Enhanced expression of fibroblast growth factors and receptor FGFR-1 during vascular remodeling in chronic obstructive pulmonary disease.

Abstract

Important characteristics of chronic obstructive pulmonary disease (COPD) include airway and vascular remodeling, the molecular mechanisms of which are poorly understood. We assessed the role of fibroblast growth factors (FGF) in pulmonary vascular remodeling by examining the expression pattern of FGF-1, FGF-2, and the FGF receptor (FGFR-1) in peripheral area of lung tissues from patients with COPD (FEV(1) < or = 75%; n = 15) and without COPD (FEV(1) > or = 85%; n = 13). Immunohistochemical staining results were evaluated by digital video image analysis as well as by manual scoring. FGF-1 and FGFR-1 were detected in vascular smooth muscle (VSM), airway smooth muscle, and airway epithelial cells. FGF-2 was localized in the cytoplasm of airway epithelium and in the nuclei of airway smooth muscle, VSM, and endothelial cells. In COPD cases, an unequivocal increase in FGF-2 expression was observed in VSM (3-fold, P = 0.001) and endothelium (2-fold, P = 0.007) of small pulmonary vessels with a luminal diameter under 200 micro m. In addition, FGFR-1 levels were elevated in the intima (1.5-fold, P = 0.05). VSM cells of large (> 200 micro m) pulmonary vessels showed increased staining for FGF-1 (1.6-fold, P < 0.03) and FGFR-1 (1.4-fold, P < 0.04) in COPD. Pulmonary vascular remodeling, assessed as the ratio of alpha-smooth muscle actin staining and vascular wall area with the lumen diameter, was increased in large vessels of patients with COPD (P = 0.007) and was inversely correlated with FEV(1) values (P < 0.007). Our results suggest an autocrine role of the FGF-FGFR-1 system in the pathogenesis of COPD-associated vascular remodeling.

Full text

Page 1

Enhanced Expression of Fibroblast Growth Factors and Receptor FGFR-1
during Vascular Remodeling in Chronic Obstructive Pulmonary Disease
Andor R. Kranenburg, Willem I. de Boer, J. Han J.M. van Krieken, Wolter J. Mooi, Jane E. Walters,
Pramod R. Saxena, Peter J. Sterk, and Hari S. Sharma
Department of Pharmacology, Erasmus University Medical Center, Rotterdam; Departments of Pulmonology and Pathology,
Leiden University Medical Center, Leiden; Department of Pathology, Netherlands Cancer Institute, Amsterdam, The
Netherlands; and School of Biological Molecular Sciences, Oxford Brookes University, Oxford, United Kingdom
Important characteristics of chronic obstructive pulmonary dis-
ease (COPD) include airway and vascular remodeling, the mo-
lecular mechanisms of which are poorly understood. We as-
sessed the role of fibroblast growth factors (FGF) in pulmonary
vascular remodeling by examining the expression pattern of
FGF-1, FGF-2, and the FGF receptor (FGFR-1) in peripheral area
of lung tissues from patients with COPD (FEV1? 75%; n ? 15)
andwithoutCOPD(FEV1?85%;n?13).Immunohistochemical
staining results were evaluated by digital video image analysis
as well as by manual scoring. FGF-1 and FGFR-1 were detected
in vascular smooth muscle (VSM), airway smooth muscle, and
airway epithelial cells. FGF-2 was localized in the cytoplasm of
airway epithelium and in the nuclei of airway smooth muscle,
VSM, and endothelial cells. In COPD cases, an unequivocal in-
crease in FGF-2 expression was observed in VSM (3-fold, P ?
0.001) and endothelium (2-fold, P ? 0.007) of small pulmonary
vessels with a luminal diameter under 200 ?m. In addition,
FGFR-1 levels were elevated in the intima (1.5-fold, P ? 0.05).
VSM cells of large (? 200 ?m) pulmonary vessels showed
increased staining for FGF-1 (1.6-fold, P ? 0.03) and FGFR-1
(1.4-fold, P ? 0.04) in COPD. Pulmonary vascular remodeling,
assessed as the ratio of ?-smooth muscle actin staining and
vascular wall area with the lumen diameter, was increased in
large vessels of patients with COPD (P ? 0.007) and was in-
versely correlated with FEV1 values (P ? 0.007). Our results
suggest an autocrine role of the FGF–FGFR-1 system in the
pathogenesis of COPD-associated vascular remodeling.
and fibroblasts, and increased deposition of extracellular
matrix(3). Inaddition,advanced COPDleads topathologic
changes in the pulmonary circulation (4, 5). At least part
of this is probably the result of alveolar hypoxia, which is
well known to cause pulmonary vasoconstriction and, if the
hypoxic stimulus persists, pulmonary vascular remodeling, of
which increased muscularization of small arterial branches
isthemoststrikingfeature(6).Withsustainedvasoconstric-
tion of pulmonary arteries, arterioles, and veins, the medial
vascular smooth muscle (VSM) extends distally to vessels
normally devoid of smooth muscle (6). Intimal thickening
due to fibrosis and emergence of smooth muscle cells within
the intima of small pulmonary arterial branches has also
been reported (5). Finally, loss of the pulmonary vascular
bedbyemphysemahasbeensuggested toleadtotheforma-
tion of new vessels (6). Thus, several phenomena acting in
concert in COPD result in pulmonary vascular remodeling.
Yet little is known about the molecular mechanisms under-
lying these processes in the context of COPD.
A variety of growth factors and cytokines released from
various sites of airway and vascular walls (VW) have the
potential to contribute to the pathogenesis of vascular re-
modeling in COPD. In view of their important role in
chronic inflammation, fibrosis, and repair of various tissues,
including the lung (9), fibroblast growth factors (FGFs)
may well play a pivotal role in airway and VW remodeling
(7, 8). FGFs exert their biologic effects via binding to four
high-affinityFGFtransmembranetyrosine-kinasereceptors
(FGFR) (9). Distinct FGF subtypes bind with different af-
finity to the various FGFR . Alternative splicing and regu-
lated protein trafficking further modulate the intracellular
events and resultant response initiated by FGF ligand–
receptorinteraction(9).Inthelungaswellasinthevascular
system, FGFs have been implicated in several pathologic
conditions. FGF-1 and FGFR-1 were shown to be upregu-
lated during the development of lung fibrosis (10). FGF-2
and platelet-derived growth factor (PDGF) have been impli-
cated in the pathogenesis of obliterative bronchiolitis after
transplantation (11). Moreover, vascular remodeling in re-
sponsetoincreasedbloodpressureisassociatedwithelevated
levels of basic FGF (12, 13).
To investigate whether the FGF–FGFR system might
be involved in the pathogenesis of COPD, we examined
the expression patterns of FGF-1, FGF-2, and FGFR-1 in
smokers and exsmokers with or without COPD and corre-
lated the expression with histologic evidence of pulmonary
vascular remodeling.
Chronic obstructive pulmonary disease (COPD) is a global
health problem with increasing morbidity and mortality (1).
One of the major causal factors is tobacco smoking (2).
However,onlytenpercentofallsmokersdevelopsymptom-
atic COPD. The causes of this variability in response of the
airways and lung parenchyma to tobacco smoke exposure
have remained largely unclear. One of the key pathologic
features of COPD is thickening of airway walls as a result
of inflammation, hyperplasia of airway smooth muscle cells
(ReceivedinoriginalformDecember15, 2000 andinrevised form May6,2002)
Address correspondence to: Hari S. Sharma, M.Phil., Ph.D., Institute of
Pharmacology,ErasmusUniversityMedicalCenter,Dr.Molewaterplein50,
3015 GE Rotterdam, The Netherlands. E-mail: sharma@farma.fgg.eur.nl
Abbreviations:chronicobstructivepulmonarydisease,COPD;forcedexpir-
atoryvolumeinonesecond,FEV1;fibroblastgrowthfactors,FGF;fibroblast
growth factor receptor, FGFR; forced vital capacity, FVC; inside diameter,
ID; carbon monoxide diffusion, Kco; platelet-derived growth factor, PDGF;
?-smooth muscle actin, ?-SMA; vascular smooth muscle, VSM; vascular
wall, VW.
Am. J. Respir. Cell Mol. Biol.Vol. 27, pp. 517–525, 2002
DOI: 10.1165/rcmb.4474
Internet address: www.atsjournals.org

Page 2

518
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 27 2002
Materials and Methods
Selection of Patients’ Specimens
We examined lung tissue specimens of subjects with or without
COPD. Peripheral parts of the lung tissue from current and ex-
smokers who underwent lobectomy or pneumonectomy for lung
cancer was obtained from the Pathology Laboratories of the
Leiden University Medical Center, Leiden, the Netherlands, and
the Zuiderziekenhuis, Rotterdam, The Netherlands. Tissue speci-
mens were taken distally from the lung hilus and predominantly
consisted of parenchyma, small airways, and vasculature. All lung
tissues were inflated by an injection syringe using formalin and
fixed for ? 24 h after which the tissues were further dehydrated
and embedded in paraffin. These tissues were subsequently pro-
cessed for immunohistochemical staining. Based on a number of
lung function data, patients were assigned to the groups with and
without COPD (14, 15).
GroupwithCOPD.FifteensubjectswereassignedtotheCOPD
group on the basis of the following parameters: forced expiratory
volume in one second (FEV1) ? 75% of predicted value (16)
before bronchodilation, FEV1/forced vital capacity (FVC) ratio ?
75%, a reversibility in FEV1? 12% of predicted after 400 ?g
inhaled salbutamol, and a transfer factor for carbon monoxide
(diffusion capacity) per liter alveolar volume (Kco) ? 80% of pre-
dicted value.
Group without COPD. Thirteen subjects were assigned to the
group without COPD on the basis of the following data; a FEV1?
85% before bronchodilatation, FEV1/FVC ratio ? 85%, and re-
versibility in FEV1? 12% of predicted after 400 ?g salbutamol
inhalation. To exclude accompanying lung disease leading to a
restrictive function disorder, the total lung capacity of each subject
included in the study was over 80% of the predicted values (16).
Clinical data of all patients were examined for possible comor-
bidity and medication usage. All patients were free of symptoms
of upper respiratory tract infection and none received antibiotics
perioperatively. Noneof thepatients receivedglucocorticosteroids
inthe3mobeforeoperation;fourpatientsreceivedoralglucocorti-
costeroids perioperatively. In addition to the rigorous criteria
based on lung function parameters, microscopic exclusion criteria
was also applied in the selection of patients for this study. After
the selection based on lung function, all the lung tissues were
subsequently examined histologically by two experienced lung pa-
thologists using the following exclusion criteria: (i) presence of
tumor in the lung tissue specimen; (ii) presence of poststenotic
pneumonia in the specimen; (iii) fibrosis of lung tissue; and (iv)
obstruction of the main bronchus of the resection specimen by
tumor (14, 15).
TABLE 1
Subject characteristics and clinical parameters
Subjects without
COPD
Patients with
COPD
FEV1
dFEV1, % predicted
FEV1/FVC, %†
TLC
RV
Kco, % predicted
Sex (male/female)
Age (yr)
Smokers/exsmokers/nonsmokers
Pack-years
Steroid use (yes/no/unknown)
99 ? 1.9
3 ? 0.6
100 ? 2.3
104 ? 2.0
115 ? 5.5
94 ? 2.0
11/2
57 ? 3.2
9/4/0
33 ? 4.7
0/12/1
53 ? 3.2*
4 ? 0.9
58 ? 5.0*
108 ? 8.8
141 ? 15.4*
55 ? 5.4*
14/1
59 ? 5.0
12/3/0
35 ? 5.2
4/9/2
Definition of abbreviations: reversibility of FEV1after 400 ?g salbutamol,
dFEV1; forced expiratory volume in 1 s, FEV1; forced vital capacity, FVC; carbon
monoxide diffusion constant, Kco; residual volume, RV.
* P ? 0.005 versus subjects without COPD.
†FEV1/FVC is given as actual ratio in %.
of staining obtained with a series of dilutions, which produced
specific and easily visible signals on paraffin sections made from
the same control tissue before performing the staining protocol
on all sections. To avoid day-to-day variations in staining intensit-
ies, the incubations of all specimens with each antibody were per-
formed in one single run. Sections were deparaffinized and rehy-
drated before incubation with specific mouse monoclonal and
affinity-purified antibodies against FGF-1 (1:2000 dilution), FGF-2
(1:200 dilution) and FGFR-1 (1:2000 dilution). The mouse IgG1
antibodyagainsthumanFGF-1wasraisedusingasyntheticpeptide
corresponding to the internal 61–99 amino acid sequence, whereas
the mouse IgG2b antibody was raised against a synthetic peptide
correspondingtothe16aminoacidsfromtheC-terminusofhuman
FGFR-1, as described previously (17, 18). FGF-2 was a mouse
(IgG1 isotype) monoclonal antibody raised against human FGF-2
(molecularweight:18–24kD),whichwasprocuredfromTransduc-
tionLaboratories(Lexington,KY).Antihumanmousemonoclonal
antibodies against ?-smooth muscle actin (?-SMA), Ki-67 and
FGF-2 were purchased from NeoMarkers (Clone 1A4; Fremont,
CA), from Biogenex (San Ramon, CA) and from Transduction
Laboratories, respectively. To block nonspecific binding, sections
were preincubated with 10% normal goat serum diluted in 5%
BSA in PBS (pH 7.4).Subsequently, sections were incubated over-
night at 4?C with the primary antibodies (FGF-1 and FGFR-1)
diluted appropriately, or for 1 h at room temperature in the case
of ?-SMA (1:1000 dilution). Secondary biotinylated anti-immuno-
globulins (Multilink, 1:75 dilution; Biogenex) and tertiary strepta-
vidin conjugated alkaline phosphatase (Label 1:50 dilution; Bio-
genex) were used to enhance the detection sensitivity. Color was
developed using new fuchsin, while endogenous alkaline phospha-
tase activity was inhibited by 0.01 M levamisole.
FGF-2 and Ki-67 immunostaining was performed on serial sec-
tions after antigen retrieval by boiling in 10 mM citrate buffer (pH
6.0) for 10 min in a microwave oven. Sections were preincubated
with 10% normal goat serum in5% BSA/PBS, followed by incuba-
tion with primary antibody (1:50 dilution) overnight at 4?C. Slides
were rinsed in PBS, incubated for 30 min with peroxidase conju-
gated streptavidin at a dilution of 1:50 (Biogenex). Subsequently,
sections were colored using 0.025% 3,3-diaminobenzidine (Sigma,
St. Louis, MO) in 0.01 molar PBS, containing 0.03% hydrogen
peroxide. Slides were counterstained with Mayer’s hematoxylin.
Positive controls consisted of human breast carcinoma and placen-
Pulmonary Function Tests
All pulmonary function tests were performed within the 3 months
beforesurgery.FEV1andFVCweremeasuredbyspirometry,total
lung capacity, and residual volume with the closed circuit helium
dilution test and the Kcousing the single breath-holding technique,
as described by Quanjer and coworkers (16). Lung function data
and other patient characteristics are shown in Table 1.
Immunohistochemistry
Sections of paraffin-embedded lung tissue were cut at 4 ?m,
mounted on Super Frost Plus microscopic slides (Menzel-Gla ¨ser,
Braunschweig, Germany) and processed for immunohistochemis-
try.SerialsectionswereusedforimmunostainingofFGF-1,FGF-2,
and FGFR-1 using human-specific antibodies. The optimal dilu-
tions for all antibodies were identified by examining the intensity

Page 3

Kranenburg, de Boer, van Krieken, et al.: Vascular Remodeling in COPD519
Figure 1. Representative examples of staining intensity pattern used for visual scoring Photomicrographs depict lung tissue sections from
patients without COPD (A and C) and with COPD (B and D), showing FGFR-1 staining (red new-fuchsine) in vascular smooth muscle
cells. A to D show representative examples of staining intensities used for visual scoring, 0–3 respectively. Scale bar ? 50 ?m; original
magnification: ?100.
tal tissue. The optimal dilutions for all antibodies were identified
by examining the intensity of staining obtained with a series of
dilutions, which produced specific and easily visible signals on
paraffin sections derived from the same control tissue. Slides were
mounted and staining results were systematically investigated (see
below). Negative controls consisted of omission of the primary
antibody.
insidediameter[ID])andintheendothelium,VSM,andadventitial
area of large (? 200 ?m ID) pulmonary arteries.
Video Image Analysis
In addition, video image analysis was performed for ?-SMA stain-
ing using Leica Qwin system version 3.0 (Leica B.V., Rijswijk, The
Netherlands). Twenty digital images (736 ? 574 pixels) from each
section were taken using a video camera. ID of blood vessels was
derived as a mean of measured vertical and horizontal diameters.
Inourstudy, weexcludedthosevesselswitha verticaltohorizontal
ID ratio ? 3. Based on the ID, pulmonary vessels were grouped
into four sizes (50–100 ?m, 100–200 ?m, 200–400 ?m and ? 400
?m). VW area, ?-SMA stained area and vessel ID were measured.
Measurements were expressed as percentages of staining per VW
(?-SMA/VWarea), for VWarea corrected for ID (VWarea/ID) and
?-SMA staining, also corrected for lumen diameter (?-SMAarea/ID).
Semiquantitative Analysis
All tissues were analyzed in a blinded fashion in random order by
two independent observers who were unaware of the clinical data
of the case under study. Semiquantitative analysis was performed
using an arbitrary visual scale with grading scores of 0, 1, 2, and 3
representing no, weak, moderate and intense staining, respectively
(14, 15). Errors within and between observers were assessed by
correlating the expression scores using Pearson’s analysis, which
indicatedaveryhighcorrelation,rangingfrom0.8to0.9.Micropho-
tographs in Figures 1A to 1D show representative examples of
stainingintensitiesusedforvisualscores0–3,respectively.Sections
weregradedfortheintensityofsignalexpressionofFGF-1,FGF-2,
and FGFR-1 in the endothelium and VSM of small (50–200 ?m
Statistical Analysis
Data were analyzed for statistical significance using the unpaired,
two-tailed Students’ t test as well as the Mann-Whitney nonpara-

Page 4

520
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 27 2002
metric test, when appropriate (14, 15, SPSS; SPSS Incorporated,
Chicago, IL). The staining score data for FGF-1, FGF-2 and
FGFR-1 were expressed as mean ? SEM. Furthermore, FGF-1,
FGF-2, and FGFR-1 staining scores for different vessels with ID ?
200 ?m and for those ? 200 ?m were correlated with FEV1using
Pearson’s correlation analysis. Furthermore, the individual FEV1
values were correlated with the vascular remodeling data (VWarea/
ID) in both the groups. Differences with P ? 0.05 were considered
to be statistically significant.
Localization and Quantification of FGF-1 and FGF-2
FGF-1 expression was detected in the media and adventitia
of large pulmonary arteries and veins, but only in the media
of small vessels. No FGF-1 staining was found in the endo-
thelial layer of any vessels. FGF-2 was localized specifically
in nuclei of endothelial and vascular smooth muscle cells.
Expression of FGF-1 and FGF-2 was also observed in epi-
thelial and bronchiolar smooth muscle cells. Representa-
tive microphotographs showing the expression patterns of
FGF-1 and FGF-2 are presented in Figure 2. A summary
of the semiquantitative data of FGF-1 and FGF-2 immuno-
staining is presented in Figure 3. In subjects with COPD,
we observed significantly increased (P ? 0.02) expression
of FGF-1 in medial VSM of larger vessels, but the level of
adventitialexpressionofFGF-1remainedunaltered(Figure
3A). In contrast to FGF-1, the expression of FGF-2 was
elevated in the group with COPD, but only in the small
vessels with an ID ? 200 ?m, where it was increased in
endothelial (P ? 0.007) and medial smooth muscle (P ?
0.001) cells (Figure 3B).
Results
Clinical Parameters
The clinical and lung function characteristics of all subjects
included in the study are listed in Table 1. FEV1and FEV1/
FVC values were significantly lower in the group with
COPD than in the group without COPD (P ? 0.001). In
the group with COPD, residual volume was increased,
whereas Kcowas reduced (P ? 0.005). The subjects in the
two groups did not differ significantly with respect to age,
total lung capacity, reversibility in FEV1, smoking status
(pack-years), or previous steroid usage (Table 1).
Figure 2. Photomicrographs of lung tissue sections from patients without COPD (A and C) and with COPD (B and D). A and B (scale
bar ? 50 ?m; original magnification: ?100) show representative examples of FGF-1 protein staining (red new-fuchsine) in vascular
smooth muscle cells of a large vessel (ID ? 200 ?m). C and D (scale bar ? 100 ?m; original magnification: ?400) show representative
examples of nuclear FGF-2 expression (brown 3,3-diaminobenzidine) in endothelium and vascular smooth muscle cells of vessels with
ID ? 200 ?m. Arrows indicate positive nuclei.

Page 5

Kranenburg, de Boer, van Krieken, et al.: Vascular Remodeling in COPD 521
Figure 3. Graphic representations of FGF-1 (A) and FGF-2 (B) expression scores (mean ? SEM) in large (ID ? 200 ?m) and small
(ID ? 200 ?m) vessels in groups without COPD (open bars) and with COPD (closed bars). * P ? 0.05 versus the group without COPD.
Localization and Quantification of FGFR-1
FGFR-1 immunoreactivity was detected in epithelial and
bronchiolar smooth muscle cells, and in the endothelium
and vascular smooth muscle of large and small vessels. No
adventitial positivity for FGFR-1 was observed. Represen-
tative microphotographs showing the expression pattern of
FGFR-1arepresentedinFigure4.Agraphicrepresentation
of the data of FGFR-1 immunostaining is displayed in
Figure 5. The expression of FGFR-1 was significantly ele-
vated in medial smooth muscle cells of large vessels (P ?
0.04) in the group with COPD as compared with the group
without COPD, whereas the staining for the receptor in
the intimal endothelium remained unaltered. Moreover, in
contrast to the FGF-1 expression in small vessels in patients
with COPD, we found significantly higher expression levels
of the receptor (P ? 0.05) in medial smooth muscle of small
vessels (Figure 5).
ingly, no significant differences were observed between the
groups with and without COPD in the percent of vascu-
lar smooth muscle, defined as ?-SMA/VWareain any vessel
type (Figure 7B). Proliferation of VSM cells, as evidenced
from Ki-67 positivity, was observed infrequently (data not
shown).
Correlation with Clinical Data
The staining scores of FGF-1, FGF-2, and FGFR-1 expres-
sion in patients with COPD and those without COPD were
analyzed using Pearson’s test. For FGF-1, we observed a
weak but significant inverse correlation (r ? ?0.39, P ?
0.038) between staining score and FEV1in the medial VSM
of vessels ? 200 ?m in ID (Figure 8A). In addition, there
wasasignificantinversecorrelationofFGF-2stainingscores
in both endothelium (r ? ?0.44, P ? 0.002) and medial
VSM (r ? ?0.55, P ? 0.0001) of vessels ? 200 ?m in ID
with FEV1(Figures 8B and 8C). However, in vessels ? 200
?m in ID, no significant correlation between FGF-2 expres-
sion and FEV1was found (data not shown). Surprisingly,
stainingscoresforFGFR-1 werenotsignificantlycorrelated
with FEV1(data not shown). When considering the associa-
tion between FEV1and medial hypertrophy (VWarea/ID ra-
tio), we observed a significant inverse correlation of ?0.42
(P ? 0.007) for vessels ? 200 ?m ID (Figure 8D). However,
no significant correlation could be established between
FEV1and VWarea/ID ratio for the vessels ? 200 ?m ID (r ?
?0.10, P ? 0.10).
Assessment of Vascular Remodeling
Toexaminepulmonaryvascularremodelingasevidencedin
variations in wall thickness and muscular medial thickness,
video imageanalysis was performed using?-SMA immuno-
staining. Four separate groups with vessels of 50–100, 100–
200, 200–400 and ? 400 ?m ID, respectively, were analyzed
(Figure 6). Measurements (mean ? SEM) were expressed
as VWarea/ID, ?-SMAarea/ID, or percentage ?-SMA staining
per VWareathat represents the volume fraction for smooth
muscle staining (?-SMA/VWarea). The graphic representa-
tion of VW is presented in Figure 7. A significant increase
was seen in VWarea/ID ratio for COPD in vessels of 100–200
?m (44.2 ? 1.9 versus 36.4 ? 2.1, P ? 0.007), 200–400 ?m
(57.9 ? 2.4 versus 44.7 ? 3.2, P ? 0.001) and ? 400 ?m
(75.6 ? 2.6 versus 56.8 ? 5.9, P ? 0.011) (Figure 7A). In
vessels ranging from 50 to 100 ?m in ID, no differences
in VWarea/ID ratio were observed. A significant increase in
?-SMAarea/ID ratio was observed for COPD in the 200–400
?m ID (26.7 ? 1.9 versus 20.3 ? 2.9, P ? 0.034) and ? 400
?m ID (38.3 ? 2.0 versus 23.5 ? 3.4, P ? 0.006) vessels,
but not in the 50–100 ?m and 100–200 ?m vessels. Surpris-
Discussion
In this study we have found that COPD is associated with
an increase in the expression of FGF-2 in small pulmonary
vessels (? 200 ?m) and FGF-1 in large (? 200 ?m) pulmo-
nary vessels, whereas FGFR-1 is increased in both vessel
types. Vascular medial thickness, assessed by video image
analysis, was significantly increased in patients with COPD
in pulmonary vessels of various sizes. Pearson’s correlation
analysis revealed a significant inverse correlation of FEV1
with FGF-1 staining in the media of large vessels and with