Action of Tropoelastin and Synthetic Elastin Sequences on Vascular Tone and on Free Ca²⁺ Level in Human Vascular Endothelial Cells

Animals

Four- to 6-month-old female Wistar rats were obtained from Iffa-Credo (L’Arbresle, France). All animals were fed with Extralabo rat chow (Pietrement). Twenty animals were used. The housing and surgery procedures were in accordance with institutional guidelines.

Preparation of Aortic Ring Segments and Perfusion Conditions

Rats were anesthetized by intraperitoneal injection of pentobarbital (60 mg/kg). The descending thoracic aorta was excised and placed in an oxygenated (O₂ 95%/CO₂ 5%) Krebs’ bicarbonate buffer containing (mmol/L) NaCl 118, KCl 5.6, CaCl₂ 2.4, MgCl₂ 1.2, KH₂PO₄ 1.2, NaHCO₃ 20, and d-glucose 11 (pH 7.4). The tissue was cleaned of adhering connective tissue and fat and then cut into 1- to 2-mm transverse segments. Aortic rings were placed horizontally between a stationary stainless steel hook and an isometric force transducer (Biologic) connected to a microcomputer. A peristaltic pump continuously delivered the medium (1.5 mL/min, 37°C) to the organ bath (0.9 mL). A passive tension of 1.5 g was applied for 60 minutes and was then readjusted back to 1.5 g for 15 minutes before starting the experiment. The aortic rings were precontracted and perfused until the end of the experiments with KbNA, which induced 80% of maximal tension developed by NA. The endothelium integrity was tested by adding 10⁻⁶ mol/L acetylcholine to KbNA before recontracting the rings with KbNA alone. In some rings, the endothelium had been removed by gentle rubbing of the intimal surface with a section of polyethylene catheter. Selective removal of the endothelium was assessed by determining the absence of vasorelaxant effect of acetylcholine (10⁻⁶ mol/L) but responsiveness to NA (10⁻⁶ mol/L). The rings were then perfused with KbNA alone for 105 minutes (control) or with increasing concentrations of the test substances (10⁻¹³ to 10⁻⁷ mol/L) for 15 minutes per dose in the absence or presence of lactose (10⁻⁴ mol/L). In this last instance, lactose was applied 15 minutes before and during the application of the elastin-related compounds.

Vasomotor Data Analysis

Results are expressed as mean±SEM of aortic ring tension (mg) after a 15-minute application of each agent concentration. Initial tensions were normalized to an average value of NA-induced precontraction (1120 mg). Statistical analysis was performed using a two-way ANOVA. In order to perform comparisons between results in the presence and in the absence of lactose, control ring tensions (with or without lactose) were normalized to 1120 mg, and the same normalization factors were then applied to ring tensions in the presence of the tested substances as described previously.¹⁶ The statistical tests were completed by Fisher’s least significant difference test for paired value comparisons. Differences were considered significant at P<.05.

HUVEC Isolation and Culture

HUVECs were obtained as previously described³³ and were maintained, at 37°C and humid atmosphere, in fibronectin-coated dishes³⁴ with medium 199 supplemented with Earle’s salts and 22% human serum. Trypsinized cells were plated on dishes (containing 13×14-mm glass coverslips) coated with fibronectin so that the cell density was in the range of 10 500 cells/cm.² Only first to fourth passage subconfluent cells were used in the experiments. The purity of the endothelial cell cultures was established using an antibody to von Willebrand factor.

Loading of HUVECs With Fura 2

Fura 2-AM loading was performed generally according to the procedure of Thomas and Delaville.³⁵ The culture dish was washed twice with a Tyrode’s solution containing (mmol/L) NaCl 118, KCl 5.6, CaCl₂ 2.4, MgCl₂ 1.2, HEPES 10, and d-glucose 11 (pH 7.4). The dish was then filled with 0.75 mL Tyrode’s solution. In the dark, 8 μL of 0.5 mmol/L fura 2-AM in DMSO was added to 250 μL of Tyrode’s solution containing 2% BSA and 0.08% pluronic F-127 (wt/vol). This solution (258 μL) was then added to the culture dish (final volume, ≈1 mL), yielding final concentrations of 4 μmol/L fura 2-AM and <1% DMSO. After a 40-minute incubation (room temperature, in the dark), the cells were washed twice and bathed with Tyrode’s solution (1 mL) for 10 minutes.

[Ca²⁺]_i Measurements

Fluorescence measurements were as described previously.^{36 37} Under low light, the HUVEC-containing glass coverslip was removed from the culture dish and fitted vertically and diagonally in a Tyrode’s solution–filled quartz cuvette (10×10 mm) with a plastic ring at the bottom that enclosed a small stir bar that mixed the solution. The coverslip was fitted so that the cells faced the dual excitation light (340 and 380 nm), and the emitted fluorescence (510 nm) was read at 90° from behind the coverslip. The cuvette was then placed at 30°C in a Deltascan 4000 spectrofluorometer (Photon Technology International), and measurements were performed after a minimum of 5 minutes and before 60 minutes. The recorded signal resulted from the integrated fluorescence from all the cells facing the excitation path. [Ca²⁺]_i was monitored for 4 minutes after the addition of elastin-related peptides to the cuvette. For investigating the inhibitory effect of lactose, fluorescence was monitored for 2 minutes in the presence of lactose (10⁻⁴ mol/L) alone, before the addition of the test compounds. Control experiments were performed by recording the HUVEC fluorescence alone for 5 minutes. Each measurement was stopped by the addition of 0.002% digitonin (maximum [Ca²⁺]_i), followed by 45 mmol/L EGTA (minimum [Ca²⁺]_i). After the autofluorescence was subtracted, [Ca²⁺]_i was calculated using standard techniques, assuming that the K_d of fura 2 for Ca²⁺ is 224 nmol/L.³⁸

[Ca²⁺]_i Measurement Data Analysis

For each measurement, initial [Ca²⁺]_i was normalized to an average measured value (50 nmol/L). Statistical analysis was performed using a one-way ANOVA for comparison of simple effects of each elastin-related compound or of lactose. A two-way ANOVA was used to assess statistical analysis of the effect of compounds in the presence or absence of lactose. When necessary, these tests were completed by the nonparametric Mann-Whitney U test for paired value comparisons. Differences were considered significant at P<.05.

Preparation of Synthetic Elastin Peptides and Control Peptide

All peptides were synthesized using 9-fluorenylmethoxycarbonyl chemistry. Purity of the peptides was confirmed by HPLC and by fast bombardment spectrometry. Purity and concentrations of the peptide stock solutions were verified by amino acid analysis.

Preparation of rTE

rTE encoded by a 2210-bp Escherichia coli–optimized synthetic gene was expressed and purified from bacterial extracts as described.³⁹ rTE molecular mass is 63490 kD.

Preparation of κ-Elastin

Elastin peptides (κ-elastin) were obtained by hydrolysis of highly purified bovine ligamentum nuchae elastin with 1 mol/L KOH in 80% (vol/vol) aqueous ethanol.⁴⁰ By adjusting the time of hydrolysis, peptides of different molecular mass can be obtained. Peptides used in these experiments had an average molecular mass of 75 kD.

Drugs and Materials

The following drugs and reagents were from Sigma: Krebs’ bicarbonate and HEPES-buffered solution components, NA, acetylcholine, lactose, fura 2-AM, DMSO, BSA, pluronic F-127, collagenase 1A, medium 199 with Earle’s salts, l-glutamine, and HEPES. Human serum was from the Centre de Transfusion Sanguine de Grenoble. Falcon cell culture dishes were from Becton-Dickinson. The 13×14-mm glass coverslips were from Vitromed.

Vasorelaxant Activity of Elastin-Related Compounds

The vasoactivity of elastin-related peptides was found to be compound and concentration dependent on aortic rings with intact endothelium (Table 1⇓, category A). Compared with the control, the peptides VGV, PGV, VGVAPG, and rTE induced dose-dependent vasorelaxation in the NA-precontracted rat aortic rings with maximum values of 32%, 21%, 26%, and 41%, respectively (Fig 1A⇓). Nevertheless, for each concentration for which several peptides were found to be active, no significant difference in activity could be detected between these peptides, except at 10⁻⁹ mol/L, where rTE was more active than VGV (Fig 1A⇓). Moreover, the peptide PGV was slightly active (14% vasorelaxation) at 10⁻¹² mol/L (Fig 1A⇓). The sequences VGVA and PVGV induced a relaxation of the aortic rings only at a concentration of 10⁻¹¹ mol/L, with values of 14% and 19%, respectively, and no activity difference could be detected between these two peptides (Fig 1B⇓). Finally, the peptides PGVGVA, VGVGVA (for which the induced aortic ring tensions presented a slightly higher variance), and GVGVA as well as the control peptide VPVGGA produced tension variations of 12% maximum compared with the control, with all values not statistically significant (P≥0.1) (Fig 1C⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Concentration-response curves showing the effects of elastin-related compounds after a 15-minute exposure on rat aortic ring tone. The rings were precontracted and were then perfused for 105 minutes with 10⁻⁶ mol/L NA (control). In some rings, increasing doses of elastin peptides were also added to the perfusion. The peptides added were as follows: rTE, VGV, VGVAPG, and PGV (A); PVGV and VGVA (B); and VPVGGA, GVGVA, VGVGVA, and PGVGVA (C). Each point represents the mean of five to nine experiments. The results are expressed as tension (mg) (mean±SEM). *P<.05 vs control tension by least significant difference test. #P<.05 vs corresponding VGV-induced tension by least significant difference test. †P<.065 (least significant difference test) and P<.05 (U test) vs corresponding control tension.

By contrast, on aortic rings without endothelium, the 10 compounds (rTE, VGV, PGV, VGVA, PVGV, GVGVA, VGVAPG, VGVGVA, PGVGVA, and VPVGGA) had no effect on the tone of NA-precontracted aortic rings for the two tested concentration ranges: 10⁻¹¹ mol/L and 10⁻⁸ to 10⁻⁷ mol/L, the active concentrations determined above (Table 2⇓). All the tested compound-induced tension values were not different from the corresponding control tension values (Table 1⇑, category B).

The effect of the elastin/laminin receptor antagonist “lactose” was then tested on the endothelium-dependent vaso-relaxant action of the active elastin-related sequences (VGV, PGV, VGVA, PVGV, VGVAPG, and rTE) for concentrations demonstrated to induce a submaximal vasorelaxation. VGV, VGVAPG, and rTE were used at 10⁻⁸ mol/L. Since only one concentration was found to be active in the range of the maximal effect for PGV (10⁻⁷ mol/L), VGVA, and PVGV (10⁻¹¹ mol/L), this concentration was used for these compounds. The direct effect of lactose (10⁻⁴ mol/L) was first studied on the NA-precontracted rat aortic rings with endothelium. Lactose (10⁻⁴ mol/L)–induced tension (mg) was identical to the control tension after 15 minutes but was lower than the control tension after 30 minutes and lower than the control tension and the initial lactose-induced tension after 45 minutes. At these times, the slight vasorelaxation induced by lactose was in the range of 5% compared with the control value (Fig 2⇓ and Table 1⇑, category C). After assessment of a second normalization of the tension values to balance the effect of lactose itself, the effect of lactose on the vasorelaxant effect induced by the active compounds (rTE, PGV, VGV, PVGV, VGVA, and VGVAPG) was investigated. As shown in Tables 1⇑ (category D) and 3⇓, the presence of lactose had a significant inhibitory effect on the elastin peptide–induced vasorelaxation, and the inhibition was independent of the active peptide used. These results indicate that lactose strongly inhibited the vaso-relaxant action of all the active elastin-related compounds.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Action of lactose (Lac, 10⁻⁴ mol/L) on aortic ring tone. The rings were precontracted with NA (10⁻⁶ mol/L), and then the tension was recorded in the presence or absence of Lac for 45 minutes. Each point represents the mean of 9 (absence of Lac) to 17 (presence of Lac) experiments. The results are expressed as tension (mg) (mean±SEM). *P<.05 vs corresponding control tension without Lac by least significant difference test. †P<.05 vs Lac-induced tension after 15 minutes by least significant difference test.

Action of Elastin-Related Compounds on HUVEC [Ca²⁺]_i

To investigate the intracellular mechanisms explaining the vasorelaxant action of rTE and certain elastin-related sequences, we studied the effect of these compounds on [Ca²⁺]_i in adherent HUVECs loaded with the fluorescent dye fura 2. We first verified that the κ-elastin–induced [Ca²⁺]_i increase previously demonstrated on pooled suspended HUVECs loaded with indo 1 or on individual adherent HUVECs loaded with fluo 3 (authors’ unpublished data, 1997) was also occurring in the fura 2–loaded adherent HUVECs stimulated with 10⁻³ mg/mL (≈10⁻⁸ mol/L) κ-elastin (Fig 3A⇓). In these conditions, the κ-elastin–induced [Ca²⁺]_i increase exhibited slow kinetics and an amplitude (average raw increase, 2.06-fold, or 1.73 times the control value) similar to those previously observed in pooled suspended HUVECs loaded with indo 1 (average raw increase, 2-fold). The control peptide VPVGGA (10⁻⁷ mol/L) had no effect on HUVEC [Ca²⁺]_i compared with cells with no peptide added (Fig 3B⇓ and 3C⇓ and Table 4⇓). The effects of the elastin-related compounds on HUVEC [Ca²⁺]_i were then investigated at concentrations in the range of 10⁻⁸ to 10⁻⁷ mol/L for the peptides found to have no activity on the vascular tone and at active concentrations for the peptides found to induce a vasorelaxation. Only rTE (10⁻⁹ to 10⁻⁷ mol/L), VGV (10⁻⁸ to 10⁻⁷ mol/L), PGV (10⁻⁸ to 10⁻⁷ mol/L), VGVA (10⁻¹¹ mol/L), PVGV (10⁻¹¹ mol/L), and VGVAPG (10⁻⁸ to 10⁻⁷ mol/L), ie, all the peptides found active on the vascular tone, induced an increase in HUVEC [Ca²⁺]_i compared with the control cells (Fig 4A⇓ and 4C⇓ and Tables 1⇑ [category E] and 4⇓). Nevertheless, no statistically significant difference (U test) between the [Ca²⁺]_i increases induced by the different active compounds could be detected. The compounds found to have no activity on the vascular tone (VPVGGA, PGVGVA, VGVGVA, and GVGVA) also exhibited no significant activity on HUVEC [Ca²⁺]_i (Fig 3B⇓ and Tables 1⇑ [category E] and 4⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Action of 10⁻³ mg/mL κ-elastin (A) and the control peptide VPVGGA at 10⁻⁷ mol/L (B) on [Ca²⁺]_i in cultured HUVECs compared with the control experiment (C). Each trace is representative of six, seven, and three experiments, respectively.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

A and B, Action of the elastin sequence VGV at 10⁻⁷ mol/L on [Ca²⁺]_i in HUVECs in the absence (A) or presence (B) of 10⁻⁴ mol/L lactose. Each trace is representative of five and three experiments, respectively. C and D, Action of 10⁻⁹ mol/L rTE on [Ca²⁺]_i in HUVECs in the absence (C) or presence (D) of 10⁻⁴ mol/L lactose. Each trace is representative of five experiments.

Furthermore, compared with the control condition, application of lactose (10⁻⁴ mol/L) to HUVECs was not found to induce a significant change in [Ca²⁺]_i over a 2-minute period (Table 1⇑, category F). However, the addition of 10⁻⁴ mol/L lactose inhibited the [Ca²⁺]_i increase induced by the active elastin-related peptides (Fig 4B⇑ and 4D⇑ and Table 4⇑). The presence of lactose had a significant inhibitory effect for all the active elastin-related compounds tested, since the inhibition was independent of the active peptide type (Table 1⇑, category G).