Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Vascular endothelial growth factor (VEGF) is the only known endothelial cell specific mitogen. Initially described as a vascular permeability factor (VPF) produced by malignant cells (1), it was later found to be identical with an endothelial cell growth factor isolated from pituitary folliculo-stellate cells (2, 3, 4). VEGF exists in four different homodimeric forms generated by alternative splicing of mRNA, the monomers consisting of 121, 165, 189, and 206 amino acids (5). In contrast to other vascular cell mitogens, VEGF contains a hydrophobic leader sequence which allows it to be secreted (4). Hypoxia is a recognized stimulus of VEGF gene induction (6). Other known inducers are platelet-derived growth factor (7), epidermal growth factor (8), transforming growth factor- (9, 10), interleukin-1 (11), and phorbol esters (12). VEGF participates in angiogenesis during embryogenesis, wound healing, ovulation, bone formation, and malignant proliferation (13). The fundamental role of VEGF in tumor growth and metastasis has been underscored by the findings that proliferation of glioblastomas in nude mice could be prevented by anti-VEGF antibodies (14) or by transfection of endothelial cells with a dominant negative mutant of the VEGF receptor flk, the murine analogue of the human VEGF receptor KDR (15).

Novel therapeutic strategies based on VEGF administration are being designed to increase collateral blood flow in systemic vascular diseases (16). Infusion of recombinant VEGF enhanced neovascularization and collateral blood flow in a canine model of myocardial ischemia (17) as well as in a rabbit model of limb ischemia (18). The effect of VEGF in the normal or remodeled lung circulation, however, has not been determined. The lung contains abundant VEGF which is present in alveolar cells, macrophages, and smooth muscle cells. Although a possible pathogenetic role for VEGF in pulmonary hypertension has not yet been elucidated, we detected increased VEGF expression in rat lungs with hypoxic pulmonary hypertension and in plexiform lesions in lungs from patients with primary pulmonary hypertension (19, 20). Relevant to a biological role of VEGF in pulmonary hypertension is our recent observation that treatment of chronically hypoxic rats with polyclonal antibodies against VEGF increased pulmonary vessel remodeling, pulmonary artery pressures and right ventricular hypertrophy (21).

Infusion of prostaglandins represents a therapeutic approach for patients with occlusive arterial disease (22) and with pulmonary hypertension (23). Although the clinical effects of prostaglandins have been attributed to their vasodilatory and antithrombotic actions (26), their long-term effects may involve alternative mechanisms (27). Since it had recently been shown that PGE₁ and PGE₂ increase VEGF production in rat osteoblasts (28) we wondered whether endogenously produced or exogenous administered prostacyclin (PGI₂) and prostaglandin E₂ (PGE₂) increase VEGF production in the lung. We tested this hypothesis in perfused rat lungs, and also in cultured human monocytic leukemia THP-1 cells, substituting for alveolar macrophages, which in vivo, when activated or exposed to hypoxia, produce VEGF (19, 20, 29).

Our data show that PGI₂, PGE₂, and the stable PGI₂-analogue iloprost induced VEGF in the human monocytic THP-1 cell line. PGI₂ and PGE₂ also increased VEGF gene expression in isolated rat lungs. The prostaglandin effects were mediated via cAMP and a cAMP-dependent protein kinase (PKA). In addition, compounds that alter intracellular calcium were also potent inducers of VEGF gene expression in THP-1 cells. The action of the prostaglandins and the maneuvers which cause alterations in intracellular calcium converge at the cAMP-PKA pathway.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

Rp-cAMP was purchased from Biomol Research Laboratories Inc. (Plymouth, PA). PGI₂, PGE₂, 3-isobutyl-1-methylxanthine, cycloheximide, dibutyryl-cAMP (db-cAMP), thapsigargin, diltiazem, A23187, ethylenediaminetetraacetic acid (EDTA), [1,2-bis(2)aminophenoxy] ethane-N,N,N',N', tetraacetic acid (BAPTA), staurosporine, and dexamethasone were from Sigma Chemical Company (St. Louis, MO). H7 was from Research Biochemical Inc. (Natick, MA), KT5823 was from Calbiochem (San Diego, CA), and iloprost was from Berlex Laboratories Inc. (Cedar Knolls, NJ).

Cell Culture and Experimental Design

The human monocytic THP-1 cell line was purchased from American Type Culture Collection. Cells were expanded in 75 mm² flasks in RPMI 1640 media (Gibco BRL, Grand Island, NY) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin-neomycin (Gibco BRL). In most studies, the medium was changed 24-48 h before the experiments to avoid serum-related influences on VEGF gene expression. For experiments designed to measure VEGF protein in conditioned media, the medium was replaced by serum-free RPMI 1640 immediately before the studies began. Human umbilical vein endothelial cells were isolated and maintained as described elsewhere (30).

Iloprost (dissolved in 0.9% saline), PGI₂, or PGE₂ (dissolved in 100% ethanol) were added to the cell suspension as indicated (the maximal ethanol concentration was 0.3%). Controls consisted of either untreated or vehicle-treated cells. Rp-cAMP, 3-isobutyl-1-methylxanthine, thapsigargin, A23187, EDTA, BAPTA, dexamethasone, diltiazem, H7, KT5823, staurosporine, or cycloheximide were added in concentrations and time intervals indicated in the RESULTS section and in the figure legends. None of the compounds used caused significant cytotoxic effects as judged by trypan-blue exclusion (cell viability in all experiments > 95%). The cells were harvested at the indicated time points, centrifuged, and the RNA was extracted as described below. Each experiment was performed at least three times.

Isolated Perfused Rat Lungs

Lungs from adult male Sprague-Dawley rats were isolated and perfused at a constant flow (0.03 ml/g/min) with a cell-free physiological salt solution as previously described (31). These lungs were not subject to increased vascular shear stress under hypoxic conditions since in lungs perfused with a salt solution, hypoxic vasoconstriction is not maintained. The lungs were kept moist in a Plexiglas chamber at 37°C and were ventilated with a room air gas mixture. After a steady-state perfusion period of 30 min, PGI₂ or PGE₂ (10⁶ M) were added to the perfusate every 60 min. Rp-cAMP was added 30 min before addition of the prostaglandin. Control lungs were perfused for the identical time periods as those treated with prostaglandins. After 3 h of perfusion, the lungs were rapidly frozen in liquid nitrogen and kept at 70°C until further processing. Each experiment was performed with three different lungs.

RNA Isolation and Northern Blot Analysis

RNA from rat lungs and THP-1 cells was extracted by the method described by Chomczinski and Sacchi (32). Total RNA (20 µg) was fractionated by electrophoresis on a 1% agarose/6% formaldehyde denaturing gel and transferred onto a positively charged nylon membrane (Sure Blot, Oncor, Gaithersburg, MD). The RNA was cross-linked to the filter by UV light. The membrane was prehybridized in Rapid-Hyb buffer (Amersham Corp., Arlington Heights, IL) for 20 min at 65°C. A plasmid containing the human VEGF cDNA probe (930 bp) was kindly provided by Dr. D. Leung. The probe for KDR was a gift from Dr. Bruce Terman (33). The probes were labeled with [-³²P] deoxy-CTP using a random prime labeling kit (Amersham Corp.) and added to the prehybridization solution at approximately 1 · 10⁶ dpm/ml. After hybridizing for 2 h at 65°C, the filters were washed three times for 15 min at room temperature in 2× SSC/0.1% SDS and two times for 30 min at 65°C in 1× SSC/0.1% SDS, followed by autoradiography and scanning densitometry. A cDNA probe for glyceraldehyde-phosphate dehydrogenase (GAPDH) was used as control for RNA loading. The quantification of VEGF mRNA was determined by scanning densitometry as described before (19). Briefly, films containing the hybridization signals were developed after 24 h and 2-3 day exposure. The 24 h exposure had a lighter signal and it was compared with a 2-3 day exposure to assure that saturation of the radiographic film had not occurred. The film with the best signal to noise ratio was then scanned using Molecular Dynamics, series 400 phosphorimager and Imagequant software (Sunnyvale, CA). The VEGF mRNA signal was then normalized to the GAPDH signal.

Reverse Transcriptase Polymerase Chain Reaction

Total RNA from control, db-cAMP-treated and iloprost-treated THP-1 cells was extracted as described above and 2.5 µg of RNA were reverse transcribed into complementary DNA with 1 unit of reverse transcriptase (Promega, Madison, WI) in 1 mM MgCl₂, Rnasin, dNTP, and random hexamers, according to the GeneAmp^® RNA PCR kit protocol (Perkin-Elmer Cetus, Norwalk, CT). The cDNA was subsequently amplified in the presence of Taq-polymerase for 35 cycles (denaturation at 95°C for 1 min, annealing and extension at 60°C for 1 min) with the human sense primer 5'-TGGGATCCATGAACTTTCTGCTG-3', and the anti-sense primer 5'-CGGAATTCTCACCGCCTCGGCTTGTCACA-3'. The final MgCl₂ and primer concentrations were 2 mM and 1 µM, respectively. The VEGF primers were designed so that the final amplification product would span the entire coding region of the various isoforms of VEGF. The predicted amplification products for VEGF 121, 165, and 189 were 460, 592, and 664 base pairs, respectively.

VEGF Protein Analysis

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Induction of VEGF by PGI₂, PGE₂ and Iloprost in THP-1 Cells

Iloprost is a carbocyclic derivative of PGI₂, with similar potency but a much higher chemical stability than the parent compound (35, 36). While PGI₂ spontaneously degrades in aqueous solutions with a half-life of 2-3 min, iloprost is active in man with a biological half-life of 20-30 min (35). Since the inactivation by -oxidation and subsequent hydroxylation and conjugation occurs primarily in the liver, the half-life of iloprost in cell culture systems can be expected to be longer. The effects of PGI₂, PGE₂, and iloprost on VEGF gene expression were investigated in the human monocytic THP-1 cell line using Northern blot analysis to determine the levels of transcripts and Western blot analysis as well as ELISA for quantification of VEGF protein. At baseline conditions, THP-1 cells expressed very low levels of VEGF mRNA. Iloprost (Figure 1) as well as PGE₂ (data not shown) led to an increase in VEGF mRNA which was concentration-dependent in the tested range of 10⁹ M to 10⁶ M. Scanning densitometry of the Northern blots revealed that a 10⁶ M concentration of iloprost or PGE₂ increased the VEGF mRNA signal between 5- and 8-fold. PGI₂ was less effective and caused reproducible induction of VEGF mRNA only at concentrations of 10⁶ M or higher (data not shown). Pretreatment with the protein synthesis inhibitor cycloheximide (5 µg/ml) before addition of iloprost or PGE₂ caused a further increase in VEGF mRNA abundance while cycloheximide alone had no effect on VEGF gene expression (Figure 1). Iloprost led to a rapid and transient increase of VEGF mRNA levels which reached a maximum between 1 and 2 h, followed by a rapid decline to baseline levels (Figure 2). Increased VEGF gene expression was accompanied by increased synthesis of VEGF protein, as shown by ELISA of conditioned media and by Western-blot analysis of cell lysates (Figure 3).

View larger version (41K): [in this window] [in a new window]   Figure 1.   Dose response of VEGF mRNA induction by iloprost in THP-1 cells. THP-1 cells were treated with the indicated concentrations of iloprost for 90 min. Cycloheximide (CHX, 5 µg/ml) was added 30 min before iloprost. Control cells (CTL) remained untreated. Total RNA was extracted and Northern blot analysis was performed with 20 µg RNA per lane. The filter was hybridized with a 32P-labeled VEGF probe followed by autoradiography. A GAPDH probe was used for control of RNA loading.

View larger version (40K): [in this window] [in a new window]   Figure 2.   Time course of VEGF mRNA induction by iloprost in THP-1 cells. THP-1 cells were incubated with 106 M iloprost for the indicated periods. The samples were further processed as described in Figure 1.

View larger version (8K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Figure 3.   Stimulation of VEGF protein synthesis by iloprost in THP-1 cells. THP-1 cells were untreated (controls, CTL) or treated with 106 M iloprost for 3 h or 6 h, respectively. For the 6  h incubation, iloprost was added at 0 h and replenished at 3 h. The cells were harvested and centrifuged. (A) ELISA of conditioned media: the conditioned media were concentrated 25-fold, resuspended in carbonate buffer and aliquoted onto multiwell plates. Standards consisted of a 2-fold serial dilution of rVEGF165. VEGF protein levels were assessed by ELISA using the monoclonal antibody 4.6.1 (Genentech, San Francisco, CA) followed by incubation with biotinylated anti-mouse IgG and streptavidin-horseradish peroxidase. Ortho-phenylenediamine (Sigma) was employed as substrate. The concentration of VEGF in experimental samples was calculated from a linear regression analysis of optical densities obtained with standard concentrations of VEGF. The optical density curve for VEGF was linear in a concentration range between 390 pg/ml and 6.25 ng/ml (correlation coefficient = 0.994). (B) Western-blot analysis of cell lysates: The cell pellets were lysed in a buffer containing triton X-100, PMSF, and leupeptin. 100 µg of total cellular protein were separated on a 12% polyacrylamide gel and transferred onto a nylon membrane. The filters were incubated with a polyclonal anti-VEGF antibody followed by biotinylated anti-rabbit IgG and streptavidin-horseradish peroxidase and developed with diaminobenzidine.

Most cell lines investigated so far express predominantly the 165 amino acid or the 189 amino acid isoform of VEGF (13). To determine which isoforms of VEGF were induced by iloprost, reverse transcriptase-PCR was performed from total RNA isolated from iloprost-stimulated THP-1 cells and compared with the isoforms induced by cAMP-treated and untreated control THP-1 cells. Gel electrophoresis of the PCR product demonstrated expression of two distinct bands, corresponding to the transcripts of VEGF₁₂₁ and VEGF₁₆₅, of which the transcript for VEGF₁₂₁ showed the highest abundance. No distinct bands corresponding to VEGF₁₈₉ or VEGF₂₀₆ were present (Figure 4). Similar isoforms of VEGF were observed in the control, iloprost and cAMP-treated cells.

View larger version (27K): [in this window] [in a new window]   Figure 4.   db-cAMP and PGI2 induces the 121 amino acid and 165 amino acid isoforms of VEGF in THP-1 cells. Reverse transcriptase-polymerase chain reaction of total RNA from control, db-cAMP, and PGI2-stimulated THP-1 cells revealed two separate bands which corresponded to the predicted transcripts for VEGF121 (expected size 460 base pairs) and VEGF165 (expected size 592 base pairs). The experimental conditions and the primers used are described under MATERIALS AND METHODS. The right lane shows a 100 base pair molecular weight marker.

The biological action of VEGF is not only regulated at the level of VEGF production but also at the level of VEGF receptor expression (15). Therefore, we wondered whether iloprost would affect the expression of transcripts for the VEGF receptor KDR. In human umbilical vein endothelial cells, neither iloprost nor Rp-cAMP did affect the KDR mRNA abundance (data not shown).

Role of cAMP in the Iloprost-induced VEGF Gene Expression in THP-1 Cells

Induction of cellular cAMP and activation of cAMP-dependent protein kinase (PKA) are the main mechanisms by which PGI₂ and PGE₂ act on their target cells (36). To test the hypothesis that the iloprost-effects on VEGF gene expression were also regulated by the cAMP-PKA pathway, we exposed THP-1 cells to various compounds which modulate this pathway (Figure 5). Dibutyryl-cAMP (db-cAMP at 5 · 10⁴ M), which penetrates cell membranes and mimics the action of cAMP led to an increase in VEGF mRNA similar to that observed with iloprost at 10⁶ M. 3-isobutyl-1-methylxanthine (IBMX) (at 10⁴ M), a phosphodiesterase inhibitor which inhibits cAMP inactivation, had no effect on VEGF gene expression when administered alone. Rp-cAMP, a Rp-diastereoisomer of adenosine 3',5'-cyclic phosphorothioate which blocks the activation of PKA by cAMP (37), abolished the induction of VEGF mRNA by iloprost (Figure 5). In addition, the effect of PGI₂ on VEGF mRNA was completely inhibited by H7 (at 3 · 10⁵ M), a nonspecific protein kinase A blocker (38) (Figure 5B), while the more specific inhibitor of protein kinase C, staurosporine (at 10⁸ M) (Figure 5B), and the specific inhibitor of cGMP-dependent protein kinase, KT5823 (at 5 · 10⁷ M [39]), had no effect on iloprost-induced VEGF gene expression (data not shown). The protein kinase C pathway is intact in THP-1 cells since the phorbol ester PMA increases VEGF gene expression (R. Tuder, unpublished observation) and PMA was shown to induce apoE gene expression by these cells (40). Thus, iloprost-induced VEGF gene expression in THP-1 cells predominantly occurred via cAMP and PKA. Since glucocorticoids can inhibit induction of VEGF and other growth factors under several experimental conditions (28, 41), we investigated the effects of glucocorticoids on VEGF gene induction by iloprost. Northern blot analysis showed that dexamethasone abolished the iloprost-induced increase of VEGF gene expression (Figure 5). Dexamethasone also inhibited the thapsigargin and A23187-induced VEGF gene activation (see below).

View larger version (44K): [in this window] [in a new window]   View larger version (43K): [in this window] [in a new window]   Figure 5.   (A) Inhibition of iloprost-induced VEGF gene expression by Rp-cAMP and dexamethasone. THP-1 cells were processed as indicated in Figure 1. Total RNA was isolated from control cells (lane 1), or from cells treated with 106 M iloprost for 90 min (lane 2), with 5 · 104 M Rp-cAMP for 30 min before 106 M iloprost (lane 3), with 5 · 104 M db-cAMP for 90 min (lane 4), with 104 M 3-isobutyl-1-methylxanthine (IBMX, lane 5), with 106 M dexamethasone alone (lane 6), or with 106 M dexamethasone for 30 min before 106 M iloprost (lane 7). (B) Effect of H7 and staurosporin on PGI2 induction of VEGF gene expression. Total mRNA was isolated from control cells, or from cells treated with 106 M iloprost for 90 min and treated with H7 (3 · 105 M), lane 2, or staurosporin (108 M), lane 3.

Role of Ca²⁺

We wondered whether the PGI₂-induced VEGF gene expression would require an increase in intracellular Ca²⁺ as demonstrated for the cAMP-dependent induction of other immediate-early genes (42). Pretreatment of THP-1 cells with diltiazem 10⁵ M, an inhibitor of calcium entry via voltage-gated calcium channels, had no effect on VEGF induction by iloprost (Figure 6A). The same was true for chelation of extracellular Ca²⁺ by 5 · 10³ M EDTA (Figure 6A) or buffering of intracellular Ca²⁺ with 10 mM [1,2-bis(2)aminophenoxy] ethane-N,N,N',N', tetracrotic acid (BAPTA) (Figure 6B). Thapsigargin (at 10⁸ M), which initially releases Ca²⁺ from intracellular stores and subsequently inhibits Ca²⁺ release and mobilization (43), increased VEGF gene expression (5/6-fold by scanning densitometry) with a maximum effect after 2 h (Figure 6C) and the VEGF mRNA levels returned to baseline at 16 h after thapsigargin. At this time, intracellular Ca²⁺ stores are depleted (43); as a consequence any stimulus that requires a release of intracellular Ca²⁺ becomes ineffective. However, when iloprost was administered 16 h after thapsigargin, VEGF gene induction was not reduced, suggesting that an increase in intracellular Ca²⁺ was not essential for PGI₂-induced VEGF gene activation. On the other hand, the early (2 h) effects of thapsigargin suggested that an increase in intracellular Ca²⁺ can induce the VEGF gene (Figure 6C). Accordingly, the calcium-ionophore A23187 also induced VEGF gene expression (Figure 6A). As the prostaglandin-effects, the effects of (short-term) thapsigargin and A23187 seemed to be mediated via the cAMP-PKA pathway since VEGF gene activation by both compounds was drastically reduced by pretreatment with Rp-cAMP (Figures 6A, 6C).

View larger version (50K): [in this window] [in a new window]   View larger version (41K): [in this window] [in a new window]   View larger version (57K): [in this window] [in a new window]   Figure 6.   Role of Ca2+ for VEGF gene induction by iloprost. THP-1 cells were processed as indicated in Figure 1. (A) Total RNA was isolated from control cells (lane 1), or from cells treated with 106 M iloprost for 90 min (lane 2), with 105 M diltiazem before 106 M iloprost (lane 3), with 5 · 103 M EDTA before 106 M iloprost (lane 4), with 104 M A23187 for 4 h (lane 5), and with 5 · 104 M Rp-cAMP for 30 min before 104 M A23187 (lane 6). (B) Northern-blot for VEGF mRNA of THP-1 cells under control or iloprost (106 M) plus Bapta (10 mm). (C) Northern blots are shown after 2 h exposure to 108 M thapsigargin (lane 1), after pretreatment with 5 · 104 M Rp-cAMP for 30 min before 108 M thapsigargin (lane 2), after pretreatment with 106 M dexamethasone before 108 M thapsigargin (lane 3), after 108 M thapsigargin for 16 h (lane 4), and after 108 M thapsigargin for 16 h plus 106 M iloprost for 90 min (lane 5).

Isolated Perfused Rat Lungs

The isolated perfused lung is a well-established system to test the effects of bioactive compounds in the intact organ (31). We used this system to address whether perfusion of rat lungs with PGI₂ or PGE₂ would affect lung VEGF gene expression as seen in vitro with the human monocytic cell line. In accordance with our previous studies (19), we found VEGF mRNA to be abundantly expressed in rat lungs (Figure 7). Doubling of the perfusate flow, which can stimulate the release of PGI₂ and other vasoactive substances, such as nitric oxide, from the endothelium (44), increased VEGF mRNA levels. Perfusion of isolated lungs for 3 h with PGI₂ or PGE₂ (10⁶ M) led to a reproducible increase in the VEGF mRNA expression (1.5-2-fold by scanning densitometry). Addition of IBMX (10⁴  M) did not significantly augment the induction of VEGF by prostaglandins. Concomitant perfusion of the lungs with Rp-cAMP (at 5 · 10⁴ M) inhibited VEGF gene induction by PGI₂. RP-cAMP alone did not change the levels of VEGF mRNA when compared with control lungs (Figure 7). No experiments were performed aiming at characterizing the VEGF isoforms present in the rat lung. The gene expression of the VEGF receptor KDR was not affected by the prostaglandins (Figure 8). Except for the double flow condition, the control and the experimental isolated lungs were perfused at constant flow and pressure, with no apparent increase in shear stress or mechanical stretch.

View larger version (60K): [in this window] [in a new window]   View larger version (18K): [in this window] [in a new window]   Figure 7.   Induction of VEGF by PGE2 and PGI2 in isolated perfused rat lungs. (A) Isolated rat lungs were ventilated with room air and perfused under (control) low-flow (0.03 ml/g/minlane 1) or high-flow (0.06 ml/g/minlane 2) conditions, with 106 M PGI2 for 3 h (lane 3), with 106 M PGE2 for 3 h (lane 4), with 106 PGI2 M plus 5 · 104 M Rp-cAMP for 3 h (lane 5), Rp-cAMP alone for 3 h (lane 6). After perfusion, the lungs were removed and immediately frozen at 70°C. Total RNA was extracted, electrophoretically separated, transferred onto a nylon membrane and hybridized for VEGF and GAPDH (4). (B) Scanning densitometry of VEGF hybridization bands seen in Figure 7A alter normalization for loading with GAPDH.

View larger version (49K): [in this window] [in a new window]   Figure 8.   Effect of cAMP, PGI2 and PGE2 on Kdr (Flk) VEGF receptor mRNA expression in isolated perfused rat lung. Isolated perfused rat lung were ventilated under normoxic (Nx) or hypoxic (Hx) conditions, and perfused with 106 M PGI2, or 106 M PGE2. Northern-blot analysis was performed as described in Figure 7.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Although VEGF is abundantly expressed in normal tissues, such as brain and lung, little is known about the role of VEGF in normal organ functions. In the lung, endogenous VEGF appears to play an important role in modulating the extent of blood vessel remodeling in rats exposed to chronic hypoxia since neutralizing antibodies to VEGF increased pulmonary artery pressures while recombinant VEGF attenuated pulmonary hypertension induced by monocrotaline (21). Because a great number of physiological responses are mediated by intracellular cAMP, we wondered whether prostaglandins (via cAMP) affect cell and lung tissue VEGF gene expression. In the present study, we demonstrate that PGI₂, PGE₂, and the stable PGI₂-analogue iloprost increase (in a cAMP-dependent manner) the VEGF gene expression in the human monocytic THP-1 cell line and in isolated perfused normal rat lungs.

The more potent induction of VEGF gene expression in the monocytic leukemia THP-1 cells by iloprost than by PGI₂ probably resulted from the higher chemical stability rather than a higher biological activity of the former substance, since both substances display similar IP receptor (prostacyclin receptor) agonist potency (36). It is not known whether prostaglandins upregulate VEGF gene expression in lung macrophages, which do express VEGF in the setting of human primary pulmonary hypertension (20).

Harada and associates demonstrated induction of VEGF by PGE₁ and PGE₂ in rat osteoblasts (28), and that Rp-cAMP, a protein kinase A inhibitor, partially blocked the effect of PGE₂ on VEGF mRNA. From our data, it appears that the iloprost effects on VEGF gene expression in THP-1 cells were predominantly or completely mediated via the cAMP-PKA pathway. This conclusion is supported by the observation that db-cAMP mimicked the prostaglandin effects and especially by the finding that the prostaglandin-induced VEGF gene expression was abolished by Rp-cAMP, a specific inhibitor of PKA activation (37), while inhibition of protein kinase C or cGMP-dependent protein kinase had no effect.

Interestingly, the iloprost-induced VEGF gene expression in THP-1 cells did not rely strictly on an increase in intracellular Ca²⁺ since neither the calcium channel blocker diltiazem nor reduction of extracellular (by EDTA) or depletion of intracellular Ca²⁺ (by BAPTA, or prolonged exposure to thapsigargin) inhibited VEGF gene activation by iloprost. This finding differs from observations concerning the regulation of other immediate early genes (such as c-fos, egr-1 and fra-1) whose gene expression was inhibited by depletion of cellular calcium induced by thapsigargin (45, 46). On the other hand, our data show that an increase in intracellular Ca²⁺ by short-term (2 h) exposure to thapsigargin as well as the calcium ionophore A23187 strongly induced VEGF gene expression. We also show that the compounds that likely increase intracellular Ca²⁺, such as thapsigargin, or A23187, increase VEGF gene expression via cAMP and PKA since Rp-cAMP did abolish the thapsigargin and A23187-induced VEGF gene activation. The exact mechanism by which an increase in intracellular Ca²⁺ activates PKA, however, is incompletely understood, but several recent publications demonstrate that Ca²⁺ is a potent inducer of at least three of the eight known adenylate cyclases (47). Based on our findings, we postulate a common signaling pathway for induction of the VEGF gene by prostaglandins or Ca²⁺-releasing agents as depicted in Figure 9.

View larger version (19K): [in this window] [in a new window]   Figure 9.   Proposed model for the induction of VEGF gene by prostaglandins and Ca2+-releasing agents. In this model, we propose that prostaglandins and Ca2+-releasing agents both activate one or several adenylyl cyclases, resulting in the formation of cAMP, followed by activation of cAMP-dependent protein kinase (PKA) and AP-2. The arrows represent activation, the bars represent inhibition. The dashed lines indicate that the mechanisms of glucocorticoid-action are unknown.

Interestingly, iloprost, PGI₂ and cAMP induced predominantly the 121 amino acid isoform of VEGF in THP-1 cells. In contrast to the longer isoforms VEGF₁₈₉ or VEGF₂₀₆ (and to a lesser extent VEGF₁₆₅), which are basic proteins, VEGF₁₂₁ is acidic and does not bind to extracellular matrix containing heparin-proteoglycans (48). Thus, VEGF₁₂₁ is freely diffusible and can act directly on endothelial cells.

The isolated perfused lung expresses abundant VEGF mRNA (19, 47, 48) and perfusion with PGI₂ or PGE₂ for 3 h caused an increase in VEGF gene expression. Since in these experiments, the flow conditions were kept constant, the increase in VEGF mRNA level in isolated lungs treated with PGI₂ or PGE₂ cannot be attributed to changes in shear stress or mechanical stretch (44). However, after doubling the perfusion flow, we noted an increase of VEGF mRNA abundance.

In conclusion, our studies show that prostaglandins and calcium can regulate VEGF gene expression in a cAMP-dependent manner. Hypothetically, administration of prostacyclin to patients with vascular diseases may enhance VEGF gene expression. VEGF may then increase collateral circulation (angiogenesis) or promote vasodilation or suppress the action of vascular growth factors, such as PDGF or endothelin via induction of nitric oxide production (50).