Pulmonary Vascular Remodeling Is an Early Feature of Fatal and Nonfatal Asthma

To the Editor:

Increased bronchial vascularization in patients with asthma is a well-documented feature of airway remodeling (1) that has been shown to correlate with airflow limitation, bronchial hyperresponsiveness, and disease severity (2, 3). It has been proposed that the increased bronchial vasculature may promote airway inflammation and remodeling in asthma through chronic inflammatory cell recruitment, pooling of inflammatory mediators, and abnormal cell growth and proliferation. It remains unknown whether both the bronchial and pulmonary vascular systems supplying the lung are remodeled and at what stage the vascular changes occur. The contribution of the current study is to quantitatively assess the structure and extracellular matrix composition (elastin and collagen) of distal pulmonary arteries and veins from pediatric and adult patients with asthma compared with healthy control subjects.

Methods

Nontransplantable donor lungs from pediatric and adult donors with asthma (n = 13, 7 fatal disease) or healthy control subjects (n = 12, Table 1) were obtained with informed consent from the next of kin through the International Institute for the Advancement of Medicine. The study was approved by the Research Ethics Board, University of British Columbia (H13-02173).

Table 1.

Demographics of Patients with Asthma and Healthy Control Subjects

	Control		Asthma
Demographics
Cases	12 (6 Ped, 6 adult)		13 (6 Ped, 7 adult)
Average age, yr	24.2 (13.7)		18.8 (8.7)
Sex, female, %	41.7		46.2
Height, cm	162.8 (19.8)		166.5 (10.2)
Weight, kg	70.3 (28.0)		63.6 (23.8)
Ethnicity, %
White	83.4		84.6
Hispanic	8.3		15.4
Others	8.3		—
Medications, %
Albuterol	—		76.9
Advair*	—		15.4
Singulair	—		7.7
Roxycodone*	—		7.7
Zyrtec	—		7.7
Cause of death, %
Asphyxiation	—		15.4
Anoxia seizure	25		—
CNS tumor	8.3		—
CVA	16.7		—
FAE	—		61.5
Head trauma	50		23.1
Morphometry
Number of cell nuclei in vessel walls	49.5 (13.1)		74.9 (25.4)†
NLOM-elastin
Vessel type	Arteries	Veins	Arteries	Veins
Intensity, a.u.	18.23 (43)	16.01 (1.33)	18.95 (1.20)	17.74 (0.83)
Entropy, a.u.	6.38 (91)	5.18 (37)	5.99 (74)	5.43 (28)

Definition of abbreviations: a.u. = arbitrary unit; CNS = central nervous system; CVA = cardiovascular attack; FAE = fatal asthma episode; NLOM = nonlinear optical microscopy; Ped = pediatric.

There was no statistical difference in age (P = 0.37) or sex (P = 0.58) between the two groups.

Demographic data are presented as mean (SD), and morphometry and NLOM-elastin data are presented as median (interquartile range).

^*Indicates medication combined with others.

^† Indicates P < 0.05.

Lungs were inflated with cryomatrix, frozen suspended, and sliced apex to base into contiguous 2-cm-thick transaxial slices as previously described (4). A uniform random sampling method was applied to select three 15 × 20-mm cylindrical tissue samples for formalin fixation and paraffin embedding. Six arteries and veins were randomly selected per case, and three regions of interest were imaged per vessel for the nonlinear optical microscopy (NLOM) analysis.

Formalin-fixed, paraffin-embedded tissue sections (5 μm) were stained with Verhoeff-van Gieson stain and digitally scanned using a ScanScopeXT slide scanner (Aperio Technologies) (Figures 1A–1D). As previously described (5), digital histological images were preprocessed with a band-pass filter to remove noise, and color segmentation was used to separate the tunica intima, media, and adventitia tissue structures. The binary masks were used to generate color-coded thickness distribution maps by fitting maximal circles to every point (Figure 1E). The mean, max, min, and variance for each thickness distribution map and vessel diameter was then obtained using the local thickness plugin in Image J software (version 1.47q; National Institutes of Health) (6).

Figure 1.

Representative images of (A) a distal pulmonary artery and (B) a vein from a healthy control subject and (C) a distal pulmonary artery and (D) a vein from a patient with asthma, stained with Verhoeff-van Gieson. The elastic laminae are shown in black, the muscular tunica media in light brown, and the collagen fibers of the tunica adventitia in red. After color segmentation, (E) binary masks (black rings) highlighting the tunica intima and media were used to generate color-coded thickness distribution maps by fitting maximal circles to every point. (F) The cross-sectional vessel area (mm²) of the vessels assessed. Based on the histograms generated from the thickness distribution maps, the median value of thickness for the (G) tunica intima and (H) adventitia were calculated for healthy control subjects (open bars) and patients with asthma (blue bars). Data for the thickness of the (I) tunica intima and (J) adventitia in patients with asthma (green bars) are also shown split by sex, pediatric and adult, and nonfatal and fatal disease. Representative nonlinear optical microscopy images of (K) a distal pulmonary artery and (L) a vein from a healthy control subject and (M) a distal pulmonary artery and (N) a vein from a patient with asthma. Elastic fibers and cells with high two-photon excitation fluorescence are shown in white and fibrillar collagen fibers with SHG are shown in blue. Using texture analysis, (O) fibrillar collagen signal intensity, (P) entropy of fibrillar collagen, (Q) fractal dimension, and (R) collagen fiber thickness are also shown for pulmonary arteries and veins of healthy control donors (open bars) and patients with asthma (blue bars). Data for (S) collagen signal intensity and (T) entropy of fibrillar collagen in patients with asthma (green bars) are shown split by sex, pediatric and adult, and nonfatal and fatal disease. Box plots represent the median [interquartile range]; white circles depict data from 12 healthy control subjects and blue circles depict data from 13 patients with asthma. Data are presented per case. Three vessels per donor and three regions of interest per vessel were assessed. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Scale bars, 50 μm. a.u.  = arbitrary unit; NLOM = nonlinear optical microscopy; ns = not significant; SHG = second-harmonic generation.

Adjacent tissue sections were stained with hematoxylin and imaged using NLOM to assess elastin with two-photon excitation fluorescence (TPEF) and collagen with second-harmonic generation as previously described (4). The signal intensity, entropy (a measure of structure organization), fractal dimension score (a measure of fiber integrity), and fiber thickness were calculated using a custom-texture analysis toolkit in MATLAB (4). The number of cell nuclei imaged with TPEF signal in vessels walls was calculated using Image J software (6).

A Shapiro-Wilk normality test was used to assess the data. Nonparametric Mann-Whitney U tests were performed using GraphPad-6.0. Linear mixed-effect models and bootstrapping power calculations were performed using the statistical software R version 3.5.0. Data are shown as the median [interquartile range]. P < 0.05 was considered significant; all Mann-Whitney U tests had greater than 0.90 power.

Results

There was no difference in the cross-sectional vessel area (mm²) for pulmonary arteries and veins between groups (Figure 1F). The thickness of the tunica intima (Figure 1G) and adventitia (Figure 1H), but not the tunica media (data not shown), of both pulmonary arteries and veins was increased in the lungs of patients with asthma compared with control lungs. In patients with asthma, there were no differences in the thickness of the tunica intima (Figure 1I) or adventitia (Figure 1J) in pulmonary vessels irrespective of age, sex, or fatal disease. Multiple-regression analysis verified asthma diagnosis was significantly associated with a thickened tunica intima (estimate, −1.21; P < 0.0013) and adventitia (estimate, −0.36; P < 0.0001) in pulmonary vessels and was not associated with age, sex, or fatal disease.

NLOM images demonstrate the TPEF signal for elastin (white) and second-harmonic generation signal for collagen (blue) in the walls of pulmonary arteries and veins (Figure 1K–N). Compared with control lungs, the pulmonary arteries and veins in lungs from patients with asthma had increased signal intensity for fibrillar collagen (Figure 1O), and the collagen fibers were more disorganized (greater entropy score, Figure 1P), fragmented (greater fractal dimension score, Figure 1Q), and thicker (Figure 1R). There were no differences in the signal intensity or organization (entropy) of elastin between groups (Table 1). In patients with asthma, there were no differences in the signal intensity of collagen (Figure 1S) or collagen disorganization (entropy score, Figure 1T) irrespective of age, sex, or fatal disease. Multiple-regression analysis verified that asthma diagnosis was significantly associated with higher collagen deposition (estimate, −2.10; P < 0.001) and disorganization (estimate, −0.22; P < 0.001) and was not associated with age, sex, or fatal disease.

The number of cell nuclei within pulmonary vessel walls was significantly greater in patients with asthma compared with healthy control subjects (Table 1).

Discussion

This report highlights that the distal pulmonary vasculature in pediatric and adult patients with asthma is remodeled with thickened walls and increased deposition of disorganized and fragmented collagen, compared with healthy control subjects. This finding is consistent with a previous study of the lamina propria in large and small airways, which also found disorganized and fragmented collagen in pediatric and adult patients with asthma (4). These data indicate that increased deposition of disorganized and fragmented collagen is a feature of remodeling that occurs in the wall of both airways and distal pulmonary vessels early in life, irrespective of sex or fatal disease.

Rydell-Tormanen and colleagues have previously reported persistent wall thickening of large and small pulmonary vessels in a house dust mite model of allergic inflammation in mice (7). Saetta and colleagues, in an autopsy study of six patients with fatal asthma, found an increase in the wall thickness of muscular pulmonary arteries adjacent to airways (8). Both studies demonstrated an increase in the number of smooth muscle cells and eosinophils in the pulmonary vasculature. In this report, increased numbers of cell nuclei were present in remodeled pulmonary vessels from patients with asthma, but future studies will be required to understand the cellular composition of remodeled pulmonary vessels in patients with asthma.

It has been proposed that “spillover” of airway inflammatory mediators including VEGF (vascular endothelial growth factor) and bFGF (basic fibroblast growth factor) may induce bronchial vasculature remodeling (9). However, in an acute allergic inflammation model, bronchial angiogenesis was demonstrated to occur in the lungs before the development of airway hyperresponsiveness or airway inflammation (10). In this report and in the study by Rydell-Tormanen and colleagues (7), the remodeled pulmonary vessels assessed were nonadjacent to the airway tree. Together, these studies suggest mechanisms other than the spillover of airway inflammation may be involved in vascular remodeling in asthma.

Remodeling of the pulmonary vasculature through the deposition of collagen and not elastin is an accepted feature of other inflammatory diseases, including chronic obstructive pulmonary disease (COPD) (11). In chronic obstructive pulmonary disease, the consequence of thickened pulmonary vascular walls is reduced distensibility of pulmonary vessels and pulmonary hypertension in end-stage disease. Several clinical studies and reviews have previously highlighted a potential link between severe allergic asthma and pulmonary hypertension (9). Recent studies in an allergic airway inflammation model have shown that pulmonary vascular remodeling is associated with the hyperreactivity of vessels to hypoxia, but not pulmonary hypertension (12). Larger validation studies will be required to understand the clinical outcomes of pulmonary vascular remodeling in asthma.

The use of computed tomography has recently shown that vascular pruning of blood vessels <10 mm² in cross-sectional area correlates with asthma severity, frequency of exacerbations, and lung function (13). In this study, microscopy enabled the resolution to assess distal pulmonary vessels 4–5 mm² in cross-sectional area, but one limitation of such two-dimensional studies is the ability to count vessels. Future studies involving three-dimensional, ultra-resolution, micro-computed tomographic imaging could be used to assess if vascular pruning continues into the distal pulmonary vasculature.

In conclusion, pulmonary vascular remodeling is a feature of asthma. Future studies are needed to understand the mechanisms driving pulmonary vascular remodeling, its contribution to disease pathophysiology, and how it may be targeted in asthma.