Mechanical stretch enhances IL-8 production in pulmonary microvascular endothelial cells

Abstract

In patients with acute respiratory distress syndrome, mechanical over-distension of the lung by a large tidal volume causes further damage and inflammation, called ventilator-induced lung injury (VILI), however, it is unclear how mechanical stretch affects the cellular functions or morphology in human pulmonary microvascular endothelial cells (HPMVECs). IL-8 has been proposed to play an important role in the progression of VILI by activating neutrophils. We demonstrated that HPMVECs exposed to cyclic uni-axial stretch produce IL-8 protein with p38 activation in strain- and time-dependent manners. The IL-8 synthesis was not regulated by other signal transduction pathways such as ERK1/2, JNK, or stretch-activated Ca²⁺ channels. Moreover, cyclic stretch enhanced IL-6 and monocyte chemoattractant protein-1 production and reoriented cell perpendicularly to the stretch axis accompanied by actin polymerization. Taken together, IL-8 production by HPMVECs due to excessive mechanical stretch may activate neutrophilic inflammation, which leads to VILI.

Introduction

Acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) are characterized by acute respiratory failure that results from pulmonary edema and inflammation. The lung in mechanically ventilated patients with ALI/ARDS may be exposed to excessive cyclic mechanical strain, which causes further lung damage called ventilator-induced lung injury (VILI) [1], [2], [3]. Damages to the alveolar-capillary barrier, alveolar edema, and inflammation due to excessive mechanical stretch have been recognized as major pathophysiologic events of VILI [2], [4], [5], [6]. Pulmonary microvascular endothelial cells (ECs) have been considered one of the primary targets of stretch during mechanical ventilation because they form the alveolar-capillary barrier with alveolar epithelial cells [7], [8]. Thus, understanding how mechanical stretch alters functions and morphology of pulmonary microvascular ECs may lead to novel treatment strategies for VILI.

Neutrophil recruitment to the airspace, which occurs mostly in pulmonary microvascular ECs, plays a critical role in the inflammatory response of ARDS [9]. Because IL-8 is the major chemoattractant CXC chemokine for neutrophils [10], [11], [12], release of IL-8 is considered to play an important role in the inflammatory response and progression of VILI [13]. Although alveolar epithelial cells and macrophages are known to be cellular sources of IL-8 in the lung [14], [15], the mechanism of the increase in IL-8 during mechanical ventilation is not well understood. Moreover, it is still unclear whether mechanical stretch induces cytokine/chemokine production in pulmonary microvascular ECs.

This study was designed to investigate the effects of cyclic stretch on the production of cytokines/chemokines, specifically IL-8, and morphological changes in human pulmonary microvascular ECs (HPMVECs). We demonstrated that mechanical stretch enhanced IL-8 production via p38 activation in HPMVECs. Furthermore, we found that cyclic stretch enhanced production of IL-6 and monocyte chemoattractant protein-1 (MCP-1), both of which are related to the pathogenesis of VILI [5], [16], and induced cell reorientation with actin polymerization.

Materials and methods

Reagents. SB203580, U0126, SP600125, and BTP-2 were obtained from Calbiochem (La Jolla, CA). GdCl₃ was from Wako (Osaka, Japan).

Cell culture. Primary cultures of HPMVECs were obtained from Cambrex (Walkersville, MD). The cells were maintained in culture medium containing 5% fetal bovine serum (FBS), human recombinant EGF (1 ng/ml), insulin (10 μg/ml), human recombinant FGF (2 ng/ml), gentamicin (50 μg/ml), and amphotericin B (50 ng/ml) (EGM-2MV; Cambrex) in an atmosphere of 5% CO₂ and 95% air at 37 °C as described previously [17].

Application of uni-axial cyclic stretch. HPMVECs were removed from the culture dish at the 4–7th passage with 0.01% EDTA-0.02% trypsin and transferred to a 4 cm² silicon chamber (2 cm long, 2 cm wide, and 1 cm deep) coated with 50 μg/ml human fibronectin (BD Biosciences, Bedford, MA) at a density of 2.0 × 10⁵  cells/cm². A uni-axial sinusoidal stretch of either 5% or 20% strain at 50 cycle/min was applied using a stretching apparatus driven by a computer-controlled stepping motor (ST-140; Strex, Osaka, Japan) in an atmosphere of 5% CO₂ and 95% air at 37 °C as described previously [18], [19], [20]. Twelve hours prior to stretching, cells were brought to a quiescent state by incubation in medium containing 1% FBS. HPMVECs incubated in a static condition on the silicone chamber were used as a time-matched control.

Measurement of cytokine/chemokine concentrations. The cell culture medium was collected 0–24 h after stretch application. IL-1β, IL-6, IL-8, and MCP-1 concentrations in the cell culture medium were quantified using commercially available enzyme-linked immunosorbent assay (ELISA) kits (Quantikine; R&D, Minneapolis, MN).

RNA isolation and real-time quantitative PCR. Total cellular RNA was extracted using RNeasy Mini Kit (Quiagen, Hilden, Germany). RNA was reverse transcribed to cDNA using a Superscript III kit (Invitrogen, Carlsbad, CA). TaqMan Gene Expression Assays for IL-8 (cat# Hs99999034_m1) and GAPDH (Hs99999905_m1) genes were purchased from Applied Biosystems (Foster City, CA, USA). Quantitative PCR was performed on 7300 Real-Time PCR system (Applied Biosystems) by using 3-stage program parameters provided by the manufacturer as follows: 2 min at 50 °C, 10 min at 95 °C, and then 40 cycles of 15 s at 95 °C and 1 min at 60 °C. The absolute value for IL-8 mRNA was normalized by the corresponding amount of GAPDH mRNA. The fold change of IL-8 mRNA was expressed as 2^−ΔΔCt, where ΔCt is the difference in threshold cycles for the IL-8 gene and GAPDH, ΔΔCt is the difference of ΔCt between stretched and non-stretched control.

Western blot analysis. Protein concentrations of cellular lysates were measured by using a Bio-Rad protein assay reagent kit (Bio-Rad, Hercules, CA). Equal amounts of lysates, adjusted for protein concentration, were resolved by SDS–PAGE using a 4–20% linear gradient running gel (Daiichi Pure Chemicals, Tokyo, Japan). Proteins were transferred to nitrocellulose membranes, and the membranes were incubated at room temperature in PBS for 1 h. Immunoblotting was performed using antibodies against phospho-p38 (anti-ACTIVE p38 pAb; Promega, Madison, WI), and p38 (p38N20; Santa Cruz, Santa Cruz, CA). Immunodetection was accomplished using a donkey anti-rabbit secondary antibody and an Enhanced Chemiluminescence kit (Amersham Biosciences, Piscataway, NJ). The intensity was quantified by using Scion image software (Scion Corp., Frederick, MD).

Immunofluorescent staining. HPMVECs subjected to cyclic stretch or static conditions in the silicone chamber were fixed with 4% formaldehyde in phosphate-buffered saline for 60 min at room temperature, and permeabilized with 0.25% Triton X-100 containing 0.5% bovine serum albumin for 60 min. F-actin and nuclei were stained with Alexa488-Phalloidin (Molecular Probes, Eugene, OR) and the DNA binding dye 4,6-diamino-2-phenylindole (DAPI; Dojin, Kumamoto, Japan) for 60 min at room temperature. The immunofluorescently stained cells were then visualized by fluorescence microscopy and an imaging-system (Biozero BZ-8000; Keyence, Osaka, Japan) [17].

Statistical analysis. All data are expressed as means ± SD. Analysis of variance followed by the Bonferroni test for post hoc analysis or paired t-test was used to evaluate the statistical significance. P  < 0.05 was considered statistically significant.

Results

Effects of cyclic stretch on cytokine production in pulmonary microvascular endothelial cells

HPMVECs were exposed to cyclic stretch (20% strain) for 12 h. Concentrations of IL-6, IL-8, and MCP-1 in the cell culture medium assessed by ELISA were significantly increased by 20% cyclic stretch (P  < 0.05) (Fig. 1, Fig. 2 ). IL-1β production was not observed in either the static or stretched condition (Fig. 1).

Fig. 1.

Effects of cyclic stretch (20% strain, 12 h) on IL-1β, IL-6, and MCP-1 production by HPMVECs (n = 6). Concentrations of IL-1β, IL-6, and MCP-1 in the cell culture medium were assessed by ELISA. Bars represent means ± SD. ∗Significantly different from the values of the time-matched control condition (P < 0.05).

Fig. 2.

(A) Strain amplitude-dependent effects of 12 h cyclic stretch (5% and 20% strain) on IL-8 production by HPMVECs (n = 5). IL-8 concentrations of cell culture medium were assessed by ELISA. (B) Time-dependent IL-8 protein production stimulated with (closed circles, stretch) or without (open circles, control) 20% cyclic stretch (n = 5). (C) Time-dependent effects of 20% cyclic stretch on the fold change in IL-8 mRNA expression over the non-stretched control (time zero) normalized by the reference gene GAPDH as assessed by real-time quantitative PCR are shown (n = 3). Bars represent means ± SD. ∗Significantly different from the values of the time-matched control condition (P < 0.05).

Time- and strain-dependent effects of cyclic stretch on IL-8 synthesis

When HPMVECs were exposed to 5% or 20% cyclic stretch for 12 h, the IL-8 concentrations were significantly increased in a strain-dependent manner (P  < 0.05) (Fig. 2A). The time-dependent (0–24 h) increases of IL-8 concentrations due to 20% stretch were significantly higher than those in time-matched control cultures (P  < 0.001) (Fig. 2B). We next determined the fold changes in IL-8 mRNA expression over the non-stretched control normalized by GAPDH mRNA using real-time quantitative PCR. The IL-8 mRNA expression was enhanced by application of 20% stretch in a time-dependent manner (Fig. 2C). In contrast, the mRNA expression was not increased in time-matched control cultures (data not shown).

Involvement of p38 activation in cyclic stretch-induced IL-8 production

To determine whether MAP-kinase pathways are involved in the cyclic stretch-evoked IL-8 production, the effects of the p38 inhibitor SB203580, ERK1/2 inhibitor U0126, and JNK inhibitor SP600125 were examined. One inhibitor was applied to the cell culture medium 30 min prior to the application of cyclic stretch. SB203580 (10 μM) significantly inhibited the IL-8 production induced by cyclic stretch (20%, 12 h) (P  < 0.01, Fig. 3 A). In contrast, neither U0126 (10 μM) nor SP600125 (10 μM) inhibited the stretch-induced increase in the IL-8 concentration (Fig. 3A). Next, the effect of cyclic stretch on p38 activation was assessed by Western blot. Application of 20% cyclic stretch to the cells transiently increased phospho-p38, an active form of p38, 5–10 min after application (Fig. 3B). Activities of p38 expressed as a phospho-p38/total-p38 ratio were significantly higher in the 5-min stretched cells than in the control cells (P  < 0.05, Fig. 3C).

Fig. 3.

(A) Effects of the p38 inhibitor SB203580 (10 μM), ERK1/2 inhibitor U0126 (10 μM), and JNK inhibitor SP600125 (10 μM) on the stretch-induced IL-8 production in HPMVECs (n = 4). (B) Time-dependent effects of cyclic stretch on p38 activation were assessed by Western blotting of its phosphorylated form (phospho-p38). (C) p38 activities expressed as the phospho (p)-p38/total (t)-p38 ratio of non-stretched (control) are compared with that of 5-min stretched cells (n = 3). (D) Effects of EGTA (500 μM), BTP-2 (10 μM), and Gd³⁺ (10 μM) on stretch-induced IL-8 production in HPMVECs (n = 4). The cells were stimulated by 20% cyclic stretch for 12 h with or without an inhibitor. Bars represent means ± SD. ∗Significantly different (P < 0.05) from the value with cyclic stretch (Str).

No involvement of Ca²⁺ influx pathways in stretch-induced IL-8 synthesis

To determine whether the stretch-induced IL-8 synthesis depends on the intracellular Ca²⁺ concentration, the effects of the Ca²⁺ chelator EGTA and Ca²⁺ channel inhibitors were examined. Pretreatment of the cells with EGTA (500 μM) did not inhibit the IL-8 production induced by cyclic stretch (20%, 12 h) (Fig. 3D). Similarly, Gd³⁺ (10 μM), an inhibitor of stretch-activated channels, or BTP-2 (10 μM), an inhibitor of both transient receptor potential family Ca²⁺ channels and store-operated channels, did not inhibit the stretch-induced IL-8 production (Fig. 3D).

Induction of endothelial cell reorientation by cyclic stretch

We examined the effects of cyclic stretch on the morphology of HPMVECs. The cells stained with Alexa488-phalloidin are shown in Fig. 4 . The cells subjected to 20% cyclic stretch for 12 h were elongated and their long axes were aligned perpendicularly to the stretch axis (Fig. 4B). In contrast, longitudinal axes of non-stretched cells were distributed randomly (Fig. 4A). Cyclic stretch induced F-actin polymerization and stress fiber formation along the axis of cell elongation (Fig. 4B). Treatment of the cells with 10 μM SB203580 did not affect the cell morphology (Fig. 4C).

Fig. 4.

Effects of cyclic stretch (20% strain, 12 h) on cell reorientation are shown. F-actin and nuclei of the cells were stained with Alexa488-phalloidin and DAPI. The cells were either in the static condition (A, control) or were exposed to cyclic stretch (B). (C) Effect of SB203580 (10 μM) on stretch-induced cell reorientation is shown.

Discussion

In general, the cellular properties of ECs are dynamically affected by physical forces such as mechanical stretch, pressure, and shear stress both in physiological and diseased conditions [18], [21], [22]. Okada et al. [23] demonstrated that cyclic stretch up-regulates IL-8 and MCP-1 productions in human umbilical vein ECs. In the microvasculature of the lung, cyclic stretch is the predominant mechanical stress during mechanical ventilation. Nevertheless, how alveoli and microvasculature react to distension during ventilation in vivo has not been fully understood because the distribution of pulmonary edema and inflammation is spatially and temporally heterogeneous in lungs with ARDS [2], [5]. When lung volume is increased by high tidal ventilation, alveoli and pulmonary ECs are expected to be stretched by 20% or more [24]. Birukov et al. [25] elongated human pulmonary artery ECs 5% and 18% and found that the larger mechanical stretch reduced the endothelial barrier function. Thus, the 5% and 20% stretches employed in our study are consistent with low and high tidal ventilation in these previous studies.

It was reported that 17–18.5% elongation of HPMVECs induces matrix metalloproteinase-2 which may contribute to the pathogenesis of VILI [7]. Recently, Pinhu et al. [26] reported that cyclic stretch (30% elongation) induces IL-8 mRNA expression in a variety of pulmonary cells including HPMVECs. In this study, we demonstrated for the first time that cyclic stretch enhanced IL-8 protein synthesis and that it was significantly inhibited by the p38 inhibitor SB203580 in HPMVECs. We also found that 20% stretch up-regulates production of IL-6 and MCP-1, which are involved in the pathogenesis of VILI [5], [16]. Taken together, pulmonary microvascular ECs enhance inflammation of the lung via cytokine/chemokine production in response to large stretching in mechanically ventilated patients with ALI/ARDS.

It has been demonstrated that IL-8 production is regulated by activation of the MAP-kinase family p38, ERK1/2 and JNK in pulmonary arterial ECs, airway epithelial cells, and airway smooth muscle cells [13], [27], [28], [29]. In HPMVECs, stretch-induced IL-8 synthesis was significantly blocked by the p38 inhibitor SB203580, whereas neither the ERK1/2 inhibitor U0126 nor JNK inhibitor SP600125 affected it (Fig. 3A). Moreover, p38 activation was transiently increased by 20% cyclic stretch within 5 min (Fig. 3B and C). These findings demonstrate that activation of p38 rather than JNK or ERK1/2 regulates IL-8 synthesis in HPMVECs following cyclic stretching. Additionally, the constitutive (baseline) IL-8 concentration in the static condition was also reduced by SB203580 (data not shown), indicating that cyclic stretch increases IL-8 production by up-regulating p38 activities which is essential for IL-8 synthesis in HPMVECs.

Mechanosensitive ion channels, specifically Ca²⁺-permeable stretch-activated (SA) channels, play an important role in mechanotransduction [30], [31]. Moreover, SA channels have been proposed as one of the key molecules related to VILI [2], [3]. Zhao et al. [32] demonstrated the crucial role of [Ca²⁺]_i elevation in the neuropeptide-induced IL-8 production in the NCM460 colon epithelial cell line. We recently reported the activation of Ca²⁺ influx pathway by a large (⩾20%) stretch in HPMVECs [17]. However, the IL-8 production was not inhibited by the blockade of Ca²⁺ influx pathways with EGTA, Gd³⁺, or BTP-2 (Fig. 3D). Collectively, the present findings indicate that a Ca²⁺ influx via the SA channel is not involved in the stretch-induced IL-8 production in HPMVECs.

It has been demonstrated that transcription of IL-8 is regulated by NF-κB [29], [33], [34]. Moreover, mechanical stretching induced activation of NF-κB in lung fibroblasts and A549 cells [20], [29], [34]. Indeed, we found that mechanical stretch induces translocation of p65/RelA into the nucleus, which indicates NF-κB activation in HPMVECs (data not shown), but the IL-8 production induced by stretch was not inhibited by NF-κB inhibitors, either SN50 or ammonium pyrrolidinedithiocarbamate (data not shown), suggesting that NF-κB is not involved in the stretch-induced IL-8 production in HPMVECs. However, because IL-8 gene expression has been reported to be regulated by multiple transcription factors [35], other transcription factors such as AP-1 and NF-IL-6 may be involved in the stretch-induced IL-8 gene expression in HPMVECs.

Cyclic stretch and fluid shear stress modulate cell morphology and migration of vascular ECs [18], [22]. We found that cyclic stretch induces cell reorientation perpendicular to the stretch axis accompanied by actin polymerization in HPMVECs (Fig. 4). The stretch-induced cell reorientation was not affected by p38 inhibition (Fig. 4C). In bovine pulmonary arterial ECs, p38 regulates shear stress-induced cell alignment but not cyclic stretch-induced cell reorientation [36], [37]. Our results are consistent with these previous findings in different EC types. Although the role of stretch-induced reorientation of pulmonary microvascular ECs in the pathophysiology of VILI remains unclear, the capacity of ECs to realign in response to the direction of physical forces is considered to contribute to vascular homeostasis, wound repair, angiogenesis, and pathogenesis [22].

In summary, mechanical stretch promoted IL-8 synthesis with activation of p38 in HPMVECs. Moreover, cyclic stretch enhanced IL-6 and MCP-1 production and induced endothelial reorientation accompanied by actin polymerization. Our findings indicate that IL-8 released from the pulmonary microvasculature due to large tidal stretch activates neutrophils, leading to increased disease severity of VILI in patients with ARDS.