The role of the endothelium in asthma and chronic obstructive pulmonary disease (COPD)

The pulmonary vasculature is critical to gas exchange in the lung, with a total pulmonary vascular surface area of 90m² [9]. The entire vascular system is lined by endothelial cells which form a continuous monolayer [9]. Endothelial cells are encased by a basement membrane, a thin protein sheet (50nm thick) that consists of laminins, collagen and proteoglycans [10]. Endothelial cells are also covered on the luminal side by the glycocalyx, a network of proteoglycans and glycoproteins involved in multiple processes such as cell-cell signalling and haemostasis [11]. Finally, embedded in the basement membrane are a non-continuous layer of pericytes which are key mediators of several microvascular processes such as endothelial cell proliferation and angiogenesis [12, 13]. A diagram of the structure of the endothelium is shown in Fig. 1.

Apoptosis (programmed cell death) is a highly ordered process which eliminates damaged or unwanted cells [30]. In the 1950s Liebow demonstrated that alveolar septa in COPD patients were almost avascular. This led to the hypothesis that vascular atrophy resulted in the destruction of alveoli [31]. Supporting this concept, increased levels of apoptotic endothelial cells have been identified in the lungs of patients with COPD [32]. Endothelial cells in COPD patients also demonstrate intranuclear staining of fragmented DNA in comparison to controls providing further evidence for this [33].

Animal studies provide support for apoptosis of endothelial cells resulting in emphysema. For example, it is possible to induce emphysema in rodents by deliberately causing endothelial apoptosis by blockade of vascular endothelial growth factor (VEGF) [34]. When the rodents were given a caspase (proteins involved in apoptosis) inhibitor VEGF inhibition no longer resulted in emphysema suggesting that apoptosis of endothelial cells may be key in emphysema development [34].

In addition to VEGF other mechanisms have been proposed for the development of endothelial apoptosis in COPD patients. For example, Noe et al. demonstrated that Cystic Fibrosis Transmembrane Regulator (CFTR) in human pulmonary endothelial cells was required for stress-induced apoptosis. CFTR inhibition resulted in the attenuation of endothelial apoptosis in response to treatment of cells with staurosporine or hydrogen peroxide [35]. Alpha-1-antitrypsin (A1AT) has also been shown in vitro to prevent caspase-3 activation and therefore apoptosis in pulmonary endothelial cells [36]. This would clearly be a possible relevant mechanism in emphysema development in alpha-1-antitrypsin disease (A1ATD) patients. However, smoking can induce post-translational modification of A1AT which reduces its activity suggesting that this mechanism may also be important in emphysema development in A1AT sufficient individuals [37].

Whole lung gene expression studies which have demonstrated that gene expression appears to differ between emphysematous tissue and normal lung have shown that angiogenesis-related genes are down-regulated in emphysema. Also, SERPINF1 (an antiprotease) was elevated in severe emphysema and this acts as an angiogenesis inhibitor by inducing endothelial cell apoptosis [38]. Therefore, perhaps changes in endothelial gene expression might underlie the development of emphysema. Another gene which might also be of importance is xanthine oxidase, a ROS-generating enzyme. Transcription of this gene is higher in vitro in pulmonary microvascular endothelial cells which are exposed to tobacco smoke condensate (TSC). The increased oxidative stress in response to xanthine oxidase upregulation could result in direct cell damage and apoptosis [39].

In conclusion, vascular atrophy due to endothelial apoptosis may contribute to the destruction of alveoli and consequently the development of emphysema. It may therefore be an important mechanism of COPD pathogenesis in this subgroup of patients.

Senescent cells are unable to progress through the cell cycle and divide, but remain metabolically active [40]. Senescence occurs due to telomere shortening (replicative senescence) and other, telomere-independent signals such as DNA damage or oxidative stress [41]. Replicative senescence is mediated via the p21 pathway and telomere-independent signals via the p16 pathway [42]. Lung tissue from COPD patients has an increased percentage of senescent endothelial cells and cultured pulmonary endothelial cells develop replicative senescence earlier [42]. Pulmonary endothelial cells from COPD patients have reduced telomerase activity, shorter telomeres and higher p21 and p16 levels earlier than similar cells from control subjects [42]. Oxidative stress and cigarette smoke both appear to be important in the pathogenesis of senescence in other pulmonary cells and thus are likely causes of senescence in the pulmonary endothelium [41, 43]. Senescent pulmonary endothelial cells also release more inflammatory markers and therefore are important in the increased level of inflammation seen in COPD patients [42].

A reduction in endothelium in patients with emphysema may be caused by reduced levels of Vascular Endothelial Growth Factor (VEGF) [44]. VEGF is a highly specific growth factor for endothelial cells that is produced in response to hypoxia [45]. It induces both cell proliferation and migration and prevents endothelial cell apoptosis [45]. VEGF levels might be reduced in such patients as Hypoxia Inducible Factor-1α (HIF-1α), a major transcription factor of VEGF, is also reduced in patients with emphysema. HIF-1α mediates cellular and systemic responses to hypoxia and binds to the hypoxia responsive element (HRE) on VEGF [46, 47]. Levels of HIF-1α and VEGF may be related to disease severity: both are correlated with FEV1 percentage predicted in patients with emphysema [46]. In addition, other studies have demonstrated that microRNAs -199a-5p and -34a (small non-coding RNAs that regulate gene expression) were both increased in lung tissue from COPD patients in comparison with controls. Transfection of human microvascular endothelial cells with these microRNAs resulted in decreased HIF-1α expression suggesting that epigenetic changes in COPD may also be important in COPD development [48].

Interestingly, similar studies looking at the expression of HIF-1α and VEGF in patients with chronic bronchitis (rather than emphysema) have shown HIF-1α and VEGF are increased in this patient group [49]. This suggests that the endothelium might be involved in different ways depending on the clinical presentation of COPD. Kanazawa et al. similarly found increased VEGF in sputum of patients with chronic bronchitis but decreased levels in patients with emphysema. There was a negative correlation between FEV1 and VEGF in the chronic bronchitis group but a positive association between VEGF and gas transfer (DLCO) in the emphysema group [50]. It is possible that the increased VEGF increases bronchial vascularity and leakage of plasma proteins resulting in airway narrowing in the chronic bronchitis group. However, increased VEGF might also prevent endothelial apoptosis and emphysema thus preserving gas exchange [51].

There is also evidence that vasculature may be altered in the airways of patients with COPD, in addition to their peripheral lung tissue – several groups have shown an increased vascular area in the airways of patients with COPD [52, 53]. It is possible that this might contribute to airway narrowing [53].

Patients with asthma also demonstrate increased VEGF expression in a similar way to chronic bronchitis patients. For example, VEGF mRNA levels are increased in endobronchial biopsies in patients with asthma in comparison to normal controls [54]. VEGF is also increased in induced sputum specimens in asthmatic patients and is negatively correlated with FEV1 [55]. Certain polymorphisms of VEGF (such as rs4711750 and rs3025038) also appear to confer an increased risk of asthma and are related to lung function [56, 57]. It is possible that this is due to different ratios of active and inhibitory isoforms of VEGF.

Again, in a similar way to chronic bronchitis, patients with asthma have increased vasculature in their airways which might be important in the development of airway narrowing [53]. Bronchial biopsies in asthma patients consist of more vessels than control patients [58]. The levels of vessels are related to disease severity suggesting that vascular remodelling increases as asthma severity worsens. These vessels are also associated with marked eosinophil recruitment [58]. As increased vasculature is seen in the airways of patients with mild asthma it is possible that vascular remodelling may be important in the early development of the disease [59]. Animal models suggest that the vascular remodelling may be in response to allergen exposure [60].

Endothelial progenitor cells (EPCs) act to repair endothelial injury and replace dysfunctional endothelium after being mobilized from the bone marrow to circulating blood [96]. Therefore, circulating EPCs provide a way to monitor endothelial damage. Animal studies have demonstrated that EPC levels are increased in rat models of emphysema in comparison to controls [96]. Conversely, in human studies, COPD patients appear to have reduced numbers of EPCs compared to controls [97]. When you look at COPD patients in isolation, however, patients with worse endothelial function have greater number of EPCs suggesting that vascular damage in these patients is stimulating the release of EPCs from the bone marrow [98]. It is possible that the endothelial function and EPC release may follow different pathways in COPD patients as similar studies in healthy controls show positive correlations between EPCs and FMD [99].

There is also evidence that EPCs do not function normally in patients with COPD. EPCs isolated from COPD patients had reduced proliferation rates and formed fewer clusters in vitro compared to control patients. EPCs from COPD patients had reduced chemotaxis levels and were less able to form tubular structures (in Matrigel angiogenesis studies) than control EPCs suggesting that their ability to repair endothelium was reduced. Expression of platelet/endothelial cell adhesion molecule-1 (PECAM-1), an adhesion molecule necessary for endothelial migration and junctional integrity on the surface of EPCs was also reduced in COPD patients further suggesting their dysfunctional nature. This was supported in animal studies by the same group who showed that fewer EPCs from COPD patients attached to injured arterial intima in mice compared to controls [100]. Human studies also support the dysfunctional nature of EPCs in COPD. One study looked at the levels of EPCs in patients with and without COPD before and after lung resection surgery. In control patients EPCs increased after surgery suggesting a normal response to injury. However, in COPD patients there was not an increase in EPCs suggesting the mobilization capacity of EPCs in COPD patients is reduced [101]. One explanation for the dysfunctional capacity of EPCs in COPD patients is the increased level of beta-2 adrenergic receptors (β2ARs) on EPCs from COPD patients. The increased β2AR level appears to alter the migration and proliferation seen in these cells [102].

In contrast to COPD EPCs appear to be upregulated in patients with asthma [103]. EPCs derived from patients with asthma also appear to have a higher proliferative capacity and an increased ability to form tubular structures in vitro in comparison to controls [103]. Recruitment of EPCs appears to be related to allergen challenge: in murine models of asthma circulating EPC levels were increased in response to an allergen challenge and mobilised to the lungs [103]. Vessel density in the lungs was also increased within 48 h of the challenge suggesting that EPC recruitment is important for the increased vascularity seen in asthma [103]. Similar findings have been seen in humans. On exposure to inhaled allergens, asthma patients show increased EPC mobilisation from the bone marrow [104]. Another study also demonstrated increased EPC levels in sputum 24 h after an allergen challenge. There was also an associated increase in the number and diameter of blood vessels in lung biopsy specimens further supporting the hypothesis than EPC mobilisation is important to the development of increased vascularity in lung tissue in asthma [105].

Smoking is known to have widespread effects on the microcirculation and may result in microvascular disease in various organs such as the eye, heart and kidney [106]. It has therefore been postulated that a similar process might occur in the lungs in COPD. In support of this, studies using magnetic resonance imaging (MRI) have demonstrated that pulmonary microvascular blood flow appears to be reduced in COPD [107]. These changes were apparent in patients with mild COPD and were worse in patients with severe COPD. This suggests that microvascular disease may represent an early part of the development of COPD and potentially is important in driving the progression of COPD to more severe disease. There is evidence that pulmonary perfusion is associated with the number of small pulmonary vessels present [108] and, in a similar fashion to perfusion, the number of small vessels appear to be reduced in the COPD lung. Studies have shown that the percentage of vessels less than 5mm (%CSA <5) on CT scans appears to be reduced in patients with emphysema and is related to disease severity [109, 110]. In a similar way to microvascular blood flow, reduction in %CSA < 5 also occurs in patients with mild disease suggesting that this may have a role in the pathogenesis of emphysema [110]. %CSA < 5 also appears to be associated with exacerbations of COPD: patients with a history of exacerbations have a significantly lower %CSA < 5 in comparison to patients without exacerbations [111]. Perhaps exacerbations result in tissue and vessel damage resulting in reduced %CSA < 5 although it is possible that a low %CSA < 5 is a risk factor for the development of exacerbations [111]. The underlying mechanism for this is not known and would require further study. The potential mechanisms for the development of small vessel disease in COPD have been investigated recently in vitro. One team looked at the expression of Krüppel-like factor 5 (KLF5) in the small pulmonary vessels in COPD patients. KLF5 is a zinc-finger transcription factor which plays a role in the vascular remodelling seen in cardiovascular diseases. In a similar way to cardiovascular disease KLF5 expression was increased in COPD pulmonary vessels suggesting a possible role in the small vessel disease seen in these patients [112].

Angiopoietins and their receptors (Tie-1 and Tie-2) are involved in the late phases of angiogenesis. Angiopoietin 1 (Ang-1) is proangiogenic and acts in the development of vascular networks [113]. Ang-2 appears to be an antagonist of Ang-1 and Tie-2, but can also act to enhance the proliferation and migration of endothelial cells [114, 115]. Studies have shown that the levels of angiopoietins appear to be altered in patients with COPD. For example, García-Lucio et al. showed that the expression of Ang-2 in pulmonary arteries appeared to be increased in COPD patients in comparison to healthy smokers [116]. Similarly, Bessa et al. showed increased Ang-2 levels in patients with COPD in induced sputum [117]. Ang-2 was also associated with vascular permeability in COPD patients suggesting that the increase in Ang-2 might stimulate leakage from vessels in COPD [117]. Interestingly, the levels of Ang-2 in the blood appear to be higher in patients with moderate COPD rather than severe COPD [113]. This could possibly suggest that an increase in Ang-2 might be important in the early stages of vascular remodelling in COPD, but not in later phases of the disease when changes in vasculature are already established. However, it is possible that levels of Ang-2 fluctuate in patients with COPD during exacerbations. Nikolakopoulou et al. demonstrated that serum Ang-2 levels are increased at the onset of COPD exacerbations and correlated with C-reactive protein (CRP) levels. The levels of Ang-2 decreased after a week of treatment. Patients with poorer outcomes also had significantly higher Ang-2 levels [118]. This suggests that Ang-2 might be a useful biomarker for COPD exacerbations and might help clinicians identify patients at risk of worse outcomes at the start of an exacerbation. Supporting this finding, another group found that blood levels of Ang-1 are reduced during COPD exacerbations and increased when patients are clinically stable [119]. As Ang-2 acts as an Ang-1 antagonist perhaps increased Ang-2 levels in these patients resulted in a reduction of Ang-1.

Angiopoietins also appear to play a role in asthma. Levels of both Ang-1 and Ang-2 in sputum are increased in stable asthmatic patients in comparison to controls. Smoking also increased angiopoietin levels in the asthma group [120]. Angiopoietins appear to be related to disease severity in asthma. Sputum levels of both Ang-1 and Ang-2 are significantly increased in severe refractory asthma patients in comparison to patients with moderate asthma [121]. This might suggest that they are important in the vascular remodelling seen in asthmatic patients. Similarly, serum Ang-1 and Ang-2 levels are also increased in asthma patients in comparison to healthy controls. However, only serum Ang-2 appears to be related to disease severity and is higher in refractory asthma cases. Serum Ang-2 (but not Ang-1) also correlates with parameters of severe asthma including number of exacerbations, emergency medical trips and number of hospitalizations [122]. Serum Ang-2 is positively correlated with exercise-induced bronchoconstriction whereas Ang-1 does not show this association. This suggests that Ang-2 might have a more important role in remodelling and disease pathogenesis seen in asthma and could potentially have use as a biomarker for severe asthma [123]. In a similar way to COPD, angiopoietins also appear to be altered acutely during asthma exacerbations. Lee et al. demonstrated that plasma Ang-2 levels were increased during exacerbations in comparison to patients with stable asthma whereas Ang-1 levels were lower during exacerbations. Ang-2 levels also correlated with the level of eosinophils and neutrophils, two important inflammatory cells involved in asthma [124].

In conclusion, the endothelium appears to behave in an abnormal fashion in COPD and asthma. Multiple pathways involving the endothelium may have importance in both conditions. In some cases such as TEM, pathways appear to behave in a similar fashion in COPD and asthma. However, other pathways behave differently depending on the phenotype of the patient. For example, increased angiogenesis in asthma/chronic bronchitis and decreased angiogenesis in emphysema. A summary of the mechanisms involved in endothelial dysfunction in COPD and asthma can be seen in Fig. 2.

There is a lack of evidence for why patients with asthma might develop endothelial dysfunction. However, one study demonstrated that vascularity in bronchial biopsy specimens was reduced in asthma patients after a 6 month course of inhaled corticosteroids (ICS) [131]. This suggests that inflammation may be an important cause of the increased vascularity and vascular remodelling seen in these patients. These findings were repeated in another study which also demonstrated that VEGF levels reduced after a 6 month ICS trial thereby suggesting that inflammation in asthma might result in increased airway vascularity through upregulation of VEGF [132].

As previously mentioned, targeting the inflammation seen in asthma using ICS can reduce VEGF expression and airway vascularity [131, 132]. Animal studies also support targeting VEGF as a potential treatment strategy in asthma. For example, two inhibitors of VEGF receptor were given to a murine asthma model. Mice who received the inhibitors demonstrated reduced VEGF airway levels, reduced airway inflammation and reduced AHR [142]. This suggests that targeting VEGF might be useful for both treating underlying vascular remodelling and symptoms in asthma. Another animal study also demonstrates that reducing angiogenesis in asthma might be effective. Vascular endothelial (VE)-cadherin antibodies were given to a mouse model of asthma. VE-cadherin is an endothelial adhesion molecule important in angiogenesis. Delivery of VE-cadherin antibodies reduced angiogenesis in the mouse model, but also reduced IgE production and eosinophil airway infiltration both hallmarks of inflammation seen in asthma [143]. AHR was also reduced in response to the antibody [143].

Another potential treatment option is to target EPCs in asthma. When a chemokine receptor antagonist (AMD3100) was given to mice sensitized to ovalbumin airway pulmonary EPCs, eosinophil accumulation, vascularity and AHR was reduced [144]. However, when AMD3100 was given to mice with established lung disease although the drug reduced EPCs, eosinophil levels and vascularity, AHR was only partially reversed [144]. This shows that it is likely to be important to reduce EPC accumulation early in asthma to prevent established airway obstruction.

Finally, reducing TEM may be another treatment option for patients with asthma. Theophylline (a methylxanthine drug with bronchodilator and anti-inflammatory properties) was added to HUVECs in vitro. This reduced adhesion of eosinophils to the HUVECs and also reduced the expression of endothelial adhesion molecules such as ICAM-1 [145]. This suggests that theophylline (which is already in use in asthma) may have beneficial effects by reduced TEM of eosinophils and consequently inflammation in asthma. A similar in vitro study was also performed by exposing HUVECs to another drug used in asthma: montelukast (a cysteinyl LT1-receptor anatagonist). Montelukast also reduced eosinophil transmigration across HUVECs and may therefore partially act by reducing TEM of eosinophils in patients [146]. There is also evidence that new drugs may be of use by targeting TEM. VUF-K-8788 (a histamine H1 antagonist) reduced eosinophil adherence to HUVEC in vitro. The same drug also reduced pulmonary eosinophil accumulation and inflammation (such as perivascular oedema) in a guinea pig model of asthma [147]. A summary of potential endothelial based treatments in COPD can be seen in Table 2.