Platelet-derived growth factor (PDGF) has been shown to act chronically on blood vessels to regulate not only proliferation but also vascular tone. These effects may be at least partly due to the chronic effect of PDGF on vascular endothelium. To evaluate this possibility, we examined the effects of PDGF on the endothelium-dependent relaxation (EDR) and total RNA for endothelial nitric oxide (NO) synthase (eNOS) using an organ culture system. In rabbit mesenteric arteries cultured in a serum-free medium for 1 wk, amplitude of the substance P-induced EDR did not change, whereas dependency of the EDR on NO (∼60.0% vs. 18.9%) and the total amounts of recoverable eNOS mRNA estimated by RT-PCR were increased compared with those in freshly isolated arteries. Culture with PDGF for 1 wk decreased the relaxant effect of substance P and ionomycin (P < 0.01 compared with the arteries without PDGF), NO production estimated by bioassay (P < 0.01), and eNOS mRNA level, whereas the sodium nitroprusside-induced relaxation did not change. These results suggest that PDGF has a chronic effect on vascular endothelium to decrease eNOS mRNA and NO production and to impair NO-dependent EDR.
- nitric oxide
- endothelial nitric oxide synthase
- nitric oxide bioassay
- reverse transcriptase-polymerase chain reaction
platelet-derived growth factor (PDGF) has been shown to play a primary role in vascular response to injury and development of atherosclerosis by stimulating cell proliferation and migration (20,22). Recent evidence indicates that, in addition to mitogenic action, PDGF has both acute and chronic effects to regulate vascular tone. As acute effects, PDGF causes contractions in smooth muscle (2, 5) and also stimulates nitric oxide (NO) production in endothelial cells (5,13). As chronic effects, PDGF inhibits NO release from interleukin-1β-stimulated vascular smooth muscle cells (6, 24). Chronic action of PDGF is believed to be important especially in pathophysiological states such as atherosclerosis, because these conditions are associated with prolonged exposure of vascular wall to PDGF (21). However, the chronic effects of PDGF on endothelial functions remain to be elucidated. In the present experiments, we examined the long-term effects of PDGF on endothelium-dependent relaxation (EDR) and total RNA for endothelial NO synthase (eNOS). For this purpose, we used organ culture of the whole vascular wall, because it is possible to incubate the tissue with a constant concentration of PDGF for a long period of time, and both morphology and functional changes of the tissue can easily be examined.
MATERIALS AND METHODS
Tissue preparation and organ culture procedure. Male Japanese White rabbits (2–3 kg) were euthanized by a sharp blow on the neck and exsanguination. Main branches of superior mesenteric arteries were isolated under sterile conditions. After the fat and adventitia were removed in sterile Hanks’ balanced salt solution containing 1% penicillin-streptomycin, each artery was cut into rings (∼2 mm wide) for measurement of muscle tension and helical strips (∼1.5 mm wide, 5 mm long) for NO bioassay experiments. Strips were then placed in 2 ml of serum-free Dulbecco’s modified Eagle medium (DMEM) supplemented with 1% penicillin-streptomycin (serum-free arteries). Some strips were placed in a similar solution with 100 ng/ml PDGF-BB (PDGF-treated arteries). PDGF-BB at a concentration of 100 ng/ml has been reported to be maximally effective in stimulating DNA synthesis in cultured endothelial cells (1) and in eliciting endothelial NO production in an isolated artery (13). Arterial preparations were maintained at 37°C in an atmosphere of 95% air-5% CO2 for up to 1 wk. Incubation medium was replaced every other day. Freshly isolated arteries (fresh arteries) were also prepared as described above but not in sterile conditions.
Histology. The freshly isolated or cultured tissue samples were fixed in 10% neutral buffered Formalin and embedded in paraffin. Four-micrometer-thick sections were stained with hematoxylin and eosin and then examined under a light microscope.
Organ chamber experiments. The rabbit mesenteric arterial rings were placed in normal physiological salt solution (PSS) that contained (mM): 136.9 NaCl, 5.4 KCl, 1.5 CaCl2, 1.0 MgCl2, 23.8 NaHCO3, and 5.5 glucose. EDTA (1 μM) was also added to remove contaminating heavy metal ions, which catalyze oxidation of organic chemicals in PSS. The high-K+ (65.4 mM) solution was prepared by replacing NaCl with equimolar KCl. These solutions were saturated with 95% O2-5% CO2 mixture at 37°C and pH 7.4. In some experiments, the endothelium was removed by gently rubbing the intimal surface with the flat face of a pair of forceps moistened with PSS. Muscle tension was recorded isometrically with a force displacement transducer (Nihon Kohden, Japan). Each muscle ring was attached to a holder under a resting tension of 10 mN. After equilibration for 30 min in a 2-ml organ bath, each strip was repeatedly exposed to the high-K+solution until responses became stable (60–90 min). Concentration-response curves were obtained by the cumulative application of relaxants after preconstriction by norepinephrine reached a steady-state level. In some experiments, a bolus dose of a relaxant was applied during a norepinephrine-induced sustained preconstriction. Data are shown as percent relaxation of the norepinephrine-induced steady-state contraction. Because 50% effective concentrations of norepinephrine were different in each group (6.72 ± 1.26 μM in the fresh artery,n = 6; 0.63 ± 0.09 μM in the serum-free artery, n = 11; and 0.67 ± 0.09 μM in the PDGF-treated artery,n = 6), we used different concentrations of norepinephrine to elicit preconstriction (10 μM for the fresh artery, 1 μM for the serum-free artery and PDGF-treated artery). These concentrations of norepinephrine caused a similar magnitude of contractions relative to the 100 μM norepinephrine-induced maximal contractions (73.7 ± 5.9% in the fresh artery, n = 6; 79.7 ± 4.8% in the serum-free artery, n = 11; and 73.3 ± 5.3% in the PDGF-treated artery,n = 6).
NO bioassay. We referred to the methods of NO bioassay previously described by Furchgott and Zawadzki (10) with slight modifications. Freshly isolated helical strips of endothelium-denuded rabbit mesenteric artery (NO recipient) and cultured helical strips of endothelium-intact muscle (NO donor) were mounted together in sandwich form with their entire intimal surfaces opposed. They were attached to a holder in an organ bath (10 ml) under a resting tension of 10 mN and the changes of the NO-recipient muscle tension were measured with an isometric force transducer connected only to the NO recipient. Thus the content of NO released from the NO-donor endothelium was bioassayed.
Quantitative RT-PCR. Total RNA was extracted from the endothelium-intact rabbit mesenteric artery by the acid guanidinium thiocyanate-phenol-chloroform method (4) using the TRIzol reagent. The concentration of total RNA was adjusted to 1 μg/μl with RNase-free distilled water. RT-PCR was performed using Takara RNA PCR Kit Version 2.1 (Takara, Japan). Briefly, the first strand of cDNA was synthesized using random 9-mer RT-primer and AMV Reverse Transcriptase XL at 30°C for 10 min, 55°C for 30 min, 99°C for 5 min, and 4°C for 5 min, followed by PCR amplification using synthetic gene-specific primers for eNOS and GAPDH. PCR amplification was performed using rTaq DNA polymerase. The oligonucleotide primers for eNOS that were designed from bovine eNOS (25) were TAC CAG CCG GGG GAC CAC (sense) and CGA GCT GAC AGA GTA GTA (antisense). The oligonucleotide primers for GAPDH were TCC CTC AAG ATT GTC AGC AA (sense) and AGA TCC ACA ACG GAT ACA TT (antisense) as described by Fort et al. (9). After initial denaturation and activation of the polymerase at 94°C for 2 min, 28–53 cycles (5-cycle interval) of amplifications at 94°C for 0.5 min, 60°C for 0.5 min, and 72°C for 1.5 min were done with a thermal cycler (Takara Thermal Cycler PERSONAL TP 240, Takara, Japan). PCR products in each cycle were electrophoresed on 2% agarose gel containing 0.1% ethidium bromide. The possible contamination of DNA was excluded by performing a PCR with total RNA without the reverse transcription step. Detectable fluorescent bands were visualized by an ultraviolet transilluminator.
Chemicals. The chemicals used were Hanks’ balanced salt solution, penicillin-streptomycin, TRIzol reagent, ethidium bromide solution (GIBCO BRL), substance P, ionomycin calcium salt, indomethacin (Sigma Chemical), norepinephrine,N G-monomethyl-l-arginine (l-NMMA), sodium nitroprusside (Wako Pure Chemical, Japan), DMEM (Nissui Pharmaceutical, Japan), EDTA (Dojindo Laboratories, Japan), recombinant human PDGF-BB (Austral Biologicals), and rTaq DNA polymerase (Takara, Japan).
Statistical analysis. The results of the experiments are expressed as means ± SE. Statistical evaluation of the data was performed by Student’st-test for comparisons between two groups. When more than two means were compared, one-way ANOVA (Bonferroni’s test) was used. A value ofP < 0.05 was taken as significant.
Morphological examination. Rabbit mesenteric arteries stained with hematoxylin and eosin were examined under a light microscope. Representative micrographs are shown in Fig.1. Endothelial cells of the fresh arteries were located along the inner surface (Fig.1 A). Endothelial cells of both the serum-free and PDGF-treated arteries also appeared to be intact after a 1-wk incubation period (Fig. 1, B andC). In the medial layer of the fresh arteries, smooth muscle cells were arranged in an orderly manner, and the shape of the nuclei was flat (Fig.1 A). In the layer of the serum-free arteries, a larger number of smooth muscle cells was arranged orderly, and the shape of the nuclei was generally flat (Fig.1 B). In the PDGF-treated arteries, however, smooth muscle cell orientation was lost and nuclear rounding was observed (Fig. 1 C).
Effects of serum-free organ culture on EDR. In the fresh arteries, contraction induced by 10 μM norepinephrine was 25.7 ± 2.7 mN/mg wet wt (n = 20). Cumulative addition of substance P relaxed the 10 μM norepinephrine-induced preconstriction in a concentration-dependent manner (Fig.2 A, open circles, n = 20). In the serum-free arteries, 1 μM norepinephrine induced a contraction (27.6 ± 2.0 mN/mg wet wt, n = 18) with a magnitude similar to that induced by 10 μM norepinephrine in the fresh arteries. The supersensitivity of organ-cultured smooth muscle to stimulants has been reported by Rogers et al. (19). Cumulative addition of substance P relaxed the 1 μM norepinephrine-induced preconstriction in a concentration-dependent manner (Fig.2 B, open circles,n = 19). The substance P-induced relaxations in the fresh and serum-free arteries were not significantly different at any concentration of substance P. In both of these arteries, substance P was ineffective when endothelium was removed (n = 4).
In both of these arteries, pretreatment with indomethacin, a cyclooxygenase inhibitor (10 μM, 30 min before the addition of norepinephrine), had no effect on the substance P-induced relaxation (Fig. 2, A andB, closed circles,n = 20–21). When these arteries were treated for 30 min with the combination of indomethacin andl-NMMA, an NOS inhibitor (300 μM), the substance P-induced relaxation was partially but significantly reduced compared with that in the arteries treated with indomethacin alone (Fig. 2, A andB, open squares,n = 12–22,P < 0.01). The inhibitory effect ofl-NMMA on the substance P-induced relaxation was significantly larger in the serum-free arteries than in the fresh arteries (P< 0.01).
Effects of PDGF on NO-mediated EDR. Norepinephrine (1 μM)-induced contraction in the PDGF-treated arteries (19.7 ± 2.3 mN/mg wet wt,n = 22) was significantly smaller than that in the serum-free arteries (39.2 ± 2.4 mN/mg wet wt,n = 54,P < 0.01; this value is larger than that described above because we performed different sets of experiments). In the endothelium-denuded arteries, norepinephrine (1 μM)-induced contraction in the PDGF-treated arteries (28.7 ± 4.1 mN/mg wet wt, n = 8) was also smaller than that in the serum-free arteries (44.6 ± 3.0 mN/mg wet wt,n = 4,P < 0.05). Figure3 demonstrates that the relaxant effect of the maximally effective concentration of substance P (0.1 μM) on the 1 μM norepinephrine-induced preconstriction was much smaller in the PDGF-treated arteries (6.4 ± 2.0%,n = 30) than in the serum-free arteries (69.9 ± 1.8%, n = 59,P < 0.01). A lower (0.01 μM) or higher (1 μM) concentration of substance P also showed almost no relaxant effect in the PDGF-treated arteries (0 ± 0% at 0.01 μM and 5.6 ± 0.5% at 1 μM, n = 4). The substance P-induced relaxation in the PDGF-treated arteries was abolished by the removal of the endothelium (n = 4).
In the serum-free arteries, ionomycin (0.01–1 μM) relaxed the 1 μM norepinephrine-induced preconstriction in a concentration-dependent manner. The effect of ionomycin was not affected by the pretreatment with 10 μM indomethacin (n = 3). The maximal relaxation caused by 1 μM ionomycin was 82.3 ± 2.0% (Fig.4, n = 19). The concentration-response curve was shifted to the right in the PDGF-treated arteries, and the maximal relaxation caused by 1 μM ionomycin was significantly attenuated (44.5 ± 5.1%,n = 17,P < 0.01).
Effects of PDGF on sodium nitroprusside-induced relaxation in smooth muscle. We also examined the relaxant effect of sodium nitroprusside on the norepinephrine-induced contraction. In both the serum-free and PDGF-treated arteries without endothelium, sodium nitroprusside (0.01–100 μM) relaxed the 1 μM norepinephrine-induced sustained preconstriction in a concentration-dependent manner. The relaxant effects of sodium nitroprusside in the serum-free and PDGF-treated arteries were not significantly different (peak relaxation was 98.5 ± 1.1% vs. 97.2 ± 1.4%, and IC50 value was 0.36 ± 0.11 μM vs. 0.61 ± 0.05 μM in the serum-free vs. PDGF-treated arteries, n = 8).
NO bioassay. To determine whether the reduced EDR was responsible for the decreased release of NO, we used a recipient-donor bioassay system (10). In the bioassay system, norepinephrine (1 μM) induced a sustained contraction in the recipient arteries. Addition of 0.1 μM substance P on the norepinephrine-induced preconstriction caused a transient relaxation in the recipient arteries as reported by Zawadzki et al. (28). With the use of the serum-free arteries as an NO donor, the substance P-induced relaxation of the recipient arteries was 23.9 ± 3.7% (n = 11). With the use of the PDGF-treated arteries as a donor, the substance P-induced relaxation was significantly reduced compared with when the serum-free arteries were used as a donor (3.6 ± 1.6%,n = 8,P < 0.01) (Fig.5).
RT-PCR analysis. RT-PCR analysis was performed on RNA extracted from the endothelium-intact rabbit mesenteric arteries. After 53 cycles of amplification, products of GAPDH (298 base pairs) were detected in all of the conditions, whereas the eNOS band (459 base pairs) was weak in the fresh arteries, strong in the serum-free arteries, and almost invisible in the PDGF-treated arteries (Fig. 6). At 48 cycles of amplification, GAPDH products were also detected in all of the conditions, whereas the eNOS band was detected only in the serum-free arteries (data not shown). Similar results were obtained in the other two sets of independent experiments. It should be noted that previous studies showed a weak but positive staining of eNOS protein in the endothelium of rabbit arteries by immunohistochemistry (14). We also obtained a positive staining of eNOS in the endothelium of rat aorta (unpublished observation) but not in rabbit mesenteric artery. We are not able to explain this discrepancy.
Acute effects of PDGF-BB on vascular endothelium. In the endothelium-intact serum-free arteries, PDGF-BB (100 ng/ml) acutely caused a relaxation of the norepinephrine (0.1 μM)-induced sustained preconstriction (28.8 ± 4.9%, n = 4). This relaxation was completely inhibited by the addition ofl-NMMA (200 μM). In the endothelium-denuded serum-free arteries, however, PDGF did not relax but slightly augmented the preconstriction (n = 4). Augmentation of contraction in the endothelium-denuded arteries may be due to the contractile effects of PDGF on smooth muscle (2, 5).
Effects of serum-free organ culture on EDR. The present results showed that the substance P-induced EDR in the fresh arteries was partly inhibited byl-NMMA. In the serum-free arteries, substance P induced a magnitude of relaxation similar to that in the fresh arteries, although the inhibitory effect ofl-NMMA was much greater (Fig. 2). We also found that the eNOS mRNA level was increased in the serum-free arteries (Fig.6). These findings suggest that the organ culture of the rabbit mesenteric artery in a serum-free condition for 1 wk upregulates an NO-dependent EDR by increasing the eNOS mRNA level.
Effects of PDGF on NO-mediated EDR.The substance P-induced NO-dependent EDR in the serum-free arteries was significantly reduced by the PDGF treatment (Fig. 3). There are at least three possible explanations for the impairment of EDR:1) the receptor that is responsible for release of NO is downregulated,2) NO production by NOS is impaired, and 3) the mechanism of cGMP-dependent relaxation in smooth muscle is impaired.
Receptor agonists produce NO by increasing the cytosolic Ca2+ level in endothelium to activate eNOS (16). Ca2+ ionophore has been shown to directly increase the endothelial Ca2+ concentration (26). In the present results, the Ca2+ionophore-induced EDR in the PDGF-treated arteries was significantly reduced (Fig. 4), suggesting that the mechanisms after an increase in endothelial Ca2+ level would be impaired in the PDGF-treated arteries. However, the possible involvement of the decreases in the number of NK1 receptors, which mediate the effect of substance P (18), could not be excluded.
NO produced by endothelial cells activates guanylate cyclase to increase cGMP content in smooth muscle (17). cGMP activates cGMP-dependent protein kinase, which leads to muscle relaxation by decreasing cytosolic Ca2+concentration and/or Ca2+sensitivity of contractile apparatus (15). In the PDGF-treated arteries, the relaxant effect of an NO releaser, sodium nitroprusside (7), was not significantly different from that in the serum-free arteries. These results suggest that the mechanism of cGMP-dependent relaxation is not impaired in the PDGF-treated arteries.
PDGF caused the morphological changes in the smooth muscle cells, including disorder of smooth muscle orientation and nuclear rounding (Fig. 1). We have previously reported that smooth muscle cells tended to proliferate in the media of the PDGF-treated arteries with the immunohistochemical analysis using an antiproliferating cell nuclear antigen antibody (23). We also found that treatment with PDGF reduced the maximum level of norepinephrine-induced contraction without changing the sensitivity to norepinephrine. Because proliferating smooth muscle cells have been shown to lose their contractility (3), contraction of the PDGF-treated arteries may be attributable to that portion of the smooth muscle cells that are not proliferating. The fact that PDGF did not change the relaxant effect of sodium nitroprusside suggests that the nonproliferating smooth muscle cells in the PDGF-treated arteries have abilities not only to contract but also to relax. Thus the nonproliferating smooth muscle cells in the PDGF-treated arteries seem to have contractility and relaxing abilities similar to those of the serum-free arteries.
To further examine the dysfunction of EDR, we measured the amounts of endothelial NO production by the NO bioassay experiments and eNOS mRNA level by RT-PCR analysis. The results demonstrated decreases in both NO production and eNOS mRNA level in the PDGF-treated arteries (Figs. 5and 6). These results suggest that treatment of rabbit mesenteric artery for 1 wk with PDGF decreases eNOS mRNA level, reduces NO production, and impairs EDR. However, PDGF did not seem to change the morphology of endothelial cells.
The existence of PDGF receptors on endothelial cells has been shown in rat aorta and mesenteric artery (5, 13). We observed that PDGF acutely caused NO-mediated EDR in the organ-cultured rabbit mesenteric artery, suggesting that PDGF acts directly on the endothelium in our model. Because it has been reported that smooth muscle cells in proliferating phenotype secrete endothelial cell-stimulating mitogens such as vascular endothelial cell growth factor and fibroblast growth factor (20), it is possible that PDGF acts on endothelium not only directly but also indirectly through the effects on smooth muscle.
The mechanism of PDGF to downregulate eNOS mRNA remains to be examined. We have recently reported that fetal bovine serum, a potent mitogen in many types of cells, also downregulated the eNOS mRNA (27). Flowers et al. (8) reported that the eNOS mRNA is downregulated in proliferating endothelial cells by a reduction in mRNA stability. These observations raise the possibility that PDGF acts indirectly on eNOS by changing the cell to a growth state.
In vascular pathophysiological states such as atherosclerosis, PDGF generated in the lesion has been shown to play a primal role in pathogenesis of the disease, including stimulation of smooth muscle cell proliferation and migration (20). It has also been reported that NO exerts important protective actions in the vascular wall by inhibiting smooth muscle proliferation (11) and platelet aggregation and adhesion (12). The effects of PDGF to impair eNOS mRNA might be responsible not only for changes in vascular tone but also for the progression of the disease process.
In summary, we found that a serum-free organ culture of the rabbit mesenteric artery for 1 wk upregulates an NO-dependent EDR by increasing eNOS mRNA. We also found that addition of PDGF to the culture medium impairs the NO-dependent EDR by decreasing eNOS mRNA and NO production.
This work was supported by a grant-in-aid for scientific research from the Health Science Foundation, the Ministry of Education, Science, Sports and Culture, The Morinaga Hoshi-kai, and The Suzuken Memorial Foundation, Japan.
Address for reprint requests and other correspondence: H. Yamawaki, Dept. of Veterinary Pharmacology, Graduate School of Agriculture and Life Sci., Univ. of Tokyo, Bunkyo-ku, Yayoi 1-1-1, Tokyo 113-8657, Japan (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society