Cholesterol Dependence of Vascular ERK1/2 Activation and Growth in Response to Stretch

Animals, Dissection, and Cholesterol Depletion

Female Sprague-Dawley rats were killed by cervical dislocation, as approved by the Animal Ethics Committee, Lund University. Portal veins were dissected under sterile conditions and cut longitudinally into halves. For cholesterol depletion, strips were incubated with 5 mmol/L methyl-β-cyclodextrin (mβcd) or vehicle in Krebs’ solution (in mmol/L: NaCl 135.5, KCl 5.9, CaCl₂ 2.5, MgCl₂ 1.2, HEPES 11.6, and glucose 11.5, pH 7.4) for 1 to 3 hours at 37°C. The 1-hour treatment was used as the standard incubation time, because longer incubations impaired contractility. Cholesterol was reloaded into depleted strips by cholesterol-saturated mβcd (water-soluble cholesterol, Sigma) in Krebs’ solution for 3 hours at 37°C. Before culture, strips were gently blotted and weighed as described.¹⁰

Organ Culture

After cholesterol depletion, strips were transferred to culture medium (Dulbecco’s modified Eagle’s medium and Ham’s F-12 medium [1:1, vol/vol], containing 2% dialyzed fetal calf serum, 10 nmol/L insulin, 50 U/mL penicillin, and 50 μg/mL streptomycin) and either left undistended or loaded with a 0.6-g gold weight, which represents a preload approximately optimal for active force development The passive length-tension relation is quite steep, so that a 10% shortening from the loaded length eliminates most of the passive force.¹⁰ The culture period was 24 hours, after which the strips were again weighed. No effects of mβcd on growth were observed with longer culture (3 days), presumably reflecting recovery from cholesterol depletion.

Mechanical Recording

Strips were mounted,¹⁰ immersed in 0.4 mL Krebs’ solution, and stretched to a passive tension of 2 mN. After equilibration (45 minutes), experimental protocols were started. Contraction was elicited by replacing NaCl with KCl to a final concentration of 140 mmol/L. The cross-sectional area was determined from the length and wet weight.

Protein and DNA Synthesis

For analysis of protein and DNA synthesis rates, [methyl-³H]thymidine or l-[4,5-³H]leucine (Amersham), at activities of 0.2 and 1 μCi/mL, respectively, were added to the culture flasks, and incorporation was determined as described previously.¹⁰

Immunoblotting

Tissue was equilibrated for 1 hour in normal Krebs’ solution with or without mβcd, loaded for 5 or 10 minutes in culture medium, and then snap-frozen. ERK phosphorylation was determined by Western blotting.¹⁰ Caveolin-1 was detected with the 6B6 antibody²⁶ (a gift from Dr J. Vinten, University of Copenhagen, Copenhagen, Denmark), and tyrosine phosphorylation (P-Tyr), with PY 20 (Transduction Laboratories). Bands were visualized by enhanced chemiluminescence (Lumi-Glo, New England Biolabs or Pierce West-pico). Phosphorylated ERK1/2 was normalized to the total amount of ERK1/2.

Tissue Fractionation by Sucrose Gradient Centrifugation

A detergent-free method to purify caveolae was adopted from that of Song et al.²⁷ Tissues were pulverized in LN₂ and suspended in 1 mL of 500 mmol/L Na₂CO₃ with protease inhibitors (5 μg/mL, Sigma), pH 11. The homogenate was adjusted to 45% sucrose with 1 mL of 90% sucrose in MBS (25 mmol/L MES and 0.15 mol/L NaCl, pH 6.5) and overlayered with 4 mL of 35% sucrose and 4 mL of 5% sucrose, both in MBS with 250 mmol/L Na₂CO₃. After centrifugation at 39 000 rpm for 22 hours (SW 41, Beckman Instruments), ten 1-mL fractions were collected. For detection of P-Tyr, protein was precipitated with 20% trichloroacetic acid, washed with acetone, and solubilized in sample buffer. When stretched and control veins were compared, the volumes loaded on the gel were adjusted to give equal protein contents in fraction 5. Protein synthesis rates are stretch dependent, but measurable differences in overall protein composition between loaded and unloaded veins do not appear during organ culture for several days,¹⁶ and hence, increases in P-Tyr during a 10-minute loading period represent increased phosphorylation of preformed proteins and not increased protein contents.

Electron Microscopy

Veins were fixed (1 to 2 hours) in 2% glutaraldehyde and 1% depolymerized paraformaldehyde in phosphate-buffered saline (pH 7.4, 300 mOsm). After being rinsed in phosphate-buffered saline, veins were postfixed in 1% OsO₄ for 60 minutes, rinsed, dehydrated in graded ethanol (70%, 95%, and 100%), and embedded in mounting medium (Durcupan ACM, Fluka). Ultrathin sections were cut, mounted on copper mesh grids, and counterstained with uranyl acetate and lead citrate. Sections were examined in a transmission electron microscope, and photomicrographs were obtained with a CCD camera (Gatan 791 MultiScan). Caveolae, including completely and partially invaginated vesicles (50- to 100-nm diameter) and those with dilated necks found within 100 nm of the plasma membrane, were counted on coded photomicrographs. Membrane length was determined with commercially available software (ImageJ). The photomicrographs used for calculation of caveolar density were obtained randomly and in a similar way for all conditions, but because we concentrated on areas containing either caveolae or their remnants, we cannot exclude a bias toward regions rich in caveolae.

Statistics

Data are presented as mean±SE. Student’s t test was used to test statistical significance. For multiple comparisons, 1-way ANOVA was used. P<0.05 was considered significant.

Stretch Rapidly Increases P-Tyr in Caveolin-1–Containing Membrane Domains

Mechanical stimuli rapidly increase P-Tyr of cellular proteins. To investigate whether stretch increases P-Tyr in caveolae of the intact rat portal vein, a detergent-free method for purification of caveolae by sucrose gradient fractionation was used. Ten minutes of stretch increased P-Tyr of several high-molecular-weight proteins (Figure 1A and 1B,a). P-Tyr increased almost exclusively in the fraction containing the caveolar marker caveolin-1 (Figure 1A). Comparisons (fold differences) of P-Tyr of several protein bands in fraction 5 of stretched versus unstretched veins are shown in Figure 1B,a. Caveolin-1 contents and protein composition in the fractions were the same (Figure 1B,b and Methods). This indicates that the stretch signal involves P-Tyr of several proteins that are located in caveolae.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Stretch increases P-Tyr in caveolin-1–containing membrane domains, and cholesterol depletion collapses smooth muscle caveolae. Caveolae were purified by sucrose density fractionation from control and stretched (10 minutes), nondenuded portal veins. After sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transfer to nitrocellulose, membranes were immunoblotted (IB) with anti–P-Tyr (A, top) and anti–caveolin-1 (A, bottom) antibodies. Loading of the gel was adjusted to give identical amounts of protein in fraction 5. The fold increases in P-Tyr of selected bands in the caveolin-1 fraction, compared with unloaded controls, are shown in B,a. Molecular-weight estimates are indicated below bars. The fold increase for smaller proteins (P43, P65) was determined after longer exposure. Blots in A are representative of 3 independent experiments. Densitometry demonstrated similar contents of caveolin-1 in fraction 5 from loaded (Lo) and unloaded (Unlo) veins (B,b). C,a–C,c show representative smooth muscle membrane areas from a control vein (C,a) and veins treated for 1 (C,b) or 2 (C,c) hours with mβcd (5 mmol/L). Arrows point to caveolae and caveolar remnants. SR indicates sarcoplasmic reticulum. Scale indicated in C,a is identical in C,b and C,c.

Cholesterol Depletion Disrupts Caveolar Structure

Cholesterol was extracted with mβcd (5 mmol/L, 1 to 3 hours),^22,28⇓ and membrane profiles were examined by electron microscopy. Caveolae were found as clusters at the cell ends and also along the longitudinal cell axes (Figure 1C,a). After incubation with mβcd for 1 hour, areas of flattened caveolae with dilated openings were present (Figure 1C,b), but in some regions, caveolae appeared normal. After longer treatment (2 to 3 hours), caveolae were more extensively flattened and dilated, and no caveolae of normal appearance were found (Figure 1C,c). The density of caveolae, irrespective of structure, was 3.8±0.4/μm membrane in controls versus 3.8±0.8, 4.3±0.2, and 1.5±0.3 (P<0.01) in preparations treated for 1, 2, and 3 hours, respectively (data from ≥8 photomicrographs per condition in 2 experiments).

Effect of Cholesterol Depletion on Stretch-Induced Growth and Protein Synthesis

Portal veins were cultured for 24 hours under loaded and unloaded conditions after cholesterol depletion for 1 hour. The weight of loaded, control vessels increased under these conditions,^10,16⇓ whereas unloaded veins lost weight (Figure 2A). Depletion of cellular cholesterol resulted in weight loss of loaded veins, similar to unloaded controls. Cholesterol-depleted, unloaded veins behaved as unloaded controls.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Cholesterol depletion abolishes stretch-induced growth of rat portal vein. Portal vein strips, treated as indicated, were subjected to mechanical loading in culture, and wet weight at the end of culture was recorded and normalized to weight before culture (A). Results are mean±SE of 5 to 13 experiments. In B and C, rat portal veins, treated as indicated, were incubated with [³H]leucine and [³H]thymidine for 24 hours. Protein and DNA synthesis rates were measured by scintillation counting. Results are mean±SE of 5 or 6 experiments. *P<0.05 vs loaded controls, †P<0.05 vs loaded and cholesterol (chol)-depleted preparations. DPM indicates disintegrations per minute.

The disruption of caveolar structure with mβcd can be reversed by incubation with cholesterol-saturated mβcd (1 mmol/L cholesterol:5 mmol/L mβcd).²⁸ Incubation with cholesterol-saturated mβcd without prior depletion did not affect weight gain in either loaded or unloaded veins (see online Figure I; available at http://atvb.ahajournals.org), demonstrating that mβcd itself has no effect on growth besides that caused by the extraction of cholesterol. Incubation with cholesterol-saturated mβcd for 3 hours after depletion (1 hour) partially restored the stretch-induced weight gain during the following culture (Figure 2A).

Protein and DNA syntheses were measured by [³H]leucine and [³H]thymidine incorporation, respectively, during culture (Figure 2B and 2C). Cholesterol depletion reduced protein and DNA syntheses in loaded veins toward the levels of unloaded controls. Synthesis recovered significantly, but not completely, when depleted veins were incubated with cholesterol-saturated mβcd before culture. Cholesterol-saturated mβcd alone did not influence leucine incorporation in loaded and unloaded veins, again excluding the possibility of nonspecific effects of mβcd on growth (see online Figure I; available at http://atvb.ahajournals.org). Neither protein synthesis nor DNA synthesis was affected by cholesterol depletion in unloaded veins. Extraction with mβcd had no effect on force in response to high-K⁺ depolarization in fresh strips (13±2 vs 11±2 mN/mm²; n=10, P=NS) or in veins cultured under load after extraction (8±3 vs 6±1 mN/mm²; n=10, P=NS). However, the amplitude of the regular spontaneous activity characteristic of the portal vein declined after cholesterol depletion (not shown).

Effect of Cholesterol Depletion on Stretch-Induced ERK1/2 Activation

Stretch-induced growth of the portal vein requires ERK1/2 phosphorylation,¹⁰ which is likely to be represented by the P-Tyr labeling in the 43-kDa region of the caveolar fraction in Figure 1A and 1B. Veins were therefore stretched for 5 minutes, and ERK1/2 phosphorylation was determined by Western blotting. We first verified that mβcd did not have any nonspecific effect on ERK1/2 activation by using cholesterol-saturated mβcd (Figure 3A). Cholesterol depletion, on the other hand, inhibited ERK1/2 phosphorylation in stretched veins by ≈70% (Figure 3B). Significant recovery of stretch activation was observed after replenishing the cholesterol in depleted veins (Figure 3B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Stretch-induced ERK1/2 phosphorylation (P-ERK1/2) is inhibited by cholesterol depletion. Phosphorylation of ERK1/2 was determined at the end of a 5-minute period in the absence or presence of load and expressed relative to phosphorylation in loaded, control veins. A demonstrates lack of effect of cholesterol (chol)-saturated mβcd (1 mmol/L cholesterol:5 mmol/L mβcd) on ERK1/2 activation by stretch. B demonstrates inhibition of ERK1/2 phosphorylation in stretched veins by cholesterol depletion. Cholesterol replenishment after prior depletion is indicated by mβcd-chol. Below the bar graphs, representative Western blots of P-ERK1/2 and total ERK1/2 are shown. Results are mean±SE of 3 (A) or 6 (B) experiments. *P<0.05 vs loaded controls, †P<0.vs loaded and cholesterol-depleted preparations.

Effect of Cholesterol Depletion on ERK1/2 Activation by Ang II and ET-1

Part of the stretch-induced ERK1/2 activation in the portal vein depends on endogenous release of Ang II.¹⁶ We therefore investigated the effect of cholesterol depletion on responses to Ang II. Addition of Ang II (10 μmol/L, 5 minutes) increased ERK1/2 phosphorylation in loaded and unloaded control as well as cholesterol-depleted veins. The effects of load and Ang II were additive (Figure 4). Although cholesterol depletion inhibited stretch-induced activation of ERK1/2, the additional response to Ang II in both loaded and unloaded veins was unaffected.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Cholesterol depletion does not impair Ang II-induced ERK1/2 phosphorylation (P-ERK1/2). Rat portal vein strips, treated as indicated, were stretched and/or stimulated with 10 μmol/L Ang II for 5 minutes followed by snap-freezing and Western blotting. Results are mean±SE of 4 experiments. Western blots from a typical experiment of P-ERK1/2 and total ERK1/2 are shown below.

Addition of ET-1 increased ERK1/2 phosphorylation in both stretched and unstretched preparations, whereas the ET-1 type A (ET_A) receptor blocker RF139317 (1 μmol/L) reduced ERK1/2 phosphorylation of both stretched and unstretched preparations below their respective control values (Figure 5). This suggests that endogenous ET-1 production contributes to ERK1/2 activity, as also shown by the finding that the endothelin-converting enzyme inhibitor phosphoramidone (PA, 10 μmol/L) inhibited activation of ERK1/2 by stretch (loaded control, 100%; unloaded, 39±4%; loaded plus PA, 40±4%; unloaded plus PA, 34±2%; n=4 for all, P<0.05 for effect of PA). Although ERK1/2 activation due to endogenous production of Ang II and ET-1 seems to play a role in the stretch response, their dependence on membrane cholesterol differs, because mβcd completely inhibited activation of ERK1/2 by exogenous ET-1 as well as by stretch (Figure 5).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Role of ET-1 receptors in stretch-induced ERK1/2 phosphorylation (P-ERK1/2). Portal vein strips were treated with 10 nmol/L ET-1, the ET_A receptor antagonist RF139317, and mβcd for 5 minutes followed by snap-freezing and Western blotting. Results are mean±SE of 4 to 8 experiments of 2 different designs, both including loaded and unloaded veins, which allowed direct comparisons. Western blots of P-ERK1/2 and total ERK1/2 from typical experiments are shown below the bar graph. *P<0.05 vs loaded controls, †P<0.05 for indicated comparisons.

Role of the Endothelium

In view of the reported role of caveolae as mechanosensors in endothelial cells, the endothelium was removed, and the effect on stretch-induced ERK1/2 activation was examined. Scraping of the luminal vessel surface abolished endothelial nitric oxide synthase labeling, whereas stretch-induced ERK1/2 activation remained unaffected (Figure 6A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Endothelial denudation does not influence stretch-induced ERK1/2 activation, and cholesterol depletion does not displace ERK1/2 from caveolae. Summarized data and Western blots of ERK1/2 phosphorylation (P-ERK1/2), total ERK1/2, and endothelial nitric oxide synthase (eNOS) from a typical experiment are shown in A. Results are mean±SE of 3 independent experiments. *P<0.05 vs loaded controls. B and C show the presence of ERK1/2 in caveolin (Cav-1)-containing fraction after sucrose density fractionation of control and cholesterol-depleted veins, respectively. Blots are representative of 3 independent experiments.

Effect of Cholesterol Extraction on Subcellular Distribution of ERK1/2

With the use of sucrose gradient fractionation, ERK1/2 was found to copurify with the caveolar marker protein caveolin-1 (Figure 6B). Cholesterol depletion did not appear to alter the distribution of either caveolin-1 or ERK1/2 (Figure 6C).