Airway and vascular remodeling may play a prominent role in the clinical severity of severe refractory asthma (SRA). Angiopoietin-1 (Ang-1) is an essential mediator of angiogenesis by establishing vascular integrity, whereas angiopoietin-2 (Ang-2) acts as its natural inhibitor.
We aimed to determine the levels of angiopoietins in sputum supernatants of patients with SRA and to investigate the possible associations with mediators and cells involved in both the inflammatory and the vascular remodeling processes.
Thirty-eight patients with SRA, 35 patients with moderate asthma, and 20 healthy subjects were studied. All participants underwent lung function tests, bronchial hyperresponsiveness assessment and sputum induction for cell count identification and Ang-1, Ang-2, VEGF, TGF-β1, Cys-LTs, MMP-2, IL-13, ECP, and IL-8 measurement in supernatants. Airway vascular permeability (AVP) index was also assessed.
Ang-1 (ng/ml) and Ang-2 (pg/ml) levels were significantly elevated in patients with SRA compared with patients with moderate asthma and control subjects [median, interquartile ranges: 30 (17–39) vs 7.5 (5–11) vs 4.7 (3.8–5.9) respectively, P < 0.001; and 506 (400–700) vs 190 (146–236) vs 96 (89–120) respectively, P < 0.001]. Regression analysis showed a significant positive association between Ang-2 and AVP index, MMP-2, Ang-1, and VEGF in SRA. A weak association was also observed between Ang-1 and sputum eosinophils% in SRA.
Our results indicate that both angiopoietins levels are higher in SRA compared with moderate asthma and healthy subjects. In SRA, Ang-2 is associated with mediators involved in both the inflammatory and the vascular remodeling processes.
Severe refractory asthma (SRA) is a heterogeneous disorder that is characterized by persistent asthma symptoms despite treatment with high-dose inhaled steroids (ICS) . Increased angiogenesis has been observed in SRA , and along with other components of tissue repair and airway inflammation, they may account for the poor steroid response in SRA .
Angiogenesis is a complex multiphase process, potentially involving a great number of growth factors, cytokines, chemokines, and numerous other mediators but the specific role of each molecule has not been clearly defined . Vascular endothelial growth factor (VEGF) is considered to be the most important angiogenetic factor [5, 6], which induces vascular endothelial cell proliferation, tubule formation and increases microvascular permeability . The latter is a common feature of vascular remodeling in asthma and is modulated by the release of different inflammatory mediators, cytokines, proteases, and growth factors [4, 8].
Angiopoietins 1 and 2 (Ang-1, Ang-2) are both ligands for the endothelial cell-specific Tie2 surface receptor and may act in a complimentary and coordinated manner along with VEGF in airway microvascular process . Ang-1 is known to promote sprouting and stabilizing of nascent vessels by promoting interactions between endothelial cells and surrounding support cells including pericytes . In contrast, Ang-2 acts as a natural antagonist of Ang-1 that competes for Tie-2 receptor and reduces vascular integrity, leading subsequently to increased vascular permeability and mucosal edema . Data derived from an animal study suggest that Ang-1 protects against airway inflammation and hyperreactivity in asthma . Ang-1 is increased in sputum supernatants of steroid-naive patients with asthma, is inversely related to the airway vascular permeability (AVP) index, and is modulated by inhaled corticosteroids (ICS) . A similar increase is also observed for Ang-2, which is positively associated with AVP index and is modulated by leukotriene receptor antagonists .
In the present study, we aimed to determine the levels of angiopoietins in sputum supernatants of patients with SRA and to investigate possible associations with mediators and cells involved in both the inflammatory and the vascular remodeling processes.
Patients were recruited from an open cohort of patients with asthma who have been followed up for at least 2 years in the asthma clinics of the 1st and 2nd Respiratory Medicine Departments of the University of Athens. The diagnosis of asthma was established according to GINA guidelines . The diagnosis of SRA was established according to the American Thoracic Society (ATS) criteria . In the initial phase of recruitment, we included 105 patients. Twelve patients did not produce adequate sputum samples for processing. Thirty-eight patients with SRA were finally recruited. Thirty-five patients with moderate asthma, diagnosed according to the classification of GINA 2005, well controlled under treatment, were also included . All patients with asthma were never smokers. Twenty healthy, nonatopic, nonsmoking subjects served as controls. Patients with any other respiratory disease or any concomitant malignant, heart, renal, liver or collagen disease as well as subjects with a respiratory tract infection or asthma exacerbation in the past 8 weeks prior to admission were excluded. Ethics committees of both hospitals approved the study, and all participants provided written informed consent.
Sputum was induced as previously described , using all the appropriate safety modifications according to the underlying asthma severity . A sample was considered adequate when the patient was able to expectorate at least 2 ml of sputum and the slides contained <10% squamous cells on differential cell counting. The sample was divided into two portions. The first one was processed using selected plugs, as previously described, for inflammatory cell identification and VEGF, TGF-β1, Cys-LTs, MMP-2, IL-13, ECP, IL-8 and albumin measurement in sputum supernatant.. Dithiothreitol (DTT) was added in a volume equal to four times the weight of the sputum specimen, and it was further diluted with phosphate-buffered saline (PBS) in a volume equal to the sputum plus DTT. At least 500 inflammatory cells were counted in each sample. Total cell count was expressed as the number of cells ×106. Sputum inflammatory cells were expressed as (%) differential counts and as the absolute count (×106/ml). The second portion was diluted with PBS without DTT, as previously described , for Ang-1 and Ang-2 measurements. Both sputum supernatants were kept at −70°C for further measurement of the mediators mentioned previously.
Forced expiratory volume in one-second (FEV1), forced vital capacity (FVC), FEV1/FVC ratio, static volumes [functional residual capacity (FRC)], and diffusion capacity (DLCO) were measured using the Master Screen Body system (Viasys Healthcare, Jaeger, Hoechberg, Germany) according to the ATS guidelines .
Bronchial hyperresponsiveness (BHR)
Bronchial hyperresponsiveness was measured as PD15 of methacholine using a commercially available system (APS; Viasys Healthcare, Jaeger), according to the ATS guidelines . In 17 of 38 patients with SRA, we did not perform a BHR assessment as they had FEV1% predicted values <60% (cutoff value as a relative contraindication). For the remaining 21, long-acting β2-agonists (LABA) were discontinued for 48 h prior to BHR testing and short-acting bronchodilators were used on demand until 8 h before the procedure.
FeNO was measured using a portable NO analyzer (NIOX MINO, airway inflammation monitor; Aerocrine, Solna, Sweden) as previously described .
A positive skin prick test to any of twenty common aeroallergens (including mites, grasses, trees, fungus, and domestic animals) was used to confirm atopy.
ECP was measured using Unicap ECP kit (detection limit 0.5ng/ml; Pharmacia diagnostics, Uppsala, Sweden). VEGF, IL-13, MMP-2 [free], and IL-8 were measured using enzyme-linked immunosorbent assay kits (ELISA; R&D systems, Minneapolis, MN, USA) with detection limits of 9 pg/ml, 32 pg/ml and 0.047 ng/mL and 3.5pg/ml, respectively. Cys-LTs were measured using an ELISA kit (detection limit 13 pg/ml; Cayman Chemical, Ann Arbor, MI, USA). TGF-β1 was measured by ELISA (detection limit 50 pg/ml; R&D systems). Levels of human Ang-1 and Ang-2 were assessed in sputum supernatant samples by ELISA according to the manufacturers’ instructions (R&D Systems). The minimum detectable dose of Ang-1 and Ang-2 ranged from 1.36 to 10.3 and 1.20 to 21.3 pg/ml, respectively. Blood was drawn for determination of serum albumin. Albumin was measured in serum and sputum supernatants by laser nephelometry. The AVP index was calculated as the ratio of albumin concentrations in induced sputum and serum . All values were expressed at pg/ml except Ang-1, which was expressed as ng/ml. For both Ang-1 and Ang-2, the intra- and interassay variability were 4 and 6.5% respectively.
Normally distributed data are presented as mean ± standard deviation (SD), whereas skewed data are presented as median (interquartile ranges). Normality of distribution was checked with Kolmogorov–Smirnov test. Statistical comparisons between groups were performed with one-way analysis of variance (anova) for normally distributed data and Kruskal–Wallis tests for skewed data, accompanied by appropriate post hoc tests for multiple comparisons (Bonferroni and Dunn's, respectively). Differences in numerical variables between two groups were evaluated with unpaired t-tests and Mann–Whitney U-tests for normally and skewed data, respectively, whereas comparisons of proportions were performed using chi-square tests. For participants’ characteristics and inflammatory variables, and to avoid an erroneous significant effect (purely by random chance), we further used the Bonferroni's method by simply dividing our desired P-value by the number of variables being conducted.
To examine the associations between Ang-1, Ang-2, mediators (ECP, IL-8, Cys-LTs, IL-13, VEGF, TGF-β1, MMP-2), sputum cells, and lung function tests, linear regression analysis was performed using Ang-1 and Ang-2 as dependent variables. All linear regressions were performed after proper adjustments for age, gender, body mass index, atopy, duration of the disease, and treatment regimens. Data were interpreted as standardized coefficients with 95% confidence intervals. A P-value <0.05 (2-sided) was considered significant. After Bonferroni correction for subjects’ characteristics, P was considered significant for values <0.0038, while for inflammatory variables it was considered significant for values <0.0033. Statistical analysis was performed using spss 16.0 (SPSS, Chicago, IL, USA) and Graph Pad Prism 5 (GraphPad Software, San Diego, CA, USA).
The demographic characteristics of the study participants are summarized in Table 1.
Table 1. Demographic characteristics of study participants
(n = 38)
(n = 35)
(n = 20)
Normally distributed data are presented as mean ± standard deviation (SD), whereas skewed data are presented as median (interquartile ranges). Bold numbers indicate significant differences, all in favor of SRA.
BMI, body mass index; FeNO, fraction of exhaled nitric oxide; BHR, bronchial hyperresponsiveness; FEV1, forced expiratory volume in one-second; PB FEV1, forced expiratory volume in one-second Postbrochodilation; FVC, forced vital capacity; DLCO, diffusion capacity; FRC, functional residual capacity; ICS, inhaled corticosteroids; LABA, long-acting β2 agonists; CS, corticosteroids; LTRA, leukotriene receptor antagonists; SRA, severe refractory asthma; n/a, not applicable; pred., predicted.
>1200 μg budesonide per day or equivalent.
≤800 μg budesonide per day or equivalent.
Nineteen were receiving 5 mg prednisolone per day while six were receiving 7.5 mg prednisolone per day.
Statistically significant difference compared with healthy subjects.
Statistically significant difference compared with moderate asthma.
Statistically significant difference compared with healthy subjects. P-values indicate differences across the three groups.
The statistical significance did not occur after Bonferroni correction.
Inflammatory variables in induced sputum according to asthma severity
Patients’ inflammatory variables are summarized in Table 2.
Table 2. Inflammatory variables of study participants
SRA&!break; (n = 38)
Moderate asthma&!break; (n = 35)
Controls&!break; (n = 20)
Normally distributed data are presented as mean ± standard deviation (SD), whereas skewed data are presented as median (interquartile ranges). Bold numbers indicate significant differences in favor of SRA.
Ang-1 (ng/ml) and Ang-2 (pg/ml) levels differed significantly among patients with SRA, patients with moderate asthma and healthy controls [30.0 (17.0–39.0) vs 7.5 (5–11) vs 4.7 (3.8–5.9) P < 0.001 and 506 (400–700) vs 190 (146–236) vs 96 (89–120), respectively P < 0.001; Table 2, Figs 1 and 2. Patients with SRA had higher Ang-1 and Ang-2 levels compared with moderate asthma ones (P < 0.001 for both comparisons, Table 2, Figs 1 and 2). Patients with moderate asthma had higher Ang-1 and Ang-2 levels compared with healthy subjects (P < 0.001 for both comparisons, Table 2, Figs 1 and 2). Ang-1 and Ang-2 levels in SRA patients treated with oral steroids (n = 25, age 54 ± 14 years) did not differ significantly from those not receiving oral steroids (n = 13, age = 53 ± 15 years), [29.5 (11–39) vs 33 (18–43), P = 0.579; and 498 (393–588) vs 623 (488–738) P = 0.097, respectively]. The AVP index was significantly higher in the SRA group compared with patients with moderate asthma and healthy controls [0.041 (0.030–0.051) vs 0.028 (0.020–0.034) vs 0.015 (0.011–0.002), respectively, P < 0.001; Table 2].
Mediators involved in the inflammatory/vascular remodeling processes were significantly higher in patients with SRA compared with the other two groups (Table 2). All the aforementioned mediators were significantly higher in patients with moderate asthma compared with healthy subjects (all comparisons with P < 0.05, Table 2).
Associations of Ang-1 and Ang-2
After already-mentioned proper adjustments, Ang-1 levels presented a significant positive association with Ang-2 levels. A weak association was observed between Ang-1 and sputum eosinophils (either absolute count × 106/ml or% differential counts) in patients with SRA (Table 3). No other significant associations were observed. Ang-2 levels were positively associated with Ang-1, AVP, VEGF, MMP-2, and TGF-β1 levels in patients with SRA (Table 4). No other significant associations were observed. Forward stepwise regression analysis showed that VEGF presented the strongest association with Ang-2 (R2 0.6, P < 0.001).
Table 3. Multivariate regression analysis (in one model) between Ang-1, inflammatory cells, mediators, and lung function tests in patients with SRA
Beta standardized coefficient (95% CI)
Regression analysis was performed after proper adjustments for age, gender, BMI, duration of the disease, atopy, and treatment regimens. Bold numbers indicate significant association.
There were no significant associations between Ang-1 or Ang-2 levels and mediators related to inflammatory/vascular remodeling or sputum inflammatory cells in patients with moderate asthma and healthy subjects (data not shown). No significant associations were observed between either Ang-1 or Ang-2 levels and both BHR and FeNO levels in all asthma groups (Tables 3 and 4 for SRA). Finally, no significant associations were observed between both Ang-1 or Ang-2 levels and lung function test variables in all study groups (Tables 3 and 4 for SRA). In a regression analysis that was performed in all study groups, a positive association between VEGF and AVP index was observed only in SRA (R2 = 0.47, P = 0.005).
In the present study, we have shown that both Ang-1 and Ang-2 levels in sputum supernatants are significantly higher in patients with SRA compared with patients with moderate asthma and healthy subjects. In patients with SRA, Ang-1 was positively associated with Ang-2 levels, whereas Ang-2 was positively associated with mediators involved in the ongoing inflammatory and vascular remodeling processes (VEGF, MMP-2, IL-13, and TGF-β1) as well as with vascular permeability, expressed by the AVP index. These associations were not observed in patients with moderate asthma. This is the first study to our knowledge that evaluated the possible role of angiopoietins 1 and 2 in patients with SRA.
Angiopoietins may play a role in vascular remodeling of asthmatic airways. In particular, Ang-2 levels can be transcriptionally and post-transcriptionally regulated by hypoxia or exposure to growth factors, such as VEGF . In our study, both Ang-2 and VEGF were increased in SRA, with a significant positive association between the two mediators. Furthermore, positive associations between Ang-2, VEGF, and the AVP index were also observed in the SRA group. Therefore, the high levels of VEGF and Ang-2 observed in our patients with SRA may indicate that blood vessels in their airways are in a hypervascularized and destabilized state, thus contributing to the up-regulation of the AVP process. The absence of such associations for Ang-2 in moderate asthma suggests a possible absence of adequate levels of either VEGF and/or Ang-2 to exert their aforementioned effects.
Ang-1 is an essential mediator of angiogenesis, which is widely expressed in normal adult tissues, resulting in the stabilization of nascent vessels and making them leak resistant . A previous study has found increased levels of Ang-1 in the induced sputum of nonoptimally treated patients with asthma compared with healthy subjects . In the same study, an inverse correlation between Ang-1 and the AVP index was observed, indicating that Ang-1 may protect the vasculature against plasma leakage induced by VEGF . Another study using an animal model of asthma supports a protective role for Ang-1 in vasculature and airway inflammation . In our study, Ang-1 levels were elevated in patients with SRA but there was no association between its levels and the levels of mediators expressing airway inflammation and remodeling. Our study did not reproduce the inverse correlation between Ang-1 and AVP index that has been previously observed, probably due to the different characteristics of the patients with asthma studied (SRA treated with steroids vs mild steroid-naïve patients). The positive association between Ang-1 and Ang-2 may suggest that both of them act in a coordinated manner, although inadequate to ameliorate both the vascular leakage and the airway inflammatory and remodeling processes. The positive association between Ang-1 and sputum eosinophils is a marginal one and is influenced by a few outliers. However, previous data suggest that Ang-1 enhances eosinophil migration in a dose-dependent manner .
In this study, Ang-2 presented positive correlations with TGF-β1, MMP-2, and IL-13, which are involved in the vascular remodeling in asthma. Increased sputum levels of MMPs in patients with asthma are associated with sputum VEGF levels . TGF-β1 and the Th2-cytokine IL-13 participate in various aspects of the remodeling process in asthma . Their positive correlation with Ang-2 may suggest that they are also involved in the angiogenic process by lengthening the existing vessels and sprouting the formation of new ones, either directly through VEGF-mediated pathways  or through secretory-dependent cells implicated in vascular inflammation and remodeling .
Inhaled steroids are currently the only treatment that may positively affect the main aspects of the vascular component of airway inflammation and remodeling [7, 27]. It is important to point out that this effect seems to be mainly mediated by the reduced expression of VEGF . Our study was not designed to evaluate the possible role of steroid intervention in SRA. However, the fact that our patients with SRA were receiving high doses of ICS implies that steroid treatment cannot suppress all the aspects of airway inflammation and remodeling in such patients or even more that the underlying pathophysiology of neo-angiogenesis in SRA may be resistant to the intervention with ICS.
A possible bias in this study may be the use of a dilution factor like DTT. There is much concern regarding the effect of DTT on the concentrations of different mediators although the majority of the published studies used the aforementioned methodology as part of their processing procedure. However, in this study we tried to eliminate the aforementioned issue by processing the samples for both Ang-1 and Ang-2 without DTT, as was previously described in the literature .
In conclusion, this study indicates that Ang-1 and Ang-2 levels are higher in SRA compared with moderate asthma and healthy subjects. In patients with SRA, Ang-2 is associated with mediators involved in the vascular permeability process. Our results provide suggestions for possible mechanisms involving angiopoietins in the pathogenesis of SRA.
The study was funded by the National Kapodistrian University of Athens.
Conflict of interest
None of the authors has any funding support or any financial relationship with a biotechnology and/or pharmaceutical manufacturer to declare.
Drs Tseliou, Hillas, Matzouranis, and Emmanouil performed sputum induction and processing. Drs Bakakos, Kostikas and Loukides recruited patients and confirmed their eligibility for entering the study. They also supervised every non- and semi-invasive procedure. Dr Simoes measured all the mediators. Dr Alchanatis and Prof Papiris coordinated the study on behalf of the two departments. The manuscript was drafted by Drs Bakakos, Kostikas, and Loukides and was approved by all authors.