Am J Physiol Lung Cell Mol Physiol 292: L769–L777, 2007.
First published November 17, 2006; doi:10.1152/ajplung.00355.2006.
Morphometry of the extremely thin pulmonary blood-gas barrier in the
chicken lung
Rebecca R. Watson, Zhenxing Fu, and John B. West
Department of Medicine, University of California San Diego, La Jolla, California
Submitted 11 September 2006; accepted in final form 10 November 2006
capillary stress failure; lung morphology; bird respiratory physiology;
air capillary; extracellular matrix; interstitium
IN THE VERTEBRATE LUNG, OPTIMAL thickness of the pulmonary
blood-gas barrier (BGB) is affected by opposing selective
pressures (65). The barrier thickness is minimized to allow
efficient exchange of oxygen and carbon dioxide, yet the
capillary walls must also be strong enough to withstand high
pulmonary capillary pressures that are incurred during extreme
physical exertion. The BGB is composed of three layers:
capillary endothelium, an interstitial layer or extracellular matrix, and an epithelial layer (Fig. 1). Combined, these cellular
layers are relatively thin, usually measuring ⬍2 ␮m in total
thickness under normal conditions (5, 33, 36). This three-ply
ultrastructure of the BGB appears to be evolutionally well
conserved over a variety of vertebrate taxa, indicating that its
basic structure is efficient and relatively immalleable (18, 25,
27, 31, 33). The strength of the pulmonary capillary walls is
likely related to the morphometry of the BGB, where the thickness of the interstitium determines the fragility of the capillaries
(5). Where the BGB is thinnest, the interstitial layer is composed
of only the fused basement membranes of the endothelium and
Address for reprint requests and other correspondence: J. B. West, UCSD,
Dept. of Medicine 0623A, 9500 Gilman Drive, La Jolla, CA 92093-0623
(e-mail: [email protected]).
http://www.ajplung.org
epithelium layers. There is indirect evidence that the basement
membrane determines the mechanical properties (e.g., wall
strength) of the pulmonary capillaries (46). For example, in
isolated perfused renal tubules of rabbits, the relationship between
tubule diameter and applied transmural pressure was identical for
the intact tubule and for a tubule consisting of only the basement
membranes (62). These findings are consistent with the observation that single alveolar walls that are stripped of their endothelial
and epithelial cellular constituents by a detergent do not show a
reduction in the force required for extension (45).
Despite the apparent strength of the BGB, under certain
conditions the barrier may fail and allow fluid to leak into the
lungs. Virtually all horses bred and trained for competitive
racing show evidence of exercise-induced pulmonary hemorrhage (EIPH), which manifests as bleeding from the nose after
exertion or the presence of red blood cells or hemosiderophages in the airways (3, 6, 21, 39, 47, 53, 58, 67). The causative
mechanism of EIPH was demonstrated experimentally when it
was shown that high pulmonary capillary pressures can cause
discrete breaks in the endothelium, epithelium, or the entire
BGB in mammals (10, 50, 64, 66). Incidences of suspected
EIPH (usually presenting as hemoptysis) have been documented in humans as well (15, 34, 35, 61), and evidence that
maximal exercise can impair the integrity of the BGB and
change the permeability of the capillary membrane was demonstrated by Hopkins et al. (16). However, submaximal exercise does not appear to affect the integrity of the BGB in
humans or horses (17, 48). Stress failure also plays a role in the
hospital setting, where overinflation of the lung via mechanical
ventilation causes ventilator-induced lung injury.
In the present study, we quantified the thicknesses of the
different components of the BGB in the chicken lung to
compare with measurements collected in the same manner
from rabbits, dogs, and horses. We selected the lung of a bird
for our studies because of the paradoxical observation that the
avian lung BGB is particularly thin despite the extremely high
aerobic capacities and maximal oxygen consumptions found in
some birds. Here, by measuring the individual components of
the BGB, we wish to determine which layers are attenuated in
birds compared with mammals with the ultimate goal of
elucidating the mechanisms whereby the avian lung is superior
in efficiency to the mammalian lung.
MATERIALS AND METHODS
Tissue sampling. The animal protocol for this experiment was
approved by the Animal Subjects Committees of the University of
California, San Diego. Animal surgery, perfusion, dissection, and
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1040-0605/07 $8.00 Copyright © 2007 the American Physiological Society
L769
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Watson RR, Fu Z, West JB. Morphometry of the extremely
thin pulmonary blood-gas barrier in the chicken lung. Am J Physiol
Lung Cell Mol Physiol 292: L769–L777, 2007. First published
November 17, 2006; doi:10.1152/ajplung.00355.2006.—The gas
exchanging region in the avian lung, although proportionally smaller
than that of the mammalian lung, efficiently manages respiration to
meet the high energetic requirements of flapping flight. Gas exchange
in the bird lung is enhanced, in part, by an extremely thin blood-gas
barrier (BGB). We measured the arithmetic mean thickness of the
different components (endothelium, interstitium, and epithelium) of
the BGB in the domestic chicken lung and compared the results with
three mammals. Morphometric analysis showed that the total BGB of
the chicken lung was significantly thinner than that of the rabbit, dog,
and horse (54, 66, and 70% thinner, respectively) and that all layers of
the BGB were significantly thinner in the chicken compared with the
mammals. The interstitial layer was strikingly thin in the chicken lung
(⬃86% thinner than the dog and horse, and 75% thinner than rabbit)
which is a paradox because the strength of the BGB is believed to come
from the interstitium. In addition, the thickness of the interstitium was
remarkably uniform, unlike the mammalian interstitium. The uniformity
of the interstitial layer in the chicken is attributable to a lack of the
supportive type I collagen cable that is found in mammalian alveolar
lungs. We propose that the surrounding air capillaries provide additional
structural support for the pulmonary capillaries in the bird lung, thus
allowing the barrier to be both very thin and extremely uniform. The net
result is to improve gas exchanging efficiency.
L770
PULMONARY BLOOD-GAS BARRIER
sample preparation were similar to procedures previously performed
in our laboratory (50). Chicken lungs were fixed using two methods,
intravascular perfusion and tracheal instillation. For perfusion, five
white leghorn chickens (Gallus gallus domesticus) were anesthetized
with intravenous pentobarbital sodium (40 mg/kg) and the chest was
opened to insert catheters into the main pulmonary artery and the left
atrium. Mean pulmonary arterial pressure of 25 cmH2O and left atrial
pressure of 20 cmH2O were maintained during vascular perfusion
with heparinized saline/dextran solution (11.1 g NaCl/l, 350 mOsm;
3% T-70 dextran and 10 IU/ml heparin). Following wash-out, the
pulmonary vasculature was perfusion fixed for 10 min with 2.5%
gluteraldehyde and 3% T-70 dextran in 0.1 M phosphate buffer (total
osmolarity 500 mOsm, pH 7.4), during which time the pulmonary
artery and left atrial pressures were maintained constant. The lungs
were then dissected free and stored in 2.5% gluteraldehyde at 4°C for
3 days. Intratracheal instillation was performed on one chicken, and
the resultant samples were used for qualitative analysis and illustrative
purposes only. During the intratracheal instillation procedure, the
chicken was placed in a supine position after anesthesia and the lungs
were infused with the fixation solution at a pressure of 20 cmH2O for
15 min. The lungs were removed, sampled as in the preceding
procedure for perfusion, and all further procedures were identical.
Lung samples were taken from the paleopulmo portion of the lung
from each chicken at about one-third the distance from the most
caudal aspect of the lung. A section of tissue 0.5-cm thick was excised
from the entire width of each lung transverse to the cranial-caudal
axis. This section was then systematically cut into ⬃10 vertical slices
and trimmed into blocks following a lung vertical section sampling
procedure for morphometric analysis (37).
Electron microscopy. Blocks were rinsed overnight in 0.1 M
phosphate buffer (350 mOsm, pH 7.4) and postfixed for 2 h in osmium
tetroxide (1% osmium tetroxide in 0.125 sodium cacodylate buffer;
400 mOsm, pH 7.4). The samples were then passed through stepwise
dehydration in increasing concentrations of ethanol (50 –100%),
rinsed with 100% propylene oxide, and embedded in Araldite.
Two blocks from each animal were randomly selected for analysis.
One-micrometer-thick sections were cut from each block with an
LKB Ultratome III, stained with 0.1% toluidine blue aqueous solution,
and examined by light microscopy to verify fixation quality. If fixation
was inadequate in a selected block, it was excluded from analysis and
another block from the same individual was randomly selected. The
blocks were then cut into ultrathin sections (50 –70 nm) and contrast
stained with saturated uranyl acetate and bismuth subnitrate. Sections
AJP-Lung Cell Mol Physiol • VOL
Fig. 2. Light micrographs of a portion of the lung from the chicken (top) and
rabbit (botom). Note the small diameter of the air capillaries in the chicken lung vs.
that of the rabbit alveoli at the same magnification (top vs. bottom). Top: in the
chicken lung, the pulmonary capillaries are supported by “struts” of epithelium
(arrows). Bottom: in the mammalian lung, the pulmonary capillaries are essentially
unsupported and appear to be suspended in the large alveolar spaces.
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Fig. 1. High-magnification micrograph of the blood-gas barrier (BGB) of a
chicken lung, showing the constituent layers of the BGB. Ac, air capillary; bc,
lumen of blood capillary; s, surface layer or surfactant layer; epi, epithelial cell
layer; int, interstitial space or extracellular matrix; end, endothelial cell layer;
rbc, red blood cell; n, nucleus of red blood cell; pv, pinocytotic vesicle or
inclusion. Bar ⫽ 0.5 ␮m.
were examined at an accelerating voltage of 60 kV using a Zeiss
EM10C transmission electron microscope. For each block, a total of
60 micrographs were taken on film plates at a magnification of
⫻2,500 (30 micrographs taken by systematic random sampling from
a single ultrathin section from each of 2 blocks). Micrographs of a
carbon grating replica were taken with each film for calibration and to
confirm that the measured magnification was within 5% of nominal
magnification. Negatives were converted to high resolution (1,200
dpi) digital images with a Microtek Scanmaker 4 flatbed scanner
(Microtek USA, Carson, CA) and saved on compact discs.
Analysis. Digital images were optimized for on-screen viewing by
adjusting the contrast and brightness of the image using Adobe
Photoshop 7.0.1 loaded onto a PC. The sites for measuring the
thickness of the layers of the BGB were determined at random by
intersection of the barrier with fixed test line intersections (up to 5
intersections per micrograph). During all digital manipulations, care
was taken to maintain the aspect ratio of the image to eliminate
distortion. Measurements of the thicknesses of the components of the
BGB were performed using MatLab 5.3. A calibration was performed
with a digitized image of the carbon replica grating each time the
program was launched. The thicknesses of the endothelial cell layer,
PULMONARY BLOOD-GAS BARRIER
L771
the interstitial layer, the epithelial cell layer, and the total BGB were
measured orthogonal to the BGB.
Thickness data were tested for normality using a KolmogorovSmirnov test in SigmaStat for Windows 2.03. All sections of the
barrier (endothelium, interstitium, epithelium, and total BGB) failed
the normality test, so we used the nonparametric Kruskal-Wallis
one-way ANOVA on Ranks test and Dunn’s method (post hoc) to
identify intraspecific differences among individuals. We also used
these nonparametric tests for interspecies comparisons. In all cases,
P ⬍ 0.05 was required for statistical significance.
RESULTS
Fig. 3. Electron micrograph of the chicken lung. The pulmonary capillaries
(identified by the presence of red blood cells) are supported by struts of
epithelial tissue (arrowheads) which allow a relatively uniform and thin BGB
(inset). leu, Leukocyte; n, endothelial cell nucleus. Bar ⫽ 2 ␮m.
AJP-Lung Cell Mol Physiol • VOL
Fig. 4. Electron micrograph of a dog lung. The mammalian lung is characterized
by pulmonary capillaries that appear “polarized” when orthogonally transected.
The arrows point to the “thick” side of the capillary that is inefficient at gas
exchange relative to the “thin” side. C, capillary; a, alveolus. Bar ⫽ 2 ␮m.
joined the blood capillaries to form a pedicle-type structure
(Figs. 3 and 5). An osmiophilic surface layer was sometimes
observed on the epithelium (4, 29) (Fig. 1), but it was usually
absent, most likely due to the fixation process (54).
Fig. 5. Electron micrographs of the fine structure of the epithelial “bridge.” A:
complete epithelial bridge spanning two blood capillaries. B: ultrastructure of
an epithelial bridge showing the “tributary” of extracellular matrix that is
continuous with the blood capillary and passes between abutting epithelial
cells to an adjacent blood capillary. Bars in both images ⫽ 0.5 ␮m.
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Qualitative observations. Light microscopy and ultrastructure of the perfusion- and tracheal instillation-fixed chicken
lungs are shown in Figs. 1, 2, and 3 (2, 9, 42). Tightly-packed
air capillaries were dispersed among blood capillaries, both of
which were of small diameter compared with the diameter of
mammalian alveoli (Fig. 2). Sections containing endothelial
cell nuclei were common, while cuboidal-shaped epithelial cell
nuclei were less frequent. The gas-exchanging tissue (BGB)
consisted primarily of the thin cytoplasmic extensions of these
cells. Fibroblasts, type I collagen fibers, and free macrophages
were infrequent in the gas-exchanging region. The three layers
of the BGB (endothelium, interstitium, and epithelium; Fig. 1)
were usually distinct (Fig. 3). The epithelium was extremely
attenuated in some areas (Fig. 1), and the transversely sectioned endothelium possessed a corrugated appearance (Figs. 1
and 3). Pinocytotic vesicles or inclusions were sometimes
visible in the endothelium and epithelium. The vesicles were
larger and more common in the endothelium (Fig. 1). Compared with the mammalian BGB, the bird BGB appeared to be
of uniform thickness around the circumference of the capillary
except near adjacent blood capillaries; in these areas, the
extracellular matrix (and total BGB) was markedly thickened
(Figs. 3, 4, 5). The tissue that separated air capillaries from one
another was composed of abutting epithelial cell extensions
(from different cells) that appeared to be separated by a thin
basement membrane, although an extracellular space was not
always visible (Fig. 5). The epithelium was usually thickened
at the base of either side of the “epithelial bridge” where it
L772
PULMONARY BLOOD-GAS BARRIER
Table 1. Morphometric data of the chicken lungs
Animal
No. of
Measurements
Endothelium
Interstitium
Epithelium
Total BGB
1
2
3
4
5
Mean (5)
295
285
299
300
300
1,479
0.143⫾0.06*
0.147⫾0.06*
0.140⫾0.07*
0.133⫾0.05*
0.114⫾0.05
0.135⫾0.06
0.047⫾0.02*
0.046⫾0.02*
0.045⫾0.02*
0.044⫾0.01
0.041⫾0.01
0.045⫾0.02
0.086⫾0.05
0.091⫾0.06
0.084⫾0.05
0.087⫾0.04
0.084⫾0.04
0.086⫾0.05
0.276⫾0.09*
0.285⫾0.09*
0.269⫾0.10*
0.264⫾0.07*
0.239⫾0.07
0.266⫾0.09
Values are averages ⫾ SD in ␮m. *Significantly greater than animal 5 (P ⬍ 0.05). BGB, blood-gas barrier.
qualitatively indicates that there was not a substantial amount
of thickness variation in any of the BGB layers measured in the
chicken. The cumulative frequency histogram is a different
graphical view of the same data (Fig. 8). It shows the accumulated percentage of data points for each thickness category,
as defined by the relative frequency histogram. In the cumulative frequency histogram, the area of the sigmoidal curve
with the steepest slope indicates that the highest percentage of
data points fall within those thickness categories. The cumulative frequency histograms for the different layers of the BGB
of the four different species show that 1) the chicken lung has
a higher proportion of thin measurements in all layers compared with that of the mammals and that 2) the interstitium
layer of the chicken has a much higher percentage of very thin
measurements in very few thickness categories; that is, the
interstitial layer in the chicken reaches nearly 100% faster than
any of the mammals (Fig. 8). Thus the interstitial layer possessed an especially high proportion (⬃72%) of thin measurements (⬍0.05 ␮m) in the chicken lung compared with the
mammalian lungs (Figs. 6 – 8).
Uniformity of BGB. Images of the avian lung ultrastructure
show that the BGB surrounding the blood capillaries in the
chicken lung are remarkably uniform, so that the walls of
transverse sections of avian blood capillaries appear nearly
radially symmetric (Fig. 3, inset). This qualitative observation
is in contrast to mammalian pulmonary capillaries, which,
when transversely sectioned, have walls with a bilaterally
symmetric appearance (Fig. 4). To quantitatively compare the
dispersion of the thicknesses of the different layers in the
mammals and the birds, we generated a coefficient of variation
table to isolate the variation in the different measurements. The
coefficients of variation showed that the variation in measurements for all layers in the chicken was substantially smaller
compared with the variation found in the mammalian measurements (Fig. 9). Thus the appearance of a virtually uniform
BGB is reflected in the relatively small dispersion of values in
the chicken. Although there was little variation in the thicknesses of all of the layers in the chicken, the interstitium
displayed a particularly low coefficient of variation, indicating
Table 2. Morphometric data of the chicken, horse, dog, and rabbit lungs
Animal (n)
No. of
Measurements
Endothelium
Interstitium
Epithelium
Total BGB
Chicken (5)
Horse (2)
Dog (3)
Rabbit (3)
1,479
662
651
714
0.135⫾0.06
0.248⫾0.13
0.248⫾0.25
0.216⫾0.14
0.045⫾0.02
0.386⫾0.64
0.319⫾0.51
0.174⫾0.23
0.086⫾0.05
0.287⫾0.23
0.230⫾0.12
0.202⫾0.17
0.266⫾0.09
0.921⫾1.00
0.796⫾0.86
0.593⫾0.40
Values are averages ⫾ SD in ␮m and are averaged for all individuals of the same species.
AJP-Lung Cell Mol Physiol • VOL
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Thickness of total BGB. We pooled all individual chicken
data to make the comparisons between our bird data and the
mammalian lung data previously collected in our lab. All data
sets showed distribution profiles that were heavily skewed to
the right (Table 1) (5, 12). However, in our chicken data,
nonparametric statistical analysis showed no significant differences in thickness between individuals for total thickness or for
each layer, with a few exceptions involving chicken 5 (Table 1).
The average total thickness of the BGB in the chicken lungs
was 0.266 ⫾ 0.09 ␮m, which was, as expected, significantly
thinner than the total thickness of the BGB of the rabbit, dog,
and horse (0.593 ⫾ 0.40, 0.796 ⫾ 0.86, and 0.921 ⫾ 1.00 ␮m,
respectively; Tables 1 and 2 and Fig. 6). Thus the BGB in the
chicken lung was 54% thinner than the BGB in the rabbit, 66%
thinner than that in the dog, and 70% thinner than that in the
horse. On average, the tissue thickness of the avian BGB is
known to be thinner than that of mammals, and our data show
that this holds true for weak flyers such as the domestic chicken
(order: Galliformes). Quantitative analysis of harmonic mean
thickness data appears to show a slight positive correlation
with body mass in mammals and birds (26, 30, 56). To test for
a similar allometric effect in our arithmetic mean thickness
data, we fitted a least squares regression line to logarithmically
transformed published data for the arithmetic mean thicknesses
of the total BGB of a variety of mammalian and avian species
ranging in body mass from 5 g to 500 kg. There was a slight
correlation between body mass and BGB thickness in birds and
mammals over a large range of masses (mammals, n ⫽ 40, F ⫽
5.1, P ⫽ 0.03, r2 ⫽ 0.12; birds n ⫽ 36, F ⫽ 1.9, P ⫽ 0.18, r2 ⫽
0.05); however, it appears that most of the variation is attributable to other, unknown, variables.
Thickness of components of BGB. The thicknesses of the
endothelium, interstitium, and epithelium layers in the chicken
lung averaged 0.135 ⫾ 0.06, 0.045 ⫾ 0.02, and 0.086 ⫾ 0.05
␮m, respectively (Table 1 and Fig. 6). All individual layers
were significantly thinner in the chicken compared with that of
the horse, dog, and rabbit, as demonstrated in the relative
frequency histogram (Fig. 7). The much smaller dispersion of
the data in the chicken compared with the mammals also
L773
PULMONARY BLOOD-GAS BARRIER
all layers of the BGB
and rabbit. Statistical
significantly different
significantly different
Fig. 7. Relative frequency histograms of
the various layers of the BGB in the
chicken, horse, dog, and rabbit. Chicken
curves are shifted to the left indicating a
higher proportion of thin measurements.
AJP-Lung Cell Mol Physiol • VOL
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Fig. 6. Bar graphs of average thicknesses ⫾ SD of
measured in the lungs of the chicken, horse, dog,
symbols: a, significantly different from chicken; b,
from horse; c, significantly different from dog; d,
from rabbit.
a remarkably uniform thin interstitial layer in the blood capillaries of the bird.
As stated above, a transverse section of the pulmonary
capillaries of the mammalian lung is characteristically polarized in appearance. One side of the capillary wall is attenuated
to minimize resistance to diffusion during gas exchange, while
the other side is thickened with connective tissue (Fig. 4). For
example, in the human lung ⬃50% of the BGB is thin and 50%
thick (13). The thick portion of the capillary wall tends to have
a larger extracellular space due, in part, to the prevalence of
type I collagen fibrils; there may also be intervening fibroblasts. This collagen cable is thought to maintain the integrity
of the large alveolar walls, which are suspended in the air
spaces of the lung (54). The thickened section of the pulmonary
capillary of the mammalian lung appears to play a necessary
“supportive” role and is therefore relatively resistant to diffusion,
thus effectively reducing the area of the capillary that is available
for gas exchange. The occurrence of thick and thin sections of the
interstitium increases the variation of thickness measurements for
both the interstitium and the total BGB, causing the high coefficient of variation in the mammals (Fig. 9). In contrast, the
chickens’ blood capillaries do not possess a thin and thick side.
Their capillaries have uniformly thin walls with subsequent low
L774
PULMONARY BLOOD-GAS BARRIER
variability and low coefficients of variation in the thickness
measurements of both the interstitium and total BGB. Functionally, the uniformly thin barrier translates into a higher proportion
of surface area that is maximized for gas exchange compared with
the mammal. Images of the mammalian pulmonary compared
with the blood capillary of the bird illustrate the differences (Figs.
3, inset, and 4).
Epithelial struts. Although birds have a semirigid lung (8,
9), in the absence of the structural support of a thick interstitium, the tissue in the avian lung is potentially susceptible to
mechanical damages sustained from elevated transmural pressures that occur with exercise and subsequent increased oxygen
consumption and elevated cardiac output. Support to the small
blood capillaries in the avian lung may be afforded by the
aforementioned epithelial “crossbraces” that are an anatomical
feature unique to the bird lung (Figs. 2 and 3) (20, 29, 43).
These epithelial struts are composed of abutting cytoplasmic
appendages from adjacent epithelial cells. They are extremely
attenuated at the midsection and then thicken at either end
where the epithelial pedicle joins to the blood capillary (Figs.
3 and 5) (20). While Maina (28) asserts that the air capillaries
are not themselves “rigid” and proposes that they are therefore
of little structural significance for support to the parabronchial
unit, we maintain that they serve a supportive roll in the tightly
AJP-Lung Cell Mol Physiol • VOL
packed air and blood capillary network within the fine structure
of the parabronchii (20, 43).
The ultrastructure of the epithelial crossbrace has not been
previously described in any great detail. Maina and King (29)
and others (20, 43) recognized the presence of this unique
structure in all bird species studied to date and observed that
the “rather uncommon sites” where two air capillaries share a
common wall are made of adjoining epithelial cell extensions.
Klika et al. (20) describes the crossbrace as being composed of
“paired retinacula”; however, the provided images display
evidence of edema and artifactual thickening. Our images of
the epithelial bridges from tracheal-fixed chickens show very
closely abutting epithelial cell extensions that are connected
via an interstitial space, but it is not always visible (Fig. 5). For
this reason, there has been some conjecture regarding the
presence or absence of an interstitial space between the two
extensions and how the cells would otherwise be joined if an
interstitium were lacking (20, 29). The epithelial extensions are
thickened where they attach at the blood capillary, and in the
area of thickening, it appears as if a portion of the blood
capillary’s interstitium diverges to form a smaller “tributary”
that runs the length of the epithelial bridge to join the interstitium of another blood capillary (Fig. 5). Although several
authors have stated that a basal lamina appears to be lacking in
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Fig. 8. Cumulative frequency plots constructed
from relative frequency histograms for all layers
of the BGB measured in the chicken, horse, dog,
and rabbit. A steeper curve indicates less variability in thickness measurements.
PULMONARY BLOOD-GAS BARRIER
the epithelial cells of the bird lung (1, 20, 29, 33, 43), we
conclude that the presence of an interstitial layer between the
epithelia of the crossbraces indicates that the epithelial layers
are anchored to a basal lamina. In mammals, the interstitial
space plays a large role in the maintenance of fluid balance in
the lung (11, 49). A similar process seems likely in birds,
although avian lymphatics are poorly understood (59, 60).
DISCUSSION
We found that all of the layers of the BGB were thinner in
the chicken compared with that in the lungs of the horse, dog,
and rabbit, including the interstitial layer. In addition, our
results show that the circumferential thickness of the interstitium layer is uniform in the chicken compared with that of the
mammal. Finally, the allometric scaling effect on the BGB
appears to be negligible in both mammals and birds. We are
able to make these comparisons because despite the considerable anatomical differences between mammalian and avian
lungs, the structural design of the BGB is essentially the same.
In the avian lung, there has been a considerable amount of
harmonic mean thickness data published. However, while this
measurement is appropriate for estimating the diffusing capacity of the lung, in the context of tissue strength, the arithmetic
mean thickness is a better measurement (33, 55, 57). In
measurements of arithmetic and harmonic mean thickness over
a range of body masses, there is little or no correlation of body
mass to barrier thickness in either birds or mammals. This lack
of correlation suggests that the thickness of the BGB has been
allometrically optimized (33). Instead, variations in BGB arithmetic mean thickness among avian species have been attributed to variations in energetic demand. For example, a highly
energetic hummingbird (Colibri coruscans) has an average
arithmetic mean thickness of 0.183 ␮m, compared with the
relatively inactive guinea fowl (Numida meleagris), whose
average thickness is 1.12 ␮m (1, 7). The parallel metabolic
correlation among mammalian species appears not to exist; for
example, the highly aerobic dog and bat have arithmetic mean
thickness values of 1.78 and 1.39 ␮m, respectively, compared
with the naked mole rat and baboon (1.09 and 1.12 ␮m,
AJP-Lung Cell Mol Physiol • VOL
respectively) (24, 30, 32, 36). Most avian species have a
thinner BGB compared with all but the smallest mammals (14,
33). A notable exception exists in the common kestrel (Falco
tinnunculus), whose BGB arithmetic mean thickness averaged
1.66 ␮m (30).
Although the thickness of the total BGB from ⬃30 species
of birds has been reported, to our knowledge there has been
only one previous study that has measured the thicknesses of
the different components of the BGB in an avian lung. Weidner
et al. (59) used a tracheal instillation fixation method and a
line-intercept counting technique to estimate endothelial cell
and septal interstitial thickness in volume-loaded chickens.
Their control (nonloaded) chickens measured endothelium and
interstitium values of 0.16 ⫾ 0.04 and 1.70 ⫾ 0.49 ␮m,
respectively (59). The much higher value for interstitial thickness is possibly explained by some edema formation during
lung fixation. Interestingly, although unquantified, the basal
lamina of the emperor penguin (Apternodytes forsteri) was
observed to be unusually thick and possessed numerous collagen fibrils. The added support to the interstitium was thought to
be an adaptation to protect the BGB from compression stress
during the high hydrostatic pressures incurred during deep
diving in these marine birds (63).
Flapping flight is an energetic form of locomotion that has
an intermediate cost-of-transport between vertebrate animals
that swim (most efficient) and those that run (least efficient)
(44, 51, 52). However, flying does require a high rate of energy
expenditure per unit time despite the behavioral adaptations
that make volant locomotion extremely efficient in many birds
(38). Exercising birds likely have elevated pulmonary capillary
pressures in a manner similar to exercising mammals, although
pulmonary vascular pressures have not been measured in flying
birds. Although the domestic chicken is a weak flyer, we found
that it has a total BGB thickness that is comparable to (or
thinner than) that of other bird species. If pulmonary capillary
pressures in exercising birds are similar or higher than that
found in exercising mammals of similar mass and the total
BGB thickness in birds is similar to or thinner than that found
in mammals, there is the potential for birds to suffer from
pulmonary stress failure as mammals do. Moreover, in mammals there appears to be a trend where the thinner the interstitium layer of the BGB, the lower the pressure required to
induce pulmonary stress failure (5). However, to the best of our
knowledge, stress failure has not been reported in birds (although there has been a report of the presence of erythrocytes
in the air spaces of bird lungs), and the evolutionary success of
this diverse taxon indicates that it is unlikely a normal occurrence. We therefore suspect that the avian BGB possesses
additional structural support in the microanatomy of the lung
parenchyma that may not have a mammalian analog.
The avian respiratory system is structurally and functionally
different from the mammalian respiratory system and arguably
more efficient. The small, flow-through lungs experience unidirectional air flow during inspiration and expiration. Ventilation is separated from gas exchange and is performed by paired
air sacs located throughout the thoracic and abdominal cavities.
The blood capillaries and air capillaries are intricately intertwined and form a meshwork of parenchyma tissue separated
by numerous atria in the parabronchi (8, 9, 23). The diameter
of the air capillaries in an avian lung is a fraction of the size of
the mammalian alveoli; the resultant surface tension forces in
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
Fig. 9. Coefficient of variation bar graph calculated from thickness measurements of
the various layers of the BGB in the chicken, dog, rabbit, and horse. The coefficient of
variation of the interstitial layer in the chicken is considerably lower than that of the
mammals.
L775
L776
PULMONARY BLOOD-GAS BARRIER
ACKNOWLEDGMENTS
We are grateful to Dr. X. Zhang for writing the morphometry MatLab
program and two anonymous reviewers for helpful comments on the manuscript. Statistical assistance was provided by B. Rowlingson and M. Olfert.
GRANTS
This study was supported by National Institutes of Health Grants 5 T32
HL-07212-29 and 5 RO1-HL-60968-06.
REFERENCES
1. Abdalla MA, Maina JN. Quantitative analysis of the exchange tissue of
the avian lung (Galliformes). J Anat 133: 677– 680, 1982.
2. Abdalla MA, Maina JN, King AS, King DZ, Henry J. Morphometrics
of the avian lung. I. The domestic fowl (Gallus gallus variant domesticus).
Respir Physiol 47: 267–278, 1982.
AJP-Lung Cell Mol Physiol • VOL
3. Araya O, Palma P, Salvi M, Bustamante H, Vits L. Endoscopic
determination of exercise-induced pulmonary hemorrhage in Chilean
Criollo horses. Vet J 169: 311–313, 2005.
4. Bernhard W, Gebert A, Vieten G, Rau GA, Hohlfeld JM, Postle AD,
Freihorst J. Pulmonary surfactant in birds: coping with surface tension in
a tubular lung. Am J Physiol Regul Integr Comp Physiol 281: R327–R337,
2001.
5. Birks EK, Mathieu-Costello O, Fu Z, Tyler WS, West JB. Comparative
aspects of the strength of pulmonary capillaries in rabbit, dog and horse.
Respir Physiol 97: 235–246, 1994.
6. Doucet MY, Viel L. Alveolar macrophage graded hemosiderin score from
bronchoalveolar lavage in horses with exercise-induced pulmonary hemorrhage and controls. J Vet Intern Med 16: 281–286, 2002.
7. Dubach M. Quantitative analysis of the respiratory system of the house
sparrow, budgerigar, and violet-eared hummingbird. Respir Physiol 46:
43– 60, 1981.
8. Duncker HR. Structure of avian lungs. Respir Physiol 14: 44 – 63, 1972.
9. Duncker HR. Structure of the avian respiratory tract. Respir Physiol 22:
1–19, 1974.
10. Erickson HH, McAvoy JL, Westfall JA. Exercise-induced changes in
the lung of Shetland ponies: ultrastructure and morphometry. J Submicrosc Cytol Pathol 29: 65–72, 1997.
11. Fishman AP. Pulmonary edema. The water-exchanging function of the
lung. Circulation 46: 390 – 408, 1972.
12. Fu Z, Heldt GP, West JB. Thickness of the blood-gas barrier in
premature and 1-day-old newborn rabbit lungs. Am J Physiol Lung Cell
Mol Physiol 285: L130 –L136, 2003.
13. Gehr P, Bachofen M, Weibel ER. The normal human lung: ultrastructure
and morphometric estimation of diffusion capacity. Respir Physiol 32:
121–140, 1978.
14. Gehr P, Sehovic S, Burri PH, Claasen H, Weibel ER. The lung of
shrews: morphometric estimation of diffusion capacity. Respir Physiol 44:
61– 86, 1980.
15. Hanel B, Law I, Mortensen J. Maximal rowing has an acute affect on the
blood-gas barrier in elite athletes. J Appl Physiol 95: 1076 –1082, 2003.
16. Hopkins SR, Schoene RB, Martin TR, Henderson WR, Spragg RG,
West JB. Intense exercise impairs the integrity of the pulmonary bloodgas barrier in elite athletes. Am J Respir Crit Care Med 155: 1090 –1094,
1997.
17. Hopkins SR, Schoene RB, Henderson WR, Spragg RG, West JB.
Sustained submaximal exercise does not alter the integrity of the lung
blood-gas barrier in elite athletes. J Appl Physiol 84: 1185–1189, 1998.
18. Hughes G. The Vertebrate Lung. Burlington, NC: Oxford Univ. Press/
Carolina Biological Supply, 1979.
19. Jones JH, Effmann EL, Schmidt-Nielsen K. Lung volume changes
during respiration in ducks. Respir Physiol 59: 15–25, 1985.
20. Klika E, Scheuermann DW, De Groodt-Lasseel MH, Bazantova I,
Switka A. Anchoring and support system of pulmonary gas-exchange
tissue in four bird species. Acta Anat (Basel) 159: 30 – 41, 1997.
21. Lapointe JM, Vrins A, McCarvill E. A survey of exercise-induced
pulmonary haemorrhage in Quebec Standardbred racehorses. Equine Vet J
26: 482– 485, 1994.
22. Macklem PT, Bouverot P, Scheid P. Measurements of the distensibility
of the parabronchi in duck lungs. Respir Physiol 38: 23–35, 1979.
23. Maina JN. A scanning EM study of the air and blood capillaries of the
lung of the domestic fowl (G domesticus). Experientia 38: 614 – 616, 1982.
24. Maina JN. The morphology and morphometry of the adult normal baboon
lung Papio anubis. J Anat 150: 229 –246, 1987.
25. Maina JN. The morphology of the lung of the black mamba Dendroaspis
polylepis (Reptilia: Ophidia: Elapidae). A scanning and transmission
electron microscopic study. J Anat 167: 31– 46, 1989.
26. Maina JN. Morphometries of the avian lung: the structural-functional
correlations in the design of the lungs of birds. Comp Biochem Physiol A
105: 397– 410, 1993.
27. Maina JN. Fundamental structural aspects and features in the bioengineering of the gas exchangers: comparative perspectives. Adv Anat Embryol Cell Biol 163: 1–108, 2002.
28. Maina JN. Spectacularly robust! Tensegrity principle explains the mechanical strength of the avian lung. Respir Physiol Neurobiol 155: 1–10,
2007.
29. Maina JN, King AS. The thickness of avian blood gas barrier: qualitative
and quantitative observations. J Anat 134: 553–562, 1982.
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
the air capillaries may be very strong and patency is ensured by
the presence of a unique surfactant-like material (4, 8, 9).
Because of the physiologically impossible forces required to
reinflate the air capillaries upon collapse, it was long held that
the avian lung was completely rigid and noncompliant (40).
We now know that this is not entirely true; during one respiration cycle in a resting bird, lung volume changes 1.4% (19)
and the parabronchi themselves show limited compliance (22).
Of interest in the latter study was the observation that the
fragile air capillaries did not show evidence of collapse upon
compression of the respiratory system; the authors remarked
that possibly “unknown factors serve to confer a remarkable
stability on these fine structures” (22). Coupled with empirical
data by Powell et al. (41) demonstrating a lack of recruitment
and distention of blood capillaries in the avian lung during
temporary unilateral pulmonary arterial occlusion, these studies provide further evidence that additional structural support
in the avian lung affords it remarkable integrity despite the thin
interstitial layer.
It has been proposed that the epithelial crossbraces of the
avian lung formed a spiderweb-like system of air capillaries
that functioned to anchor and support the air capillaries to the
lung parenchyma (20, 43). This theory was suggested because
of the noticeable lack of significant extracellular space and
supportive connective tissue surrounding the blood capillaries
(20, 29). However, the provided transmission electron microscope images of the crossbraces show considerable edema. We
agree that the epithelial crossbraces provide support to the gas
exchange tissue, and we further propose that the epithelial
extensions serve to reduce or absorb the transmural stresses
that occur within the blood capillaries upon exercise. The
presence of these epithelial bridges creates a honeycomb-like
structure in the bird lung upon cross-section of the parenchyma, and the honeycomb shape, much like a geodesic dome,
efficiently transmits stress throughout the entire structure and
has a high ratio of volume to weight, therefore requiring a
minimum amount of tissue for metabolic maintenance. Additionally, our images show a continuous extracellular matrix
from blood capillary to air capillary, which we believe provides evidence that the epithelial struts may play an important
role in lung fluid balance. Finally, the epithelial bridges may
themselves be afforded support by the anchoring system to the
thicker blood capillaries. The anchoring of the air capillaries to
the blood capillaries [as proposed by Klika et al. (20)] may
stabilize the very thin epithelial extensions during the small in
volume changes of the avian lung during respiration and aid in
the maintenance of small airway patency.
PULMONARY BLOOD-GAS BARRIER
AJP-Lung Cell Mol Physiol • VOL
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
ogy, edited by Snow DH, Persson SGB, and Rose BJ. Cambridge: Granta
editions, 1983.
Szidon JP, Pietra GG, Fishman AP. The alveolar-capillary membrane
and pulmonary edema. N Engl J Med 286: 1200 –1204, 1972.
Tsukimoto K, Mathieu-Costello O, Prediletto R, Elliott AR, West JB.
Ultrastructural appearances of pulmonary capillaries at high transmural
pressures. J Appl Physiol 71: 573–582, 1991.
Tucker VA. Energetic cost of locomotion in animals. Comp Biochem
Physiol A 34: 841– 846, 1970.
Tucker VA. Metabolism during flight in the laughing gull, Larus atricilla.
Am J Physiol 222: 237–245, 1972.
Voynick BT, Sweeney CR. Exercise-induced pulmonary hemorrhage in
polo and racing horses. J Am Vet Med Assoc 188: 301–302, 1986.
Weibel ER. The ultrastructure of the alveolar-capillary membrane or
barrier. In: The Pulmonary Circulation and Interstitial Space, edited by
Fishman AP and Hecht HH. Chicago: The University of Chicago Press,
1969.
Weibel ER. Morphometric estimation of pulmonary diffusing capacity. I.
Model and method. Respir Physiol 11: 54 –75, 1970/71.
Weibel ER. Morphometric estimation of pulmonary diffusing capacity. V.
Comparative analysis of mammalian lungs. Respir Physiol 14: 26 – 43,
1972.
Weibel ER, Knight BW. A morphometric study on the thickness of the
pulmonary air-blood barrier. J Cell Biol 21: 367–384, 1964.
Weideman H, Schoeman SJ, Jordaan GF, Kidd M. Epistaxis related to
exercise-induced pulmonary haemorrhage in South African Thoroughbreds. J S Afr Vet Assoc 74: 127–131, 2003.
Weidner WJ, Selna LA, McClure DE, DeFouw DO. Effect of extracellular fluid volume expansion on avian lung fluid balance. Respir Physiol
91: 125–136, 1993.
Weidner WJ, Kinnison JR. Effect of extracellular fluid volume expansion on the interparabronchial septum of the avian lung. J Comp Pathol
127: 219 –222, 2002.
Weiler-Ravell D, Shupak A, Goldenberg I, Halpern P, Shoshani O,
Hirschhorn G, Margulis A. Pulmonary oedema and haemoptysis induced
by strenuous swimming. BMJ 311: 361–362, 1995.
Welling LW, Grantham JJ. Physical properties of isolated perfused renal
tubules and tubular basement membranes. J Clin Invest 51: 1063–1075,
1972.
Welsch U, Aschauer B. Ultrastructural observations on the lung of the
emperor penguin (Apternodytes forsteri). Cell Tissue Res 243: 137–144,
1986.
West JB, Tsukimoto K, Mathieu-Costello O, Prediletto R. Stress
failure in pulmonary capillaries. J Appl Physiol 70: 1731–1742, 1991.
West JB, Mathieu-Costello O. Strength of the pulmonary blood-gas
barrier. Respir Physiol 88: 141–148, 1992.
West JB, Mathieu-Costello O, Jones JH, Birks EK, Logemann RB,
Pascoe JR, Tyler WS. Stress failure of pulmonary capillaries in racehorses with exercise-induced pulmonary hemorrhage. J Appl Physiol 75:
1097–1109, 1993.
Whitwell KE, Greet TR. Collection and evaluation of tracheobronchial
washes in the horse. Equine Vet J 16: 499 –508, 1984.
292 • MARCH 2007 •
www.ajplung.org
Downloaded from http://ajplung.physiology.org/ by 10.220.33.6 on June 18, 2017
30. Maina JN, King AS, Settle JG. An allometric study of the pulmonary
morphometric patterns of birds, with mammalian comparison. Philos
Trans R Soc Lond B Biol Sci 326: 1–57, 1989.
31. Maina JN, Maloiy GMO. The morphometry of the lung of the African
lungfish (Protopterus-Aethiopicus): its structural-functional correlations.
Proc Roy Soc Lond Ser B 224: 399 – 420, 1985.
32. Maina JN, Maloiy GMO, Makanya AN. Morphology and morphometry
of the lungs of two East African mole rats, Tachyoryctes splendens and
Heterocephalus glaber (Mammalia, Rodentia). Zoomorphology 112: 167–
179, 1992.
33. Maina JN, West JB. Thin and strong! The bioengineering dilemma in the
structural and functional design of the blood-gas barrier. Physiol Rev 85:
811– 844, 2005.
34. Maxwell D, Rees PJ. Exertion-induced haemoptysis. Lancet 341: 375,
1993.
35. McKechnie JK, Leary WP, Noakes TD, Kallmeyer JC, MacSerraigh
ETM, Olivier LR. Acute pulmonary oedema in two athletes during a
90-km running race. S Afr Med J 56: 261–265, 1979.
36. Meban C. Thickness of the air-blood barriers in vertebrate lungs. J Anat
131: 299 –307, 1980.
37. Michel RP, Cruz-Orive LM. Application of the Cavalieri principle and
vertical sections method to lung: estimation of volume and pleural surface
area. J Microsc 150: 117–136, 1988.
38. Norberg UM. Energetics of flight. In: Avian Energetics and Nutritional
Ecology, edited by Carey C. New York: Chapman and Hall, 1996.
39. Pascoe JR, Ferraro GL, Cannon DH, Arthur RM, Wheat JD. Exercise-induced pulmonary hemorrhage in racing thoroughbreds: a preliminary study. Am J Vet Res 42: 703–707, 1981.
40. Pattle RE. Lung surfactant and lung lining in birds. In: Proceedings in
Life Sciences: Respiratory Function in Birds, Adult and Embryonic.
Satellite symposium of the 27th International Congress of Physiological
Sciences, Paris, 1977, edited by Piiper J. Berlin: Springer-Verlag, 1978.
41. Powell FL, Hastings RH, Mazzone RW. Pulmonary vascular resistance
during unilateral pulmonary artery occlusion in ducks. Am J Physiol Regul
Integr Comp Physiol 249: R34 –R43, 1985.
42. Powell FL, Mazzone RW. Morphometrics of rapidly frozen goose lung.
Respir Physiol 51: 319 –332, 1983.
43. Scheuermann DW, Klika E, DeGroot-Lasseel MHA, Bazantova I,
Switka A. An electron microscope study of the parabronchial epithelium
in the mature lung of four bird species. Anat Rec 249: 213–225, 1997.
44. Schmidt-Nielsen K. Locomotion: energy cost of swimming, running, and
flying. Science 177: 222–228, 1972.
45. Shimura S, Martin CJ, Boatman ES, Dhand R. A role for interstitial
matrix in tissue tension of alveolar wall. Respir Physiol 62: 293–303,
1985.
46. Swayne GT, Smaje LH, Bergel DH. Distensibility of single capillaries
and venules in the rat and frog mesentery. Int J Microcirc Clin Exp 8:
25– 42, 1989.
47. Sweeney CR, Humber KA. Cytologic findings of tracheobronchial aspirates from 66 Thoroughbred racehorses. Am J Vet Res 53: 1172–1175,
1992.
48. Sweeney CR, Soma LR. Exercise-induced pulmonary haemorrhage in
horses after different competitive exercises. In: Equine Exercise Physiol-
L777