INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

VASCULAR CELLS in vivo are primarily subjected to two hemodynamic forces: shear stress from flowing blood and cyclical strain from pulsations in pressure with each cardiac cycle. During the cardiac cycle, large arteries can stretch by 9-10%, which can be increased to 15% with hypertension (15, 20, 45, 46). Many vascular adaptations to hypertension, such as matrix remodeling and vascular smooth muscle (VSM) hypertrophy, can be prevented if the arterial bed is protected from a rise in pressure (5). This suggests that VSM cells are capable of sensing and responding to increased stretch. However, the cellular mechanisms underlying the mechanotransduction of stretch are incompletely understood.

Extracellular matrix (ECM) receptors, when coupling cells to the matrix, are plausible transducers of stretch. These receptors are transmembrane proteins, capable of linking the ECM to the actin cytoskeleton; thus deformation of the cytoskeleton is transmitted intracellularly via these linkages. The integrin family of receptors has been studied extensively as mechanotransducers of stretch in VSM (23, 44, 51). Many integrins, e.g., _v1, _v5, _v3, ₅₁, and _IIb3, are capable of binding ECM proteins through a conserved arginine-glycine-aspartic acid (RGD) motif (40). Other matrix proteins, such as laminin, can bind to integrins in VSM (₃₁) through non-RGD motifs (9, 40). The binding of integrins to the matrix proteins initiates coupling to focal adhesion contacts, which trigger cellular signaling pathways (23, 24, 31). Although the integrins are clearly involved in myriad cellular functions, including acculturation to stretch, other transmembrane receptors may play pivotal roles as well. We make this point because two major constituents of the ECM are elastin and laminin. Although laminin can bind to VSM cells via the ₃₁-integrin (9), elastin lacks known amino acid sequences necessary for binding to integrins. However, elastin and laminin can adhere to cells by a distinct transmembrane protein, the elastin-laminin receptor (ELR), which recognizes hydrophobic motifs in the proteins (16). This receptor is distinct from the integrin family and shows sequence homology to the alternatively spliced variant of human -galactosidase (19).

Because of the promiscuity of elastin in the vessel wall, we postulate that the ELR may mechanotransduce stretch in VSM. The ELR is a heterotrimeric transmembrane receptor that binds specific hydrophobic sequences in elastin and laminin and induces various signal transduction pathways (16, 22, 35, 43). Within the vessel wall, the ELR is located on both endothelial and VSM cells where it has been shown to participate in endothelial control of vascular tone (11, 12). It seems probable that the ELR, in addition to the previously described roles, can function as a mechanotransducer in VSM cells. Indeed, previous work in our laboratory (27) and others (51) has shown that gene expression and proliferation in stretched VSM is different in cells grown on collagen vs. elastin. Whereas several laboratories have focused on the role of integrins in mediating these adaptations, to date none have examined a role for the ELR.

Under basal conditions, the expression of the early response gene, c-fos, is very low in many cell types but is quickly and transiently induced by multiple extracellular signals (6, 30, 39, 48). The c-Fos protein acts as a transcription factor at the activator protein-1 site where it controls expression of many target genes. These characteristics provided the foundation for our hypothesis that the expression of this gene would be sensitive to mechanical stretch and, based on our previous studies (27), likely to be modulated in a matrix-dependent fashion. Likewise, cellular proliferation, as a biological output of altered gene expression and signaling cascades, would also be affected by stretch in a substrate-dependent manner. Accordingly, we further hypothesized that the responses to stretch on elastin matrixes would signal through the ELR and could therefore be abrogated by blockade of this receptor.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture. Adult VSM was isolated from canine coronary arteries cleaned of fat, adventitia, and endothelium. With the use of an explant technique, freshly isolated left anterior descending arteries were placed luminal side down in 35-mm cell culture dishes containing DMEM supplemented with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 200 IU/ml penicillin, and 200 µg/ml streptomycin (GIBCO-BRL, Rockville, MD). To decrease variability seen with phenotypic drift characteristic of smooth muscle cells cultured for extended periods of time, we used cells from early passages (passage 2 to passage 9). Cells were maintained in a humidified atmosphere at 37°C in the presence of 5% CO₂ until reaching confluence. Immunocytochemistry for smooth muscle -actin and calponin (Sigma, St. Louis, MO) was used to verify smooth muscle cells, and the purity of the cultures was absolute. Cells were seeded at 2.5 × 10⁴ cells/well (for proliferation studies) or 1 × 10⁵ cells/well (for gene expression studies) in six-well silicone elastomer-bottomed Flexercell dishes (Flexercell International, McKeesport, PA) coated with type I collagen or elastin. Control dishes were coated with identical silicone elastomer and matrix on rigid plates. Once adherent, the cells were growth arrested by serum withdrawal (the FBS is decreased from 10 to 0.1%). We maintained the cells at 0.1% FBS during growth arrest rather than totally withdrawing serum because in our experience total withdrawal produces some apoptotic cell death, whereas in 0.1% FBS the population is stable. The low serum media was replaced daily for 3 days to render the cells quiescent.

Cellular stretch. The stress unit is a modification of the Flexercell Strain Unit described in detail by Banes et al. (3). It consists of a computer-controlled vacuum and base plate to hold the culture dishes. Intermittent application (1 Hz) of 10-15 kPa of pressure using a solenoid-controlled vacuum to the underside of the plate resulted in deformation of the plate. When the vacuum was released, the membrane returned to the unstretched position. The computer controlled the frequency of deformation and the negative pressure applied to the culture dishes. Cycling parameters were chosen for comparability with previous work (25, 27, 36, 37) and result in cyclical stretch of ~10% at the periphery.

Cellular proliferation. Proliferation was assessed by direct determination of the cell numbers. Trypsinized cells were plated (2.5 × 10⁴ cells/well) on type I collagen- or elastin-coated Flexercell dishes with rigid or flexible bottoms. Cells were growth arrested in 0.1% FBS for 3 days, with media replacement each day. After growth arrest, 10% serum was added, and cells were stretched for 48 h. At the end of the stretching period, cells were trypsinized and counted using a hemacytometer. To confirm complete trypsinization, cells were visualized before counting. Cellular viability was assessed by trypan blue exclusion. Comparisons were made to the number of cells seeded at the beginning of the experiment and reported as a percent increase in cell number.

RNA isolation/Northern blot analysis. RNA was isolated from cells using TRIzol reagent (GIBCO-BRL) according to the manufacturer's directions. Total RNA was quantified using a spectrophotometer, and 20 µg were loaded on a denaturing 1.2% agarose gel with formaldehyde. The RNA was electrophoresed at 4 V/cm and transferred to a positively charged nylon membrane (Boehringer Mannheim, Indianapolis, IN) by capillary blotting. The membrane was cross-linked with ultraviolet light and prehybridized using PerfectHybe reagent (Sigma) according to the manufacturer's directions. cDNA probes from murine c-fos and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; 2.2 and 1.4 kb, respectively) were radiolabeled with [³²P]dCTP (Amersham Pharmacia, Piscataway, NJ) and hybridized overnight (5 × 10⁶ counts · min¹ · ml¹) at 68°C. The blot was washed and placed on Kodak BioMax MS film with a BioMax intensifying screen at 80°C for 4-16 h. The film was developed and scanned, and the relative band intensity of each signal was quantified using image-processing software (Image 1.28c; National Institutes of Health Research Services Branch, Bethesda, MD). Densitometric units were normalized to the ethidium bromide-stained 28S ribosomal subunit or GAPDH signal. There is continued debate in the literature as to the best housekeeping gene for normalization of molecular data (49, 54). In this study, the early time points (15 and 30 min) were studied using 28S, whereas the latter time points (60 and 120 min) were studied using GAPDH. To help discern whether there was a difference in the data sets based on which gene was used as the housekeeping gene, we repeated some of the early time points with GAPDH. We found no significant difference in the percent change from static control between the data sets normalized to either gene. This suggests that stretching in the presence of serum does not uniformly alter the expression of c-fos and GAPDH, genes both transcribed by RNA polymerase II.

Protocols. Cells were stretched for 48 h to assess changes in proliferation or for 15, 30, 60, or 120 min to assess changes in c-fos expression. The ELR antagonist hexapeptide valine-glycine-valine-alanine-proline-glycine (VGVAPG; Sigma) and the biologically inactive control peptide valine-serine-leucine-serine-proline-glycine (VSLSPG; custom synthesis, Medical College of Wisconsin Protein & Nucleic Acid Facility) were solubilized in DMSO (final concentration 0.02% vol/vol). Both peptides (10⁵ M) were added to cells in low serum media for a 2-h preincubation period before the initiation of stretch. Once the preincubation was completed, serum was added, and the cells were stretched. Previous studies have revealed that the optimal concentration of VGVAPG to block elastin-ELR binding was 10⁵ M, which elicited biological blockade of the receptor without causing cells to lift from the culture dish. The control peptide, VSLSPG, was found to be ineffective at eluting the ELR from elastin affinity columns, suggesting that a sequence-specific interaction beyond similar charge, hydrophobicity, and secondary structure is necessary for elastin-ELR interactions (29). In pilot experiments, the basal expression of c-fos was undetectable under growth arrest conditions, and we could not discern any effects of stretch. However, we found that, when stretch and 10% serum were applied simultaneously after 3 days of growth arrest, we could adequately detect and modulate the c-fos message.

Statistics. Dimensionless quantities (relative densitometric units) of the c-fos and GAPDH signals from multiple similar experiments were calculated as a percent change from unstretched control under each condition. Combined data were expressed as means ± SE. For comparisons of two groups, we used the Student's unpaired t-test (two tail). We used ANOVA followed by Scheffé's multiple comparison test for comparisons of three or more groups. Significance was set at P  0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Stretch-mediated changes in c-fos expression are matrix dependent. When cells were grown on type I collagen, the serum-induced expression of c-fos (normalized for differences in loading by GAPDH or 28S rRNA) did not differ between stretched cells and the unstretched control at any time point studied (Fig. 1). Under both conditions, c-fos expression increased with time to a peak value (30 min) and then decreased to undetectable values (120 min), which is characteristic of c-fos expression kinetics. However, when cells were grown on elastin matrix, the expression of c-fos was significantly decreased at 15 and 30 min of stretch (17.1 ± 5.0% and 31.1 ± 6.9%, respectively). At 60 min, expression was not significantly different between stretched and static groups, and by 120 min c-fos was undetectable (Fig. 2).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   c-fos expression in cyclically stretched vascular smooth muscle (VSM) cells grown on a collagen matrix. A: Northern blot analyses from total RNA in stretched (S) or unstretched (US) cultures for 15-120 min. B: normalized [c-fos/28S or c-fos/glyceraldehyde-3-phosphate dehydrogenase (GAPDH)] expression as percent change from unstretched at each time point studied. Data are means ± SE. No significant differences from unstretched group or over time; n, no. of experiments.

View larger version (25K): [in this window] [in a new window]   Fig. 2.   c-fos expression in cyclically stretched VSM cells grown on an elastin matrix. A: Northern blot analyses from total RNA in stretched or unstretched cultures for 15-120 min. B: normalized (c-fos/28S or c-fos/GAPDH) expression as percent change from unstretched at each time point studied. Data are means ± SE. *P  0.05 compared with unstretched group.

Stretch-mediated changes in c-fos expression on elastin are dependent on the ELR. To examine the role of the ELR in mediating changes in gene expression, cells were treated with 10⁵ M VGVAPG for 2 h before stretch and during the stretching period. Our results show that, in the presence of ELR blockade, c-fos expression was not different from the unstretched control (10.83 ± 10.8% at 15 min and 1.01 ± 4.2% at 30 min), suggesting paramount importance for this receptor in mediating stretch-induced reductions in c-fos expression on elastin matrixes (Fig. 3).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   c-fos expression in cyclically stretched VSM cells grown on an elastin matrix in the presence of the elastin-laminin receptor antagonist valine-glycine-valine-alanine-proline-glycine (VGVAPG; 105 M) or under control (no antagonist) conditions. A: Northern blot analyses from total RNA in stretched or unstretched VSM cultures for 15-30 min. B: normalized (c-fos/28S) expression as percent change from unstretched group at each time point studied. Data are means ± SE for 5 independent experiments. *P  0.05 compared with unstretched group.

Proliferation of stretched VSM cells is matrix dependent. The increase in cell number, as measured by percent change from numbers of cells plated, was used to determine biological adaptations to stretch. On elastin under static conditions, serum induces robust proliferation (1,189 ± 82% increase in cell number), but this serum effect was reduced by ~50% (621 ± 96%) when cells were stretched on elastin (Fig. 4A). In contrast, there was not a significant difference in serum-induced proliferation between stretched and unstretched cells on collagen (1,073 ± 111% vs. 1,051 ± 88%, Fig. 4B).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   A: proliferation of cultured VSM cells stretched on elastin matrixes for 48 h after growth arrest. Proliferation rates after 3 days of growth arrest are denoted baseline. VGVAPG cultures were stretched in the presence of 105 M elastin-laminin receptor antagonist. Data are means ± SE for 5 independent experiments. *P  0.05 compared with static group. B: proliferation of cultured VSM cells stretched on collagen matrixes for 48 h after growth arrest. Proliferation rates after 3 days of growth arrest are denoted baseline. VSLSPG, valine-serine-leucine-serine-proline-glycine. Data are means ± SE for 5 independent experiments.

Decreased proliferation with stretch on elastin matrix is mediated by the ELR. To assess the role of the ELR in mediating stretch-dependent decreases in proliferation on elastin matrix, we antagonized the normal elastin receptor interaction with the competitive elastin hexapeptide VGVAPG. Treatment with the ELR inhibitor restored the proliferative capacity of VSM cells stretched on elastin, suggesting a prominent role for this receptor in signaling stretch to VSM cells grown on elastin matrixes. The control hydrophobic hexapeptide VSLSPG did not attenuate the decreased proliferation with stretch on elastin matrixes (Fig. 4A).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

These data provide the first direct evidence that the ELR functions as a mechanotransducer of stretch in VSM. The major findings of our study are 1) modulated expression of c-fos with stretch is matrix dependent, 2) decreases in c-fos expression with stretch on elastin matrix are dependent on ELR signaling, 3) stretch-mediated proliferation is matrix dependent, and 4) decreased proliferation with stretch on elastin matrix is mediated by the ELR. To place our findings and conclusions in perspective, our discussion will focus on aspects of the methodology and cogent literature relating to heterogeneity of VSM and mechanotransduction.

Methodological considerations. Mechanical stress is a prominent feature of the VSM cell environment in vivo. Physiological stretch of the vessel wall may play an important role in both embryonic development of the cardiovascular system and maintenance of vascular homeostasis. Increased levels of stretch may act as inductive or progressive stimuli for vascular pathologies. With these processes in mind, understanding the cellular adaptations to cyclical stretch are of paramount importance.

Heterogeneity of VSM. We have previously reported that stretch-mediated gene expression is differentially regulated in VSM from different arterial beds (27), which may parallel functional properties of those cells in vivo. Likewise, it has been shown that neonatal smooth muscle cells behave differently than adult cells, even within the same vascular bed (13, 42, 50). There are conflicting reports about whether proliferation of VSM cells is decreased (10, 47) or increased with stretch (8, 26, 50). Some of these disparate experimental findings may be explained by VSM cell origin (splanchnic mesoderm vs. neuroectoderm), species (human vs. rat), vascular site (vein vs. conduit artery vs. resistance artery), or location (tunica intima vs. tunica media). Supporting this theory were previous results that showed proliferative responses of VSM to stretch were dependent on the artery from which the cells were derived (26, 52). Clearly, more studies are needed to determine how physical forces influence cellular behavior in different vascular domains.

Receptors as mechanotransducers of stretch. Ingber and Jamieson (21) proposed the tensegrity paradigm to account for force transmission across cellular membranes, which leads to adaptive responses. This paradigm emphasizes integrin receptors linking ECM domains to intracellular cytoskeletal elements but does not consider the possibility of ELR signaling. Several studies have illustrated the importance of matrix composition on cellular responses to stretch. Wilson et al. (50) have shown that late-passage (passage 16 to passage 29) neonatal rat VSM cell grown on vitronectin or fibronectin proliferate with stretch via autocrine platelet-derived growth factor production. The stretch-induced proliferation was not present when cells were grown on laminin matrixes, suggesting specific cell-matrix interactions may differentially signal the transduction of stretch (50). In another study, Reusch et al. (37) showed that expression of smooth muscle myosin heavy chains (SM-1 and SM-2) was induced in neonatal VSM cells stretched on laminin but not pronectin (poly-RGD). These investigators also found that stretched neonatal VSM cells activated different mitogen-activated protein kinase pathways, depending on the matrix. Interestingly, the result was not seen in adult cells, suggesting a functional difference in cells derived from animals of different ages (36). Our study extends these previous observations by showing that stretch-induced, matrix-specific gene expression and cellular proliferation occur on elastin in addition to collagen matrixes. Importantly, we found that a control hexapeptide (VSLSPG) with a similar sequence, charge, hydrophobicity, and secondary structure as VGVAPG did not restore stretch-mediated proliferation and gene expression patterns seen in the unstretched controls. This minimizes the possibility that the ELR binds nonspecifically to any hydrophobic sequence.

Pathological implications. The ELR may have a role in the pathogenesis of hypertension and atherosclerosis. Two hallmark features of atherosclerosis are migration of smooth muscle cells through the internal elastic laminae of the vessel wall and neointimal thickening (38). It is interesting to speculate that degradation of elastin and the subsequent loss of elastin/ELR interactions may be a prerequisite for the formation of the lesion. Although vascular lesions are rich in collagen, they contain negligible amounts of elastin; thus VSM may be able to proliferate in lesions because the normal static influence of elastin is absent. Likewise, ductus arteriosus smooth muscle cells, which are similar to cells from atherosclerotic plaques, lack functional ELR (17, 18). Moreover, the elastase inhibitor elafin prevents VSM cell migration and proliferation after viral myocarditis (53). The hypertrophic response of coronary arterioles in hypertension includes interstitial fibrosis and perivascular fibrosis and increased collagen biosynthesis by vascular myocytes (4, 41). This complex process changes the vascular ECM composition, which strengthens the vessel. As with atherosclerotic processes, this alteration is likely to corrupt normal signaling pathways via increased collagen signaling and decreased elastin signaling. Collectively, these data suggest that disruption of the normal matrix-receptor interaction could disrupt homeostatic signal transduction pathways and may contribute to SMC migration and proliferation. If, on the other hand, normal elastin-elastin receptor interactions occur, VSM cells may be prevented from initiating migratory and proliferative processes common in vascular diseases such as atherosclerosis and hypertension.