| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
* Department of Human Pathology, University of Messina, Messina, Italy;
Institute of Experimental Pathology, University of Ancona, Italy;
Department of Medicine and Aging, University ‘G. D’Annunzio’, Chieti, Italy;
|| DNA Unit Hospital Clinic, Barcelona, Spain; and
¶ Department of Pathology, City Hospital, Alessandria, Italy
1Correspondence: Dipartimento di Medicina e Scienze dell’Invecchiamento, Cattedra di Ematologia, Università ‘G. D’Annunzio’, Via dei Vestini, 31, 66013, Chieti, Italy. E-mail: mromano{at}unich.it
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: mesothelioma • arachidonic acid • apoptosis • angiogenesis • eicosanoids
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
, and it has been shown that nonsteroidal anti-inflammatory drugs possess antineoplastic activity (3
, 4)
. Arachidonic acid can be also metabolized by the lipoxygenases (LO). The 5, 12, and 15 are considered the main LO isoforms catalyzing the biosynthesis of biologically active compounds such as leukotrienes and hydroxyeicosatetraenoic acids (HETEs) (5)
. Interaction between LO originates the lipoxins (LX), which possess a potent anti-inflammatory property (6)
. Recent studies have shown that LO are expressed in cancer cells, where they regulate proliferation and survival. 5- and 12-LO were found in human pancreatic and prostatic cancer cells, and their inhibition resulted in proliferative block and apoptotic cell death (7
, 8)
. On the other hand, a rat carcinosarcoma cell line underwent apoptosis when exposed to a 12-LO antisense oligonucleotide (9)
. Thus, LO inhibition may represent, at least in certain settings, a potential novel approach to cancer treatment.
Malignant pleural mesothelioma is an aggressive neoplasm that in most cases is fatal (10)
. An association between exposure to asbestos fibers and incidence of pleural mesothelioma has been established (11)
. Asbestos fibers have a profound effect on multiple mechanisms of the inflammatory reaction. They inhibit the activity of lymphokine-activated killer cells, which possess a strong lytic activity against malignant pleural mesothelial (MM) cells (12)
, and asbestos exposure is associated with decreased natural killer cell activity (13)
. Recently, the presence of SV-40 DNA sequences and the expression of SV-40 large T antigen (Tag) have been observed in human pleural mesothelioma (14
, 15)
. It has been proposed that SV-40 Tag may play a major role in the pathogenesis of mesothelioma (16
, 17)
. In vitro studies have shown that MM cells stimulate their own growth in autocrine/paracrine fashion by releasing numerous growth factors, including the proangiogenic vascular endothelial growth factor (VEGF). We have observed that MM cells produce large amounts of VEGF and express the VEGF receptors Flt-1 and KDR, and that neutralizing antibodies against VEGF and its receptors significantly reduce MM cell proliferation (18)
. Increased VEGF levels can be measured in pleural effusions of patients with mesothelioma (19)
. Thus, VEGF up-regulation appears to be a crucial event in mesothelial cell transformation. The mechanisms that regulate VEGF release and gene expression by mesothelioma cells are not completely known.
The presence of LO has not been reported in mesothelial cells, and the possibility that LO may regulate mesothelioma cell proliferation and survival has never been explored. In this report, we show for the first time that MM cells selectively express a catalytically active 5-LO and that this enzyme is a key regulator of MM cell survival through the modulation of VEGF release and gene expression.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. After 2 wk in culture, 100% of NM cells stained positive for calretinin. They were then expanded and used for the experiments. NM cells were used between the third and seventh passage. All cells were maintained in RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum, 1% L-glutamine, and 1% penicillin-streptomycin (a complete medium) (all from HyClone, Rome, Italy) at 37°C and 5% CO2.
Analysis of LO products
LO products were extracted from 48 to 72 h conditioned media of mesothelial cells. Media were mixed with two volumes of cold methanol containing prostaglandin B2 as internal standard and placed at -20°C for at least 2 h. Precipitates were removed by centrifugation at 1200 g for 15 min at 4°C. Materials were dried using a rotavapor and solid-phase extracted as described (20)
. The methyl formate fractions were dried in a SC 110 Speed Vac apparatus (Savant, Rome, Italy) and suspended in a small volume of methanol. Products were analyzed using a dual pump reversed-phase (RP) -HPLC gradient system equipped with a 996 Photodiode Array Detector (Waters, Milford, MA). The column was a Waters Symmetry C18, 3 µm, 2.1 x 150 mm. Post-run analysis was performed with the Millennium 32 chromatography manager. Mono-HETEs were identified by online UV spectral analysis and a comparison of their retention times vs. those of authentic standards (all from Cascade Biochem Ltd., Reading, England, except for 5(S)-HpETE, which was obtained from Cayman Chemical Company, Ann Arbor, MI). LTA4-methyl ester was converted to free acid by saponification as described (20)
.
LO antisense
Phosphorotioate antisense oligodeoxynucleotides and PCR primers were prepared by solid-phase phosphoramidite chemistry in a 394 DNA/RNA synthesizer (Applied Biosystems, Foster City, CA). 5- and 12-LO antisense and their relative sense and scrambled sequences were derived from previous studies (7
, 9)
. In our setting, a concentration of 50–100 nM of each antisense was sufficient to fully inhibit the accumulation of the correspondent LO mRNA, determined by RT-PCR. For cell treatment, 
5 x 104 cells/ml were exposed to the oligonucleotides as described (21)
. Cells were washed three times with prewarmed (37°C) serum-free RPMI and incubated with the antisense or its corresponding scrambled oligonucleotide previously mixed with 10 µg/ml of Lipofectin (Life Technologies, Milano, Italy) in serum-free RPMI for 4 h at 37°C, 5% CO2. Complete medium was then added to the cells for different periods.
Proliferation assays
For time course studies, 104 cells/ml were incubated with LO inhibitors (Biomol Research Laboratories, Plymouth Meeting, PA). Cells were collected after 24, 48, and 72 h, stained with trypan blue, and counted. For analysis of DNA synthesis, 4–5 x 104 cells/ml were cultured for 24 h in complete RPMI 1640 medium. Cells were then kept for 24 h in serum-free medium, which was replaced by complete medium containing LO inhibitors. Alternatively, cells were incubated with LO antisense oligonucleotides as indicated above. At the end of the incubation, [3H]-thymidine (0.5 µCi/ml) was added for an additional 4 h. Cells were then washed with PBS2+, incubated with ice-cold trichloroacetic acid (5%) for 10 min, washed again with ethanol/ether (2:1), and lysed with 0.5 ml of PBS2+ containing 0.5% triton, 200 mM NH4OH, 0.1% bovine serum albumin. Scintillation fluid (3 ml) was added to lysates and radioactivity was determined for 1 min in a ß-scintillation analyzer (Tri-Carb 2100TR, Packard, Downers Grove, IL).
Apoptosis
Apoptosis was assessed by DNA laddering and nucleosome formation. 106 cells were cultured in complete medium for 24 h, then washed twice with fresh complete medium and exposed to LO inhibitors or antisense oligonucleotides in complete medium for the times indicated. For evaluation of DNA laddering, DNA was phenol extracted and precipitated with ethanol. Low molecular weight DNA fragments were visualized by 2% agarose gel electrophoresis and ethidium bromide staining. Nucleosome formation was determined using a cell death detection ELISA kit (Roche Molecular Biochemicals, Milano, Italy) following the manufacturer’s instructions. Rescue experiments were carried out by adding arachidonic acid, LO metabolites, or growth factors (R&D System Europe Ltd., Abingdon, UK) simultaneously with LO inhibitors.
VEGF immunoassay
VEGF was measured in conditioned media of 70–80% confluent MM and NM cells exposed for varying times to arachidonic acid or LO inhibitors, metabolites, or antisense oligonucleotides using an ELISA assay that recognizes all the main VEGF isoforms (R&D System Europe) according to the manufacturer’s instructions. Media were kept at -80°C until assayed. Aliquots were centrifuged for 10 min at 800 g before measurements. Results were normalized for protein concentration determined using the Bradford reagent (Bio-Rad Lab., Hercules, CA).
Transfections
Transient transfections were performed using the lipofectAMINE plus reagent (Gibco BRL Ltd., Paisley, UK) according to the manufacturer’s instructions. Briefly, 2 x 105 cells/ml were maintained for 24 h in complete medium. Monolayers were then washed twice with serum-free RPMI and exposed for 3 h to 4 µg of each of the following plasmids in serum-free medium: green fluorescent protein/5-LO (GFP-5LO), 5-LO mutants H367Q, R131/132Q, H372Q, H550Q (22)
(kindly provided by Dr. C. Funk), 2068 bp fragment of the VEGF promoter cloned into the luciferase reporter vector pAH (23)
(kindly provided by Dr. G. Finkenzeller with permission from Dr. Judith Abraham from Scios, Sunnyvale, CA). Monolayers were then washed with serum-free medium, maintained in complete medium for 24 h, washed twice with serum-free medium, and exposed to test agents for an additional 24–48 h. For luciferase activity measurements, 1 µg of the ß-galactosidase expression plasmid pCDNA3 was cotransfected with the VEGF reporter plasmid. After transfection, cells were washed twice with PBS (2-)
and incubated with reporter lysis buffer (Promega, Madison, WI). Luciferase activity in cell extracts was determined using a Lumat LB 9507 Eg&G Berthold luminometer (Bad Widbad, Germany); results were normalized for transfection efficiency by determining ß-galactosidase activity with the Galacto-Light PlusTM kit (Tropix, Bedford, MA) according to the manufacturer’s instructions.
RT-PCR
Total RNA was extracted using the SV Total RNA Isolation System (Promega). RNA was denatured for 10 min at 70°C, then reverse-transcribed for 1 h at 42°C using StrataScript II (50 U/ml) (Stratagene, La Jolla, CA) in a 50 µl reaction. cDNA was incubated for 5 min at 99°C to inactivate the reverse transcriptase; 10 µl of cDNA was used for each amplification reaction. PCR was performed with Taq DNA polymerase, using the buffer supplied by the manufacturer (Promega). Preliminary PCR (from 20 to 50 cycles) was carried out to determine the optimal number of amplification cycles for quantitative evaluation. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers were used for PCR reaction control and semiquantitative analysis. Other oligonucleotide primers were 5-LO: 5'-GAAGACCTGATGTTTGGCTACC-3' (sense), 5'-AGGGTTCTCATCTCCCGG-3' (antisense) generating a 326 bp fragment (annealing temperature 56°C for 45 cycles); platelet-type 12-LO: 5'-GTTGAGACGCTTGACCTCTC-3' (sense) and 5'-GCAGCCAGGTATTGCTTCTC-3' (antisense) generating a 173 bp product (annealing temperature 58°C for 35 cycles); VEGF: 5'-GAAGTGGTGAAGTTCATGGATGTC-3' (sense), 5'-CGATCGTTCTGTATCAGTCTTTCC-3' (antisense) amplifying 408, 541 and 613 bp fragments corresponding to the VEGF121, VEGF165, and VEGF185 isoforms (annealing temperature 58°C for 25 cycles); platelet-derived growth factor B (PDGF-B): 5'-GAAGGAGCCTGGGTTCCCTG-3' (sense), 5' TTTCTCACCTGGACAGGTCG-3' (antisense) (226 bp fragment, annealing temperature 60°C for 40 cycles); transforming growth factor ß1 (TGF-ß1): 5'-GCCCTGGACACCAACTATTGC-3' (sense), 5'-TCAGCTGCACTTGCAGGAGC-3' (antisense) (338 bp fragment, annealing temperature 60°C for 30 cycles); insulin growth factor 1 (IGF-1): 5'-TGAAGATGCACACCATGTCC-3' (sense), 5'-GCAATACATCTCCAGCCTCC-3' (antisense) (263 bp fragment, annealing temperature 60°C for 40 cycles); fibroblast growth factor 2 (FGF-2): 5'-GTGTGTGCTAACCGTTACCT-3' (sense), 5'-GCTCTTAGCAGACATTGGAAG-3' (antisense) (250 bp fragment, annealing temperature 60°C for 35 cycles). Nonretrotranscribed RNA was always used as a negative control. PCR products were separated by 1–1.5% agarose gel electrophoresis and visualized by ethidium bromide staining.
Western blotting
Protein (100 µg) from cell lysates was resolved using 8% SDS-polyacrylamide gel electrophoresis. Proteins were transferred to polyvinylidene difluoride membranes that were incubated overnight at 4°C with a 1:1000 dilution of a 5-LO rabbit antiserum (kindly provided by Dr. J. Evans, Merck & Co., West Point, PA). Membranes were subsequently exposed to a horseradish peroxidase-conjugated secondary antibody for 1 h. Bands were visualized using the ECL Plus detection kit (Amersham Corp., Milan, Italy). As an internal control, the upper portion of the blot was cut and reprobed with an anti-ß-actin antibody.
Statistical analysis
Results are expressed as mean ± SD. The two-sided Student’s t test was used for statistical comparison. A P value < 0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, three different primary normal mesothelial cells (NM, NM1, NM2) gave one amplification product of 173 bp corresponding to the human platelet 12-LO, whereas three cell lines derived from patients affected by malignant pleural mesothelioma (MM, MM1, MM2) displayed two amplification products of 173 and 326 bp, corresponding to human 12- and 5-LO, respectively. Thus, the 5-LO gene appeared to be selectively expressed by MM cells. Accordingly, 12-HETE, but not 5-HETE, was detected in NM cell-conditioned media whereas 5- and 12-HETE were found in MM cell-conditioned media (Table 1
), indicating that the mesothelial LO were catalytically active. The presence of the 5-LO protein in MM cells was confirmed by Western blotting (results not shown).
|
|
Effect of LO on mesothelial cell proliferation and apoptosis
The functional effect of LO on mesothelial cell proliferation and survival was analyzed by exposing cells to either LO inhibitors or LO antisense oligonucleotides. NDGA, a nonselective antioxidant also used as a general LO inhibitor, and AA-861, a specific 5-LO inhibitor, time-dependently inhibited MM cell proliferation (-72% and -92%, respectively, after 72 h of incubation). Baicalein, a specific 12-LO inhibitor, gave a less pronounced inhibition (-19%) (Fig. 2
A). The concentration of the inhibitors corresponded to approximately twice the previously determined ID50. To rule out nonspecific effects of LO inhibitors, mesothelial cells were exposed to 5- and 12-LO antisense oligonucleotides at concentrations that suppressed accumulation of the correspondent mRNA. As shown in Fig. 2B
, the 5-LO antisense potently reduced [3H]-thymidine uptake by MM cells (-83%) after 48 h of exposure. The 12-LO antisense also yielded inhibition (-56%). In contrast, neither 5- nor 12-LO antisense significantly altered [3H]-thymidine incorporation by NM cells. Similar results were obtained when exposure to the antisense was extended to 72 h and when LO inhibitors were used instead of the antisense with NM cells (results not shown).
|
NDGA and AA-861, but not baicalein, induced a time-dependent nucleosome formation in MM cells that was significant after 48 h and maximal after 72 h of treatment (Fig. 3
A), with a DNA fragmentation pattern typical of apoptosis (Fig. 3B
). The 5-LO antisense oligonucleotide greatly increased (+275%) nucleosome formation in MM cells after 48 h, whereas the 12-LO antisense had no appreciable effect. None of the antisense or LO inhibitors induced nucleosome formation in NM cells (Fig. 3C
and results not shown).
|
MM cells also expressed COX-1 and 2 mRNA. However, indomethacin (<100 µM), aspirin (<150 mM), or sulindac (<50 µM) did not significantly increase nucleosome formation after 48 h of exposure (results not shown).
5(S)-HETE rescues MM cells from NDGA-induced apoptosis
To better assess the metabolic pathways involved in apoptosis induction of MM cell by 5-LO inhibition, cells were exposed to NDGA in the absence or presence of exogenously added arachidonic acid or LO metabolites. As displayed in Fig. 4
, the 5-LO product 5(S)-HETE (1 µM) abrogated NDGA-induced nucleosome formation (-90%). An equimolar concentration of the 12-LO product 12(S)-HETE or the 5-LO/LTA4-hydrolase metabolite LTB4 also yielded inhibition (-30%), but did not reach statistical significance. The 15-LO metabolite 15(S)-HETE and native arachidonic acid were ineffective. Similarly, 1 µM of 5(S)-HETE restored the ability of MM cells to proliferate in the presence of NDGA (+60%) and an equimolar concentration of 12(S)-HETE gave an 16% increment; 15(S)-HETE and LTB4 were ineffective (results not shown). Along with results in Figs. 2
and 3
, these findings strongly indicate that a catalytically active 5-LO and, to a much lesser extent, 12-LO regulate MM cell proliferation and survival.
|
LO regulates VEGF release and gene expression in mesothelial cells
Since MM cells release considerable amounts of VEGF and we recently observed that VEGF is a potent stimulator of mesothelioma cell proliferation (18)
, we asked whether VEGF was involved in the 5-LO-dependent regulation of MM survival. As shown in Fig. 5
A, a 5-LO antisense time-dependently inhibited VEGF release by MM cells, with a significant reduction after 24 h and an almost complete suppression after 48 h. A maximum of 40% inhibition was also seen with a 12-LO antisense. Similar results were obtained when NDGA, AA-861, and baicalein were used to block LO activity (results not shown). Consistent with results in Figs. 2
and 3
, NDGA, AA-861, and baicalein did not alter VEGF release by NM cells (Fig. 5B
). After 48 h, either the 5-LO antisense or the 5-LO inhibitor AA-861 drastically reduced (-80% by scanning densitometry) VEGF mRNA levels relative to the isoforms 121 and 165 (Fig. 5
, inserts). 5(S)-HETE, but not 12(S)-HETE, 15(S)-HETE, or native arachidonic acid, almost completely reversed inhibition of VEGF release by NDGA (Table 2
). VEGF release and gene expression in NM and human umbilical vein endothelial cells were unaffected by LO antisense or inhibitors (results not shown).
|
|
Transfection of a 5-LO construct increases VEGF release and [3H]-thymidine uptake by NM cells
To obtain a more direct evidence of the relationship between 5-LO expression and VEGF formation, NM cells were transiently transfected with a human 5-LO cDNA. After transfection, the basal VEGF release was increased by sixfold. Moreover, arachidonic acid potently stimulated VEGF accumulation in the conditioned media of 5-LO transfected cells, but did not change VEGF release by either untransfected or mock-transfected cells. This increment corresponded to 12-fold above the VEGF levels measured in the conditioned media of either untransfected or mock-transfected cells exposed to arachidonic acid (Fig. 6
A). Transfection of the previously characterized 5-LO mutant H367Q, which has nuclear localization but is devoid of catalytic activity (22)
, did not change VEGF release by mesothelial cells. Similar results were obtained with other mutants—R131/132Q, H372Q, H550Q (22)
—having different intracellular localization but all unable to metabolize arachidonic acid (results not shown). 5-LO transfection of MM cells increased 5-LO protein expression and gave a twofold increment in VEGF release (Fig. 6B
), but did not significantly alter the profile of 5-LO metabolites accumulated in the 48 h conditioned medium of MM cells; 5-HETE was a more abundant product (results not shown).
|
Since VEGF is considered a main autocrine growth factor for mesothelial cells, an increase in its production might stimulate mesothelial cell growth. Indeed, NM cells transfected with the 5-LO cDNA incorporated 60% more [3H]-thymidine vs. untransfected, mock-transfected, or H367Q-transfected cells. Moreover, exposure of the 5-LO-transfected cells to arachidonic acid gave a 60% increase in [3H]-thymidine uptake compared with the untreated transfected cells and a 120% increment vs. untransfected, mock-transfected, or H367Q-transfected cells treated with arachidonic acid. An anti-VEGF antibody almost completely abrogated the increase in [3H]-thymidine incorporation observed with the 5-LO transfected cells in either the presence or absence of arachidonic acid, whereas an irrelevant antibody was ineffective (Fig. 6C
). These results indicate that 5-LO confers to NM cells the ability, typical of MM cells, to stimulate their own proliferation by producing VEGF.
5-LO activity regulates VEGF gene expression at the transcriptional level
The molecular mechanisms of LO regulation of VEGF gene expression were further investigated using a VEGF promoter-luciferase construct covering the promoter region -2018 to +50 (23)
. As shown in Fig. 7
the activity of the VEGF promoter in MM cells was stimulated by arachidonic acid and 5(S)-HETE (+168 and 154%, respectively) to an extent similar to PDGF (+195%), which is considered a potent inducer of this construct (23)
. 12(S)-HETE gave a 29% increment, but 15(S)-HETE was ineffective. LTA4 also activated the VEGF promoter with a potency similar to that of 5(S)-HETE (+140%), whereas 5(S)-HpETE was ineffective. NDGA and AA-861 suppressed the basal promoter activity; baicalein reduced it by 44% and indomethacin was ineffective. Thus, 5-LO and (to a considerably lesser extent) 12-LO regulate the transcriptional activity of the VEGF gene in MM cells.
|
Effect of 5-LO activity on other mesothelioma-derived growth factors
In addition to VEGF, MM cells release factors that are potential autocrine/paracrine growth stimulators (24
25
26)
; it has been shown that antisense oligonucleotides specific for TGF-ß1 and ß2 inhibit mesothelioma cell growth (27)
. Therefore, we asked whether 5-LO activity could regulate the expression of other mesothelioma-derived growth factors and what the effect of these factors was on MM cell apoptosis triggered by 5-LO inhibition. Total RNA from MM cells exposed for 48 h to the 5-LO antisense was subjected to semiquantitative RT-PCR using oligonucleotides corresponding to specific sequences of PDGF-B, IGF-1, FGF-2, TGF-ß1. As shown in Fig. 8
A, the 5-LO antisense did not significantly change TGF-ß1 and FGF-2, but slightly reduced PDGF-B and IGF-1 (20–40% by scanning densitometry) mRNA levels. On the other hand, VEGF, but not PDGF-AB, IGF, FGF-2, or TGF-ß1, rescued MM cells from AA-861-induced apoptosis (Fig. 8B
). Thus, it appears that the selective regulation of VEGF formation is a key mechanism of the 5-LO-dependent apoptosis of MM cells.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and Table 1
). Varying levels of 5-LO expression were observed in several MM cell lines with different phenotypes and degrees of invasiveness (epithelioid, sarcomatous, biphasic). Although an apparently higher 5-LO expression was denoted in the epithelioid and biphasic phenotypes, additional observations are required to establish whether there is a correlation between MM histotype and 5-LO expression. Pleural fluids and tumor sections from patients with different types of mesothelioma are being examined to evaluate 5-HETE production and 5-LO expression. The reason 5-LO mRNA and 5-LO activity can be better detected in MM cells than in NM cells is not completely understood. Analysis of the human 5-LO gene promoter has revealed the presence of a region (-212 to -88 from ATG) that contains putative cis-acting sequences bearing binding motifs for transcription factors including Sp1, Egr-1, NF-
B, and AP-2 (28)
. Different cell lines of the myeloid lineage may contain different sets of trans-acting factors, resulting in different patterns of 5-LO promoter activity and endogenous 5-LO mRNA expression (28)
. Whether differences in the content and/or availability of transcription factors between NM and MM cells account for the difference in 5-LO mRNA levels remains to be determined. Our laboratory is examining the relationship between events leading to mesothelial cell transformation and 5-LO expression.
That 5-LO could be involved in the mechanisms of mesothelial cell carcinogenesis is suggested by the evidence presented here that this enzyme has a profound effect on MM proliferation and survival. We observed that 5-LO inhibition resulted in growth arrest and apoptosis even in the presence of 10% serum (Figs. 2
and 3)
. 5-LO inhibition was attained either pharmacologically by using NDGA, a nonselective LO inhibitor, and AA-861, a specific 5-LO inhibitor, or at the gene level with an antisense oligonucleotide. These agents yielded comparable results, ruling out possible nonspecific pharmacological effects, particularly by NDGA (29)
. In contrast, 5-LO inhibition did not alter NM cell proliferating capability and viability. These results extend previous observations with human pancreatic and prostate cancer cells (7
, 8)
and indicate that the involvement of 5-LO in the regulation of cancer cell growth is broader than expected. However, in this model of mesothelioma, the selectivity of action of 5-LO inhibitors toward the malignant phenotype of mesothelial cells represents a novel finding that may have considerable clinical implications.
12-LO gene blockage by an antisense oligonucleotide gave a 50% reduction in MM cell DNA synthesis (Fig. 2)
, but did not significantly induce nucleosome formation or DNA fragmentation (Fig. 3)
. Thus, 12-LO activity also has an effect on MM cell proliferation, although it may not be a survival factor. This suggests that 5-and 12-LO products may activate different signal transduction pathways and have distinct intranuclear targets and functions in MM cells. Since esterification into cellular lipids is considered an important event in monohydroxy acid signaling (30)
, 5- and 12-LO products may follow differential routes of esterification in MM cells. We limited our investigation to a restricted number of LO metabolites. Thus, we cannot exclude the possibility that other 5-LO products such as 5-oxo-HETE, as in the case of prostate cancer cells (8)
, or the cysteinyl-leukotrienes may influence MM cell functions. Expressing both 5 and 12-LO, MM cells possess the enzymatic machinery to form LX from endogenous substrate. By converting the 5-LO product LTA4 into LXA4 and B4, the platelet type 12-LO, in fact, has LX synthase activity, (31)
. Moreover, we have detected the presence of the LXA4 receptor mRNA in MM cells (M. Romano et al., unpublished results). Whether LX may also play a role in MM cell functions is currently under investigation. LO metabolites released by MM cells may also contribute to alter the host immune inflammatory response. LO products are centrally involved in a complex cytokine-regulated network that influences various pathophysiologic events (32)
.
In the case of mesothelioma, COX isoforms appear to be less relevant for cell survival, since the COX inhibitors indomethacin, aspirin, or sulindac did not induce apoptosis of MM cells and indomethacin did not alter VEGF release or VEGF promoter activity (results not shown and Fig. 7
). These results are consistent with previous findings showing that mesothelioma tissue express constitutive COX-2 and that the selective inhibition of this enzyme does not induce apoptosis of mesothelioma cells, but does have antiproliferative activity (33)
. Thus, the involvement of COX isoforms in mesothelioma survival should be further investigated.
The fact that the effect of 5-LO inhibition on MM cell proliferation and survival became significant after 24 h (Figs. 2
and 3)
suggested that a sequence of events involving additional factors might be following the block of 5-LO. We present here direct evidence indicating that VEGF release and mRNA levels are regulated by 5-LO activity in MM cells and that this regulation is a crucial mechanism of 5-LO actions on proliferation and apoptosis (Figs. 5
6
7
8)
. Although the relationships between the COX or the 12-LO pathway of arachidonic acid metabolism and tumor angiogenesis have been examined (34
35
36
37)
and it has been reported that COX products can stimulate VEGF release in selected settings (38
, 39)
, the present results are the first to document a direct effect of 5-LO activity on VEGF release and gene expression. Whether this finding represents a peculiarity of MM cells or whether it can be observed in other normal or malignant VEGF-releasing cells that also express 5-LO remains to be determined. However, it is striking that upon 5-LO transfection, NM cells, which normally release quite low amounts of VEGF and proliferate at a low rate, displayed some similarity in behavior with MM cells (Fig. 6A
). The proliferative advantage conferred by 5-LO to NM cells depended primarily on VEGF up-regulation, as it was abolished by an anti-VEGF antibody (Fig. 6C
). These results are strengthened by data in Fig. 6B
showing that 5-LO overexpression in MM cells, which produce considerable amounts of VEGF, determined a twofold increase in VEGF release. Together with data in Fig. 8
, these results strongly indicate that VEGF is a highly selective target of 5-LO. The interaction between 5-LO and VEGF becomes especially relevant in the case of mesothelioma. VEGF may in fact promote mesothelioma growth either indirectly, by stimulating endothelial cell-dependent neo-angiogenesis, or directly by activating the specific Flt-1 and KDR receptors also expressed in mesothelioma cells (18)
. Thus, inhibition of VEGF release by 5-LO antagonists may profoundly hamper mesothelioma progression by acting on multiple mechanisms of tumor growth. Both 5(S)- and 12(S)-HETE potently stimulated VEGF release by human umbilical vein endothelial cells (A. Catalano et al., unpublished observation). This may further contribute to neo-angiogenesis in mesothelioma, since these metabolites are constitutively released by MM cells (Table 1)
. Stable 5-LO transfectants are being raised to determine whether 5-LO expression can confer a neoplastic phenotype to NM cells. In vivo experiments are under way to obtain direct evidence of the potential antineoplastic/antiangiogenic properties of 5-LO inhibitors in mesothelioma.
The 5-LO product 5(S)-HETE potently stimulated the activity of a VEGF promoter construct that encompasses a 2068 bp region of the human VEGF gene, from -2018 to +50, containing consensus binding sites for transcription factors including Sp1, AP-1, AP-2, and Egr-1 (23)
. Moreover, the specific 5-LO inhibitor AA-861 suppressed activity of the VEGF construct (Fig. 7)
. 5(S)-HETE and arachidonic acid stimulated the VEGF promoter similar to PDGF, which is considered a potent inducer of VEGF transcription (23)
. However, a significant direct effect of native arachidonic acid on VEGF expression is unlikely. Arachidonic acid was unable to counteract the effect of NDGA on VEGF release (Table 2)
. Moreover, it did not stimulate VEGF production by NM cells transfected with a catalytically inactive 5-LO (Fig. 6A
). Thus, the stimulatory effect of arachidonic acid on VEGF formation depends primarily on conversion by 5-LO. However, it appears that initial arachidonic acid oxygenation by 5-LO is not sufficient to stimulate VEGF transcription since 5(S)HpETE, the hydroperoxide generated by 5-LO-governed stereospecific insertion of molecular oxygen into arachidonate, did not activate the VEGF promoter (Fig. 7)
. Instead, LTA4 was as potent an inducer of the VEGF promoter as 5(S)-HETE (Fig. 7)
, indicating that after initial arachidonate oxygenation, additional steps (i.e., reduction or dehydration) are required to induce VEGF transcription. We do not know the mechanisms of 5(S)-HETE or LTA4 up-regulation of VEGF transcription. It would be interesting to identify cis-acting elements and trans-acting factors involved in the transcriptional activation of VEGF by 5-LO products.
In conclusion, we present evidence of a functional and molecular interplay between 5-LO metabolism of arachidonic acid, VEGF formation, and MM cell proliferation and apoptosis. These results may contribute to a better understanding of mesothelial cell pathophysiology and, as a consequence, to the design of novel strategies for the treatment of malignant mesothelioma.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 12, 2001.
Revision received July 13, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87,803-809[Medline]
This article has been cited by other articles:
|
|
 |
|
  M. Peters-Golden and W. R. Henderson Jr. Leukotrienes N. Engl. J. Med., November 1, 2007; 357(18): 1841 - 1854. [Full Text] [PDF]
|
|
|
|
 |
|
 D. Nie, S. Krishnamoorthy, R. Jin, K. Tang, Y. Chen, Y. Qiao, A. Zacharek, Y. Guo, J. Milanini, G. Pages, et al. Mechanisms Regulating Tumor Angiogenesis by 12-Lipoxygenase in Prostate Cancer Cells J. Biol. Chem., July 7, 2006; 281(27): 18601 - 18609. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Zhang, H. Cao, and G. N. Rao Fibroblast Growth Factor-2 Is a Downstream Mediator of Phosphatidylinositol 3-Kinase-Akt Signaling in 14,15-Epoxyeicosatrienoic Acid-induced Angiogenesis J. Biol. Chem., January 13, 2006; 281(2): 905 - 914. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Demirag, E. Unsal, A. Yilmaz, and A. Caglar Prognostic Significance of Vascular Endothelial Growth Factor, Tumor Necrosis, and Mitotic Activity Index in Malignant Pleural Mesothelioma Chest, November 1, 2005; 128(5): 3382 - 3387. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Titos, J. Claria, A. Planaguma, M. Lopez-Parra, A. Gonzalez-Periz, J. Gaya, R. Miquel, V. Arroyo, and J. Rodes Inhibition of 5-lipoxygenase-activating protein abrogates experimental liver injury: role of Kupffer cells J. Leukoc. Biol., October 1, 2005; 78(4): 871 - 878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Zhang, H. Cao, and G. N. Rao 15(S)-Hydroxyeicosatetraenoic Acid Induces Angiogenesis via Activation of PI3K-Akt-mTOR-S6K1 Signaling Cancer Res., August 15, 2005; 65(16): 7283 - 7291. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.N. Ye, W.K.K. Wu, V.Y. Shin, I.C. Bruce, B.C.Y. Wong, and C.H. Cho Dual inhibition of 5-LOX and COX-2 suppresses colon cancer formation promoted by cigarette smoke Carcinogenesis, April 1, 2005; 26(4): 827 - 834. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Shinji, T. Tsukui, A. Tatsuguchi, K. Shinoki, M. Kusunoki, K. Suzuki, T. Hiratsuka, K. Wada, S. Futagami, K. Miyake, et al. Induced microsomal PGE synthase-1 is involved in cyclooxygenase-2-dependent PGE2 production in gastric fibroblasts Am J Physiol Gastrointest Liver Physiol, February 1, 2005; 288(2): G308 - G315. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
), 2 µM of baicalein (•), 1 µM of AA-861 (
), or 5 µM of NDGA (
) in complete RPMI medium for the times indicated. Cells were then collected, stained with trypan blue, and counted. Results represent the mean ± SD from 3 separate experiments with triplicate determinations. B) NM (
) and MM ([
