Published ahead of print on January 30, 2004, doi:10.1164/rccm.200308-1111OC
© 2004 American Thoracic Society
Heterogeneous Increase in CD34-positive Alveolar Capillaries in Idiopathic Pulmonary FibrosisDepartment of Respiratory Oncology and Molecular Medicine and Department of Thoracic Surgery, Institute of Development, Aging and Cancer, Tohoku University; and Department of Pathology, Tohoku University School of Medicine, Sendai, Japan Correspondence and requests for reprints should be addressed to Masahito Ebina, M.D., Ph.D., Department of Respiratory Oncology and Molecular Medicine, Institute of Development, Aging and Cancer, Tohoku University, 4-1 Seiryo, Aoba-ku, Sendai 980-8575, Japan. E-mail: ebinam{at}idac.tohoku.ac.jp
To elucidate the apparent contradictions in vascular remodeling in the lungs of patients with idiopathic pulmonary fibrosis, we evaluated alveolar vascularity in relation to the various degrees of fibrosis in surgically biopsied lungs of usual interstitial pneumonia. Alveolar capillary endothelial cells were intensely immunoreactive with CD34 but not with von Willebrand factor. Vascular density, that is, the relative ratio of capillary area to total area of alveolar walls, was significantly higher at low grades of fibrosis than in control lungs, whereas vascular density gradually decreased as the degree of fibrosis increased and was lower than that of control lungs in the most extensively fibrotic lesions. No vessels were observed inside fibroblastic foci. The potent angiogenic factors vascular endothelial growth factor and interleukin-8 were abundantly produced by capillary endothelial cells and alveolar epithelial cells in highly vascularized alveolar walls. In contrast, venules with CD34-negative but von Willebrand factor-positive endothelial cells localized in the center of the fibrotic lesions were slightly increased and identified as postcapillary venules by three-dimensional reconstructed images. These results indicate the presence of heterogeneous vascular remodeling in usual interstitial pneumonia.
Key Words: alveolar capillary • endothelial cells • idiopathic pulmonary fibrosis Idiopathic pulmonary fibrosis (IPF) is a progressive, diffuse parenchymal lung disease of unknown etiology with significant morbidity and mortality despite aggressive therapy. IPF is associated with the histologic appearance of usual interstitial pneumonia (UIP) on surgical lung biopsy (1). A strikingly heterogeneous and nonuniform fibrosing process is one of the most characteristic features of UIP, with alternating zones of fibrosis, honeycomb change, and intervening patches of normal lung (2). Although apoptosis of the alveolar epithelial cells has been reported to be involved in the pathogenesis of IPF (3), there is only limited information about the alveolar capillaries in the UIP lungs of patients with IPF. Regarding vascular remodeling in IPF, there are contradictory reports of it being highly vascularized (4–6) or less vascularized (7, 8). We hypothesized that these discrepancies were due to the temporal heterogeneity of the lesions in UIP lungs. Therefore, we evaluated vascular density within the alveolar walls with variegated fibrosis in surgically biopsied lungs of seven patients diagnosed with IPF/UIP. As a pilot study, we compared alveolar capillary endothelial cells according to their immunoreactivity for CD34, von Willebrand factor (vWF), and thrombomodulin (TM) because of the phenotypic heterogeneity of pulmonary vascular endothelial cells (9, 10). We confirmed first that alveolar capillary endothelial cells were intensely immunoreactive with CD34, but rarely with vWF. Thrombomodulin immunoreactivity of alveolar capillary endothelial cells was decreased around fibrotic lesions. We applied double immunohistochemical staining for CD34 and vWF (CD34/vWF) with counterstaining by elastica-Goldner stain (11), which clarified the relation between vascularity and the degree of fibrosis. We observed that the alveolar capillaries with CD34-positive endothelial cells were remarkably increased in the nonfibrotic lesions of UIP lungs. A subset of these endothelial cells was immunoreactive with Ki-67, a marker of proliferation (12). Alveolar Type II epithelial cells in close contact with increased capillary endothelial cells were intensely immunoreactive with the potent angiogenic factors vascular endothelial growth factor (VEGF) (13) and interleukin-8 (IL-8) (14). In contrast, CD34-positive capillaries were decreased in fibrotic lesions where the venules with vWF-positive endothelial cells were localized in the center. Three-dimensional images demonstrated that these vWF-positive venules connected CD34-positive alveolar capillaries with pulmonary veins in the fibrotic lesions (10). The possible role of these increased CD34-positive alveolar capillaries in the regeneration of alveolar septa in IPF is discussed.
For details, see the online supplement.
Lung Tissues Examined
Antibodies
Immunohistochemistry
Double Immunostaining
Quantitative Evaluation of Vascularity
Three-dimensional Reconstruction
Statistical Analysis
Heterogeneous Immunoreactivity of Endothelial Cells The lung specimens from seven patients with IPF had a variegated appearance, with alternating zones of fibrosis and intervening patches of normal lung at low magnification. Fibroblastic foci were easily observed in these specimens, consistent with the characteristic features of UIP. To clarify the involvement of endothelial cells in UIP lesions, we examined the immunoreactivity for three endothelial markers: CD34, von Willebrand factor (vWF), and thrombomodulin (15). We observed heterogeneous immunolocalization of CD34 and vWF in endothelial cells of these UIP lungs. To compare the heterogeneous immunolocalization of these endothelial cells among lesions with various degrees of fibrosis, the sections for immunohistochemistry were counterstained with elastica-Goldner stain, which showed collagen tissues in green and elastic fibers in dark purple (11). Almost all the alveolar capillary endothelial cells were intensely positive for CD34, and these were distributed densely in the nonfibrotic lesions (Figure 1A) . In contrast, endothelial cells positive for vWF were localized mainly in the larger vessels, and little immunoreactivity for vWF was observed in alveolar capillaries (Figure 1B). Immunoreactivity for the thrombomodulin of endothelial cells was also examined in consecutive sections as a positive control (10). Variable immunoreactivity for thrombomodulin was observed in the endothelial cells of both CD34-positive alveolar capillaries and vWF-positive pulmonary veins (Figure 1C). Double immunohistochemical staining for CD34 and vWF (CD34/vWF) clearly showed the heterogeneous distribution of CD34-positive and vWF-positive endothelial cells (Figure 1D). Fibroblastic foci were free from any endothelial cells immunoreactive for CD34, vWF, or thrombomodulin (Figure 1).
Remodeling of Alveolar Capillaries in UIP We compared the relation between capillary remodeling and the distribution of alveolar Type II epithelial cells producing surfactant protein A (SP-A) in various degrees of fibrosis in UIP. The small fibrous tissues arising from nonvascularized alveolar septa contained no capillaries inside (Figure 2A) . Double staining for CD34 and SP-A of consecutive sections revealed the close contact of alveolar Type II cells with alveolar capillaries (Figure 2B). No capillaries were distributed in the side of fibrous tissues without alveolar Type II cells (Figure 2B). This tendency was also shown in massive fibrotic lesions (Figure 2C). In fibrotic lesions, CD34-positive capillaries were distributed only on the edges of fibrotic lesions, where there were alveolar Type II cells. In the walls of honeycomb lesions, venules with vWF-positive endothelial cells were sparsely distributed but dominant compared with CD34-positive capillaries (Figure 2D).
Proliferation of CD34-positive Endothelial Cells and Distribution of Angiogenic Factors in UIP The proliferation of capillary endothelial cells was examined by double staining for CD34 and Ki-67, a cell proliferation-associated human nuclear antigen (12) (Figures 3A and 3B) . In contrast to the alveolar walls of normal control lungs (Figure 3A), alveolar capillaries in the nonfibrous lesions of UIP lungs were dilated and remarkably increased (Figure 3B). A subset of the endothelial cells of these increased alveolar capillaries was immunoreactive with Ki-67 (Figure 3B). The distribution of Ki-67–positive capillary endothelial cells in the area of a lung tissue section (cm2) was 2.6 ± 0.3 cells/cm2 in UIP lungs (n = 7) and 0 ± 0 cells/cm2 in control lungs (n = 3, p < 0.05). Augmented expression of both VEGF and IL-8 was detected in capillary endothelial cells and alveolar Type II epithelial cells in highly vascularized alveolar septa (Figures 3C–3F). In fibrotic lesions, fibroblasts and leukocytes were faintly immunoreactive with VEGF and IL-8 (Figures 3G and 3H), with few endothelial cells positive for Ki-67 (Figure 3I).
Reverse Correlation between Vascular Density and the Degree of Fibrosis in UIP To examine the heterogeneous remodeling of CD34-positive capillary endothelial cells in UIP by morphometric analysis, we quantified the vascular density of alveolar walls with various degrees of fibrosis, using an image analyzer system. Figure 4A shows a field of thickening alveolar septum in UIP lung tissues, monitored with a CAS 200 image analysis system (15). Vascular density was determined as the percentage of the total area surrounded by CD34-positive capillary endothelial cells per area of alveolar septa in each field, scored according to the degree of fibrosis from 1 (minimal) to 8 (severe) (17). The relation between vascular density and the fibrotic degree in each field of one patient with IPF/UIP is shown in Figure 4B. The vascular density of minimally fibrotic lesions, scored as 1 or 2, was 25 ± 2.6% (mean ± SEM, n = 10), and decreased gradually toward more severe fibrosis. In the moderately fibrotic area, scored as 5 or 6, vascular density was 11 ± 2.2% (n = 8), which was lower than that of minimally fibrotic lesions (p < 0.01). The vascular density of more severe lesions, scored as 7 or 8, was the lowest (6.8 ± 1.6%, n = 8, p < 0.01).
This tendency toward high vascularity in minimally fibrotic lesions and low vascularity in severely fibrotic lesions was also apparent when the mean values of vascular density at each fibrotic score of the seven patients with IPF were pooled (Figure 4C). The control level of vascular density at score 0 (12 ± 1.6%) was obtained from the mean values of vascular density of three control lungs without lung disease. The value of vascular density at a score of 1 (22 ± 1.7%) was remarkably higher than that of the control level (p < 0.01), and the vascular density was the highest at score 2 (25 ± 1.6%, p < 0.01). As the fibrosis score increased to more than 3, however, the vascular density gradually decreased. A significant difference in the decrease from the control level of vascular density was revealed at scores of 6 (8.0 ± 0.5%, p < 0.05), 7 (5.7 ± 0.6%, p < 0.01), and 8 (3.1 ± 0.7%, p < 0.01).
Three-dimensional Reconstruction of Capillaries and Vessels
We reconstructed three-dimensional images of CD34-positive capillaries and vWF-positive venules to compare the relation between alveolar capillaries and vWF-positive venules. Each three-dimensional image was reconstructed from serial two-dimensional images with double immunostaining for CD34 and vWF (Figures 6B and 6C) , on the basis of our previous studies (18, 19). Figure 6A shows a three-dimensionally reconstructed image of fibrotic lesions in UIP. CD34-positive alveolar capillaries appear in red, pulmonary veins and venules with vWF-positive endothelial cells are light green, and the contours of fibrotic lesions are blue. This reconstructed image revealed the distribution of CD34-positive capillary endothelial cells on the edges of fibrotic lesions and the connection of vWF-positive venules between CD34-positive capillaries and pulmonary veins located in the center of fibrotic lesions. These small vWF-positive venules, seen as green dots in Figure 6A, were slightly increased compared with those venules shown in the three-dimensional image of a control lung (Figure 6D).
This is the first study to show the remarkable increase in alveolar capillaries with CD34+vWF– endothelial cells in nonfibrotic lesions in UIP. The highly vascular nature of the lesions in pulmonary fibrosis was noted first by Golden and Bronk in 1953 (4). The vascular supply was found to derive from the systemic circulation through systemic–pulmonary anastomoses in the neighborhood of the vascular proliferation (5). However, few studies have identified clearly the vessels increased in UIP lungs with heterogeneous endothelial cells. In this study, we demonstrated the proliferation of CD34-positive alveolar capillaries in nonfibrotic lesions in UIP. In close contact with these increased CD34-positive alveolar capillaries, alveolar Type II cells produced the potent angiogenic factors VEGF (13) and IL-8 (14). These angiogenic factors were also detected in increased endothelial cells in nonfibrotic lesions. In fibrotic lesions, CD34-positive capillaries were distributed only on the edges where there were alveolar Type II cells. No capillaries were observed in fibrotic tissues without the attachment of alveolar cells. The important role of VEGF for maintaining alveolar structures (20) would account for the increased expression of VEGF by these alveolar epithelial cells. In this context, increased CD34-positive capillaries may contribute to the regeneration of alveolar walls lost in IPF. A report showing that transforming growth factor-ß1 and hypoxia increased the production of VEGF by epithelial cells also supports the possible role of VEGF in the endothelial repair and angiogenesis that follow lung injury (21). In contrast to the increase in capillaries in nonfibrotic lesions, we observed a decrease of CD34-positive capillary endothelial cells in fibrotic lesions in UIP. Gracey and coworkers pointed out the striking feature, as revealed by microscopy, of a decrease in the number of capillaries within the alveolar septa and the total absence of septal capillaries in the lungs of patients with pulmonary fibrosis (7). The decrease in lumenal size of vessels within remodeled alveolar walls has also been shown in biopsied specimens of pulmonary fibrosis (8). Decreased vessel density was also reported in cryptogenic fibrosing alveolitis (22). Given the lack of vascularization in fibroblastic foci, the process of fibrosis appears not to require neovascularization. The development of small fibrous tissues from nonfibrotic septa without an increase in capillaries suggested that these fibrous tissues were early lesions in UIP. In this regard, the increase in alveolar capillaries in nonfibrotic lesions is thought to represent a secondary change following the fibrotic process. A reduction in the microvasculature of intralumenal fibromyxoid lesions in UIP compared with that in organizing pneumonia was also reported (23). Keane and coworkers demonstrated an increase in vessels with vWF-positive endothelial cells in fibrous lesions and the regulatory roles of IL-8 and IP-10 in the angiogenesis and fibrosis of IPF (6). These results are consistent in part with our results that showed the dominant distribution of venules with vWF-positive endothelial cells in the centers of fibrous lesions. These vWF-positive venules are assumed to be postcapillary venules because these venules connect CD34-positive alveolar capillaries with pulmonary veins. Our three-dimensionally reconstructed images suggested that there was a slight increase in vWF-positive venules in fibrotic lesions. Because most of the endothelial cells in fibrotic lesions were not immunoreactive with Ki-67, a marker of proliferation, this increase in vWF-positive venules may have been caused by the increased immunoreactivity of vWF in CD34-positive endothelial cells, as reported in injured endothelial cells (24). In this sense, the apparent vascular density, based on the immunoreactivity with CD34 for endothelial cells, may have been biased in these fibrotic lesions. Another possible mechanism would be the contribution of circulating bone marrow–derived endothelial cells (25) to the increase in these vWF-positive venules or even to the increase in CD34-positive capillaries. In conclusion, we demonstrated a remarkable increase in CD34-positive alveolar capillaries in nonfibrotic lesions and a corresponding decrease in fibrotic lesions. vWF-positive venules, identified as postcapillary venules, were slightly increased in fibrotic lesions. These results account for the apparent contradictions concerning vascularity in IPF. The increase in CD34-positive capillaries in nonfibrotic lesions appears to be a subsequent event in the fibrotic process, which may contribute to the regeneration of alveolar walls damaged by fibrosis in IPF.
Supported by a grant for scientific research from the Ministry of Education, Science, Sports and Culture of Japan (to M.E.). This study was presented in part at the annual meeting of the American Thoracic Society in 2003. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: cM.E. has no declared conflict of interest; M.S. has no declared conflict of interest; N.S. has no declared conflict of interest; Y.K. has no declared conflict of interest; T.S. has no declared conflict of interest; M.E. has no declared conflict of interest; H.S. has no declared conflict of interest; T.K. has no declared conflict of interest; T.N. has no declared conflict of interest. Received in original form August 11, 2003; accepted in final form January 26, 2004
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
