Effects of cigarette smoke on pulmonary endothelial cells

Abstract

Cigarette smoking is the leading cause of preventable disease and death in the United States. Cardiovascular comorbidities associated with both active and secondhand cigarette smoking indicate the vascular toxicity of smoke exposure. Growing evidence supports the injurious effect of cigarette smoke on pulmonary endothelial cells and the roles of endothelial cell injury in development of acute respiratory distress syndrome (ARDS), emphysema, and pulmonary hypertension. This review summarizes results from studies of humans, preclinical animal models, and cultured endothelial cells that document toxicities of cigarette smoke exposure on pulmonary endothelial cell functions, including barrier dysfunction, endothelial activation and inflammation, apoptosis, and vasoactive mediator production. The discussion is focused on effects of cigarette smoke-induced endothelial injury in the development of ARDS, emphysema, and vascular remodeling in chronic obstructive pulmonary disease.

INTRODUCTION

Cigarette smoking is the leading cause of preventable disease and death in the US, causing >480,000 deaths per year in the US, according to the 2014 Surgeon General’s report (136a). Worldwide, the World Health Organization estimates that tobacco use kills >7 million people each year, 6 million of whom are smokers while 890,000 die from the effects of secondhand smoke exposure [World Health Organization Tobacco Fact Sheet (150a)]. Although the incidence of tobacco use in the US has decreased, 20.1% of adults still continued to use tobacco products in 2015 (106). Thus cigarette smoking remains a significant cause of morbidity and mortality.

Cigarette smoking has systemic effects, in addition to altering lung function. Causes of smoking-related deaths include cancers, cardiovascular and metabolic diseases, pulmonary diseases, and conditions related to pregnancy and birth [2014 Surgeon General’s report (136a)]. Chronic lung diseases caused by cigarette smoking include chronic obstructive pulmonary disease (COPD), idiopathic pulmonary fibrosis, and asthma. COPD in particular is associated with systemic comorbidities (“the comorbidome”; 35). Among the systemic comorbidities, cardiovascular diseases, including pulmonary hypertension, are the most prevalent and carry a high risk of mortality (32). Furthermore, secondhand smoke exposure is also associated with increased risk of coronary artery disease and other cardiovascular diseases (13). These cardiovascular comorbidities of both mainstream and secondhand smoke exposure suggest that toxic cigarette smoke constituents are absorbed from the airways into the blood and that the inhalation of cigarette smoke causes systemic toxicity.

A growing body of evidence indicates that cigarette smoke alters vascular function, in addition to airways and immune function (127). Voelkel and colleagues were among the first to demonstrate lung endothelial cell apoptosis in human emphysema [66; Imai et al. (59)] and in animal models of cigarette smoke exposure [Marwick et al. (88)] and suggested the “vascular hypothesis” for pathogenesis of emphysema (140). Endothelial progenitor cell dysfunction has been recognized in acute exacerbation of COPD, suggesting that cigarette smoke (CS)-induced inadequate repair of the endothelium may also contribute to COPD mortality (80).

CS is complex and includes ~4,500 components. Among these, CO, nicotine, oxidants, fine particulate matter, aryl hydrocarbons, and aldehydes such as formaldehyde and acrolein have potential for causing endothelial cell injury. Acrolein is of particular interest since increased levels have been found in blood and urine of smokers (7). Intraperitoneal administration of acrolein causes emphysema in rats (71), and intratracheal instillation increases lung vascular permeability in mice (82). A recent review summarized the known effects on lung function of nicotine administered by electronic cigarettes (30).

It is now recognized that the lung endothelium is not a passive barrier to gas exchange, but that it is metabolically active and is a key mediator of inflammation. Among the functions of lung endothelium, barrier function is critical in preventing extravasation of water, protein, and cells from the circulation, thus maintaining normal gas exchange. A consequence of injury to the lung endothelium is loss of barrier function and increased-permeability pulmonary edema, the pathophysiologic hallmark of acute respiratory distress syndrome (ARDS). ARDS is an important cause of mortality that does not have specific treatment, despite great improvements in supportive care. In this review we consider growing epidemiologic evidence that cigarette smokers are at increased risk for ARDS and are more susceptible to loss of lung endothelial barrier function from other events associated with development of ARDS (Table 1). In addition, we summarize evidence for cigarette smoke effects on alveolar loss in emphysema and on vascular remodeling and reactivity in pulmonary hypertension associated with COPD. We review studies using cell and animal models of lung endothelial toxicity caused by cigarette smoke exposure, resulting in abnormal barrier function, increased permeability pulmonary edema, endothelial cell apoptosis, endothelial activation and inflammation, and endothelial-dependent vascular remodeling.

Table 1.

Human studies on effects of cigarette smoke on development of ARDS

Authors	Year	Cohort	Smoking Status	Association
Iribarren et al. (60)	2000	HMO database, extracted ARDS admissions	Patient report	Dose-response effect between cigarette smoking and risk of ARDS
Tandon et al. (132)	2001	Esophagectomy patients in the United Kingdom	Patient report	Smoking was associated with postesophagectomy ARDS
Calfee et al. (26)	2011	Emergency department visits for blunt trauma	Cotinine levels measured at admission	Cotinine level correlated with risk of ARDS
Hsieh et al. (57)	2014	ARDSNET cohort (2 studies)	4-(Methylnitrosamino)-1-(3-pyridyl)-1-butanol levels	Similar risk of ARDS despite being younger and healthier
Ware et al. (146)	2014	Brain-dead donors for Beta-Agonists for Oxygenation in Lung Donors (BOLD) study	Medical record and interviews with next of kin	Current smokers had higher lung weights than ever smokers, and ever smokers had higher lung weights than lifelong nonsmokers (measure of edema)
Ware et al. (147)	2016	Validating Acute Lung Injury Biomarkers for Diagnosis (VALID) study cohort	Medical record/patient report	Ozone exposure significantly associated with ARDS risk only in current smokers
Wacharasint et al. (141)	2015 (poster)	Thai-Surgical Intensive Care Unit (Thai-SICU) study databases	Medical record/patient report	Smokers with higher number of pack years and heavier smokers had higher incidence of ARDS and longer ICU length of stay
Moazed et al. (94)	2016	Volunteer smokers and nonsmokers	Self-report	Greater increase in total protein in smokers (measure of permeability) after exposure to LPS, multiple biomarkers
García-Lucio et al. (43)	2016	Lobectomy/pneumonectomy patients for solitary pulmonary nodule	Patient report	Elevated ANGPT-2 in smokers
Gajic et al. (42)	2011	Admitted patients with risk factors for ARDS	Patient report	No association between smoking and ARDS
Ferro et al. (38)	2010	Admitted trauma patients	Patient report	No difference in adverse outcomes when stratified by smoking status

ANGPT-2, angiopoietin-2; ARDS, acute respiratory distress syndrome; ARDSNET, ARDS Network; HMO, health maintenance organization; ICU, intensive care unit.

EFFECT OF CIGARETTE SMOKING ON ENDOTHELIAL PERMEABILITY AND ACTIVATION IN ARDS

CS Is a Priming Factor for ARDS in Human Subjects

CS increases susceptibility to lung infections and development of ARDS in human subjects.

Respiratory tract infections and complications are common causes of morbidity and mortality among smokers (136a). Smokers are more susceptible to infections by human immunodeficiency virus (HIV; 107, 113), Mycobacterium tuberculosis (27), Streptococcus pneumoniae (15), and influenza A (150). Current smokers with community-acquired pneumonia often progress to severe sepsis (15). In addition to increased risk of infections, recent work has implicated primary and secondhand smoke exposure as a risk factor for the development of ARDS. Iribarren et al. extracted ARDS cases from a Kaiser Permanente database of hospitalizations (60). They found that current heavy smokers (>20 cigarettes per day) had 5.7 times the risk of developing ARDS, with a dose-response association in a multivariate regression. They estimated that 50% of the risk of ARDS is attributable to cigarette smoking. In a study of esophagectomy patients, a history of smoking was one of two preoperative factors associated with the later development of ARDS (132). Wacharasint et al. found that in a cohort of patients admitted to a surgical intensive care unit in Thailand, those who were active smokers had a higher incidence of ARDS than former smokers, who in turn had a higher incidence than never smokers (141). This study also demonstrated a dose-response association in which pack years correlated with risk of ARDS. Ware et al. evaluated the cohort from the Validating Acute Lung Injury Biomarkers for Diagnosis (VALID) study (147). Although they were testing for an association between ambient ozone levels and risk of ARDS, they found that in this critically ill population, ozone exposure was associated with ARDS only in patients who smoked (147). Furthermore, when ARDS survivors were followed as in the ARDS Network (ARDSNET) Long-Term Outcomes Study (ALTOS), those who smoked had a poorer health-related quality of life (22).

There are a few studies that have not shown smoking to be a significant factor. Gajic et al. examined a variety of variables in a prospective study to determine which were associated with the development of mild and severe ARDS (42). While characteristics including alcohol abuse, obesity, and male sex were significant, smoking was not associated with higher rates of ARDS (42). In a retrospective study of trauma patients, Ferro et al. showed no association between smoking and various adverse outcomes including ARDS, as well as sepsis, septic shock, and mortality (38). Smoking also did not impact the number of days of mechanical ventilation or intensive care unit length of stay (38).

Notably, all the aforementioned studies relied on self-report and/or chart review to determine smoking status. To better avoid bias inherent in self-report and to capture secondhand smoke exposure, a few studies have used biomarkers such as cotinine and NNAL [4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol] to assess smoking status. Calfee et al. measured levels of plasma cotinine and tracked the subsequent incidence of ARDS (26). They found that patients in the highest quartile of cotinine concentrations were 3.25 times as likely as those in the lowest quartile to develop ARDS (26). Hsieh et al. performed a secondary analysis of the ARDSNET Albuterol for the Treatment of Acute Lung Injury (ALTA) and Omega-3 Fatty Acid/Antioxidant Supplementation for Treating People with Acute Lung Injury or Acute Respiratory Distress Syndrome (Omega) studies. Smoking status was determined on the basis of urine NNAL concentrations at the time of randomization in the primary study (57). Smokers tended to be younger and healthier and have fewer comorbidities and lower Acute Physiology and Chronic Health Evaluation (APACHE) scores than patients with undetectable NNAL, but they nonetheless had equally severe lung injury. In these patients, smoking nullified the protective benefit of overall better health when they developed ARDS. Smokers were also overrepresented in this sample of patients with ARDS compared with national smoking rates (57).

CS increases alveolar-capillary barrier permeability and inflammation in humans.

The poor outcomes of ARDS among smokers are likely due to inflammation, alveolar epithelial and endothelial injury, and increased alveolar-capillary permeability, all of which predispose to development of pulmonary edema. In a study of donor lungs harvested with the intention of transplant, Ware et al. (146) showed that lungs from current smokers weighed more than lungs of former smokers and that these were heavier than the lungs of never smokers, all presumably because of edema. They found the same pattern for efficiency of oxygenation [arterial partial pressure of O₂ ($\mathrm{P}{\mathrm{a}}_{\mathrm{O}}{}_{{}_2}$)/inspired O₂ fraction (${\text{FI}}_{{\text{O}}_{\text{2}}}$)] before harvest. Additionally, they found that a more extensive smoking history was associated with worse results in a laboratory test of alveolar fluid clearance (146), a measure of alveolar epithelial cell function. Relative to the more permeable pulmonary capillary endothelial barrier (pore radius of intercellular junctions is ~40–80 Å), the alveolar epithelial barrier is less permeable because of tight junctions of an estimated pore radius of 4–10 Å (133). Pulmonary clearance of inhaled ^99mtechnetium-labeled diethylenetriamine pentaacetic acid (^99mTc-DTPA) has been used to noninvasively assess alveolar epithelial barrier permeability in humans (14). Using this technique, Jones et al. have documented increased alveolar epithelial barrier permeability in cigarette smokers (63). Taken together, CS increases alveolar epithelial barrier permeability.

Whether CS exposure increases pulmonary endothelial barrier permeability in humans has been controversial. A pilot study on four smokers suggests that smokers may have decreased pulmonary capillary barrier permeability to urea (145). In contrast, pulmonary endothelial denudation has been documented in smokers with COPD (9). Altered endothelial function has also been suggested by biomarkers in human subjects. García-Lucio et al. (43) obtained pneumonectomy specimens from a sample of patients with solitary nodules, who either had COPD, were smokers with normal lung function, or were nonsmokers with normal lung function. In these patients, expression of angiopoietin-2 (ANGPT-2), as measured by immunohistochemistry, was associated with increased pulmonary artery vessel wall remodeling. In another cohort, they found that plasma levels of ANGPT-2 were higher in current smokers than in former smokers and that COPD patients overall had higher levels of ANGPT-2 than nonsmokers. Again, they demonstrated a dose-response association with pack years (43). A similar association was demonstrated by Petta et al. in healthy smokers with asthma (105). Since ANGPT-2 is secreted by vascular endothelial cells in Weibel-Palade bodies, these findings support the notion that cigarette smoking alters endothelial cell function in humans.

Moazed et al. (94) demonstrated experimentally that after exposure to inhaled lipopolysaccharide (LPS), smokers had a greater increase in plasma IL-8, a known mediator of inflammation in ARDS, compared with nonsmokers. These smokers had increased bronchoalveolar lavage (BAL) protein, suggesting increased alveolar-capillary permeability. They also had elevated plasma matrix metalloproteinase-8 (MMP-8), a potential cause of lung matrix breakdown. MMP-8 is secreted by neutrophils, which were elevated in total numbers and percentages in BAL fluid. LPS inhalation caused an exaggerated decrease in BAL VEGF, which has been shown to be decreased in ARDS patients and to increase with recovery (1). Smokers had increased levels of soluble VEGF receptor 1 (VEGFR1), which inhibits VEGF and may contribute to increased permeability (94). A recent study of mechanically ventilated trauma patients also showed differences in the lung microbiota of smokers and nonsmokers, at admission and after 48 h, which in turn affected expression of inflammatory markers and was associated with differences in incidence and severity of ARDS (101).

In summary, with some exceptions, these clinical studies present significant evidence that cigarette smoking potentiates ARDS and worsens outcomes. Cigarette smoke exposure increases both epithelial and endothelial barrier permeability at baseline and upregulates inflammatory pathways in humans. The role of inflammation in ARDS may differ among cases; however, evidence from several ARDSNET trials suggests that there are at least two subphenotypes of ARDS with distinct biochemical and inflammatory profiles (25). This is clinically significant because these subphenotypes respond differently to liberal vs. conservative fluid strategies and to high vs. low positive end-expiratory pressure (25, 36, 122). In clinical practice, subphenotype determination is limited by the biochemical assays available at the bedside. Future studies should incorporate smoking history as a factor in the characterization of ARDS subphenotypes and determine whether it can be used to identify the subphenotype and influence treatment strategies (25, 36, 122).

Both Brief and Subacute CS Exposures Increase Alveolar-Capillary Barrier Permeability in Animal Models

Loss of alveolar-capillary barrier function is a hallmark of ARDS. By assessing pulmonary clearance of fluorescein isothiocyanate-dextran (FITC-D; molecular radius at 22.2 Å), Burns et al. reported that CS exposure increases alveolar epithelial barrier permeability in guinea pigs (23). When the alveolar epithelial barrier is compromised, components of CS may pass through the leaky epithelial barrier and potentially result in endothelial cell injury and increased permeability. By assessing blood concentrations of intratracheally instilled iodine-125 bovine serum albumin (¹²⁵I-BSA), which has dimensions of 140 × 40 × 40 Å and is impermeable to a healthy capillary endothelial barrier, Li et al. reported that intratracheal instillation of cigarette smoke condensate increases pulmonary capillary barrier permeability in rats (78). They also suggested that CS-induced increase in alveolar-capillary barrier permeability was due to oxidative stress but not a consequence of lung inflammation (78). We have demonstrated that lung microvascular endothelial cells (LMVEC) isolated from mice exposed to CS had a greater permeability and incomplete recovery after challenges by LPS or thrombin (Fig. 1, A and B; 20). Our results suggest that CS exposure may cause epigenetic changes in endothelial cells, which make these cells more vulnerable to a second injury. Both brief (hours) and subacute (4 wk) CS exposures increased BAL protein levels in guinea pigs (74). We have shown that acute CS preexposure followed by saline (control) caused lung edema; this effect was augmented in response to a second injury (LPS or Pseudomonas) in mice (Fig. 2; 20, 84). To our surprise, CS exposure alone without saline did not increase BAL protein levels or lung wet-to-dry weight ratio (data not shown). We suspect that the increased alveolar-capillary permeability by CS exposure alone is so covert in mice that it is not detectable by BAL protein content or wet-to-dry lung weight ratio. We speculate that more sensitive methods may be necessary to assess the effect of CS alone on alveolar-capillary permeability in mice. Nevertheless, there is strong evidence suggesting that CS exposure increases lung alveolar-capillary barrier permeability both in humans and in other animal models.

Fig. 1.

Cigarette smoke (CS) increased lung microvascular endothelial cell (LMVEC) permeability in vitro. A and B: 6-wk-old male C57BL/6 mice were exposed to room air (RA) or CS for 6 h. After overnight rest, LMVEC were isolated from cortical lung tissues. Isolated and characterized LMVEC at passage 1 were plated onto electric cell-substrate impedance sensing (ECIS) arrays at equal numbers of cells (2.5 × 10⁵ cells per well). After overnight attachment, cells were treated with vehicle (V), LPS (1 µg/ml), or thrombin (2 U/ml) for 20 h, and monolayer permeability was assessed by measuring electric resistance across the monolayers. Four independent experiments with duplicated ECIS wells for each condition at each time were conducted. C: primary human LMVEC were treated with vehicle (10% PBS) or varying concentrations of cigarette smoke extract (CSE) for indicated times, and monolayer permeability was assessed by ECIS. Three independent experiments with duplicated ECIS wells for each condition at each time were conducted. Data are presented as the means ± SE of the normalized electrical resistance at the selected time points relative to their initial resistance. ANOVA and Tukey-Kramer post hoc test were used to determine statistically significant difference across means among groups. *P < 0.05 vs. RA+V; ξP < 0.05 vs. RA+LPS (A) or RA+thrombin (B). Arrows indicate the time for addition of treatments. [Reprinted with permission of the American Journal of Respiratory Cell and Molecular Biology (20).]

Fig. 2.

Effects of a brief cigarette smoke (CS) exposure on LPS- and Pseudomonas-induced lung edema. Male 6-wk-old C57BL/6 were exposed to room air (RA) or CS for 6 h. One hour after CS exposure, mice were intratracheally administered with 2.5 mg/kg of LPS or 5 × 10³ colony-forming units of P. aeruginosa (strain PA103) or equal volume of saline as a control (ctrl). After 18 h, bronchoalveolar lavage (BAL) protein content (A and C) and lung wet-to-dry weight ratio (B) were assessed. Four to six mice per group were used. εP < 0.05 CS/ctrl vs. RA/ctrl; *P < 0.05 RA/LPS vs. RA/ctrl; €P < 0.05 CS/LPS vs. RA/LPS or RA/PA103. [Reprinted with permission of the American Journal of Physiology Lung Cellular and Molecular Physiology (84) and the American Journal of Respiratory Cell and Molecular Biology (20).]

Both Brief and Subacute CS Exposures Increase Susceptibility to Acute Lung Injury in Animal Models

Similarly to the human studies described above, CS preexposure delayed bacterial clearance and promoted lung inflammation induced by Pseudomonas aeruginosa infection in mice (33). Two weeks of CS exposure also enhanced alveolar inflammation and apoptosis induced by poly(I:C), a viral pathogen-associated molecular pattern (PAMP) in mice (65). Four weeks of CS exposure exacerbated lung inflammation and compromised adaptive immunity to infection by Hemophilus influenzae in mice (86). Two weeks of CS exposure did not change basal BAL protein levels or neutrophil count but significantly exacerbated mechanical ventilation-induced increase in BAL protein content and neutrophil influx in rats (54). Similarly, CS exposure for 3–4 wk aggravated blunt chest trauma-induced lung inflammation and lung cell apoptosis in mice (142).

Using two strains of inbred mice with different susceptibility to CS-induced emphysema (48), we found that as little as 3–6 h of exposure to CS potentiated LPS-induced lung edema in both C57BL/6 (Fig. 2A) and AKR mice (116). Similar exacerbating effects were observed after 3 wk of preexposure to CS (Fig. 3A). We also noted a strain difference in susceptibility to CS priming for acute lung injury, with a greater effect in AKR mice (Fig. 3A; 116).

Fig. 3.

Effects of prolonged cigarette smoke (CS) exposure on LPS-induced lung edema. Male 6-wk-old C57BL/6 and AKR mice were exposed to room air (RA) or CS for 3 wk. One hour after the last CS exposure, mice were intratracheally administered with 2.5 mg/kg of LPS or equal volume of saline as a control (ctrl). After 18 h, bronchoalveolar lavage (BAL) protein content (A), lung wet-to-dry weight ratio (B), and lung extravasation of albumin-conjugated Evans blue dye (C) were assessed. Three to four C57BL/6 mice per group and 4–6 AKR mice per group were used. εP < 0.05 CS/ctrl vs. RA/ctrl; *P < 0.05 RA/LPS vs. RA/ctrl; €P < 0.05 CS/LPS vs. RA/LPS. [Reprinted with permission of the American Journal of Physiology Lung Cellular and Molecular Physiology (116).]

Taken together, studies of both human and preclinical animal models indicate that increased susceptibility to and severity of acute lung injury/ARDS are important adverse health consequences of CS exposure that need to be taken into consideration when treating critically ill individuals. In addition, there is evidence from humans and from animal studies that CS exposure increases lung barrier permeability, through effects on both epithelial and endothelial cells. In this review we focus on mechanisms of CS effects on lung endothelial cell permeability.

Cigarette Smoke Extract Increases Endothelial Monolayer Permeability in Vitro

Although the nature and concentrations of smoke components may differ between CS inhalation in vivo and aqueous cigarette smoke extract (CSE) exposure of culture cells in vitro, CSE exposure is a widely used and acceptable in vitro system to address effects of CS on specific cells and underlying mechanisms (92, 155). As has been shown in human and animal studies, exposure to CSE in vitro increases pulmonary macrovascular and microvascular endothelial monolayer permeability (Fig. 1C; 20, 55). The underlying mechanism involves inhibition of RhoA and focal adhesion kinase (FAK) signaling (84), microtubule depolymerization (16), histone deacetylase-6 (HDAC6) activation (20), p38 activation (81), ceramide upregulation (120), and disruption of intercellular adhesion molecules, as depicted in Fig. 4.

Fig. 4.

Proposed model of cigarette smoke-induced pulmonary endothelial cell injury. Cigarette smoke (CS) exposure causes lung endothelial cell (EC) activation, resulting in inflammation, via activation of NF-κB signaling; CS exposure directly increases EC permeability likely through multiple signaling pathways, including inhibition of RhoA, focal adhesion kinase (FAK), and intercellular adhesion molecules (IAM) and activation/upregulation of histone deacetylase-6 (HDAC6), p38, and ceramide; CS exposure also causes EC apoptosis through signaling pathways involving downregulation of VEGF, FAK, α₁-antitrypsin (AAT), and unfolded protein response (UPR) and upregulation of ceramide, adenosine, p38, and p53. Excessive EC apoptosis can lead to necrosis and autoimmunity because of release of mitochondrial damage-associated molecular patterns (DAMPs), endothelial microparticles (eMPs), and other cellular contents. CS-induced increases in EC activation, permeability, apoptosis, and necrosis collectively contribute to increased risk of development of acute respiratory distress syndrome (ARDS). CS-induced lung inflammation, EC apoptosis and necrosis, and autoimmunity cause emphysema. CS exposure increases EC endothelin (ET)-1 levels, leading to increased smooth muscle cell (SMC) proliferation and vasoconstriction. CS exposure also decreases vasodilation because of reduction in EC NO. All these changes, in combination with increased blood flow in remaining vessels due to EC apoptosis and loss of vessels, cause vascular remodeling and pulmonary hypertension (PH) in chronic obstructive pulmonary disease (COPD) patients.

Endothelial barrier integrity is regulated by adherens junctions (AJs), F-actin cytoskeleton, microtubules (102), and focal adhesion complexes (FAC; 138). It has been well documented that RhoA is required for coordinated assembly of F-actin fibers and that AJs and FAC are critical for establishment of cell-cell contacts (21). RhoA activation also mediates increased endothelial permeability induced by various edemagenic agents, such as thrombin and histamine, by increasing contractile F-actin stress fiber formation (136). The role of RhoA signaling in CSE-induced endothelial monolayer permeability has been controversial. Inhibition of Rho kinase has been reported to attenuate CSE-induced increases in lung microvascular endothelial monolayer permeability (120). In contrast, we have shown that CSE inactivated RhoA via oxidative stress in a dose- and time-dependent manner (84). In support of our results, reactive oxygen species (ROS) have been shown to inhibit RhoA, leading to structural disruption of FAC and F-actin fibers in other cells (24). We further demonstrated that overexpression of constitutively active RhoA blunted CSE-induced disassembly of AJs, FAC, and cortical F-actin fibers in pulmonary endothelial cells (84). Our data suggest that inhibition of RhoA contributes to CSE-induced endothelial barrier dysfunction. The apparent discrepancy of these results may be due to different preparations and concentrations of CSE used. RhoA contains a redox-sensitive motif [GXXXCGK(S/T)C]. It has been suggested that physiological levels of ROS and high reduction potential may directly oxidize and activate RhoA through direct oxidation of two cysteine residues located in the redox-sensitive motif (2). Conversely, high levels of oxidants inhibit RhoA through formation of an intramolecular disulfide bridge that prevents GTP binding (53). Whether CSE inhibits RhoA via intramolecular disulfide bridge formation needs further investigation.

FAK is a key component of FAC and an essential survival factor for anchorage-dependent cells (83). Although FAK activation has been implicated in endothelial barrier dysfunction in some settings (49, 153), FAK activity is also required for maintenance of endothelial barrier function (72). Endothelial cells isolated from FAK knockout mouse embryos or embryos expressing kinase-defective FAK displayed increased monolayer permeability (160). Knockdown of FAK by small interfering RNA (siRNA) also caused disruption of tight junctions in the testis (126). FAK activity is also required for resealing of endothelial adhesion junctions after exposure to H₂O₂ (112) and thrombin (56). We have shown that CSE dose and time dependently inhibits FAK activity via oxidative stress and that overexpression of FAK attenuates CSE-induced disruption of AJs, FAC, and F-actin fibers (84). The reactive aldehyde, 4-hydroxy-2-nonenal (4-HNE), a component of CSE, also decreases FAK activity and endothelial barrier function via oxidative stress (137). Taken together, oxidative stress-induced FAK inhibition contributes to CSE-induced endothelial barrier dysfunction.

Intact microtubules are critical for maintenance of endothelial barrier integrity (17, 91, 139). α-Tubulin is a building block of microtubules, and acetylation of α-tubulin is characteristic of stable microtubules (89, 159). CSE has been shown to cause microtubule depolymerization in human umbilical vein endothelial cells (HUVEC; 16). We have shown that CSE decreases α-tubulin acetylation and causes microtubule depolymerization in pulmonary endothelial cells (20). The microtubule stabilizer paclitaxel elevates α-tubulin acetylation and prevents CSE-induced pulmonary endothelial barrier dysfunction (20). These results indicate that microtubule instability contributes to CSE-induced endothelial barrier dysfunction.

Acetylated α-tubulin is deacetylated by HDAC6 (158). HDAC6 is a microtubule-associated protein deacetylase that is ubiquitously expressed and predominantly located in the cytoplasm (58). HDAC6 knockout mice do not have a visible abnormal phenotype (158) but have better survival and reduced lung injury after LPS challenge (28, 144). Histone deacetylase inhibitors attenuate lung injury induced by LPS (96), P. aeruginosa (20), mechanical ventilation (29), and cecal ligation and puncture (157). HDAC6 inhibitors also blunt increased permeability of lung endothelial monolayers induced by thrombin (114) and LPS (28). In addition, inhibition of HDAC6 has been proposed as a novel promising therapeutic option for a variety of disorders, including cardiac dysfunction, depression, Alzheimer’s disease, COPD-associated cilia dysfunction, and diabetes (31, 41, 46, 62, 75, 149, 154).

The role of HDAC6 in CS-induced lung injury is less well understood. We have recently shown that HDAC6 activity is enhanced in cultured lung endothelial cells exposed to CSE likely via oxidative stress-mediated Akt inactivation and subsequent GSK-3β-mediated phosphorylation of HDAC6 at serine-22 (20). Tubacin, a specific HDAC6 inhibitor (50), and HDAC6 siRNA abolished CSE-induced barrier dysfunction in primary human LMVEC in vitro (20). Importantly, tubacin prevented CS priming for lung injury after infection with P. aeruginosa in mice (20). We speculated that HDAC6 inhibitors may become a novel therapeutic option for CS-associated lung injury.

Endothelial intercellular adhesion molecule CD146, also known as cell surface glycoprotein MUC18, is decreased in lung tissue of smokers with COPD and of rats exposed to CS, as well as in cultured lung macrovascular and microvascular endothelial cells exposed to CSE (73). CD146 knockout mice exhibit perivascular edema and lung inflammation (73). However, it remains unknown whether loss of CD146 plays a role in CS-induced endothelial barrier dysfunction and increased vascular permeability in vivo. Notably, soluble sCD146, which lacks transmembrane and intracellular domains of CD146, is increased in plasma and BAL of COPD patients as well as in BAL of rats exposed to CS (73). Further investigation is needed to determine whether circulating or BAL sCD146 can be used as biomarkers of CS-induced lung injury.

Platelet endothelial cell adhesion molecule-1 (PECAM-1) regulates endothelial barrier permeability by facilitating dephosphorylation of β-catenin (18). PECAM-1 null mice exhibit prolonged and increased permeability after inflammatory insults (18). CSE causes a 10-fold increase in tyrosine phosphorylation of PECAM-1 in HUVEC (124). However, the role of PECAM-1 in CS-induced lung endothelial injury remains unknown.

It is likely that multiple components of cigarette smoke are responsible for increased endothelial permeability. We have demonstrated that acrolein increased endothelial monolayer and lung microvascular permeability in vivo (82). Schweitzer et al. reported that nicotine, e-cigarette solution, and condensed e-cigarette vapor increased endothelial monolayer permeability (119). In addition, they demonstrated nicotine-independent effects of e-cigarette solutions on endothelial permeability. Acrolein is among the components of combustion of e-cigarette solutions.

CS Causes Lung Endothelial Cell Activation and Inflammation

CS activates the lung endothelium and causes inflammatory cell accumulation. Sharma et al. showed in human LMVEC that this mechanism involves the inhibition of platelet aggregating factor acetylhydrolase (PAF-AH), which degrades platelet aggregating factor (PAF), leading to increased levels of PAF (123). Increased PAF is in turn associated with neutrophil [polymorphonuclear neutrophil (PMN)] adherence to the endothelium. They also showed that CSE stimulates expression of other endothelium-activating molecules, including P-selectin, E-selectin, intercellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1 (VCAM-1; 123). The same group subsequently showed that increased PAF is associated with increased adherence of metastatic breast cancer cells to the lung endothelium and that CSE stimulates the upregulation of the PAF receptor on the breast cancer cells as well (70).

Acute CS exposure also activates xanthine oxidase and elicits the rolling and adhesion of leukocytes to endothelium of arterioles and postcapillary venules of striated muscle microcirculation and aortic endothelium in hamsters in vivo; these effects were prevented by superoxide dismutase and water-soluble antioxidants but not by lipid-soluble antioxidants (76, 77). CSE increases the surface expression of adhesion molecules, including ICAM-1, endothelial leukocyte adhesion molecule-1 (ELAM-1), VCAM-1, and E-selectin, as well as cytokines and chemokines, including tumor necrosis factor-α, IL-6, and IL-1β, via NADPH oxidase-dependent NF-κB transcriptional activation in endothelial cells (98) CSE also increases adherence of monocytes to the endothelium and transendothelial migration (124), as well as neutrophil transmigration across HUVEC (99). CS-induced upregulation of the C-X-C motif chemokine receptor 3 (CXCR3) receptor in endothelial cells may mediate endothelial cell apoptosis (47). CSE synergizes with the inflammatory cytokine IL-1β to increase vascular permeability and endothelial dysfunction via ROS/p38/phosphatase and tensin homolog (PTEN)-mediated tyrosine phosphorylation of vascular endothelial cadherin and β-catenin, as well as subsequent β-catenin nuclear translocation and expression of inflammatory genes, such as cyclooxygenase-2 (COX-2) in cardiac endothelial cells (11, 12). Taken together, CS causes lung inflammation via its direct effect on endothelial cell activation, leading to enhanced endothelial cell apoptosis, increased barrier permeability, and endothelial dysfunction.

Damage-associated molecular patterns (DAMPs) enhance PMN adherence to endothelial cells by increasing surface expression of adhesion molecules in both PMN and human pulmonary artery endothelial cells in vitro (130). CS exposure induces necrosis of bronchial epithelial cells and neutrophils, leading to DAMP release and proinflammatory responses (51, 111). Levels of DAMPs, including high-mobility group box-1 (HMGB1), are increased in extracellular lung fluids of COPD patients (109), and increased DAMPs have been implicated in inflammation in COPD (110). Future studies are needed to address the sources and roles of DAMPs in CS-induced endothelial activation and related lung diseases.

EFFECT OF CIGARETTE SMOKE ON ENDOTHELIAL APOPTOSIS IN COPD

Robust pulmonary endothelial cell apoptosis has been observed in patients with severe ARDS (1) and in mice with mild ARDS induced by LPS (40). Pulmonary microvascular endothelial cell apoptosis has been suggested as a cause of barrier dysfunction and edema in septic mice (44, 45). Inhibition of apoptosis by a broad-spectrum caspase inhibitor prolonged survival of mice exposed to LPS (67). These studies suggest that lung endothelial cell apoptosis contributes to the pathogenesis of ARDS.

Recent clinical studies also implicate vascular endothelial damage in the pathogenesis of COPD. This has been observed outside the lungs as well, most notably with the loss of glomerular integrity in patients with COPD. Polverino et al. (108) showed that COPD patients had lower estimated glomerular filtration rates than smokers without COPD and nonsmoking controls and that lower forced expiratory volume in 1 s (FEV₁) %predicted was associated with double contouring of the glomerular basement membrane. Most smokers had albuminuria as well, but the quantity did not correlate with FEV₁. In mouse studies, cigarette smoke also increased albuminuria. This group also showed that levels of oxidative stress-advanced glycation end products (AGEs) and their receptors (RAGEs) increased in mice with cigarette smoke exposure. Treatment with enalapril blunted these effects in mice, suggesting a role of angiotensin I-converting enzyme (ACE) inhibitors for renal protection in patients with COPD (108). Agrawal et al. also demonstrated increased albuminuria in patients with acute exacerbations of COPD (3).

Lung tissue from patients with emphysema displays increased apoptosis of alveolar epithelial and endothelial cells (59, 66). We also observed an increase in lung endothelial cell apoptosis in AKR mice exposed to CS for 3 wk (115). Circulating endothelial microparticles (eMPs) are thought to be shed into the bloodstream from activated, apoptotic, or necrotic endothelial cells. Blood eMPs are significantly elevated in healthy smokers (128) and patients with COPD (131, 134). Circulating eMP levels remain elevated in COPD patients despite smoking cessation but return to normal low levels in healthy former smokers, suggesting that pulmonary endothelial cell injury is reversible only in healthy smokers who quit smoking (128). Circulating eMPs are also increased in rats exposed to CS for 2–6 mo (79). Further studies are needed to determine whether circulating eMPs are markers of vascular endothelial injury or part of the pathogenesis of diseases linked to endothelial injury and inflammation in smokers (121).

CSE triggers apoptosis of pulmonary endothelial cells (115, 135) and induces necrosis and inhibits apoptosis of alveolar epithelial cells and HUVEC (148). The potential mechanisms underlying CS-induced endothelial apoptosis include inhibition of VEGF, FAK, and α₁-antitrypsin (AAT) signaling and upregulation of ceramide and adenosine, as well as impairment of unfolded protein response (UPR), as depicted in Fig. 4.

VEGF/VEGFR2 signaling is reduced in the lungs of smokers with and without COPD (59, 66) and of rodents exposed to CS (88). CSE decreases expression of VEGF in pulmonary endothelial cells in vitro (135). We observed that FAK activity is reduced in AKR mice exposed to CS for 3 wk and in lung endothelial cells exposed to CSE; these effects were associated with enhanced pulmonary endothelial cell apoptosis in vivo and in vitro (115). Overexpression of FAK prevented CSE-induced endothelial cell apoptosis, suggesting that reduced FAK activity may contribute to CSE-induced endothelial cell apoptosis (115). Inhibition of FAK also causes emphysema-like changes in rat lungs (93). Further studies are necessary to address whether reduced FAK activity contributes to CS-induced lung endothelial cell apoptosis in vivo. AAT inhibits apoptosis of cultured lung endothelial cells by blocking interaction of caspases-3, -6, and -7 with their substrates (103). Overexpression of AAT inhibits apoptosis of cultured lung endothelial cells and attenuates emphysema caused by active caspase-3 and blockade of VEGF signaling in vivo (103). Exogenous AAT protects cultured pulmonary endothelial cells from CSE-induced apoptosis in vitro (6). Whether exogenous AAT or overexpression of AAT prevents or reverses CS-induced lung endothelial cell apoptosis and emphysema in vivo remains to be determined. Ceramide is upregulated in emphysematous lungs of patients and animal models, as well as in cultured pulmonary endothelial cells exposed to CSE (104). This increase in ceramide enhances alveolar cell apoptosis (104).

CS exposure increases lung tissue adenosine levels in mice, an effect associated with lung endothelial cell apoptosis and early emphysema (85). Sustained increased adenosine in adenosine deaminase (ADA)-deficient mice enhances alveolar cell apoptosis and causes emphysema-like changes in mice (161). ADA expression and activity are reduced in lungs of smokers with COPD (162). Whether chronically elevated adenosine contributes to CS-induced lung endothelial cell apoptosis and development of emphysema remains to be investigated. Other investigators have reported that CS increased xanthine oxidoreductase (XOR) expression and activity and that this increase was sufficient and necessary for p53 induction and subsequent endothelial cell apoptosis (69).

The UPR is an important mechanism of elimination of endoplasmic reticulum stress and enhanced cell survival (118). The UPR is activated in lung tissue of smokers without emphysema (68). UPR is also activated by CSE in cultured human bronchial epithelial cells and 3T3 fibroblasts (52, 64) and in cultured pulmonary endothelial cells (115). Using mouse models of CS exposure, we have observed a strong link between impairment of eukaryotic initiation factor 2α (eIF2α) signaling and lung endothelial cell apoptosis in a mouse strain susceptible to CS-induced emphysema (115). Future studies are necessary to determine whether impaired eIF2α signaling contributes to lung endothelial cell apoptosis and emphysema.

In summary, CS exposure causes lung endothelial cell apoptosis. The underlying mechanism is rather complicated, as summarized in Fig. 4. This may be due to the complexity of toxins in smoke. Therefore it would be difficult to prevent and treat cigarette smoke-induced lung diseases by using a single approach/drug. It may be important to target multiple pathways.

EFFECT OF CIGARETTE SMOKE ON ENDOTHELIAL DYSFUNCTION IN PULMONARY HYPERTENSION ASSOCIATED WITH COPD

Pulmonary hypertension is a complication of smoking-induced COPD with a prevalence of ~19% in COPD patients (39). The causes of pulmonary hypertension in COPD are complex and include vasoconstriction caused by acute hypoxia and acidosis, loss of microcirculation in emphysema, increased blood viscosity caused by polycythemia, and vascular remodeling caused by chronic hypoxia.

Pulmonary vasoconstriction and vascular remodeling have been recognized in smokers with normal lung function and in patients with mild COPD without hypoxemia (117). Pulmonary endothelial dysfunction has also been reported in otherwise healthy young smokers (100). In addition, pulmonary endothelial dysfunction and pulmonary vascular smooth muscle cell proliferation precede emphysema in guinea pigs exposed to cigarette smoke (37). These results suggest that cigarette smoke directly alters the structure and function of pulmonary vessels at early stage COPD. CS-induced pulmonary endothelial cell dysfunction might predispose patients with COPD to further smooth muscle hypertrophy, thus contributing to development of pulmonary hypertension in COPD. Furthermore, changes in endogenous pulmonary vasoconstrictors and vasodilators produced by lung endothelium can cause vasoconstriction. The potential role of endothelial dysfunction in the pathophysiology of pulmonary hypertension in COPD suggests that therapies targeting the endothelium may be effective in treating these patients.

Endothelin (ET)-1 and nitric oxide (NO) are endothelium-derived mediators with opposite effects on vascular tone and cell growth. ET-1 is a potent vasoconstrictor and mitogenic agent (19). Plasma ET-1 levels are elevated in smokers, and patients with both COPD and pulmonary hypertension have increased expression of ET-1 in pulmonary endothelial cells (156). ET-1 levels are also elevated in hypertensive intrapulmonary arteries of guinea pigs exposed to CS for 6 mo (151). CSE promotes ET receptor expression in pulmonary artery endothelial cells (90). Future studies are needed to address the role of ET-1 in pulmonary vascular remodeling in COPD patients with pulmonary hypertension.

NO, a major vasodilator, is produced by endothelial NO synthase (eNOS). Expression of eNOS is reduced in pulmonary arteries of healthy smokers (10). Pulmonary arterial relaxation was also reduced in otherwise healthy smokers; this effect was associated with decreased eNOS activity and reduced NO in exhaled air (34). Diminished lung eNOS levels, lower NO, and reduced endothelium-dependent vasodilatation in pulmonary arteries precede the development of emphysema in guinea pigs (37) and mice (125). After 6 mo of cigarette smoke exposure, eNOS knockout mice had enhanced vascular remodeling and pulmonary hypertension (152). CSE irreversibly decreased eNOS mRNA and protein levels and inhibited eNOS enzymatic activity in pulmonary endothelial cells (129, 143). CSE also increases eNOS acetylation via oxidative stress-dependent downregulation of sirtuin 1, leading to blunted eNOS activity (8). In addition, CSE-induced ROS may interact with NO to produce peroxynitrite (ONOO⁻), thereby reducing NO bioactivity (61). In fact, reactive nitrogen species are increased in lung vessels of smokers with COPD and further increased in smokers with combined COPD and pulmonary hypertension (152). Taken together, oxidative stress-mediated reduction in eNOS-NO signaling is a significant contributor to CS-induced vascular endothelial dysfunction.

Nana-Sinkam et al. demonstrated decreased expression of prostacyclin synthase (PGI₂S) and lower levels of prostacyclin, which prevents apoptosis, in lung tissue of patients with COPD (95). In cultured human pulmonary microvascular endothelial cells, treatment with CSE directly decreased expression of COX-1 and increased expression of COX-2 and cytosolic phospholipase A₂ (CPLA₂), potential causes of oxidative stress. This was prevented by pretreatment with the antioxidant N-acetylcysteine (95). These results suggest that decreased prostacyclin and increased oxidative stress also contribute to CS-induced pulmonary endothelial dysfunction and pulmonary hypertension in patients with COPD.

PERSPECTIVES AND FUTURE DIRECTIONS

Growing evidence from clinical studies supports an adverse effect of cigarette smoke exposure as a risk factor for ARDS and poor outcome. Numerous studies of preclinical animal models and cultured cells have demonstrated that cigarette smoke exposure increases pulmonary endothelial barrier permeability and causes endothelial activation resulting in enhanced lung inflammation. Cigarette smoke exposure also causes pulmonary endothelial cell apoptosis; this may initiate and promote development of the lung destruction characteristic of emphysema. Cigarette smoke exposure directly causes endothelial dysfunction associated with vascular remodeling and vasoconstriction in smokers.

We hypothesize that the normal alveolar epithelial barrier protects endothelial cells from exposure to toxins in cigarette smoke. However, as illustrated in Fig. 4, after smoke-induced injury to and disruption of the tight alveolar epithelial barrier, components of smoke and/or DAMPs released from injured airway epithelial cells can reach and cause injury to endothelial cells and subsequently enter the circulation via a more permeable endothelial barrier. Future studies should include developing strategies to protect endothelial cells in order to prevent and treat cigarette smoking-associated lung diseases, such as ARDS, emphysema, and pulmonary hypertension associated with COPD. Since components of cigarette smoke and/or DAMPs released from injured epithelial and endothelial cells can cross the injured endothelial barrier and enter the systemic circulation, it would be reasonable to strengthen the pulmonary endothelial barrier as a measure to prevent cardiovascular and other cigarette smoke-induced systemic diseases.

GRANTS

This study was supported using facilities at the Providence Veterans Affairs Medical Center and by Veterans Affairs Merit Review (S. Rounds), National Heart, Lung, and Blood Institute Grant R01-HL-130230 (Q. Lu), a Brown University Dean’s Emerging Areas of New Science (DEANS) award (S. Rounds), National Institute of General Medical Sciences (NIGMS) Grant U54-GM-115677 (S. Rounds), and an Institutional Development award from NIGMS under Grant P20-GM-103652 (S. Rounds; project 1 to Q. Lu).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Q.L. and E.G. prepared figures; Q.L., E.G., and S.R. drafted manuscript; Q.L., E.G., and S.R. edited and revised manuscript; Q.L., E.G., and S.R. approved final version of manuscript.