Endothelial Cells Provide Feedback Control for Vascular Remodeling Through a Mechanosensitive Autocrine TGF-β Signaling Pathway

Abstract

Mechanical forces are potent modulators of the growth and hypertrophy of vascular cells. We examined the molecular mechanisms through which mechanical force and hypertension modulate endothelial cell regulation of vascular homeostasis. Exposure to mechanical strain increased the paracrine inhibition of vascular smooth muscle cells (vSMCs) by endothelial cells. Mechanical strain stimulated the production of perlecan and heparan sulfate glycosaminoglycans by endothelial cells. By inhibiting the expression of perlecan with an antisense vector we demonstrated that perlecan was essential to the strain-mediated effects on endothelial cell growth control. Mechanical regulation of perlecan expression in endothelial cells was governed by a mechanotransduction pathway requiring autocrine TGF-β signaling and intracellular signaling through the ERK pathway. Immunohistochemical staining of the aortae of spontaneously hypertensive rats demonstrated strong correlations between endothelial TGF-β, phosphorylated signaling intermediates, and arterial thickening. Further, studies on ex-vivo arteries exposed to varying levels of pressure demonstrated that ERK and TGF-β signaling were required for pressure-induced upregulation of endothelial HSPG. Our findings suggest a novel feedback control mechanism in which net arterial remodeling to hemodynamic forces is controlled by a dynamic interplay between growth stimulatory signals from vSMCs and growth inhibitory signals from endothelial cells.

The endothelium is a dynamic constituent of the vascular system that exerts remarkable control over diverse process such as hemostasis, inflammation and the regulation of vascular tone^{1, 2}. In addition to interacting with blood constituents and circulating cells, the vascular endothelium is exposed to a distinct mechanical environment consisting of hemodynamic shear stress and mechanical stretch from blood pressure. As a result, endothelial cells are optimally situated as both sensors and effectors in the vascular remodeling process. An extensive body of scientific work supports the concept that vascular cells respond to their mechanical environment^{3, 4}. In vitro, shear stress and stretch stimuli regulate endothelial cell production of regulatory molecules including nitric oxide⁵, reactive oxygen species⁶, inflammatory cell adhesion molecules⁷, and extracellular matrix molecules⁸. Similarly, vascular smooth muscle cells (vSMCs) respond to mechanical stimuli by altering their expression of cell adhesion and extracellular matrix molecules⁹ as well as increasing proliferation and migration^{10, 11}.

An essential function of the intact endothelium is the paracrine regulation of the growth and phenotype of underlying vSMCs. Soluble heparan sulfate proteoglycans (HSPGs) are intimately involved in endothelial inhibition of vSMC proliferation^12-14. These complex molecules are composed of a core protein covalently coupled to one or more heparan sulfate glycosaminoglycan chains. The heparan sulfate chains consist of a linear polymer of alternating disaccharide units heterogenenously modified by epimerization, deacetylation, and sulfation¹⁵ which attain an intricate molecular structure whose information capacity far exceeds that of an equivalent mass of nucleic acid polymers¹⁶. Here, we show that HSPGs are a key component in an integrated feedback control loop regulating vascular remodeling through the modulation of paracrine endothelial inhibition of vSMC growth. Further, we define an endothelial mechanotransduction pathway that is dependent on strain-induced autocrine TGF-β signaling in combination with the activation of the ERK signaling pathways, and leads to alterations in expression of growth inhibitory HSPGs.

Methods

Cell Culture

Human umbilical vascular endothelial cells (HUVECs) and vascular smooth muscle cells (vSMCs) were purchased from Cambrex, Walkersville, MD. All vSMCs were used between passage 4 and 5 and all endothelial cultures were used at passage 3 to 5. All cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. Stable cell lines of transfected bovine endothelial cells were as previously described^{14, 17}.

Application of Mechanical Strain to Cultured Cells

The device to apply mechanical strain to cells was as described previously ¹⁸. Briefly, 0.005″ thick silastic membranes (Specialty Manufacturing, Inc., Saginaw, MI) were sterilized and coated with collagen I. Endothelial cells were grown on the collagen-coated silastic membranes and cyclic mechanical strain was applied with maximal strain of 3% or 5% at a frequency of 1 Hz. All loading experiments were performed on cells that were at least 2 days post-confluence.

Smooth Muscle Cell Proliferation Assay

Rat or human smooth muscle cells were passaged into six-well plates at low density. Endothelial cell conditioned media was isolated and used in an assay of vSMC proliferation. Smooth muscle cells were serum starved in 0.5% calf serum for 24 hours, washed with PBS, and incubated in conditioned media with 1 μCi/ml ³H-thymidine for 24 hours. The cells were then washed three times with PBS at 4°C. The cells were then incubated with 10% TCA for 30 min at 4°C, washed twice in 95% ethanol, and solubilized in 1 ml of 0.25 M NaOH with 0.1% SDS for 1 hour. The samples were then added to scintillation cocktail and radioactivity was measured using a liquid scintillation counter.

Cell Lysis and Western Blotting

Cells were lysed in 20 mM Tris with 150 mM NaCl and protease inhibitors. For perlecan, the samples were digested with heparitinase III and chondroitinase ABC (1 U/ml; Associates of Cape Cod, East Falmouth, MA) for 4 hours prior to western blotting. Digested or non-digested samples were run on 4-15% polyacrylamide gradient gels and transferred to PVDF membranes. The membranes were blocked for 1 hr in 5% non-fat milk in PBS with 0.01% tween-20 (PBST) and exposed to the following antibodies at 4°C overnight in 1% non-fat milk: anti-phospho-Smad 2 (Ser465/467), anti-Smad 2, anti-phospho-p38 MAPK (Thr180/Tyr182), anti-p38 MAPK, anti-phospho-ERK (Thr202/Tyr204), anti-ERK (Cell Signaling, Danvers, MA), mouse anti-perlecan (Invitrogen) or rat anti-perlecan (Lab Vision, Fremont, CA). The membranes were washed with PBST, incubated at room temperature for 2 hrs with a secondary antibody and detected using chemiluminescence (Perkin Elmer, Boston, MA).

Metabolic Labeling and Isolation of Proteoglycans

At the time of mechanical loading, 100 μCi of ³⁵SO₄ and 80 μCi of ³H-glucosamine was added to each plate. After loading, the conditioned media was collected and combined with guanidine-HCl to a final concentration of 4 M. The cell layers were then washed three times with cold PBS then PBS with 1.0 mM EDTA was added and the cells were allowed to detach. The cells were spun down and washed with PBS twice and then resuspended in 1 ml of 0.05% trypsin-EDTA solution (Invitrogen, Carlsbad, CA) and placed on ice for 10 min. After the trypsin digestion, the samples were centrifuged and 1 ml of DMEM with 5% calf serum was added to neutralize the trypsin. This solution was then brought to a final concentration of 4 M guanidine-HCl. Proteoglycans were extracted from the extracellular matrix for 48 hrs with a solution of 50 mM sodium acetate (pH = 6.0), 4 M guanidine-HCl, 2% triton x-100 and protease inhibitors.

ELISA Measurement of TGF-β1

The concentration of TGF-β1 was measured using ELISA (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions.

Animal Model of Hypertension

All experimental procedures and protocols used in this investigation were reviewed and approved by the Animal Care and Use Committee of the Massachusetts Institute of Technology and conformed to the “Guiding Priniciples in the Care and Use of Animals” of the American Physiological Society and the NIH Guide for the Care and Use of Laboratory Animals. Age matched Wild-Type Wistar-Kyoto Rats and Spontaneously Hypertensive (SHR; NTac:SHR) rats were obtained from Taconic (Germantown, NY). At 20 weeks of age the animals were sacrificed and the aortae were harvested.

Immunohistochemistry

The abdominal aorta from hypertensive rats was formalin fixed and sectioned using standard methods. For staining of HSPGs the slides were treated for 2 hrs with a 0.4 U/ml solution of heparitinase III (Ibex, Canada). Primary antibodies were diluted in PBS containing 1% BSA, applied to slides, and incubated in a humid chamber overnight at 4°C. A monoclonal antibody recognizing “stubs” of digested heparan sulfate (3G10; Seikagaku) was used. Secondary antibody staining at detection was performed using the LSAB 2 kit (DakoCytomation, Carpinteria, CA) according to the manufacturer's instructions. An AEC substrate (DakoCytomation) was used for detection of the HRP conjugate. The samples were counterstained in Mayer's hematoxylin for 3 min, washed with tap water, and mounted in aqueous mounting medium (DakoCytomation). Quantitative analysis was performed using Adobe Photoshop.

Ex-Vivo Perfusion System

Thoracic aortae were harvested from rats and rinsed several times in modified Krebs-Henseleit buffer with 1% penicillin-streptomycin (Sigma). The aortae were mounted in physiological orientation between two cannulas within the perfusion system (Figure 1). The flow system maintained the artery in a modified Krebs-Henseleit buffer with sodium biocarbonate and 5% dextran to maintain physiologic viscosity. Fluid flow was driven by a peristaltic pump (Cole Palmer, Vernon Hills, IL) and set to produce a physiologic shear stress of 15 dyn/cm² on the aortic wall. Pressure was controlled by changing the height of the fluid column following the aortic segment. High pressure was set at 200 mmHg and normal pressure at 100 mmHg, as measured using a pressure sensor (Harvard Apparatus, South Natick, MA). At the end of the experiment the aortic segments were frozen in isopentane cooled with liquid nitrogen.

Figure 1.

Diagram of an ex-vivo system for culturing aortic segments from rats. The system utilizes a peristaltic pump to apply pulsatile flow and pressure to an ex-vivo aortic segment. The pressure is set through changing the height of upper reservoir into which the fluid is pumped. A mixture of 5% CO₂ and 95% O₂ was bubbled into the main reservoir to maintain oxygenation and pH. A water bath was used to maintain the temperature of the buffer at 37°C.

Statistics

Data are presented as mean ± SEM. A Student t-test was used to analyze for differences between two groups. Multiple comparisons between groups were analyzed by one-way ANOVA followed by Student-Newman–Keuls post hoc test. A two-tailed p value <0.05 was considered statistical significant.

Results

Cyclic Mechanical Strain Increases Endothelial Inhibition of vSMC Proliferation through Increased Perlecan Production

One major functional role of endothelial cells is the production of soluble factors controlling vSMC growth. A fundamental question addressed by this work is how mechanical strain alters paracrine inhibition of vSMCs by endothelial cells. Endothelial cells were grown on silastic membranes for two days postconfluence and then exposed to uniform, cyclic mechanical strain (3% or 5% maximal strain, 1 Hz) for 1-24 hrs. Conditioned media was harvested from endothelial cells under non-strained or strained conditions and applied to low density cultures of vSMCs (Figure 2a). In these experiments the 0% strain samples are endothelial cell conditioned media (i.e. media exposed to endothelial cells without strain for the period of the experiment). Thus, the difference in DNA synthesis between control media and 0% strain samples represents the baseline inhibitory capacity of confluent endothelial cells. Brief periods of mechanical strain induced slightly stimulatory media (Figure 2b). This stimulatory effect decreased with duration of exposure to become inhibitory after 8 hours. After 24 hrs of loading, mechanical strain induced a 2.3 fold increase in the inhibitory properties of the endothelial conditioned media (19.5±0.06% versus 45.6±12.7% of no strain samples; Figure 2b). Overall, conditioned media from strained endothelial cells inhibited vSMC proliferation about 80% greater than control growth media. We confirmed this effect in another endothelial cell type and found that this effect increased with the magnitude of the load (Figure 2c).

Figure 2.

Prolonged cyclic mechanical strain causes an increase in endothelial inhibition of vascular smooth muscle cell growth through a perlecan-mediated pathway. (a) Endothelial cells were exposed to various regimes of mechanical loading and the conditioned media was assayed for growth inhibition of vSMCs. Control samples are normal growth media and 0% strain samples are endothelial cell conditioned media. (b) After short periods of exposure to cyclic mechanical strain, human umbilical vein endothelial cells (HUVECs) do not show increased inhibition of vSMC growth. Twenty four hours of cyclic mechanical strain causes HUVECs to produce conditioned medium with two-fold greater inhibitory properties towards vSMC cell growth. *p < 0.05 versus control samples; **p < 0.05 versus no strain samples. (c) Endothelial cells have increased inhibition due to prolonged mechanical strain. This effect increases with magnitude of the cyclic strain applied (all points shown are after 24 hrs). Control media is growth media that has not been exposed to endothelial cells. *p < 0.05 versus control samples; **p < 0.05 versus all other samples. (d) Western blot analysis showed an increase in perlecan protein levels in the conditioned media of mechanically loaded endothelial cells. *p < 0.05 versus no strain samples. (e) Diagram of perlecan antisense construct. Stable cell lines expressing either a perlecan antisense vector or an empty expression vector were exposed to 24 hrs of load and the conditioned media was assayed for smooth muscle cell growth inhibition. (f) Perlecan antisense (Perl-AS) reduces perlecan in endothelial cell conditioned media. (g) A stable cell line of bovine endothelial cells expressing a perlecan antisense construct does not produce conditioned media that is inhibitory towards vSMCs and does not have an induced increase of inhibitory properties by exposure to 24 hrs of cyclic mechanical strain. A similar effect was achieved by depleting the conditioned media of perlecan by affinity chromatography. Black bars = control cell transfected with pcDNA3. Grey bars = perlecan antisense transfected cells. White bars = media depleted of perlecan by affinity chromatography. *p < 0.05 versus control samples; **p < 0.05 versus all other samples.

Previous studies have suggested that perlecan can control vSMC growth^{13, 14}. We examined the perlecan production in the conditioned media of strained and non-strained endothelial cultures. Western blotting of the conditioned media revealed increased soluble perlecan core protein after 24 hours of loading (Figure 2d). To establish a mechanistic link between strain-induced changes in perlecan expression and inhibition of vSMCs, we applied 5% cyclic mechanical strain for 24 hours to stably transfected endothelial cell lines with either an expression vector (pcDNA3) or a vector expressing an antisense construct targeted to perlecan (Figure 2e). Transfection of this construct into endothelial cells led to a 10-fold reduction in perlecan in the conditioned media of the cells (Figure 2f). The increase in paracrine inhibition of vSMCs by endothelial cells with mechanical load was retained in endothelial cells transfected with the empty vector and eliminated in endothelial cells expressing the perlecan antisense construct (Figure 2g). To further confirm the specificity of our results we depleted endothelial conditioned media of perlecan using an antibody-based affinity column. These results also showed a reduction in the inhibition of vSMC growth after perlecan depletion (Figure 2g).

Mechanical Strain Increases Endothelial Cell Production of Soluble Heparan Sulfate Proteoglycans

The cellular production of proteoglycans can be regulated on multiple levels including direct regulation of the core protein and through alterations in the assembly the glycosaminoglycan sugar chains. We metabolically labeled the glycosaminoglycan chains during exposure to mechanical strain and found an increase of 24% in total soluble glycosaminoglycan production by HUVECs (Figure 3a and 3b). A 38% increase in soluble heparan sulfate was also observed. With strain, no change in cell surface total glycosaminoglycan was observed, but a decrease in cell-associated heparan sulfate was found. Total glycosaminoglycans and heparan sulfate glycosaminoglycans in the extracellular matrix were increased 85% and 20%, respectively, for mechanically strained versus non-strained cultures.

Figure 3.

Mechanical strain increased extracellular glycosaminoglycan production in endothelial cells. Proteoglycans were isolated from endothelial cells metabolically labeled with ³H-glucosamine and ³⁵SO₄ and exposed to cyclic mechanical strain of 5% strain amplitude at 1 Hz for 24 hrs. Cell surface proteoglycans were isolated by mild trypsin digestion and heparan sulfate proteoglycans (HSPGs) remained after chondroitinase ABC digestion. Matrix bound proteoglycans were isolated by extraction with 4M guanidine HCl for 48 hrs. Proteoglycans were separated by ion exchange chromatography using a Q-ion column and a linear NaCl concentration gradient. Values shown are for liquid scintillation readings of ³H-glucosamine incorporation only. (a) Samples treated with 5% strain (dashed line) and no strain (solid line) are shown with the applied salt gradient (dotted line) on the second axis. (b) Average glycosaminoglycan and HSPG amounts for endothelial cells exposed to mechanical strain and non-strained controls. *p < 0.05 versus non-loaded samples.

Mechanical Strain Controls Perlecan Expression Through ERK and p38 MAPK-Dependent Autocrine TGF-β Production

Transforming growth factor-β (TGF-β) is intimately involved in regulating cell growth and production of extracellular matrix¹⁹. We examined the amount of TGF-β produced by endothelial cells under mechanical strain in the presence of inhibitors to the ERK and p38 MAPK signaling pathways. In endothelial cells, TGF-β production was increased by exposure to mechanical strain (Figure 4a). This effect was significantly blocked by both inhibitors to MEK1/2 (U0126) and p38 MAPK (SB029063). We also hypothesized that since mechanical strain induced TGF-β production, an autocrine signaling pathway through this growth factor may be an important control mechanism. To examine the intracellular signaling pathways responsible for mechanical strain-mediated modulation of HSPGs, we pretreated cells with inhibitors to the ERK and p38 MAPK signaling pathways. Following inhibitor or neutralization antibody treatment we used Western blotting to measure the levels of phosphorylated signaling intermediates, including Smad 2 (a downstream effector of TGF-β signaling), p38 MAPK, and ERK1/2 (Figure 4b). Mechanical strain led to activation of these intermediates after 24 hours. Interestingly, there was crosstalk between the signaling pathways. Maximal Smad 2 phosphorylation only occurred in the absence of both inhibitors to the ERK and p38 MAPK signaling cascades. Further, phosphorylation of ERK1/2 and p38 MAPK was partially blocked by inhibitors to each of the other pathways. We performed a similar Western blotting analysis in mechanically stimulated cultures pretreated with a neutralizing antibody to TGF-β These studies demonstrated that TGF-β was required for optimal phosphorylation of the intracellular signaling pathways (Figure 4c).

Figure 4.

Regulation of perlecan requires autocrine TGF-β signaling, p38 MAPK and ERK signaling pathways. (a) Mechanical strain induces increased TGF-β production by human endothelial cells. Results shown are an ELISA assay after subtracting TGF-β in growth media. This effect is blocked by inhibitors of the ERK signaling pathway (U0126) and p38 MAPK signaling pathway (SB 293063). *p < 0.05 versus all other groups; **p < 0.05 significantly different from all groups except unloaded samples with SB293062; other comparisons not shown for simplicity. (b) The p38 and ERK signaling pathways are activated by 24 hrs of mechanical strain. Endothelial cells were treated with inhibitors one hour prior to exposure to 24 hrs of mechanical strain. Western blotting analysis revealed increased phosphorylation of smad 2, ERK1/2 and p38 MAPK in the absence of inhibitors. (c) Autocrine TGF-β signaling is needed for maximal phosphorylation of p38, ERK1/2 and Smad 2. Endothelial cells were treated with a neutralizing antibody to TGF-β for 1 hr prior to 24 hrs of cyclic mechanical strain. Western blotting analysis of intracellular signaling intermediates showed reduced activation of Smad 2, ERK1/2, and p38 by mechanical strain in the presence of a neutralizing antibody to TGF-β. (d) Inhibitors to p38 or MEK block load-mediated increases in soluble perlecan. (e) A neutralizing antibody to TGF-β blocks load-induced increase in soluble perlecan. (f) Addition of 100 pg/ml TGF-β increased perlecan expression even in the absence of mechanical strain. An inhibitor to TGF-β receptor I signaling, 200 nM of [3-(Pyridin-2-yl)-4-(4-quinonyl)] -1H-pyrazole, blocked perlecan upregulation by mechanical load.

We also examined the role of these pathways in the production of the heparan sulfate proteoglycan core proteins in response to mechanical strain. Inhibitors to p38 MAPK, MEK1/2, and the neutralizing antibody to TGF-β attenuated the load-induced increase in perlecan production (Figure 4d and 4e). To confirm the involvement of the TGF-β signaling pathway, we added TGF-β added prior to mechanical strain and found that TGF-β increased perlecan expression independent of mechanical load (Figure 4f). Further, we blocked TGF-β signaling using a TGF-β receptor I inhibitor and found no increase in perlecan expression in the presence of this inhibitor with mechanical strain (Figure 4f). Measurements of cellular cytotoxicity were made under the various treatment conditions using an LDH release assay indicating no cellular toxicity under long term treatment with these inhibitors and mechanical strain (data not shown).

Spontaneously Hypertensive Rats have Increased Endothelial HSPGs, TGF-β and ERK Signaling

To further extend our work to in-vivo models of altered hemodynamics, we examined the association between vascular remodeling, heparan sulfate and intracellular signaling pathway intermediates in an animal model of hypertension (Figure 5). Histological staining and morphological analysis revealed increased medial thickening in spontaneously hypertensive rats (SHR) in comparison to wild type (WKY) rats. In SHR and WKY animals the tunica media and endothelial layer stained heavily for HSPGs. The specificity of staining for HSPGs was confirmed by lack of staining in arterial sections not digested with heparitinase but stained with primary and secondary antibodies (data not shown). The endothelial cell HSPG staining increased by 58% in SHR rats compared to WKY controls. We also found increases in endothelial TGF-β, phospho-Smad 2, phospho-ERK, and phospho-p38. Interestingly, TGF-β and phospho-Smad 2 were reduced in the medial layers of the aorta. We then examined the association between arterial remodeling (media thickening) and the various signaling pathways (Table 1). This analysis revealed strong correlations between medial thickness, p38, ERK, and TGF-β in the endothelium.

Figure 5.

Immunohistochemical analysis of the aorta from wild-type (WKY) and spontaneously hypertensive rats (SHR) revealed increased endothelial heparan sulfate proteoglycan (HSPG) and endothelial staining for TGF-β, phospho-ERK, phospho-p38, and phospho-Smad 2. Rats were aged to 20 weeks and the aortae harvested and processed for paraffin sections (Bar = 100 μm). Staining for HSPGs was performed by deparaffinizing the sections and digesting for 2 hrs with 48 mU/ml heparitinase III. An antibody was used that recognizes the heparan sulfate “stub” following digestion with heparitinase (3G10). *p < 0.05 versus wild-type arteries (n = 8).

Table 1. Correlation Analysis of Immunohistochemical Staining of the Endothelium.

	Medial Thickness	Phospho-ERK	Phospho-p38
Phospho-ERK	0.729*	1	0.817*
Phospho-p38	0.890**	0.817*	1
TGF-β	0.902**	0.740*	0.793*

^*p < 0.05,

^**p < 0.005

Pressure-Induced Changes in Endothelial HSPGs Require Autocrine TGF-β and ERK Signaling

To further verify that these pathways were acting in an intact arterial system we created a culture system allowing the application of pulsatile pressure and flow to ex-vivo rat aortae. As the maintenance of endothelial integrity was an important factor in our experiments, we confirmed endothelial integrity by immunostaining for PECAM in the arteries after 24 hours of culture (Figure 6a). In this system we demonstrated that 24 hrs of high pulsatile pressure (mean pressure of 200 mmHg) induced increased expression of endothelial HSPG, TGF-β, phospho-Smad 2 and phospho-ERK in comparison to arteries exposed to low pulatile pressure (mean pressure of 100 mmHg; Figure 6b and 6c). Inhibitors to TGF-β and MEK1/2 (U0126) blocked high pressure induced expression of endothelial HSPG. Interestingly, the inhibitors of the ERK pathway also inhibited TGF-β production and phospho-Smad 2 upregulation. Similarly, the neutralizing antibody to TGF-β blocked the phosphorylation of ERK. In contrast to our in-vitro studies on HUVECs and in hypertensive animals, the p38 pathway did not appear to be activated in ex-vivo arteries by 24 hrs of load (data not shown). In addition, inhibitors to this pathway did not inhibit load induced HSPG expression or activation of the TGF-β and ERK pathways. Together these results are consistent with and support a model of autocrine TGF-β signaling linked to ERK signaling as the control mechanism for pressure induced modulation of endothelial HSPGs.

Figure 6.

Measurement of arterial cell signaling in an ex-vivo arterial culture system. (a) Endothelial integrity is maintained in the ex-vivo culture system. Immunofluorescence staining for PECAM-1 is shown in red and staining for nuclei in blue (Bar = 200 μm). (b) Immunohistochemical analysis of ex-vivo aortae treated with inhibitors and exposed to pulsatile pressure and flow for 24 hrs. (c) In the absence of inhibitors high pressure led to an increase in endothelial HSPG, TGF-β, phospho-Smad 2 and phospho-ERK. Upregulation of endothelial HSPG, TGF-β, phospho-Smad 2 and phospho-ERK by high pressure was blocked by a neutralizing antibody to TGF-β or U0126. In this system, phospho-p38 was not significantly upregulated by high pressure and inhibition of p38 activity (SB293063) did not affect the upregulation of endothelial HSPG, TGF-β, phospho-Smad 2 and phospho-ERK by high pressure. Bar = 100 μm. *p < 0.05 versus normal pressure, control arteries.

Discussion

Mechanical forces have long been known to be potent regulators of vascular endothelial function³. Endothelial cells have evolved sophisticated sensory and regulatory ability to maintain vascular homeostasis through adaptive remodeling²⁰. This study addresses the question of how endothelial cells respond to mechanical strain to control the growth of the underlying vSMCs. Previously, it was known that endothelial cells can regulate vSMC proliferation²¹. In particular, heparin and endothelial cell HSPGs are potent inhibitors of vSMC proliferation and FGF-2 induced mitogenesis^{13, 22-24}. This regulation is growth state dependent, with subconfluent cultures of endothelial cells stimulating vSMC growth and postconfluent cultures inhibiting vSMC growth^{12, 25-28}. Similarly, perlecan and endothelial derived HSPGs have been shown to be essential in inhibiting the neointimal response to vascular injury^{14, 29-31}. Our study adds a new dimension to these results, demonstrating that the regulation of perlecan by mechanical strain is an important mechanism in altering endothelial paracrine inhibition of vSMC proliferation in response to changes in mechanical environment. Further, we have also demonstrated a mechanotransduction pathway in endothelial cells controlling their production of perlecan in response to mechanical forces.

Our in vitro studies with cultured endothelial cells indicate that optimal stimulation of perlecan production requires both ERK and p38 MAPK signaling as well as autocrine TGF-β signalingIn turn, p38 MAPK signaling is dependent upon ERK and autocrine TGF-β. TGF-β production and subsequent stimulation of phospho-Smad 2 is partially blocked by inhibitors of both of these pathways. We therefore propose an expanded model of endothelial cell control of vascular homeostasis in which ERK and p38 signaling are important for activation of TGF-β and TGF-β drives the continued stimulation of ERK and p38 pathways. Consistent with this model, inhibitors of these intracellular signaling pathways or TGF-β signaling significantly reduced mechanical strain-stimulated increases in perlecan production. This paradigm is also supported by the correlations found between TGF-β and ERK/p38 in the endothelium of the SHR rats. Further correlations between members of a TGF-β autocrine signaling loop and vascular wall thickness are consistent with a shared mechanistic pathway.

Ex-vivo studies demonstrated that pressure-induced upregulation in HSPG expression in the endothelium was also blocked by inhibitors to both ERK and TGF-β. Interestingly, in our ex-vivo system the p38 pathway did not appear to be regulated by high pressure and was not required for load-induced increases in HSPG expression. Several factors could account for the relative insensitivity of the p38 pathway to pressure and its non-essential role in mechanical regulation of HSPG expression. Acute strain ex-vivo and wall strain in chronic hypertensive rat models may differentially affect p38, the drugs used to inhibit p38 may be less stable in the ex-vivo setting or may act on vSMCs as well as endothelial cells in the excised artery. In-vitro, in-vivo and ex-vivo studies were consistent in supporting the central nature of TGF-β autocrine signaling in combination with ERK signaling to control HSPG production and ultimately modulation of vSMC proliferation.

In the context of the intact arterial system, the finding that endothelial cells produce more inhibitory factors to vSMCs with mechanical load is in contrast with the behavior of vSMC under similar mechanical strain conditions. Mechanical load induces vSMC proliferation^{10, 11}, release of FGF-2³², production of PDGF³³ and elaboration of cell surface associated HSPG^{34, 35}. Consequently, we propose that under mechanical stimulation the endothelium provides negative feedback control inhibiting the pro-growth signals produced by vSMCs under mechanical strain. In the absence of an intact endothelium, as might arise following vascular injury or disease, mechanical forces would induce vSMCs to produce growth stimulatory factors that are unchecked by load-enhanced endothelial cell paracrine growth inhibition. Once the endothelium is restored, additional means of responding to hemodynamic forces can be applied including regulation of vascular tone and inhibition of vSMC proliferation. In this context, our findings contribute to the understanding of the complex coregulation of these two cell types under hemodynamic stimuli and during aberrant vascular remodeling in the face of persistent injury imposed by chronic hypertension and indwelling endovascular devices.