Vascular endothelial growth factor (VEGF) and basic (b) fibroblast growth factor (FGF-2/bFGF) are involved in vascular development and angiogenesis. Pulmonary artery smooth muscle cells express VEGF and FGF-2 and are subjected to mechanical forces during pulsatile blood flow. The effect of stretch on growth factor expression in these cells is not well characterized. We investigated the effect of cyclic stretch on the expression of VEGF and FGF-2 in ovine pulmonary artery smooth muscle cells. Primary confluent cells from 6-wk-old lambs were cultured on flexible silicon membranes and subjected to cyclic biaxial stretch (1 Hz; 5–25% stretch; 4–48 h). Nonstretched cells served as controls. Expression of VEGF and FGF-2 was determined by Northern blot analysis. Cyclic stretch induced expression of both VEGF and FGF-2 mRNA in a time- and amplitude-dependent manner. Maximum expression was found at 24 h and 15% stretch (VEGF: 1.8-fold; FGF-2: 1.9-fold). These results demonstrate that mechanical stretch regulates VEGF and FGF-2 gene expression, which could play a role in pulmonary vascular development or in postnatal pulmonary artery function or disease.
- vascular endothelial growth factor
- basic fibroblast growth factor
- gene expression
- mechanical stretch
- pulmonary artery smooth muscle cells
vascular smooth muscle cells in the arterial wall are subject to hemodynamic forces, including mechanical stretch resulting from pulsatile blood flow. In the aorta, smooth muscle cells are stretched 9–12% during systole (13). Although blood pressures are normally lower in the pulmonary arteries than the aorta, stretching likely occurs in the pulmonary arteries as well during systole. Several diseases are characterized by elevated pulmonary arterial pressures, including primary pulmonary hypertension and congenital heart diseases with increased pulmonary blood flow. In these conditions, the pulmonary arterial vasculature undergoes remodeling (20, 44, 54). In the proximal pulmonary artery, the media thickens because of hypertrophy and hyperplasia of smooth muscle cells and because of an increase in extracellular matrix protein synthesis. In distal arterioles, muscularization of normally nonmuscular vessels occurs (44, 51). A lamb model that mimics congenital heart disease with increased pulmonary blood flow shows similar histopathological changes (45). Animal models of pulmonary hypertension (using chronic hypoxia or monocrotaline) also demonstrate arteriolar muscularization (10, 43). The pathophysiological basis for arterial and arteriolar remodeling in the setting of elevated pulmonary artery pressure is not well understood. One hypothesis is that increased mechanical stretch resulting from elevated hemodynamic pressures increases the expression of peptide growth factors that are mitogenic or chemotactic for vascular smooth muscle cells.
Two growth factors that are potential mediators of arterial remodeling in the setting of persistent pulmonary hypertension are vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF)-2. VEGF (also known as VEGF-A or vascular permeability factor) is a secreted, heparin-binding 45-kDa glycoprotein that stimulates angiogenesis in numerous physiological and pathological settings and is chemotactic for vascular smooth muscle cells (16, 22). VEGF is expressed in pulmonary artery smooth muscle cells, and expression is increased in congenital heart diseases with increased pulmonary blood flow (20). Higher levels of VEGF correlated with increased histological features of arteriopathy, suggesting a role for VEGF as a mediator of remodeling (20). FGF-2 [also known as basic FGF (bFGF)] is a vascular smooth muscle mitogen that is also expressed in pulmonary artery smooth muscle cells. Like VEGF, FGF-2 is upregulated in animal models of pulmonary hypertension (1).
The effect of mechanical stretch on VEGF expression has been investigated in cardiac myocytes, renal mesangial cells, aortic smooth muscle cells, and in mixed pulmonary cell cultures (15, 23, 37,48). In all of these cells types, mechanical stretch upregulated VEGF mRNA or protein expression. Mechanical stretch stimulated FGF-2 release from vascular cells (6). These results suggest that upregulation or release of VEGF or FGF by mechanical stretch may play a role in pulmonary remodeling that occurs in the setting of chronic pulmonary hypertension. In this study, we demonstrate that mechanical stretch increases expression of VEGF and FGF-2 genes in cultured pulmonary artery smooth muscle cells.
Ovine pulmonary artery vascular smooth muscle cells were harvested by the explant technique. After 4-wk-old lambs were killed, the main and branch arteries were removed and dissected free. The exterior of the vessel was rinsed with 70% ethanol. The artery was then opened longitudinally, the interior was rinsed with PBS to remove any blood, and the endothelium was vigorously removed using a cell scraper. The inner media was separated, minced into small pieces (1 × 1 mm), placed on tissue culture plastic, covered with a small amount of medium (DME-H16 medium with 10% FCS, 1 g/l glucose, penicillin, and streptomycin), and incubated at 37°C in 5% CO2. The next day, more of the medium was added. When the vascular smooth muscle cells had grown out on the plates, the pieces of tissue were removed. Vascular smooth muscle cell identity was confirmed by their elongated fusiform shape, formation of “hills and valleys,” and positive staining for smooth muscle actin (Enzo Biochem, Farmingdale, NY). Cells were frozen and used at low passage. For stretch experiments, cells were seeded on fibronectin-coated Silastic membranes in six-well plates (Flexcell, McKeesport, PA) at a density of 5 × 105cells/well and were grown to confluence.
Cells were subjected to biaxial cyclical stretch by using the Flexcell 3000 Strain Unit (Flexcell). Membranes were placed on a loading station and stretched by applying an oscillating vacuum to the underside of the membranes. A computer controlled the duration, amplitude, and frequency of the applied stretch. To determine the duration of applied stimulus necessary to produce a change in VEGF and bFGF expression, cells were stretched at 1 Hz at an amplitude of 15% for varying periods of time. To determine the minimal amplitude of stretch necessary to produce a change in gene expression, cells were stretched at a frequency of 1 Hz for 24 h at amplitudes of 5, 15, and 25%.
Cell injury was determined by measuring lactate dehydrogenase (LDH) activity in media from control and stretched cells (Bio-analytics, Palm City, FL). Direct measurement of LDH activity was determined by adding 50 μl of media to a solution containing 55 mM lithium lactate and 7.5 mM NAD. Absorbance at 340 nm was measured after 30 s and then again 60 s later. Mean absorbance difference per minute was determined, and LDH activity in units per liter was calculated using the following equation where ΔAbs/min is the absorbance difference per minute, TV is total assay volume, 1,000 is used for the conversion of units per milliliter to units per liter, MMA is millimolar absorptivity of NADH, SV is sample volume, and LP is the light path (cm). Interexperimental variability for control cells was ∼5%.
Preparation of RNA, Northern blotting, and hybridization.
Three six-well plates of confluent cells were subjected to stretch for the indicated times and amplitudes or were left unstretched as controls. Cells were harvested in RNA-STAT (Tel-Test, Friendswood, TX), and lysates were pooled for each plate, yielding three independent samples for each stretch condition and three for the contemporaneous unstretched control. Total cellular RNA from each sample was extracted with phenol/chloroform, precipitated with isopropanol, and quantitated spectrophotometrically. RNA integrity was assessed by electrophoresis and ethidium bromide staining for rRNA. Total RNA (10 μg/sample) was separated electrophoretically on 1% agarose gels. Total RNA was transferred to nylon membranes by downward capillary action and cross-linked with ultraviolet light (UV Stratalinker 2400; Stratagene). Filters were probed with cDNAs for ovine bFGF, VEGF, and 18S rRNA labeled with [α-32P]dCTP (NEN Research Products, Boston, MA) by random-primer second-strand synthesis (Random Primer Labeling Kit; GIBCO-BRL, Gaithersburg, MD). Filters were prehybridized for 10 min in QuikHyb Hybridization Solution (Stratagene, La Jolla, CA) at 68°C. Filters were hybridized in 10 ml of QuikHyb solution containing 1.25 × 106disintegrations · min−1(dpm) · ml−1 of [α-32P]dCTP for 18 h. Hybridized filters were washed under high-stringency conditions and subjected to autoradiography (Hyperfilm; Amersham) before quantification of radiolabeled bands by volume integration of pixels measured by PhosphorImage analysis (Imagequant Software; Molecular Dynamics, Sunnyvale, CA). The use of 18S rRNA as a control ensured equal loading. Stretched samples were always run next to and quantitatively compared with their contemporaneous nonstretched controls. Samples (stretched plus nonstretched controls) from different stretch conditions were usually run as separate experiments. Because the specific radioactivity of riboprobes (dpm/mol) is never constant across all experiments, it is not accurate to quantitatively compare raw data from different experiments.
Measurement of protein.
Protein content was measured by the bicinchoninic acid method (Pierce, Rockford, IL).
Quantitation of VEGF protein by dot blotting.
We used a dot-blot assay to measure VEGF. Serial dilutions (5, 2.5, 1.25, and 0.625 μg) of protein from cell lysates were diluted 1:100 with 50 mM NaHCO3, pH 9.0, and dot blotted on nitrocellulose. Endogenous peroxidase activity was quenched by treatment with 15% hydrogen peroxide for 5 min. Blots were blocked for 1 h in a solution containing 1% nonfat milk, 0.4% gelatin, 0.1% BSA, 0.9% NaCl, and 10 mM Tris-buffered saline, pH 7.2 (“blocking buffer”). The blots were incubated in blocking buffer containing anti-VEGF antibody (1 μg/ml; Santa Cruz Biotechnology) for 20 min. The blots were washed 20 times with 20 mM Tris-buffered saline, pH 7.4, containing 0.05% Tween (TBS-T), incubated for an additional 20 min in a solution containing peroxidase-labeled goat anti-mouse IgG (Amersham, Buckinghamshire, UK) in TBS-T (1:5,000), and washed ten times in TBS-T. Bound secondary antibody was detected by exposure to luminol (ECL Light Detection System; Amersham) for 1 min followed by autoradiography. Autoradiographs were scanned, and the density of each dot was quantitated using Image-Pro Plus software (Media Cybernetics, Silver Spring, MD).
Comparisons between controls and stretched cells were made using Student's t-test. P < 0.05 was considered significant.
Stretch is not cytotoxic.
Biaxial cyclic stretch at 1 Hz and at amplitudes of 5, 15, or 25% produced no changes in cell morphology or detachment of cells. Cellular injury was evaluated by measuring LDH activity in cell culture media (Fig. 1). LDH activity in media from cells subjected to cyclic stretch for 24 h was not significantly different from nonstretched control cells [5% amplitude: 86 ± 30% (n = 6, P = 0.476); 15% amplitude: 90 ± 24% (n = 6, P = 0.367); 25% amplitude: 117 ± 18% (n = 6,P = 0.06); Fig. 1 A]. A time-course experiment also found no significant cellular injury (15% amplitude for up to 48 h; Fig. 1 B).
Cyclic stretch increases VEGF and FGF-2 mRNA expression in a time-dependent manner.
Compared with nonstretched control cells, cyclic stretch increased VEGF mRNA and FGF-2 mRNA expression [VEGF: 179 ± 50% (n = 3, P < 0.05; Fig.2 A); FGF-2: 189 ± 50% (n = 3, P < 0.05; Fig. 2 B)]. For both growth factor genes, mRNA expression peaked at 24 h and declined toward control levels at 48 h. We have found similar quantitative changes in growth factor mRNA expression in four independent experiments. (Raw data from radiographs from different time points should not be quantitatively compared with each other because they were analyzed in separate experiments using different riboprobe preparations.)
Cyclic stretch increases VEGF and FGF-2 mRNA expression in an amplitude-dependent manner.
Cyclic stretch at 5 and 15% amplitude increased VEGF and FGF-2 mRNA expression (Fig. 3, A andB). Expression was highest at 15% amplitude for both growth factor genes [VEGF: 179 ± 50% (n = 3, P < 0.05); FGF-2: 189 ± 50% (n = 3, P < 0.05)]. At 25% amplitude, VEGF mRNA expression was unchanged vs. nonstretched control cells, and FGF-2 mRNA expression was decreased vs. control cells [50 ± 20% (n = 3, P < 0.05)].
Cyclic stretch increases VEGF protein.
Cyclical stretch at an amplitude of 15% for 24 h resulted in a 35 ± 2% (n = 3, P < .005) increase in VEGF protein by quantitative dot-blot analysis (Fig.4). A small increase in VEGF protein was seen by Western blot analysis (data not shown).
These studies are the first to demonstrate that mechanical stretch regulates expression of VEGF and FGF-2 mRNA in pulmonary artery smooth muscle cells. Previous studies have examined the effect of mechanical stretch on VEGF expression in other cell types. In renal mesangial cells, stretch increased VEGF mRNA expression 2.4-fold, peaking at 6 h (23). In cardiac myocytes, a 1.8-fold increase in VEGF mRNA that peaked at 2 h was found (48). In a mixed pulmonary cell culture model, stretch increased VEGF mRNA expression approximately twofold, peaking at 2–4 h (37). Our results, using primary ovine pulmonary smooth muscle cells, demonstrate that stretch increased VEGF mRNA expression to a similar magnitude (1.8-fold), but with a slower onset and later peak (24 h) than was seen in other cell types.
Transient stretch of vascular smooth muscle cells (1.5 min at 1 Hz) stimulated release of intracellular FGF-2 (6), but the effect of mechanical stretch on FGF-2 gene expression in vascular smooth muscle cells has not been investigated previously. In skeletal muscle cells and osteoblast-like cells, mechanical stretch increased FGF-2 expression (7, 35), but other models that generate mechanical stretch did not find an effect on FGF-2 expression in the heart, in skeletal muscle, or in mesangial cells (14, 30,55). Our results are the first to demonstrate that mechanical stretch increases the expression of FGF-2 mRNA in vascular smooth muscle cells.
Mechanical stimuli alter the expression of several genes in different cell types. Shear stress stimulates synthesis of mRNAs for platelet-derived growth factor (PDGF)-A and PDGF-B (28), tissue plasminogen activator (12), and intercellular adhesion molecule-1 (38) in human vascular endothelial cells. Mechanical stretching also increases the amount of mRNA for atrial natriuretic factor in neonatal rat cardiac atriocytes (19). The recent identification of stress-responsive elements (SSREs) in the promoter regions of some of these genes has provided a direct link between physical force and gene expression (46, 47). In endothelial cells, one pathway by which shear stress regulates gene expression is the binding of transcription factors to a specific 6- to 12-bp SSRE found upstream of the start site of shear stress-sensitive genes (46). At least three othercis-acting elements functional in mechanosensitive gene transcription have been identified. The SSRE is not, however, involved in the regulation of some mechanically sensitive genes. The shear phorbol ester tissue-responsive element (TRE) described by Shyy et al. (49) is a divergent TRE that can also transduce mechanical signals to transcriptional events. More recent data suggest that mechanotransduction may also occur via binding and/or displacement of the surfactant protein (SP-1; see Ref. 33) and Egr-1 (50) families of transcription factors and their recognition elements. Upregulation of VEGF expression by hypoxia involves both transcriptional and posttranscriptional processes mediated by SP-1, hypoxia-inducible transcription factors, the von Hippel-Lindau protein, and protein kinase C (PKC) members (21,34, 36, 41). Although the presence of mechanosensitive regulatory elements does not necessarily indicate functional importance, their presence makes it more likely that changes in response to mechanical stimuli are at least in part the result of changes at the transcriptional level.
Induction of VEGF and FGF-2 expression by stretch could play a role in vascular development or in postnatal vascular homeostasis, remodeling, or disease. Arterial smooth muscle cells are exposed to stretch resulting from pulsatile blood flow during the cardiac cycle, and both VEGF and FGF-2 have multiple effects on different cells in the vascular system. VEGF stimulates endothelial cell proliferation, permeability, chemotaxis, secretion of proteases that degrade extracellular matrix, and invasion (16). In vascular smooth muscle cells and monocytes, VEGF stimulates chemotaxis but not proliferation (8,22). FGF-2 stimulates endothelial cell and vascular smooth muscle cell proliferation and chemotaxis as well as vasodilation (11, 26, 53).
The roles of VEGF and FGF-2 in embryonic vascular development have been investigated in mice with targeted gene deletions. Mice lacking FGF-2 are viable with apparently normal vasculature, demonstrating that FGF-2 is not required for embryonic vascular development (40). Mice with targeted deletions of VEGF die in utero at day 7.5because of failure of endothelial cell differentiation (3,17). Because embryonic lethality occurs so early, it is impossible to assess the role of VEGF in vascular development later in development. Other evidence implicates VEGF in pulmonary vascular development. In ovine models of congenital heart disease with increased or decreased pulmonary artery blood flow, increased VEGF expression in the lung correlates with increased pulmonary arterial pressure, increased blood flow, and increased vascularization (S. J. Soifer and J. R. Fineman, unpublished observation). Although physiological stimuli other than increased stretch may also be present in this setting (e.g., changes in shear stress or oxygenation), these findings are consistent with a model in which stretch regulates VEGF expression, thereby influencing pulmonary development.
It is not clear what role VEGF and FGF-2 may play in postnatal vascular homeostasis or vascular disease. The chemotactic effect of VEGF on vascular smooth muscle cells and monocytes suggests that VEGF could be important in the development of atherosclerotic lesions (8,22). Increased expression of VEGF in pulmonary artery smooth muscle cells in patients with congenital heart disease with increased pulmonary blood flow suggests a role for VEGF as a mediator of pulmonary arteriopathy (20). However, adenoviral delivery of the VEGF gene to the pulmonary epithelium by intratracheal administration decreased the severity of hypoxia-induced pulmonary hypertension, suggesting that, at least when VEGF is expressed in the epithelium, the effect is protective (42). FGF-2 has been implicated as an important factor in intimal proliferation after experimental arterial injury and in venous bypass grafts, in atherosclerosis and unstable angina, and in experimental pulmonary hypertension models (1, 4, 18, 25, 29, 39, 52). Regulation of VEGF and FGF-2 expression by stretch may contribute to the development of these arterial diseases, along with other factors that may be present (e.g., hypoxia, direct mechanical injury, endothelial denudation, or lipid accumulation).
Mechanical stretch alters the regulation of numerous genes in vascular smooth muscle cells, including growth factors, components of extracellular matrix, and gap junction proteins (5, 9, 15,31). Stretch also activates intracellular signaling pathways, including the Jun NH2-terminal kinase/stress-activated protein kinase and nuclear factor-κB pathways (24,27). The signaling pathways activated by stretch that mediate upregulation of VEGF and FGF-2 gene expression in pulmonary artery smooth muscle cells are unknown. Stretch stimulates the generation of inositol trisphosphate and calcium flux in vascular smooth muscle cells (2, 32) and induces VEGF gene expression by PKC-dependent mechanisms in mesangial cells (23). These findings suggest that, in pulmonary artery smooth muscle cells, stretch may upregulate VEGF expression by generating inositol phosphates, which stimulate calcium flux and activate PKC enzymes. Further investigation is required to investigate this hypothesis.
Address for reprint requests and other correspondence: T. P. Quinn, Box 1245, Dept. of Pediatrics, Univ. of California San Francisco, San Francisco, CA 94118 (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. Section 1734 solely to indicate this fact.
First published October 26, 2001;10.1152/ajplung.00044.2001
- Copyright © 2002 the American Physiological Society