Effect of Mechanical Forces on Growth and Matrix Protein Synthesis in the In Vitro Pulmonary Artery
Analysis of the Role of Individual Cell Types
Abstract The effect of mechanical stimuli on pulmonary artery growth and matrix tissue synthesis (and how individual cell types in the vessel wall respond to such stimuli) is incompletely characterized. Rabbit pulmonary arteries were placed in tissue culture medium and subjected to varying magnitudes of stretch or hydrostatic pressure (separately) for 4 days. The rate of protein synthesis in smooth muscle cells (by quantitative autoradiography) was positively related to the magnitude of stretch, as were the percentage of procollagen type I–positive cells and the rate of cell replication. In adventitial fibroblasts, stretch increased the rate of replication but not of protein synthesis. Hydrostatic pressure had little or no effect on the variables measured in either smooth muscle cells or fibroblasts. Stretch also increased the rate of elastin and collagen synthesis in the whole pulmonary artery segment, and after 4 days of stretch, the contents of actin and elastin were increased. Removal of the endothelium did not affect stretch-induced protein, collagen, or elastin synthesis but augmented stretch-induced smooth muscle replication. These data suggest that in the intact pulmonary artery, stretch, but not pressure, can stimulate hypertrophy and hyperplasia in smooth muscle cells and hyperplasia in fibroblasts. Matrix protein synthesis and accumulation are also increased by stretch. Neither stretch-mediated growth nor matrix protein synthesis required endothelium in this model.
Several pieces of data lend support to the notion that mechanical stimuli are important in modulating growth and matrix protein synthesis in pulmonary blood vessels: (1) Patients with heart lesions that increase PA pressure and flow develop pathological remodeling of the pulmonary vessels, whereas those with increased pulmonary blood flow but normal PA pressures are much less likely to do so, suggesting that increased intravascular pressure per se may stimulate medial hypertrophy and fibrosis.1 (2) Similar observations have been made in somewhat better-defined systems. For example, in the developing lamb, the pattern of tropoelastin mRNA expression in the PA closely parallels PA pressure.2 (3) Research has clearly shown stretch to modulate growth and phenotype in striated muscle in vitro.3 4 5 These observations, while suggestive, neither clearly establish a link between mechanical stimuli and alterations in pulmonary vascular biology nor specify in what way(s) such forces may alter PA structure and physiology.
In fact, several studies of cultured PA endothelial cells and SMCs suggest that mechanical forces may have little effect on growth and even inhibit it.6 7 8 However, such “negative” findings could be due to alterations in the biology of the cell related to culture: If the cell’s response to mechanical forces depends on its shape, its orientation relative to the vector of the force, the matrix to which it adheres, and/or its interaction with other cell types, cell culture systems may be especially poor surrogates for in vivo conditions.4 9 There are only two published reports on the effect of mechanical stimuli in the intact PA in vitro.10 11 Both indicate that stretch of short duration (4 hours) causes an endothelium-dependent increase in matrix protein synthesis in PA segments in vitro. These experiments confirm that stretch can affect collagen and elastin expression in the intact PA but leave multiple essential questions unanswered: (1) The effect of mechanical forces on growth (hypertrophy and hyperplasia) of the cellular constituents of the intact PA is unreported. (2) Because previous analyses have not separated the media from the adventitia, it is not known whether SMCs or fibroblasts (or both) account for the stretch-induced increase in matrix protein synthesis. In vivo studies demonstrate that PA SMCs and adventitial fibroblasts respond in temporally and spatially distinct ways to hypoxia-induced PA hypertension,12 13 14 suggesting that SMCs and fibroblasts may respond in different ways to mechanical forces. How the individual cellular components of the intact vessel respond to mechanical stimuli is unexplored. (3) Increased intraluminal pressure is associated with an increase in both hydrostatic pressure and wall stress (given a fixed vessel radius and wall thickness), and the biological effect(s) of these forces may differ. Hydrostatic pressure can affect the release of growth regulators in cultured PA endothelial cells6 15 (and cultured systemic arterial endothelial cells16 and SMCs17 ), yet the effect of hydrostatic pressure on PA growth and matrix protein synthesis is unreported.
The present study was therefore undertaken to approach four important but largely unaddressed questions: (1) Does stretch, or hydrostatic pressure, increase growth (as determined by the relative rate of total protein synthesis, the percentage of cells synthesizing DNA, and accumulation of total cellular actin) in the intact PA segment in vitro? (2) Does stretch increase the relative rate of synthesis of collagen and/or elastin or the accumulation of elastin in the intact PA segment in vitro? (3) Which cells (SMCs, fibroblasts) in the PA are affected by stretch or pressure? (4) What effect does removal of endothelium have on the response to mechanical stimulation in the PA?
Materials and Methods
Procurement of PA Vessel Strips: Tissue Culture Setup
Male New Zealand rabbits (2.3 to 2.7 kg) were killed by use of inhaled ether, and before cessation of cardiac activity, heparin (100 U/kg IV) was given. The right and left proximal PAs were sterilely removed. During the procedure, the vessels were kept moist with tissue culture medium (medium 199 [Earle’s salts] with l-glutamine [2 mmol/L], penicillin [100 U/mL], streptomycin [0.1 mg/mL], amphotericin B [0.25 μg/mL], and 10% fetal calf serum; all were obtained from Sigma Chemical Co) containing papaverine (1 mmol/L). These cylinders of PAs were incised longitudinally to convert them into strips (4 mm wide, ≈5 mm long). Plexiglas “clamps” were affixed to the two ends of the strip; one clamp was used to suspend the strip, and a weight was attached to the other clamp so that the strip could be subjected to the desired level of wall stress (Fig 1⇓). The strips were stretched in the direction of the long axis of the SMCs, presuming that these cells are circumferentially arranged in the vessel wall.
Three or four PA strips were removed from each animal (each proximal PA is divided into one or two strips), put into culture (using the culture medium noted above, lacking papaverine), and maintained under identical conditions except for wall stress or hydrostatic pressure. For studies of the effect of wall stress on PA growth, different strips (from the same rabbit) were exposed to loads comparable to intraluminal pressures of 12, 25, and 45 mm Hg, all at a hydrostatic pressure of 12 mm Hg. (For some experiments, as noted, only two levels of stress were used [12 and 45 mm Hg].)
For studies of the effect of hydrostatic pressure on vessel growth, PA strips were subjected to hydrostatic pressures of 12, 25, and 45 mm Hg, all at a wall stress of 12 mm Hg. Hydrostatic pressure was adjusted by altering the height of a column of fluid (culture medium), which was in continuity with the (closed) culture chamber (Fig 1⇑). The hydrostatic pressure (in millimeters of mercury) was taken as the distance (in millimeters) between the midpoint of the PA strip and the top of the fluid column divided by 13.6 (the specific gravity of mercury).
The strips were suspended vertically in custom-built culture chambers of ≈10 mL volume (one strip per chamber), which were kept in a humidified incubator (37°C, 5% CO2/balance air) (see Fig 1⇑). The culture medium was changed daily. Intentional removal of endothelium was performed by gently abrading the inner surface of the vessel with a cotton-tipped rod. In all experiments, the presence (or absence) of endothelium was confirmed by using en face silver staining18 at the end of the experiment.
The protocol for these studies was approved by the University Committee on Use and Care of Animals, the University of Michigan.
Setting the Wall Stress
Wall stresses were chosen to correspond to physiologically relevant intravascular PA pressures. As noted above, a weight was hung from each PA strip in order to subject it to a given load. The load corresponding to a given intravascular pressure—and hence wall stress—was calculated as follows: ς=r/t(P), where for a cylinder, ς is tangential wall stress, r is internal radius, t is wall thickness, and P is intraluminal pressure.19 The ς for a given P can be related to the load necessary to cause that ς:w=ς · l · t, where w is tangential load (dynes) and l is the width of vessel strip.
In preliminary studies, the average radius and wall thickness of PAs from rabbits the same size as those used for the present experiments were measured by using cut sections of paraffin-embedded and -fixed PAs. Ten vessels were measured as follows: r=0.90 to 1.20 mm (mean=1.0±0.1 mm): t=0.114 to 0.177 mm (mean=0.141±0.02 mm). These mean dimensions were used to calculate the load necessary for the desired wall stress (for 12 mm Hg, 0.67 g; for 25 mm Hg, 1.40 g; and for 45 mm Hg, 2.52 g).
Measurement of Relative Rates of Total Protein Synthesis and Cell Replication
For measuring the relative rate of protein synthesis, PA segments were placed in culture under the wall stress or hydrostatic pressure noted above. After 4 days of incubation, the culture medium was replaced with identical media, lacking antibiotics and containing 3 μCi/mL l-[4,5-3H]leucine (60.0 Ci/mmol; DuPont NEN). The previous stretch conditions were maintained during the 4-hour labeling period, after which time the vessels were washed three times with PBS (4°C) containing “cold” leucine (10 mmol/L), fixed for 10 minutes in 6% formalin/0.5% TCA with 0.76 mol/L cold leucine,20 and processed for quantitative autoradiography as detailed below.
For measurement of cell replication, BrdU (0.010 mmol/L, Boehringer Mannheim or Amersham) was added to the culture media 24 hours before the termination of the experiment. Fixation was by the above-noted formalin-TCA fixative. Five-micrometer-thick sections of the strips were immunostained by use of anti-BrdU antibodies and secondary antibodies as per directions of the manufacturer of the kit (either Boehringer Mannheim or Amersham). Secondary antibodies were applied as detailed below. Cell nuclei were counterstained with hematoxylin.
Immunostaining for Procollagen Type I
Mouse anti-sheep procollagen I amino-terminal monoclonal antibody SPI.D8 was obtained from the Developmental Studies Hybridoma Bank maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, Md, and the Department of Biological Sciences, University of Iowa, Iowa City, under contract NO1-HD-6-2915 from the National Institute of Child Health and Human Development. This antibody recognizes only newly synthesized procollagen (before cleavage of the amino terminal end)21 and hence does not reflect collagen accumulation.
The PA strips were fixed as above and embedded in paraffin. Five-micrometer-thick sections of vessels were deparaffinized, and the antibody (dilution, 1:20) was applied overnight at 4°C. Nonimmune mouse or rabbit serum was used as a negative control. The secondary antibody (biotinylated horse anti-mouse antibody; dilution, 1:500; Vector Laboratories) was applied for 3 hours at 4°C, followed by avidin-biotin amplification (ABC Elite Kit, Vector Laboratories) for 30 minutes. Incubation with 0.1% 3,3′-diaminobenzidine (Sigma) and H2O2 at room temperature for 5 to 10 minutes produced a brown reaction product. Gill’s hematoxylin was used for nuclear counterstaining, followed by dehydration and coverslip mounting. As a positive tissue control for procollagen type I protein expression, human hypertrophic scar tissue was used, which yielded a pattern of well-localized immunostaining among reactive dermal fibroblasts, with negative staining of skin epithelial cells (data not shown).
The procedures followed were generally those of Baserga and Malamud.20 The vessel segments were kept in fixative at 4°C for 3 days (to wash out all unincorporated 3H label), rinsed in running tap water for 12 hours, and embedded in paraffin. After the first 500 μm of the vessel (which, being immediately adjacent to the clamp, may have suffered damage) was discarded, multiple 5-micrometer-thick sections were cut and mounted on silane-coated slides. Sections from each of the two or three vessel strips from each experiment were mounted side by side on a single slide so that they would be exposed to identical conditions of emulsion, exposure, developing, etc. Previously 3H-labeled (hot) and unlabeled (cold) vessel segments were included on each slide for controls. After deparaffinization, the slides were coated with undiluted Kodak NTB 2 emulsion (Kodak) (at 40°C). The slides were allowed to dry for 15 minutes, kept in a humidified chamber at 37°C for 60 minutes (to reduce background), and then stored in a light-tight box with desiccant at 4°C for 7 days. Preliminary experiments showed that a 7-day exposure results in a large enough number of grains to be easily discriminated from the background (<1% of total grains) but that the grain distribution is so sparse that there is relatively little overlap of grains. The slides were developed by use of Kodak D 19 developer (4 minutes at 15°C), fixed by use of Kodak fixer (5 minutes at 15°C), and rinsed in distilled H2O. The cytoplasm was lightly counterstained by using eosin.
The percent area of media and adventitial fibroblasts covered by silver grains was used as a measure of the relative rate of protein synthesis and was determined by using the Image I system of computer color image analysis (Universal Imaging Corp). This system can readily distinguish silver grains from the underlying cellular material and can precisely quantify the relative area of silver grains. For determining protein synthesis in the media, at least 10 contiguous nonoverlapping microscopic fields (magnification, ×1000) from at least five sections (total, ≥50 fields) of each artery segment were analyzed. For measuring protein synthesis in adventitial fibroblasts, the cluster of silver grains overlying the fibroblast is encircled using the Image I system’s area delineation system, and the percent area of silver grains within that area determined. As noted in preliminary observations of the organ culture system (below), fibroblasts at the outer edge of the adventitia appear to be highly activated, presumably because of an injury response related to dissection of the PA, and were therefore excluded from analysis.
For cell replication in medial SMCs, the number of BrdU-labeled nuclei per total nuclei in each of at least five sections was determined. At least 500 cells were counted. The rate of cell proliferation in the adventitia was determined by counting the number of BrdU-labeled fibroblasts within the interior of the adventitia; fibroblasts at the outer edge of the adventitia were excluded for the reason noted above.
The number of cells synthesizing procollagen type I was quantified by counting the number of procollagen-positive cells in the media and normalizing to the total number of cells (determined by counting nuclei). Because virtually all adventitial fibroblasts stained positively in the PA strips, no attempt was made to quantify procollagen production in these cells.
Validation of the Use of Quantitative Autoradiography for Measuring the Relative Rate of Protein Synthesis in This System
Although quantitative autoradiography is a well-described technique for measuring the relative rate of protein synthesis in similar systems,13 20 we conducted initial studies to confirm its suitability for use in our model. Segments of rabbit PA and aorta (n=4 experiments) were incubated in tissue culture medium containing 3 μCi/mL of [3H]leucine for 30, 60, and 90 minutes and for 90 minutes in identical medium with cycloheximide (0.010 mmol/L) to inhibit protein synthesis. The segments were divided into two parts: one was prepared for autoradiography (as described above); in the other, the relative rate of protein synthesis was determined by measuring the number of TCA-precipitated counts (in disintegrations per minute) normalized to total protein, which was measured according to the Bradford assay.8
Measurement of the Relative Rate of Elastin Synthesis
PA strips were treated as for determination of total protein synthesis, but they were subjected to only two different wall stresses (12 and 45 mm Hg, 4 days) and were labeled (4 hours) with 3 μCi/mL media of l-[2,3,4,5-3H]proline (127 Ci/mmol; DuPont NEN). At the end of the labeling period, the strips were removed from culture, rinsed in cold PBS three times, and weighed. The strips were digested with CNBr (5% CNBr in 70% formic acid [both from Sigma] under N2 at 20°C for 24 hours) to digest all nonelastin protein.11 (In a preliminary experiment, amino acid analysis confirmed that only elastin residue remained after the CNBr digestion.) The elastin was washed in 100°C H2O three times and dissolved in Soluble (DuPont NEN), liquid scintillation fluid was added, and rate of disintegrations per minute was measured. The value of disintegrations per minute (representing the relative rate of elastin synthesis) was normalized to the wet weight of the strip.
Measurement of the Relative Rate of Collagen Synthesis
A technique previously described was used.8 Paired PA strips were placed in culture under 12 and 45 mm Hg–equivalent stresses for 4 days and were labeled with [3H]prolene (3 μCi/mL) for the last 4 hours of culture (n=4 independent experiments). The strips were then homogenized in cold buffer (0.65 mol/L NaCl, 0.1 mol/L Tris [pH 7.4], 4.7 mmol/L CaCl2, and 2.5 mg/mL N-ethylmaleimide) and freeze-thawed. Bovine serum albumin (100 μg/mL) was added as a carrier. An aliquot was removed, 10% TCA was added, and the material was allowed to flocculate for 30 minutes at 4°C. The TCA-precipitated material was pelleted, washed twice with 5% TCA and twice with cold 95% ethanol, dried, and dissolved in 0.1N NaOH, and the rate of disintegrations per minute was determined with a liquid scintillation counter. A second aliquot was digested with a highly specific collagenase (collagenase form III, 10 U/mL media, Advance Biofactures Corp) for 90 minutes at 37°C and then treated identically as the nondigested aliquot. The rates of disintegrations per minute representing total protein and total protein minus collagen were normalized to the total protein in the sample (Bradford assay). The relative rate of collagen synthesis was determined, assuming that the ratio of proline residues in collagen relative to noncollagen protein is 5.4.22
Measurement of the Total Elastin Content of PA Strips
Paired PA rings were weighed and then placed in culture under 12 and 45 mm Hg–equivalent stresses for 4 days. (PA rings rather than strips were used for this and the experiment described below but were subjected to loads and culture conditions identical to the PA strips in the other experiments.) The rings were then digested with CNBr as described above, and the resulting elastin residue was washed with water, dried, and weighed.23 The weight of the elastin was normalized to the (preculture) weight of the PA strip (n=4 independent experiments).
Measurement of the Total Actin Content of PA Strips by Western Blotting
Paired PA rings were placed in culture under 12 and 45 mm Hg–equivalent stresses for 4 days; they were then divided into two halves, and each one was weighed. One of the halves of the strip was homogenized in 0.5 mL of lysis buffer (9.5 mol/L urea, 1.6% Triton X-100, and 5% β-mercaptoethanol) and freeze-thawed, and undissolved material was removed by centrifugation. Equal aliquots of the supernatant were boiled in 1% SDS for 10 minutes, added to 5′ Laemmli loading buffer, size-fractionated with an SDS–polyacrylamide gel (3.6% stacking gel, 10% running gel), and transferred to an Immobilon-P (Millipore) membrane. The membrane was blocked by using 5% bovine serum albumin in TBS (20 mmol/L Tris [pH 8.0] and 150 mmol/L NaCl) for 1 hour, washed, and then incubated with a monoclonal antibody to all actin isoforms (Sigma, clone AC-40) for 1 hour. After it was washed, the membrane was incubated with the secondary antibody (anti-mouse IgG conjugated to horseradish peroxidase, Jackson Immuno Research Lab Inc) for 1 hour, and the immunoreactive band was detected by using the Amersham enhanced chemiluminescence system. The autoradiographic signals were quantified by using the Eagle Eye II still video system (Stratagene) and the National Institutes of Health Image 1.54 program. The relative amount of actin (arbitrary units) was normalized to the DNA content of the strip (see below) (n=7 independent experiments).
For measuring DNA, one half of the PA strip was homogenized in buffer (1′ standard saline citrate with 0.25% SDS), freeze-thawed, and incubated overnight at 37°C. Insoluble material was removed by centrifugation. The amount of DNA in an aliquot of the supernatant was measured by use of bisbenzimide H33258, with calf thymus DNA used for standards,24 and normalized to the weight of the strip.
For each independent experiment, tissue from a single rabbit was divided into two or three portions, and each was subjected to different magnitudes of stretch or pressure. Hence, for statistical analysis of multiple experiments in which only two comparisons were made, Student’s paired t test was used. When three different measurements were compared, a two-tailed ANOVA with post hoc comparisons using the Tukey test was performed.25 A significant difference was taken at P<.05. Data are given as mean±SD.
Preliminary Observations of the Organ Culture System
Preliminary observations were as follows: (1) Rabbit PA segments can be maintained in organ culture for at least 5 days with excellent viability, as judged by their histological appearance (Fig 2⇓) and ability to synthesize protein. En face silver staining (n=7) showed endothelium to cover ≥≈80% of the vessel segment in all cases after 4 days in culture. (2) In [3H]leucine-labeled strips, silver grains were uniformly arrayed over SMCs in the media. In the adventitia, clusters of grains were seen over individual fibroblasts. At the outer edge of the adventitia (which was, of course, separated from the surrounding tissue during the dissection of the PA), there was a very dense display of silver grains, representing a high rate of protein synthesis of fibroblasts, which are presumably reacting to injury related to PA dissection (Fig 3⇓). (3) Application of load to the PA segments caused them to elongate (segment lengths were ≈3 mm before application of load, ≈4 mm at 12 mm Hg–equivalent stress, ≈5 mm at 25 mm Hg–equivalent stress, and ≈6.5 mm at 45 mm Hg–equivalent stress); hence, both stress and strain differed with different loads.
Validation of the Use of Quantitative Autoradiography in This System
The background level of silver grains was very low (area of background grains was 0.01±0.002%, with the area of grains over the media being 3% to 10%). PA strips fixed and treated identically to the experimental strips but not [3H]leucine-labeled (negative controls) showed no silver grains other than background. There was good correspondence between the number of silver grains per unit area and the TCA-precipitated disintegrations per minute per total protein (Fig 4⇓), indicating that the relative rate of protein synthesis as determined by autoradiography is proportional to that measured by TCA-precipitated counts.
Effect of Wall Stress and Hydrostatic Pressure on the Relative Rate of Protein Synthesis
In the media, the relative rate of protein synthesis, as measured by the percent area of silver grains, was positively and significantly related to the magnitude of wall stress. Removal of the endothelium did not affect the stress-related increase in protein synthesis (Fig 5⇓).
In adventitial fibroblasts, the relative rate of protein synthesis was considerably higher than in medial SMCs but was the same regardless of the wall stress: The percent area of silver grains over fibroblasts in strips with endothelium was 9.35±0.32% at a wall stress of 12 mm Hg, 9.46±0.12% at a wall stress of 25 mm Hg, and 9.65±0.17% at a wall stress of 45 mm Hg (P=NS for all, n=7 independent experiments). For deendothelialized strips, the percent area of silver grains over fibroblasts was 9.14±0.40% at a wall stress of 12 mm Hg, 9.36±0.33% at a wall stress of 25 mm Hg, and 9.84±0.29% at a wall stress of 45 mm Hg (P=NS for all, n=7 independent experiments). The relative rate of total protein synthesis was not related to hydrostatic pressure in either the media (Fig 5⇑) or the adventitia (data not shown).
Effect of Wall Stress and Hydrostatic Pressure on Cell Replication
The percentage of BrdU-labeled nuclei in the media was the same in endothelialized strips at 12 and 25 mm Hg–equivalent wall stresses but increased significantly at a wall stress of 45 mm Hg. In the deendothelialized strips, the number of medial SMCs undergoing DNA synthesis at the 12 mm Hg load was the same as in endothelialized strips, but the percent BrdU-positive cells was increased in both the 25 and 45 mm Hg–equivalent strips relative to 12 mm Hg deendothelialized strips (Figs 6⇓ and 7⇓). In addition, the percentage of BrdU-positive cells was significantly higher in the deendothelialized strips at loads of 25 and 45 mm Hg than in strips with intact endothelium, suggesting that the endothelium may serve to negatively regulate the stretch-induced proliferation of SMCs (Fig 7⇓).
The percentage of BrdU-labeled adventitial fibroblasts was significantly and positively related to the magnitude of wall stress in both endothelialized and deendothelialized strips (Fig 8⇓).
The percentage of BrdU-labeled cells was only weakly related to the level of hydrostatic pressure: A small but significant increase in BrdU-positive cells was seen in the media at the highest pressure (Fig 7⇑) and in fibroblasts when 12 mm Hg pressure was compared with 25 mm Hg (Fig 8⇑).
Effect of Wall Stress and Hydrostatic Pressure on the Number of Procollagen Type I–Positive SMCs
The percentage of SMCs positive for procollagen type I was positively and significantly related to wall stress in the media of the endothelialized strips. The percentage of procollagen-positive SMCs was considerably higher in the deendothelialized strips than in those with the endothelium intact at all levels of stress, and the number of procollagen-positive cells was greater in the 45 mm Hg–equivalent strips than those at lower stress (Figs 9⇓ and 10⇓).
Effect of Wall Stress on the Relative Rate of Collagen Synthesis
This experiment was performed to determine if the increase in procollagen type I synthesis as measured by immunostaining was also reflected in a change in the rate of total collagen synthesis measured biochemically. The relative rate of collagen synthesis was 62±6% greater in the strips maintained at the 45 mm Hg–equivalent load than at the 12 mm Hg load (P<.05).
Effect of Wall Stress on the Relative Rate of Elastin Synthesis
The values of disintegrations per minute per milligram wet tissue weight were significantly increased (P<.02) in the 45 mm Hg–equivalent strips relative to the 12 mm Hg strips in both the endothelialized and deendothelialized strips (n=6 independent experiments in each of the two groups), indicating that stretch increased the relative rate of elastin synthesis. For endothelialized strips, values were 131±42 dpm/mg for 12 mm Hg–equivalent strips and 197±68 dpm/mg for 45 mm Hg strips (+50%). For deendothelialized strips, values were 137±78 dpm/mg for 12 mm Hg–equivalent strips and 183±67 dpm/mg for 45 mm Hg strips (+33%). The percent increase in elastin synthesis evoked by stretch (12 mm Hg–equivalent stress versus 45 mm Hg) was roughly the same as the percent increase in total protein synthesis (measured by autoradiography) in the media of similar strips at identical stresses (+59% in endothelialized strips and +65% in deendothelialized strips).
Effect of Wall Stress on PA Elastin Content After 4 Days in Culture
The elastin content of the 12 mm Hg–equivalent PA rings was 0.077±0.004 mg elastin per milligram wet weight of tissue; the elastin content of the 45 mm Hg–equivalent rings was 0.084±0.004 mg elastin per milligram wet weight of tissue (+9%, P=.001). Since the weight of elastin was normalized to the weight of the PA rings before culture, this implies a significantly greater accumulation (after 4 days of culture) of elastin in the rings subjected to a load of 45 mm Hg than the more lightly loaded rings.
Effect of Wall Stress on PA Actin Content
The mean quantity of actin (arbitrary units) normalized to micrograms of DNA in the 12 mm Hg–equivalent PA rings was 73±32; the actin content of the 45 mm Hg–equivalent rings was 117±47 (+60%, P<.005).
Although the idea that mechanical stimuli can alter PA growth has come to be widely accepted, the available data are actually very limited and do not clearly support this notion.6 7 8 Even investigations of the more frequently studied aortic SMCs are incomplete and somewhat inconsistent with regard to the effect of mechanical stimuli on vascular growth (References 26 through 2926 27 28 29 and discussion in Reference 88 ). The experiments described here (summarized in the Table⇓) confirm previous studies showing stretch to increase matrix protein synthesis in the intact PA10 11 but also for the first time demonstrate the following: (1) Stretch can increase the rate of total protein synthesis in medial SMCs and the total actin content (normalized to DNA) of the PA strip, indicating that it is a hypertrophic stimulus in the intact PA in vitro. Adventitial fibroblasts did not increase total protein synthesis with stretch, although the baseline rate of protein synthesis was considerably higher in these cells and thus may not have been capable of upregulation. It is unclear to what extent this high rate of protein synthesis in the fibroblasts may have been related to adventitial injury caused by vessel dissection and therefore perhaps not reflective of in vivo biology. (2) Stretch can increase the rate of proliferation of medial SMCs and adventitial fibroblasts, suggesting that it can also act as a hyperplastic stimulus in the intact PA. The effect of stretch on replication was more marked in fibroblasts than in SMCs, which is reminiscent of in vivo models of PA hypertension (see below). (3) Stretch can increase the rate of matrix protein synthesis and accumulation in the in vitro PA. We took advantage of the availability of procollagen type I antibodies in an effort to discover which type(s) of cells upregulates collagen synthesis with stretch, but uniformly positive immunostaining of fibroblasts made it impossible to determine if stretch affected procollagen synthesis in these cells. These experiments do, however, suggest that stretch upregulates procollagen synthesis in SMCs (at least) in the intact PA, which has not been previously established.10 11 In addition, they provide the novel observation that collagen synthesis may be upregulated at least in part via an increase in the number of cells synthesizing detectable amounts of procollagen rather than solely through an increase in collagen produced by each cell. Because the procollagen immunostaining provided only an index of the number of SMCs synthesizing collagen, the relative rate of total collagen synthesis in the PAs was also quantified biochemically. The rate of collagen synthesis was also positively related to the magnitude of stretch, which is consistent with the immunostaining data, although it should be noted that the entire PA segment was analyzed (including adventitia) and that these data may thus reflect collagen synthesis by fibroblasts as well as SMCs. The relative rate of elastin synthesis (measured biochemically) and the total amount of elastin in the PA segment after 4 days of culture were also increased by stretch. These data further suggest that stretch may be an important regulator of matrix protein synthesis (and accumulation) in the PA. (4) The magnitude of hydrostatic pressure had little effect on total protein or collagen synthesis and was only weakly related to cell proliferation in the adventitia, suggesting that this force may be less important in PA remodeling than stress/strain. However, if the endothelium releases a growth-inhibiting factor(s) with pressure (see below), it is possible that any direct effect of pressure on SMCs may be negated by such an inhibitory substance. The experiments reported here therefore do not eliminate the possibility of a direct effect of hydrostatic pressure on either SMCs or endothelial cells but suggest that it has little net effect on the intact vessel.
Role of Endothelium in Modulating Stretch-Induced Growth and Matrix Protein Synthesis
Two previous reports showed an endothelium-dependent stretch-induced increase in matrix protein synthesis in PA segments,10 11 but we found that endothelium was not necessary for the stretch-induced increase in growth or matrix protein synthesis. Although it is conceivable that the strips were not completely deendothelialized in our experiments, this is unlikely for two reasons: (1) Absence of the endothelium was confirmed in every experiment using en face silver staining. (2) As discussed below, the deendothelialized strips showed an augmented rate of SMC replication and number of procollagen-positive cells, suggesting that they behaved in a way biologically distinct from the endothelialized strips.
The augmented effect of stretch on SMC replication in the deendothelialized strips suggests that mechanical stimulation may cause PA endothelial cells to elaborate a substance that inhibits cell replication, which is consistent with data showing stretch/pressure to cause cultured PA endothelial cells to elaborate a growth inhibitor.6 Deendothelialized strips also had an increased percentage of procollagen-positive cells relative to endothelialized strips, suggesting also that the endothelium may help regulate collagen synthesis in this system, but it is also possible that this could be due to mechanical injury of SMCs resulting from the process of deendothelialization.
Why the results of the present study differ from previous reports regarding the endothelium dependence of stretch-induced matrix protein synthesis is unclear, but the model used here differs in multiple ways from those previously reported.10 11 The species of animal used, the solution in which the PA segments were incubated, the time between PA removal and study, and the duration of stretch (4 hours versus 4 days) were all significantly different, and one or more of these factors may have played a role.
Comparison of These In Vitro Data With Those Derived From In Vivo Systems
The system described in the present study differs considerably from in vivo models of PA hypertension: the latter are associated with a variety of physiological alterations that may affect PA growth (such as hypoxia, activation of the sympathetic nervous system, etc), whereas only the magnitude of the mechanical stimulus is altered in the in vitro system. Nevertheless, to put these findings in perspective, it is perhaps useful to compare them with in vivo models.
Our finding of a stretch-mediated increase in total protein and matrix protein synthesis is generally consistent with in vivo experiments. Increased elastin and collagen synthesis occurred in large PAs after as little as 4 days of hypoxia in the calf30 and within 1 day in the hypoxic rat.31 Total protein synthesis in large PAs in the hyperoxic rat is decreased after 3 days of hyperoxia but markedly increased after 7 days,32 findings not inconsistent with the data reported here, given that a 4-day period of stretch was studied. On the other hand, McKenzie et al13 showed increased protein synthesis in both the media and adventitia after 3 days of hypoxia in rats, which is at variance with our findings, perhaps for the reason noted above.
Our data regarding cell replication, which showed a greater stretch-induced increase in the replication of fibroblasts than of SMCs, are also largely congruent with those obtained from in vivo models of PA hypertension. Hilar PAs of hypoxic rats show a marked increase in fibroblast replication after 3 to 5 days of hypoxia but a more modest increase in SMC replication.12 13 Hilar PAs in the hyperoxic rat (day 4) also had increased fibroblast replication but little change in SMC replication until day 7.33
Implications for Further Studies of the Biological Effects of Mechanical Forces on PAs
Although, as noted above, the findings reported in the present study are in many ways consistent with those observed with experimental models of PA hypertension, extrapolation to in vivo biology must be done with caution. The near impossibility of selectively altering the mechanical forces acting on PAs in vivo dictates the use of in vitro systems for studying these phenomena, yet at least two factors potentially confound the present studies: (1) The process of removing the PAs and the fact of culture itself doubtlessly alter the biology of the PA strips in many and incompletely understood ways. (2) When vascular tissue is stretched, both stress and strain are increased; whether the biological effects of these two stimuli are the same is unknown. Increased PA pressure in both experimental and clinical settings probably does not cause significant PA dilation (and hence increased strain) but appears to be associated with decreased PA diameter.34 35 Models such as the one described in the present study may therefore imperfectly mimic the physical forces that impact PA endothelial cells, SMCs, and fibroblasts in vivo. Nevertheless, these experiments provide substantial support for the hypothesis that mechanical forces mediate growth and matrix protein synthesis in the pulmonary circulation and increase our knowledge of which cells are affected by these forces. In addition, they suggest that stretch, at least under some circumstances, can act directly on SMCs and fibroblasts to alter their growth and synthetic characteristics.
Selected Abbreviations and Acronyms
|SMC||=||smooth muscle cell|
This study was funded by National Institutes of Health grant HL-42908 and a grant from the Michigan Affiliate of the American Heart Association. The authors thank Brett Phinney for technical assistance and Jun-ichi Sadoshima for discussion and assistance.
- Received December 29, 1994.
- Accepted June 9, 1995.
- © 1995 American Heart Association, Inc.
Kulik TJ. Pulmonary hypertension. In: Fyler DC, ed. Nadas’ Pediatric Cardiology. Philadelphia, Pa: Hanley & Belfus Inc; 1992:87-88.
Stenmark KR, Durmowicz AG, Roby JD, Mecham RP, Parks WC. Persistence of the fetal pattern of tropoelastin gene expression in severe neonatal bovine pulmonary hypertension. J Clin Invest. 1994;93:1234-1242.
Cooper G IV. Cardiocyte adaptation to chronically altered load. Annu Rev Physiol. 1987;49:510-518.
Vandenburgh HH. Mechanical forces and their second messengers in stimulating cell growth in vitro. Am J Physiol. 1992;262:R350-R355.
Sadoshima J, Jahn L, Takahashi T, Kulik TJ, Izumo S. Molecular characterization of the stretch-induced adaptation of cultured cardiac cells. J Biol Chem. 1992;267:10551-10560.
Bothwell, T, Rabinovitch M. A pulmonary artery smooth muscle growth inhibitor released from endothelial cells stretched at high pulsatile pressure. FASEB J. 1988;2:A1577. Abstract.
Farmer SR, Dike LE. Cell shape and growth control: role of cytoskeleton-extracellular matrix interactions. In: Stein WD, Bronner F, ed. Cell Shape: Determinants, Regulation, and Regulatory Role. San Diego, Calif: Academic Press Inc; 1989:173-202.
Tozzi CA, Poiani GJ, Harangozo AM, Boyd CD, Riley DJ. Pressure-induced connective tissue synthesis in pulmonary artery segments is dependent on intact endothelium. J Clin Invest. 1989;84:1005-1012.
Belik J, Keeley FW, Baldwin F, Rabinovitch M. Pulmonary hypertension and vascular remodeling in fetal sheep. Am J Physiol. 1994;266:H2303-H2309.
Stenmark KR, Aldashev AA, Orton EC, Durmowicz AG, Badesch DB, Parks WC, Mecham RP, Voelkel NF, Reeves JT. Cellular adaptation during chronic neonatal hypoxic pulmonary hypertension. Am J Physiol. 1991;261(suppl):97-104.
Hishikawa K, Nakaki T, Marumo T, Hayachi M, Suzuki H, Kato R, Saruta T. Pressure promotes DNA synthesis in rat cultured vascular smooth muscle cells. J Clin Invest. 1994;93:1975-1980.
Cox RH. Comparison of arterial wall mechanics using ring and cylindrical segments. Am J Physiol. 1983;244:H298-H303.
Baserga R, Malamud D. Autoradiography. New York, NY: Harper & Row; 1969.
Bendeck MP, Langille BL. Rapid accumulation of elastin and collagen in the aortas of sheep in the immediate perinatal period. Circ Res. 1991;69:1165-1169.
Snedecor GW, Cochran WG. Statistical Methods. 8th ed. Ames, Iowa: Iowa State University Press; 1989:282.
Boonen HCM, Schiffers PHM, Fazzi GE, Janssen GMJ, Daemen MJAP, De May JGR. DNA synthesis in isolated arteries: kinetics and structural consequences. Am J Physiol. 1991;260:H210-H217.
Wilson E, Mai Q, Sidhir K, Weiss RH, Ives HE. Mechanical strain induces growth of vascular smooth muscle cells via autocrine action of PDGF. J Cell Biol. 1993;123:741-747.
Stenmark KR, Orton EC, Reeves JT, Voelkel NF, Crouch EC, Parks WC, Mecham RP. Vascular remodeling in neonatal pulmonary hypertension: role of the smooth muscle cell. Chest. 1988;93:127s-133s.
Poiani GJ, Tozzi CA, Yohn SE, Pierce RA, Belsky SA, Berg RA, Yu SY, Deak SB, Riley DJ. Collagen and elastin metabolism in hypertensive pulmonary arteries of rats. Circ Res. 1990;66:968-978.
Stirewalt WS, Coflesky JT, Young LH, Evans JN. Protein synthesis in pulmonary arteries from rats exposed to hyperoxia. Am J Physiol. 1993;264:L74-L79.
Haworth SG. Pulmonary vascular remodeling in neonatal pulmonary hypertension: state of the art. Chest. 1988;93:133s-138s.