 cells, and the mice were then split into three groups of equal size 3 days after tumor graft. The first group (control) received daily intraperitoneal administrations of 0.2 mL of saline, whereas the second and third groups received daily administrations (10 mg/kg) of either L364,718 or L365,260. The administrations started the 5th day after tumor graft. Each animal was euthanized when it had lost 10% of its weight, and, as detailed above for the nude rats, the brains were processed to determine the levels of angiogenesis.
All of the in vivo experiments described were performed with the authorization of the Animal Ethics Committee of the Faculty of Medicine of the Université Libre de Bruxelles (Agreement No. 55/LA 1230342).
Determination of the mRNA Expression of Gastrin/CCK Receptors, ILs, and IL-Rs in HUVECs and Vessels Microdissected from a Human Glioblastoma
Total RNA Extraction from HUVECs.
Using the TRIzol isolation reagent (Life Technologies, Inc., Merelbeke, Belgium) according to the manufacturer’s instructions, total RNA was extracted from the HUVEC lines under study. The RNA extracted was treated with DNase I (Life Technologies, Inc.) to eliminate any remaining genomic DNA. The quantity of RNA was measured by spectrophotometric analysis at 260 nm (Beckman Coulter DU640; Analis, Ghent, Belgium). The quality and integrity of the extracted RNA was assessed by both BioAnalyzer 2100 (Agilent, Toulouse, France) and gel electrophoresis in 1.2% agarose Tris-Acetate-EDTA (TAE) gels and visualized by ethidium bromide staining under ultraviolet (UV) light. This verification was completed by an analysis of ß-actin gene expression by means of a standard reverse transcription-PCR (RT-PCR) method (see below).
Standard RT-PCR Analyses.
All reverse transcription and PCR reactions were carried out in a thermal cycler (Thermocycler, Westburg, Leusden, the Netherlands). The purification of the cDNAs produced was carried out using the High Pure PCR Product Purification Kit (Roche Diagnostics, Mannheim, Germany) in accordance with the manufacturer’s instructions.
As mentioned above, the integrity of the cDNA was confirmed by an analysis of ß-actin gene expression on the basis of a 25-cycle PCR analysis in a total volume of 50 µL with 20 ng of loaded cDNA. All of the PCR analyses were performed on the basis of the same quantity of purified cDNA (total amount, 20 ng). The products amplified by means of the standard PCR reaction [40 cycles, except for CCKB (50 cycles)] were resolved by gel electrophoresis in 1.2% agarose TAE gels in parallel with a 1-kb plus DNA ladder (Invitrogen, Carlsbad, CA) and visualized by ethidium bromide staining under UV light. The primers used for all of the PCR analyses (including those carried out on the microdissected vessels described below) are listed in Table 1
. They were provided by Invitrogen and selected using the HYBSIMULATOR software (Advanced Gene Computing Technology, Irvine, CA). The sequencing of the PCR products was performed by GenoScreen (Lille, France) using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA).
|
The total RNA isolation from the samples obtained by microdissection was carried out by means of the RNA Isolation Kit (Biozym; Gentra Systems, Minneapolis, MN) according to the manufacturer’s instructions.
RT-PCR Analysis for Microdissected Vessels.
The RNA extracted was transcribed into cDNA using a SensiScript Reverse Transcriptase kit (Qiagen, Westburg; Leusden, the Netherlands) according to the manufacturer’s recommendation. This procedure is specifically recommended for first-strand cDNA synthesis using less than 50 ng of RNA. The reaction was performed in a total volume of 20 µL containing 2 µL of supplied 10x concentrated buffer, 1 µL of oligo(dT)12–18 primers (10 µmol/L), 2 µL of a supplied mix of deoxynucleotide triphosphates (5 mmol/L each), 1 µL of supplied enzyme, 1 µL of RNase inhibitor (10 units/µL), 9 µL of supplied RNase-free water, and 4 µL of extracted RNA. The mix was incubated for 60 minutes at 37°C, and the enzyme was inactivated by heating the reaction mixture at 93°C for 5 minutes before rapid cooling on ice. The products of 10 reverse transcription reactions were pooled and stored at –80°C.
The integrity of the cDNA was confirmed by an analysis of ß-actin gene expression on the basis of a 45-cycle PCR method in a total volume of 25 µL containing 5 µL (one fourth of one reverse transcription reaction) of loaded cDNA. The absence of contamination by genomic DNA was verified by means of the PCR analysis of ß-actin gene expression in a nontranscribed RNA sample. The evaluation of the expression of the different mRNAs under study was performed by means of a 50-cycle PCR method in a total volume of 25 µL containing 5 µL (one fourth of one reverse transcription reaction) of loaded cDNA. The remaining steps (MgCl2 concentration, annealing temperature, electrophoresis, and visualization) were similar to those performed in standard RT-PCR analyses (see above).
Complementary DNA Microarray Assays.
We used two types of microarrays provided by Superarray (Bethesda, MD) and referenced as the GEArray Original Series Human Common Cytokine Gene Array and the GEArray Original Series Human Interleukin Receptor Gene Array, respectively. Whereas the former was designed to assess the levels of expression of 23 common cytokines, the latter was specifically concerned with the level of expression of 23 interleukin receptors (spotted in duplicate). A complete description of these arrays can be found in files xpd_hGEA9912090.pdf and xpd_hGEA9913020.pdf on-line.6
The array kits were used in accordance with the procedures recommended by the manufacturer on-line7
to assess the levels of gene expression in equal amounts of total RNA extracted from untreated HUVECs. The hybridization results were visualized using a Fuji-BAS5000 scanner and AIDA image analyzer software (Raytest Benelux, Tilburg, the Netherlands).
Enzyme-Linked Immunosorbent Assay Tests
HUVEC culture supernatants were collected after different treatments and periods of time. Separate aliquots of these supernatants were stored at –20°C. Two different human enzyme-linked immunosorbent assay (ELISA) kits were used in our study (in accordance with the procedures recommended by the manufacturers), namely, Human IL-1ß Quantikine and Human IL-8 Quantikine Parameter ELISA Kits (R&D Systems). Each sample was assessed in triplicate.
NF-B DNA Binding Assay
NF-B DNA binding activity was assessed with trans-active motif (trans-AM) NF-B family transcription factor assay kits (Active Motif Europe, Rixensart, Belgium) according to the manufacturer’s instructions. This ELISA-like test measures the level of the active form of NF-B contained in cell extracts specifically able to bind to an oligonucleotide containing the NF-B consensus site (5'-GGGACTTTCC-3') attached to a 96-well plate (45)
. Whole cell lysates were prepared after G17 treatment, and 20-µg extracts were added to the 96-well plates. The binding of NF-B to the DNA was visualized by anti-p50, anti-p52, anti–p65/Rel-A, anti–Rel-B, and anti–c-Rel antibodies that specifically recognize activated NF-B (45)
. Antibody binding was determined as absorbance values at 450 nm (45)
.
HUVEC Capillary-Like Structure Formation
Twenty-four–well culture plates (Nunc, Nalge Europe; Neerijse, Belgium) were coated with 250 µL of Matrigel (BD Biosciences), a basement membrane matrix liquid at 4°C, which was then allowed to solidify at 37°C for 30 minutes. The HUVECs growing as primocultures in 25-mm2 flasks (Nunc) were trypsinized, counted, suspended in a culture medium, and added to the Matrigel-coated wells (100,000 cells per well in 900 µL of medium). Different compounds (see Results) were added to the HUVEC suspensions just before seeding onto the Matrigel support. The HUVECs were incubated for 10 hours at 37°C, and digitized pictures were made (with a computer-assisted phase-contrast Olympus IX50 microscope; Omnilabo SA, Antwerp, Belgium) every 2 minutes (during this 10-hour period of observation) to determine the influence of these compounds on the ability of the HUVECs to form tubule-like structures. Each experimental condition was carried out in triplicate.
The E-selectin antisense and scrambled oligonucleotides took the form of 5'-fluo-GCTAAGTTAATGGAAGT-phosphorothioate-3' (antisense) and 5'-fluo-AGTAGGGCATAAGTTTA-phosphorothioate-3' (scrambled), whereas 5'-fluo-GCTATTGGCTTATTAAC-phosphorothioate-3' (antisense) and 5'-fluo-CAATTATTCGGTTATCG-phosphorothioate-3' (scrambled) constituted the P-selectin antisense and scrambled oligonucleotides. These four oligonucleotides, purchased from BioSource Europe (Nivelles, Belgium), were also selected using HYBSIMULATOR software (Advanced Gene Computing Technology).
Quantitative Determination of G17- and IL-8–Induced Chemotactic Effects on HUVECs
Chemotaxis was assessed by the direct observation and recording of cell behavior in stable concentration gradients of G17 or IL-8 established in Dunn chemotaxis chambers (Weber Scientific International Ltd., Teddington, United Kingdom). As detailed elsewhere (47)
, these chambers have two concentric wells separated by an annular platform and enable radially directed linear diffusion gradients to be established. The cells were cultured on a coverslip that was then inverted onto a glass slide. The cells attached to the annular platform were observed under a phase-contrast microscope and recorded automatically by an acquisition system described elsewhere (47)
. A software that we had previously developed enables the MRDO parameter to be quantitatively determined; MRDO represents the maximum relative distance traveled by each cell from its point of origin during a 24-hour period of observation, whereas MRDO is the vector joining the original point of a cell trajectory to the farthest point reached. We analyzed the distribution of the angular directions of these vectors to evidence a possible chemotactic effect of G17 as compared with IL-8 (see Statistical Analyses).
Computer-Assisted Fluorescence Microscopy for Determination of the Levels of Expression of E- and P-Selectin in HUVECs
The levels of expression and/or patterns of activation of E-selectin and P-selectin were quantitatively determined by means of computer-assisted fluorescence microscopy (as detailed elsewhere; ref. 52
) carried out on HUVECs cultured on gelatin substrates and treated with either G17, IL-8, or IL-1ß (as detailed in the figure legends) or left untreated (control). The anti–E-selectin and anti–P-selectin antibodies were purchased from R&D Systems. The HUVEC suspensions were seeded onto glass coverslips 48 hours before immunofluorescence staining; the cells were then fixed with 4% formaldehyde in PBS (pH 7.4) for 20 minutes. Three coverslips were available for each experimental condition. The cells were not permeabilized and were incubated for 1 hour at room temperature with the various primary and secondary antibodies.
The levels of expression and/or patterns of activation of the two selectins (relative to the fluorescence staining intensity) were determined quantitatively by means of a PROVIS Olympus Microscope (Omnilabo SA) coupled to a Megaview 2 camera (Omnilabo SA) feeding digitized information to a computer using AnalySIS software (Soft Imaging System, GMbH, Munster, Germany). One hundred cells were analyzed per cell line for each of the markers under study.
Statistical Analyses
Statistical comparisons between the control and the treated groups were made by first carrying out the Kruskal-Wallis test (a nonparametric one-way analysis of variance), and where this test revealed significant differences, we investigated whether any of the treated groups differed from control. For this purpose, we applied the Dunn multiple comparison procedure (two-sided test) adapted to the special case of comparisons of treatments and control, i.e., where only (k – 1) comparisons were made among the k groups tested by the Kruskal-Wallis test [instead of the possible k(k – 1)/2 comparisons considered in the general procedure (55
, 56)
].
Survival analysis was performed by using Kaplan-Meier curves and Gehan’s generalized Wilcoxon test. All these statistical analyses were carried out using Statistica (Statsoft, Tulsa, OK).
To study the impact of a potentially chemoattractive agent (say G17 or IL-8) on cell displacements, we analyzed its influence on the distribution of the MRDO vectors used to characterize the directions of the cell trajectories (see above). The standard Rayleigh test (57)
was used for this purpose (using an algorithm that we developed).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
illustrates the histopathological pattern of a human U373 glioma (T) implanted into the brain (B) of a nude rat. Fig. 1B
shows the presence of pseudopalisading processes with necrosis (N), and the arrow points to a microvessel further magnified in Fig. 1C
. As illustrated in Fig. 1C
, determination of the levels of angiogenesis depends on the count of this type of microvessel. In sharp contrast, we did not take into account the types of vessels illustrated in Fig. 1G
for the reasons explained below.
The continuous delivery by means of a micropump of G17 (Fig. 1D
, S+G) into the surgical resection cavity significantly increased neoangiogenesis in both the tumor bulk (Fig. 1D
, 
) and the peritumoral areas (Fig. 1D
,
) of the U373 gliomas implanted into the brains of the nude rats, but not in normal brain tissues (Fig. 1D
, 
), as compared with saline (Fig. 1D
, S+S) or with an actual control group of untreated U373 gliomas (Fig. 1D
, CT).
Twelve intracranial administrations of G17 (Fig. 1E
, G17) significantly increased the levels of neoangiogenesis in the U373 experimental GBMs as compared with the same type of delivery for saline (Fig. 1E
, Control). Twelve intracranial administrations of either L365,260 or L364,718 did not significantly modify the levels of neoangiogenesis in the orthotopic U373 gliomas. In contrast, the daily intraperitoneal administration of L365,260, but not L364,718, for about 2 months (thus about 40 intraperitoneal administrations of L365,260) significantly decreased these angiogenesis levels (Fig. 1I)
, with a concomitant L365,260-induced increase in the survival periods of the U373 orthotopic xenograft-bearing nude mice, a feature not observed in the case of L364,718 (Fig. 1H)
. L365,260 significantly antagonized the G17-induced proangiogenic effects (Fig. 1E
, L365 + G17), a feature not observed with L364,718 (Fig. 1E
, L364 + G17). CCK-8 induced only weak proangiogenic effects as compared with G17 (Fig. 1E)
, a fact that encouraged us to pursue our investigations with G17 alone.
As illustrated in Fig. 1G
, in many of the gastrin-treated tumors (about 70%), we observed large lacunae full of blood, a feature that was not observed in the control tumors (Fig. 1F)
. These lacunae could correspond to gastrin-induced sprouting of endothelial cells, but with the incomplete vessel formation that can be seen in many types of tumors (58)
.
Characterization of Expression of CCKA, CCKB, and CCKC Receptor, IL and IL-R mRNAs, and CXCR1 and CXCR2 mRNAs in HUVECs and in Vessels Microdissected from a Human Glioblastoma.
As illustrated in Fig. 2A–C
, we observed strong mRNA expression for the CCKA, CCKB, and CCKC receptors in three independent HUVEC primocultures; the mRNA expression for these receptors was also evidenced in the microvessels microdissected from a human glioblastoma (Fig. 2E–G)
. All of the PCR products were checked by means of sequencing (data not shown).
|
shows that the IL-8 mRNAs were markedly present in HUVECs (as they were in the GBM microdissected microvessels; Fig. 2P
), whereas Fig. 2J
indicates that the IL-8 receptor [i.e., CXCR1 (IL-8 receptor ) and CXCR2 (IL-8 receptor ß)] mRNAs were not present in the HUVECs. The absence (or only very weak presence) of CXCR1 and CXCR2 mRNAs was further confirmed by means of standard PCR analyses (Fig. 2K and L)
.
TNF- (Fig. 2N)
and IL-8 (Fig. 2P)
mRNAs were detected in the vessels microdissected from a human glioblastoma, but no IL-6 (Fig. 2O)
, gp130-oncostatin-M receptor chain (Fig. 2Q)
, or oncostatin M (Fig. 2R)
mRNA was detected.
Characterization of the Influence of G17 and CCKB (and CCKA) Receptor Antagonist on the Ability of HUVECs to Form Capillary-Like Structures In vitro.
HUVECs form capillary-like structures when they are plated on Matrigel, as illustrated in Fig. 3A
(t = 0 h) and B (t = 2 h), but they remain individual when plated in a Dunn chamber (without Matrigel, but on gelatin), as illustrated in Fig. 3C
(t = 0 h) and D (t = 8 h).
|
illustrates our method of analysis of angiogenesis [i.e., by counting the number of branches (b) present on a 1,000-µm2 surface at x20 magnification]. Fig. 3B
shows that 11 branches (b1–b11) were present in the field that we analyzed. For each experimental condition, we analyzed six fields like the one illustrated in Fig. 3B
. Each experiment was performed in triplicate.
Fig. 3E
shows that 10 nmol/L G17 () significantly increased the rate at which the HUVECs formed tubulogenesis as compared with control (
). Whereas G17 at 0.01 nmol/L caused no statistically significant effects on HUVEC capillary-like formation as compared with control, the data obtained with 0.1 and 1 nmol/L were similar to but weaker than those illustrated in Fig. 3B
(at 10 nmol/L; data not shown). The G17-mediated effects on HUVEC tubulogenesis at 100 nmol/L were not higher (P < 0.05) than those observed at 10 nmol/L (see Fig. 4A
). At 10 nmol/L, L365,260 (a CCKB receptor antagonist) did not have any effect by itself, nor did it antagonize the G17-mediated effects (data not shown). In contrast, 100 nmol/L L365,260 completely antagonized the G17-mediated proangiogenic effects, a feature not observed with 100 nmol/L L364,718 (Fig. 3F)
.
|
; the effects observed at 100 nmol/L were similar (P > 0.05) to those observed at 10 nmol/L (Fig. 4A)
. Whereas 10 µmol/L LY294,002 induced no statistically significant effects on the HUVEC tubulogenesis, 50 µmol/L LY294,002 slightly, but nevertheless significantly (P < 0.05), decreased it (Fig. 4A)
. Whereas the addition of 10 µmol/L LY294,002 one hour before 10 nmol/L G17 brought about a slight decrease in HUVEC tubulogenesis, the addition of 50 µmol/L LY294,002 one hour before 10 nmol/L G17 provoked a significant decrease (Fig. 4A)
. Increasing the concentration of G17 from 10 to 100 nmol/L blocked the L294,002-induced inhibitory effects on HUVEC tubulogenesis (Fig. 4A)
. These data suggest that PI3K plays a major role in the G17-mediated effects on HUVEC tubulogenesis features. Fig. 4B
illustrates the morphologic appearance of HUVEC tubulogenesis 7 hours after the addition of 10 nmol/L G17 to the culture media, whereas Fig. 4C
illustrates the morphologic appearance of the HUVECs coated on Matrigel, where 50 µmol/L LY294,002 has been added to the culture medium 1 hour before 10 nmol/L G17. Fig. 4C
clearly shows that the HUVECs treated with 50 µmol/L LY294,002 one hour before 10 nmol/L G17 were alive, but they were no longer capable of undergoing tubulogenesis. Fig. 4D
shows that Ro320,432 significantly decreased HUVEC tubulogenesis, probably through inhibition of the culture medium growth factor-mediated activation of this tubulogenesis. The fact that the concomitant addition of Ro320,432 and G17 only slightly decreased the G17-mediated effects on the HUVEC tubulogenesis suggests that the PKC signaling pathways were only marginally involved in the G17-mediated effects on HUVEC tubulogenesis. On the contrary, G17 was potent enough to antagonize the Ro320,432-induced inhibitory effects on HUVEC tubulogenesis (Fig. 4D)
.
Characterization of G17-Induced Effects on IL-8 Secretion by HUVECs and on NF-B–Mediated IL-8 Secretion and Potential NF-B–Mediated Effects on HUVEC Tubulogenesis.
G17 did not significantly (P > 0.05) increase IL-8 release (secretion) from HUVECs 2 hours after the addition of G17 to the HUVEC culture (Fig. 5A
,
). In contrast, G17 markedly (P < 0.001) stimulated IL-8 release from HUVECs 12 hours after the addition of G17 to the culture media (Fig. 5A
, ). Thus, the G17-mediated increase in the HUVEC tubulogenesis process (see Fig. 3
) cannot be related to a G17-induced release of IL-8 from the HUVECs because G17 had already significantly increased HUVEC tubulogenesis 2 hours after the addition of G17 to the culture media (Fig. 3D)
, whereas G17 had not significantly increased IL-8 release from the HUVECs by then (Fig. 5A)
. G17 (10 nmol/L) did not significantly activate NF-B DNA binding activity during the first 2 hours after addition of G17 to the HUVEC culture media (Fig. 5B)
.
|
schematically illustrates where on the NF-B pathway (see explanations in Discussion) the three inhibitors Parthenolide, BAY-11-7085, and CAPE, specifically have an inhibitory effect. Fig. 5D
shows that G17 significantly increased HUVEC tubulogenesis 10 hours after its addition to the culture media. Whereas the G17-induced increase in HUVEC tubulogenesis was not antagonized by CAPE, it was slightly antagonized by Parthenolide and BAY-11-7085 only (Fig. 5D)
.
Characterization of G17–, IL-1ß–, and IL-8–Induced Chemotactic Effects on HUVEC Migration.
The chemotactic-related experiments into G17 and IL-8 were carried out independently and in triplicate, with about 300 cell trajectories analyzed in each of the experimental conditions (after pooling). These trajectories are illustrated in Fig. 6A
, in which each colored line corresponds to a path traveled by an individual HUVEC (see Fig. 3C and D
) over an 8-hour observation period. In Fig. 6B
, all of the initial cell positions are set to the origin (0, 0) of the axes, and the trajectories are reported in terms of gradient direction. The gradient direction is arbitrarily fixed in an easterly direction (the red dot highlighted by the red arrow). The red arrows in Fig. 6C and D
thus indicate the point where either saline (control) or 10 nmol/L G17 was added. Each individual HUVEC trajectory is characterized by means of its MRDO vector (joining the initial and most distant cell positions) in Fig. 6C and D
(blue arrows). These figures also compare the resulting direction of all of the MRDO vectors (green arrows) with the gradient direction (red arrows). Fig. 6C and D
clearly show that G17 had a chemotactic effect on the HUVECs. Rayleigh statistical test analyses applied to the computer-assisted microscope quantification of these directional motility features revealed that G17 significantly chemoattracted the HUVECs at 0.1 (P < 0.001), 1 (P < 0.001), and 10 nmol/L (P < 0.01), but not at 100 nmol/L (P > 0.05). We did not test whether concentrations of <0.1 nmol/L G17 still chemoattracted the HUVECs. Neither IL-8 nor IL-1ß significantly chemoattracted the HUVECs over concentrations ranging from 0.01 to 10 nmol/L (data not shown).
|
illustrates the morphologic aspect of the actin cytoskeleton in the untreated HUVECs (the 0 experimental condition in Fig. 6H
). The presence of fibrillary actin was revealed by green fluorescence, and that of globular actin was revealed by red fluorescence. Fig. 6F and G
illustrate the appearance of the actin cytoskeleton of HUVECs cultured in the presence of 10 nmol/L G17 for 3 and 6 hours, respectively, a feature indicating a marked G17-induced depolymerization of the fibrillary actin. The use of computer-assisted fluorescence microscopy made it possible to compute the ratio of fibrillary/globular actin, as shown in Fig. 6H
. This demonstrates that the G17-induced modifications to the actin cytoskeleton organization occurred mainly during the first 6 hours after the addition of G17 to the HUVEC culture media.
Characterization of G17–, IL-8–, and IL-1ß–Induced Effects on E- and P-Selectin Expression in HUVECs, with Emphasis on the Potential Roles of E- and P-Selectins in HUVEC Migration Features.
The influence of G17, IL-1ß, and IL-8 on the levels of expression of E- and P-selectin in the HUVECs was determined quantitatively by means of computer-assisted fluorescence microscopy. Fig. 7A and B
illustrate the patterns of E-selectin expression in an untreated HUVEC and in a HUVEC treated with 10 nmol/L G17 for 6 hours. In accordance with these patterns, quantitative immunofluorescence revealed that at 10 nmol/L, G17 (Fig. 7C
, red bars) dramatically but transiently decreased the levels of E-selectin in the HUVECs. Because we did not permeabilize HUVEC membrane during the E-selectin cytochemical staining process, this G17-induced transient disappearance of E-selectin on the surface of the HUVECs could correspond to a G17-induced internalization process of E-selectin rather than to an actual G17-induced decrease in E-selectin expression. G17 (Fig. 7C
, blue bars) began by slightly decreasing the levels of expression (or by slightly activating the internalization process) of P-selectin in the HUVECs and then dramatically stimulated P-selectin expression and/or the activation processes. As revealed by an ELISA (data not shown), the G17-induced decrease in E- and P-selectin levels of expression in the HUVECs (during the first 6 hours after their addition to the culture medium; see Fig. 7C
) did not correspond to a G17-induced increase in E- and P-selectin secretion by the HUVECs into their culture media. These G17-induced transient decreases in E- and P-selectin might therefore correspond to a G17-induced internalization of E- and P-selectin, as already suggested above in connection with the fact that no permeabilization procedure was used during the cytochemical staining process.
|
IL-1ß (Fig. 7D
,
) and IL-8 (Fig. 7D
, ) significantly increased the levels of expression of E-selectin in the HUVECs, but did not significantly increase the levels of expression P-selectin in the HUVECs (data not shown).
We used an antisense oligonucleotide approach (Fig. 7E and F)
to decrease the levels of E- and P-selectin (Fig. 7G)
membrane expression in the HUVECs to investigate the influence of this decrease on the ability of the HUVECs to form capillary-like structures. We made use of green fluorescent scrambled or antisense oligonucleotides and showed the presence of E- or P-selectin under red fluorescence (Fig. 7E)
. We then quantitatively determined (by means of computer-assisted fluorescent microscopy) the levels of E- or P-selectin in the transfected (red and green = yellow) HUVECs (Fig. 7F)
. Fig. 7G
illustrates the E-selectin (red bars) and P-selectin (blue bars) expression in untreated HUVECs (Ct) or in HUVECs transfected with a scrambled oligonucleotide (Scram.) or an actual antisense oligonucleotide (As_Ol.). Fig. 7G
shows that a concentration of 0.01 µmol/L E (red bars)- or P-selectin antisense oligonucleotides (blue bars) was able to decrease the levels of E- or P-selectin membrane expression in the HUVECs both selectively (as compared with scrambled oligonucleotides) and significantly. As illustrated in Fig. 7H
, the use of this concentration of E (red bars)- and P-selectin antisense oligonucleotides (blue bars) significantly decreased the ability of the HUVECs to form capillary-like structures, as compared with control (absence of oligonucleotides; ) and with conditions involving scrambled oligonucleotides for E- (open red bars)- and P-selectins (open blue bars).
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Whereas G17 increased neoangiogenesis in orthotopic xenografts of U373 human gliomas, CCK had marginal effects only (see Fig. 1
), and this is the reason why we decided to remain focused on G17-mediated effects on angiogenesis alone. A CCKB receptor antagonist (L365,260) significantly antagonized the in vivo G17-mediated proangiogenic effects in the U373 gliomas, whereas a CCKA receptor antagonist (L364,718) did not. In the same way, whereas daily delivery of L365,260 over several weeks to orthotopic U373 glioma-bearing nude mice significantly increased their survival periods as compared with control mice, L364,718 did not induce any such therapeutic benefit. The G17-induced increase in the survival periods of the U373 glioma-bearing nude mice was paralleled by a significant G17-induced decrease in neoangiogenesis in these gliomas. All these data therefore strongly suggest that, in vivo, G17 has a proangiogenic effect on malignant experimental gliomas, an effect that could be (at least partly) mediated by the CCKB receptor. We also observed that G17 induced large lacunae full of blood, as illustrated in Fig. 1G
. This process could correspond to a marked G17-mediated proangiogeneic effect leading to endothelial cell sprouting, but with incomplete vessel formation, a feature that is commonly observed in tumor neoangiogenesis (58)
.
We also observed that the daily administration of L365,260 to nude mice bearing orthotopic xenografts of human BxPC-3 pancreas cancer cells significantly decreased neoangiogenesis levels in these xenografts (data not shown). In the same manner, we observed that the daily administration of both L365,260 and L364,718 to nude mice bearing subcutaneous xenografts of human C32 melanoma cells significantly decreased angiogenesis in these xenografts, with a concomitant decrease in the growth rates of these experimental melanomas.8
The in vitro data obtained on the HUVECs indicate that G17 speeds up the HUVEC-related tubulogenesis process during the first hours of HUVEC culture on Matrigel, but without modifying the subsequent pattern of HUVEC tubulogenesis. These effects seem to be largely dependent on the CCKB receptor (see Fig. 3F
). The fact that the concomitant administration of L365,260 and L364,718 induced slightly, but nevertheless significantly, more marked antagonistic effects against G17-induced proangiogenic influences than L365,260 alone could relate to heterodimerization processes between CCKB and CCKA receptors (31)
.
The G17-mediated acceleration of HUVEC tubulogenesis seems to be partly mediated by PI3K because a PI3K antagonist (LY294,002) was able to significantly antagonize the G17-mediated proangiogenic effects on HUVEC tubulogenesis (see Fig. 4A
). In contrast, PKC seems to be less involved in G17-mediated effects on HUVEC tubulogenesis because the PKC inhibitor Ro320,432 did not antagonize the G17-induced proangiogenic effects on HUVEC tubulogenesis (Fig. 4D)
. Pagliocca et al. (46)
observe that when AGS gastric cancer cells (which are CCKB receptor positive) are cultured on plastic, gastrin stimulates cell adhesion, the formation of lamellipodia, and the extension of the already long processes in part by the activation of PKC and PI3K as observed here. Branching morphogenesis was not observed in these circumstances (46)
, as in the case of the gelatin-cultured HUVECs. When AGS cells were cultured on an artificial basement membrane, the same stimuli increased the formation of organized multicellular arrays exhibiting branching morphogenesis (46)
. These effects were reversed by PKC inhibitors, but not by PI3K inhibitors (46)
. In our study, we observed the reverse features in the case of the HUVECs, a fact that could relate to different intracellular signaling pathways activated by gastrin in gastric cancer cells as compared with normal endothelial cells. We observed slight, but nevertheless significant, modifications induced by gastrin at the levels of expression of both PKC and PKCµ (data not shown). These two PKC isoforms are involved in gastrin-mediated modifications of the actin cytoskeleton (59
, 60)
.
G17 stimulates the activities of various genes (including IL-8 and E-selectin) via activation of the NF-B pathway (43)
, and the levels of activation of the NF-B pathway play important roles in survival processes of endothelial cells against cytotoxic influences (48)
. Three of the four regulatory elements found in the human E-selectin promoter are NF-B binding sites (49)
.
NF-B is a collective designation for a family of highly regulated dimer transcription factors (61)
. Virtually all vertebrate cells express at least one of five Rel/NF-B members, namely, p50/p105 (NF-B1), p52/100 (NF-B2), c-Rel, p65 (RelA), and RelB, which are assembled into homo- and heterodimers (ref. 62
; Fig. 5C
). Almost every kinase pathway explored was found to participate in NF-B activation [examples include PKC, Akt/protein kinase B, c-Jun-NH2-terminal kinase, mitogen-activated protein kinase kinase kinase, and mitogen-activated protein/kinase/ERG kinase kinase (Fig. 5C)
], but few have been corroborated by genetic data (63)
. With few exceptions (e.g., H2O2 and UV radiation), all of the signals converge to activate IB kinase, leading to IB phosphorylation, ubiquitination, and degradation, a process that enables the nuclear translocation of NF-B dimers and their binding to DNA (ref. 64
; Fig. 5C
). Once activated, IKK phosphorylates the bulk of the cytoplasmic IB pool, promoting its degradation and NF-B activation (ref. 64
; Fig. 5C
). IB represents a family of NF-B inhibitors, of which IB, IBß, and IB
are the chief regulators in mammalians (ref. 64
; Fig. 5C
). We made use of three NF-B inhibitors, i.e., Parthenolide, BAY-11-7085, and CAPE (see Fig. 5C
), and we observed that these inhibitors exerted no effects or only weak antagonistic effects against G17-mediated proangiogenic effects with respect to the HUVEC tubulogenesis process (see Fig. 5D
). These data fit in with those illustrated in Fig. 5B
, which shows that G17 did not activate NF-B DNA binding activity in HUVECs during the first 2 hours after the addition of G17 to the HUVEC culture media.
The use of a cDNA microarray approach showed that, like GBM microvessels, HUVECs express mRNAs for both TNF- and IL-8, two major regulators of angiogenesis (9, 10, 11)
. We observed that G17 markedly stimulates IL-8 release by HUVECs. However, the G17-induced increase in IL-8 release by the HUVECs (Fig. 5A)
occurred after the G17-induced proangiogenic effects on HUVEC tubulogenesis (Fig. 3E)
, a fact that strongly suggests that this G17-induced proangiogenic effect is IL-8 independent. In the same way, whereas G17 induced chemotactic effects on individual HUVECs, IL-8 did not. The absence of IL-8–mediated chemotactic effects on the HUVECs can be at least partly explained by the absence of CXCR1 (IL-8 receptor ) and CXCR2 (IL-8 receptor ß) in the HUVEC primocultures that we used. The presence of CXCR1 and CXCR2 receptors varies larg
