We have investigated the basis and
implications of pneumoconstriction by measuring disposition and
quantities of 
-smooth muscle actin in rat and guinea pig lungs and
modeling its effects on lung recoil and compliance. A robust marker of
contractility, -smooth muscle actin appears in smooth muscle or
myofibroblast-like cells in pleura, airways, blood vessels, and
alveolar ductal tissues. In each site, we measured its transected area
by immunofluorescent staining and frequency-modulated scanning confocal
microscopy. We incorporated these data in a model of the parenchyma
consisting of an extensive elastic network with embedded contractile
structures. We conclude that contraction at any one of these sites
alone can decrease parenchymal compliance by 20-30% during tidal
breathing. This is due mostly to the stiffness of activated contractile
elements undergoing passive cycling; constant muscle tension would have little effect. The magnitude of the effect corresponds with known responses of the lung to hypocapnia, consistent with a homeostatic function in which gas exchange is defended by redistributing
ventilation away from overventilated units.
alveolar duct; intensity-modulated multiple-wavelength scanning
confocal microscopy; lung compliance; ventilation-perfusion ratio; hypocapnia
 |
INTRODUCTION |
A VARIETY OF CONSTRICTING
AGENTS can increase the lung's elastic recoil pressure and
elastance (3, 6, 7, 25, 30, 42), and conversely relaxing
agents can decrease its elastic recoil pressure (9).
The static and passive dynamic properties of strips cut from lung
parenchyma are influenced by smooth muscle agonists (10, 17, 32,
38, 50). Hypocapnia can reduce the lung's dynamic compliance on
the order of 30% (1, 18, 41, 43, 48, 49), and this may
represent a mechanism for control of ventilation (6, 18).
The source of this contractility, however, is unclear. Structural
considerations suggest that contractile elements in the airways, blood
vessels, or alveolar ducts can draw in on the surrounding parenchyma,
thereby increasing its tensions. Contractile interstitial cells located
within the alveolar walls themselves might increase the tensions of the
alveolar parenchyma (19, 50). Airways acting alone may
have an effect, because selective pharmacological constriction of
conducting airways substantially reduces quasi-static and dynamic lung
compliances (30). Other locations may also have
independent effects, because agonists can change tissue dynamic properties independently of airway resistance (17).
Furthermore, it has been argued that the contractility observed in
subpleural strips of lung parenchyma may not be due to vascular or
airway smooth muscle contraction (10, 19, 38).
We thought that some insight might be gained from knowing more about
the contractile material in each location and estimating the effects of
each on lung recoil and compliance. Accordingly, we chose -smooth
muscle actin as a major marker of contractile capability, and we
applied a variety of microscopic techniques to sections of rat and
guinea pig lungs to characterize the cells that contained it and to
determine its locations and quantities in the lung parenchyma. We then
modeled its effects by considering the lung parenchyma as an elastic
network with embedded contractile elements. In the model, the elastic
network was given the known elastic properties of relaxed lung, and the
contractile elements were given the plastoelastic properties of
passively cycled, activated smooth muscle. This model, incorporating
the anatomical data from the first part of the study, estimates the
effects of activation of the contractile elements in each location on
regional parenchymal recoil and compliance during tidal breathing.
Our results show, in the two species examined, that any one of
the locations alone has sufficient contractile material to account for
modest but significant reduction of parenchymal compliance under
dynamic conditions. This is consistent with the hypothesis that this
mechanism, locally responsive to carbon dioxide levels, defends the
overall efficiency of gas exchange by modulating local ventilation-perfusion (
/
) ratios.
Glossary
Primary symbols
Forces
| F |
force
|
| T |
tension (force per length)
|
| P |
pressure (force per area)
|
 |
nominal stress (force per area at lo)
|
Dimensions
| A |
area
|
| V |
volume
|
 |
density of embedded structure in the lung unit
|
| TLC |
maximal physiological lung unit volume
|
| FRC |
relaxed physiological lung unit volume
|
Mechanical properties
 |
bulk modulus
|
| µ |
shear modulus
|
| lo |
length for maximal muscle stress
|
Geometry
 |
angle of muscle fiber to the axis of airway cylinder or equator of
ductal sphere
|
| a |
radius of sphere or cylinder
|
| u |
displacement of a boundary
|
Modifiers
Geometry
| ax |
axial
|
| circ |
circumferential
|
| tm |
transmural
|
| 1 |
inner radius of surrounding (thick-walled) parenchyma, or equivalently
outer radius of (thin-walled) duct or airway
|
| 2 |
outer radius of lung unit
|
Structures
| ce |
contractile element
|
| tiss |
tissue
|
| L |
lung
|
| duct |
alveolar duct
|
| aw |
airway
|
| par |
parenchyma
|
| pl |
pleura
|
Condition
| active |
contractile elements activated
|
| passive |
contractile elements relaxed
|
| |
METHODS |
Immunostaining
We used a monoclonal antibody to mark -smooth muscle actin
(44). It is highly specific, having been raised against
the highly conserved amino terminal 10 amino acids of the smooth muscle isoform of -actin. It is nonetheless very broad in its species reactivity, including human, rat, and chicken (44), cow,
frog, goat, guinea pig, mouse, rabbit, dog, sheep, and snake (Sigma Chemical), and hamster and pig (personal observations of an author, C. Kuhn III).
The antibody was visualized by three different techniques.
1) To determine the disposition of -smooth muscle actin
on large sections with moderately well preserved septal configuration, we examined frozen sections by transmission light microscopy with the
antibody made visible by diaminobenzidine-horseradish peroxidase (DAB-HRP). 2) To quantify -smooth muscle actin in various
sites over large fields of view, we prepared paraffin sections and used intensity-modulated multiple-wavelength scanning confocal microscopy (IMS) to measure the areas of fluorescence of Cy-3 linked to the antibody and to obtain anatomical context from fluorescent images of
BODIPY-SE (Molecular Probes, Eugene, OR) staining lung tissue generally. 3) To determine the subcellular patterns
and concentrations of -smooth muscle actin, we examined epoxy
sections by transmission electron microscopy with the antibody
visualized by immunogold.
Organ Isolation, Fixation, and Sample Preparation
Experimental animals comprised six female, viral antigen-free
Sprague-Dawley rats (241-348 g) and three guinea pigs
(423-523 g). Each was briefly sedated with carbon dioxide (dry ice
vapor in a desiccator) and then anesthetized with pentobarbital sodium (30-41.5 mg/kg body wt im). Tracheal intubation was
immediately performed, and the animal was maintained on a ventilator
until the heart and lungs could be rapidly removed en bloc.
Transmission light microscopy.
Rat lungs were filled with a 1:1 mixture of cryostabilizing compound
(Tissue-Tek OCT compound; Sakura Finetek, Torrance, CA) and saline.
Blocks of tissue were placed in pools of OCT compound on copper chucks
and frozen by immersion in 2-methyl butane cooled in dry ice. Frozen
sections were prepared on a cryotome with section thickness of 6 µm.
Sections were thaw-mounted on silanized slides and dried for 30-60
min. After fixation for 5 min in 80% ethanol, the sections were
hydrated in phosphate buffered saline-horse serum albumin-horse serum
and incubated with anti--smooth muscle actin antibody 1:1,000 for
1 h. The bound antibody was visualized by an immunoperoxidase
reaction by using the avidin-biotin complex method with commercial kits
(Vectastain, Vector Laboratories, Burlingame, CA).
3,3'-Diaminobenzidine was used as a chromogen.
Confocal fluorescence light microscopy.
Rat and guinea pig lungs were degassed under vacuum, filled with
fixative of methanol-chloroform-glacial acetic acid 60:30:10 by volume
(29), and submerged in a bath of the same fixative for
2 h, with the meniscus of the liquid column continuous with the
trachea maintained at 25 cm above the bath, which we consider to be
close to maximal physiological volume (total lung capacity; TLC). The
left (unlobated) lung was sliced transversely into three to eight
slices ~2-3 mm thick, dehydrated in absolute ethanol, embedded
in paraffin, cut ~7 µm thick, deparaffinized by standard techniques, and affixed to slides. For fluorescent staining of -smooth muscle actin, slides were incubated in 0.1% Triton X-100 in
0.1 M Tris-buffered saline (TBS, pH 7.6) for 10 min, washed in
TBS-bovine serum albumin (TBS-BSA) three times for 10 min each, incubated in TBS-BSA containing 1% horse serum for 30 min, incubated in mouse anti-human -smooth muscle actin antibody (1:100 dilution in
TBS-BSA, containing 1% horse serum and 0.1% Tween 20, Sigma Chemical,
St. Louis, MO) overnight at 4°C, washed with TBS three times for 10 min each, incubated with biotinylated goat anti-mouse IgG (1:100
dilution with PBS-BSA-HS; Sigma Chemical) for 2 h, washed in PBS
three times for 10 min each, incubated in the dark with ExtraAvidin
Cy-3 complex (1:100 dilution with 0.1 M PBS, pH 8.2; Sigma Chemical)
for 1 h, and washed in distilled water two times for 5 min each.
For fluorescent counterstaining of tissue in general, slides were
incubated in 1:40,000 dilution of BODIPY-SE in 0.1 M NaHCO3
(Molecular Probes) for 30 min and washed in distilled water two times
for 5 min each.
Transmission electron microscopy.
Rat lungs were similarly prepared, except that the intratracheal
fixative was 2% paraformaldehyde, 0.1% glutaraldehyde in 0.1 M
cacodylate buffer, pH 7.4. The slices, after dehydration in absolute
ethanol, were embedded in LR white resin (Polysciences; Warrenton, PA).
Ultrathin sections were cut and mounted on gold grids (Polaron
Equipment; Watford, UK). Grids were incubated overnight with -smooth
muscle actin antibody as described above, washed, incubated with goat
anti-mouse IgG gold (1:40 dilution with the same diluent as used for
the primary antibody, Sigma Chemical) for 2 h, washed with TBS-BSA
two times for 5 min and then with distilled water two times for 5 min;
air-dried overnight; and finally counterstained with uranyl acetate and
then lead nitrate.
Imaging and Field of View Selection
Transmission light microscopy.
Sections were viewed with a Zeiss Photomicroscope I with ×25
objective, 1.25 intermediate lens. A Dage-MTI CCD video camera fed a
signal to a Sony Trinitron monitor and Sony Video Printer. The images
on the monitor were ×375 final magnification. Forty-two contiguous,
nonoverlapping fields of view covering the entire usable area of the
section were acquired.
Confocal fluorescence light microscopy.
Sections were examined with ×63 (rat) or ×100 (guinea pig)
objectives. The optimal pinhole size in the confocal detection unit was
selected so as to collect as much light as possible while still
obtaining nearly maximal depth resolution. This depends on the ratio of
magnification to numerical aperture. It corresponded to a diameter of
20 µm (rat) or 40 µm (guinea pig) in the image plane of the
objective. Cy-3 was excited by a laser light beam at 568-nm wavelength
and detected at 620 ± 30 nm (red). BODIPY-SE was excited at 488 nm and detected at 525 ± 25 nm (green). Because of overlap
between the Cy-3 and BODIPY-SE images, we used a recently developed
adaptation of IMS and morphology-based image processing (see
DISCUSSION). The energies and intensity modulations of the two excitation beams were monitored before imaging of each sample. Pixels represented areas 0.32 µm on a side for the rat and 0.2 µm
for the guinea pig.
As seen in the raw 620-nm images (Fig.
1A), Cy-3 stained large masses
of -smooth muscle actin in the smooth muscle bands in walls of
airways and blood vessels and smaller groupings of -smooth muscle
actin in the alveolar ductal structures. Images were processed with the
sequence of morphological operations "2× dilate, 4× convex hull
creation, 2× erode, and 1× isolated pixel removal" to remove
isolated pixels and to collect pixels enfolded by a group of -smooth
muscle actin pixels or between two or more neighboring groups
(35). Those pixels preserved after thresholding and image
processing were considered -smooth muscle actin (Fig. 1B). As seen on the raw 525-nm images (Fig. 1C),
BODIPY-SE stained all tissue, including -smooth muscle
actin-containing structures. Tissue was readily distinguished from
generally dimmer air and lumen pixels by thresholding at an
interactively chosen intensity. The same sequence of morphological
operations that separated -smooth muscle actin from tissue on the
620-nm images was also effective on the 525-nm images in smoothing the
edges of tissue profiles, filling internal holes due to noise, and
repairing occasional gaps in thin capillary wall profiles. Small
capillary lumens were also filled, although lumens of arterioles and
venules were not, in general. The pixels preserved after this operation
(both -smooth muscle actin- and non--smooth muscle
actin-containing) were considered tissue (Fig. 1D). Each
animal generated three to five lung slices. Systematic sampling was
performed on a rectilinear grid with constant spacing and random
starting point; 2.7 and 1.5% of the areas of the sections of rat and
guinea pig lung, respectively, were recorded.
View larger version (90K):
[in this window]
[in a new window]
|
Fig. 1.
Four intensity-modulated multiple-wavelength scanning
confocal micrographs of the same field of view of a rat lung section.
A: raw image for Cy-3 immunofluorescence (620 nm).
B: same image after thresholding and morphology-based image
processing. C: raw image for BODIPY-SE fluorescence (525 nm). D: same image after thresholding and morphology-based
image processing. Pixels in A and C are shown
with intensity registered on a color scale. White pixels in
B and D are those selected as -smooth muscle
actin and tissue respectively. Areal densities (Table 2) are obtained
from the counts of pixels in B, categorized by the
anatomical context of D.
|
|
Transmission electron microscopy.
Micrographs of sections stained with immunogold were viewed at an
instrumental magnification of ×20,000 and were printed at a final
magnification of ×68,000. All micrographs from a single animal were of
tissue sections stained under the same conditions. Gold particles
decorated smooth muscle cells in all airway and blood vessel walls and
were very rare in airspaces or vessel lumens, indicating adequate
sensitivity and specificity.
Quantifying -Smooth Muscle Actin in Anatomical Context
We first addressed the quantitative distribution of -smooth
muscle actin among three sites: blood vessels, airways, and alveolar structures. On the confocal fluorescence light micrographs, we identified all airways (terminal bronchioles and larger) on each field-of-view's original image pair by referring principally to the
525-nm channel's images, but also to the 620-nm channel's images for
clarification of smooth muscle morphology. A polygon was drawn closely
around the airway, and the tissue within was categorized as airway.
Analogous images were produced for blood vessel (arterioles, venules,
and larger) and unknown regions (recognizable as airway or blood vessel
but not further classifiable). All remaining tissue was considered
alveolar. All pixels selected as -smooth muscle actin or
non--smooth muscle actin in each of these anatomic locations were
registered from 790 image pairs from the rats and 240 from the guinea
pigs. Areal densities of -smooth muscle actin or tissue were
calculated from the fractions of pixels of each category in each
location. For example, the areal density of airway contractile elements
(Ace) at TLC
[(Aceaw/AL)TLC]
is the product of the morphometrically observed ratio of the number of
pixels recording -smooth muscle actin in airways to the number of
pixels recording lung tissue and the areal density of tissue in the
fully distended lung. The latter, being stereologically identical with the corresponding volumetric density, was obtained as the mass density
of the lung at TLC divided by the mass density of lung tissue, with the
assumptions of 0.067 and 1 g/ml, respectively. The use of summed
intensities instead of summed areas did not materially affect the
estimates of the quantitative distribution of -smooth muscle actin.
Our measurements of the fraction of muscle in the airway wall by the
immunofluorescent IMS technique (0.53/6.9 = 8%; see Table 2) were
similar to previous measurements obtained by tracing relevant areas on
digitized light micrographs (Table 1 of Ref. 15). This
tends to validate our quantitative approach, although the species were
not the same.
We next addressed the prevalence of -smooth muscle actin among
alveolar structures. We examined transmission light micrographs using a
Richards surgical microscope at four locations of particular mechanical
significance (see DISCUSSION). These are 1) the
border of an alveolar septum that abuts two other septa, appearing on section as a three-way septal junction; 2) the border of an
alveolar septum that abuts one other septum, forming a ridge between
the two septa and appearing on section as a distinct bend of the septal trace; 3) the free border of an alveolar septum, as at an
alveolar entrance ring, appearing on section as the free end of a
septal trace; and 4) the alveolar septum away from its
borders. We identified and circled all the septal ends and bends on
each transmission micrograph (the other two locations turned out to
have no identifiable -smooth muscle actin), and recorded each as
showing -smooth muscle actin stain or no stain. When we applied a
similar approach to the confocal light micrographs, we obtained data
for the septal ends, but could not obtain satisfactory data for the
septal bends, because their characteristic configuration was not well
enough preserved in these paraffin preparations for secure identification.
To quantify the distribution of -smooth muscle actin among the
septal ends, we examined confocal fluorescence light micrographs from
three rats. A polygon was drawn closely around every septal end on each
section studied. The transected areas of -smooth muscle actin in
each of 350 septal ends were obtained from the number of -smooth
muscle actin pixels.
The subcellular concentration and configuration of -smooth muscle
actin were studied on transmission electron micrographs of rat lungs.
Fields-of-view centering on -smooth muscle actin-containing material
were collected from airway smooth muscle, blood vessel smooth muscle,
and septal ends (36, 35, and 33 fields-of-view, respectively). The
outlines of all cells tagged with gold particles were emphasized with a
fine-pointed marker. Each micrograph was overlaid with a transparent
grid with lines spaced at distances equivalent to 0.4 µm. The number
of gold particles visible in each grid square was recorded on the grid,
and the mean densities (particles per grid square) were calculated.
For examining the visceral pleura, we abandoned the approach of random
sampling because of the discreteness of the pleural structure and the
frequent stripping of pleura from sections. On two sections of guinea
pig lung, we measured pleural thicknesses, -actin layer thicknesses,
and lung slice perimeters, using calipers on light micrographs. From
these we calculated the transected areas of pleural tissue, pleural
-smooth muscle actin, and lung.
| |
MODEL |
Overview
Our goal was to predict how activation of the contractile
elements in each site affects lung distensibility during periodic breathing, specifically the effects on constriction, elastic recoil, and compliance. Our approach was to model a region of lung as an
extensive network of elastic elements within which are embedded various
contractile structures, namely airways, blood vessels, and alveolar
ducts. During breathing, the entire region is cyclically extended and
released, and the lengths and tensions of its elastic and contractile
components are cycled according to their mechanical properties and
their mechanical linkage.
The model respects the important distinction between elasticity and
plastoelasticity. We assume that the lung behaves reasonably elastically, i.e., its stress-strain relationship is single valued. By
contrast, the activated contractile elements are plastoelastic, i.e.,
they can change their stress-strain relationship substantially, e.g.,
after being stretched (11, 14, 39, 42). Although the two
systems have this fundamental difference, they are each reasonably predictable.
The model also respects fundamental differences in the mechanical
linkage of the elastic and contractile elements. When two elements are
linked in parallel, i.e., side by side, their lengths are the same,
whereas the net tension is the sum of their individual tensions. By
contrast, when they are linked in series, i.e., end to end, their
tensions are the same, whereas their lengths sum to the net length. The
present context suggests both models. The anatomies of the contractile
elements of the pleura and of the axially oriented elements in the
airways and blood vessels suggest mechanically parallel linkages with
the elastic network of the lung. Because the dimensions of both elastic
and contractile components depend only on lung volume, the
contributions of each component to lung recoil can be predicted from
its characteristic mechanical properties, the level of activation, and
lung volume, lung volume excursions, and timing. Their contributions
can be simply summed. By contrast, although the circumferentially
oriented elastic and contractile elements of the ducts, airways, and
blood vessels are mechanically in parallel with each other, they in
turn are linked mechanically in series with the surrounding alveolar
parenchyma. Because of the series linkage, the activated contractile
elements at any given lung volume are shorter; the duct, airway, or
vessel is narrower; and the surrounding alveolar parenchyma is drawn in. The governing constraint here is that the additional radial tension
due to the contractile element matches the net increase in opposing
radial tensions in the elastic structures, i.e., in the surrounding
parenchyma and the passive elastic elements of the duct, airway, or
blood vessel. Calculations for serial linkage are more complex than for
parallel linkage, because the elastic solution must include the
distortion from contraction of the embedded structures, and the
dimensions and tensions of elastic and contractile components must be
resolved to be compatible with each other, with each following its
respective properties.
Responses of the Elastic Structure to Internal Contractile
Stress
The major assumptions are: 1) that the noncontractile
structure of the lung is linearly elastic over the range of distortion under consideration (22), which includes the airways,
blood vessels, and alveolar ducts as well as the network of alveolar septa that surrounds them and does not separate the properties of their
different tension-bearing components; 2) that, for serial linkage, predictions of small deformation elasticity theory are reasonable approximations for the magnitude of deformation of concern;
and 3) that constriction of the alveolar duct is reasonably spherical and that of airways or blood vessels reasonably cylindrical.
Spherical model for constriction of the alveolar duct.
Consider a spherical ventilatory unit, comprising a spherical alveolar
duct and its associated alveoli. The structural elements of the duct
consist of a net of connective tissue "cables" and their
accompanying contractile elements. In the model, we replace that
spherical net by a thin-walled spherical shell with the same mechanical
properties as the net, and we express the radial stress at the
interface as a stress jump across the wall of the shell. The
contribution of the activated contractile elements to that transmural
stress jump (Ptmactive) is obtained from the
Laplace relationship, in which the contractile tensions are those of
passively cycled, activated smooth muscle and the stereologically
determined density of -smooth muscle actin in the unit. The
contribution of the passive elements (Ptmpassive), is
taken, for computational purposes, at the radial stress at the surface
of a solid spherical elastic continuum that has the same bulk modulus
as that of the whole lung (22).
The alveolar network forms a thick-walled spherical shell lying outside
the duct. Its mechanical properties are similar to those of the whole
lung, on the bases of the relatively homogeneous expansion of the duct
and parenchyma during inflation of the relaxed lung (2,
13) and in addition the similarity in shear of the macro and
micro strains (4) (assumption 1). This
thick-walled shell is modeled as an elastic continuum with the same
bulk and shear moduli as those of the whole lung (22).
A region of lung comprises a number of such spherical ventilatory units
with identical properties. We assume that the small departures of the
units from spherical shape to pack them together do not materially
change their behavior. When considering a region that is homogeneously
activated, we represent the behavior of the whole by the behavior of a
single unit.
We first derive the relationship between stress in the contractile
elements and constriction of the duct at fixed lung unit volume and
then the effects of that constriction on the recoil of the unit. We
base the analysis on small deformation elasticity theory for a
prestressed elastic structure, the constraint that the radial stresses
at the interface between the duct and the surrounding parenchyma remain
balanced, and the Laplace relationship. As developed in the
APPENDIX, constriction is a function of the elastic moduli,
the volumetric density of the relaxed duct relative to its ventilatory
unit, and the incremental tension developed by the contractile elements
of the alveolar duct
|
(1)
|
The increment in regional lung recoil is a function only of
volumetric density and contractile tension
|
(2)
|
Eqs. 1 and 2 appear in the
APPENDIX as Eqs. A5 and A7.
Cylindrical model for constriction of the airway or blood vessel.
The analysis for the cylindrical structures differs from the above in
two ways. First, there is an axial linkage to consider. We assume that,
because of the treelike structure of the airways and blood vessels, the
lengths of airway and blood vessel segments are absolutely constrained
by lung volume and vary closely with its cube root, regardless of
activation. The linear linkage of the segments suggests that axial
tensions can be transmitted along the branches of the trees,
independent of the tensions in the alveolar parenchyma. The axial
components, thus, are mechanically in parallel with the parenchyma.
Second, there is an asymmetry to the serial linkage because at a fixed
lung volume the segments can narrow without changing length by drawing
in on the surrounding alveolar network. Absent an elastic solution for
such a deformation, we take the approach of an earlier report from this
laboratory (45) and consider the effects of radial
constriction at fixed lung unit volume as a volumetric constriction
within an extensive elastic medium.
We consider the effects of circumferential and axial tensions
separately. As developed in the APPENDIX, radial
constriction of the airway (or blood
vessel)1 is a function of the
shear modulus and the increment of circumferential tension
|
(3)
|
and the increment in regional lung recoil is a function of the
elastic moduli, volumetric density, and increments of both circumferential and axial tensions
|
(4)
|
Eqs. 3 and 4 appear in the
APPENDIX as Eqs. A11 and A14.
Incorporation of stereologic data.
The increment of ductal tension is a function of the volume reference
for scaling the normal stress of the contractile element, the
fractional volume of the lung relative to that at maximal physiological
volume, the density of the embedded structures in the lung, and the
stereological densities of contractile material.
|
(5)
|
The increments of the circumferential and axial tensions of the
airway (or blood vessel) are functions in addition of the angulation of
the contractile elements to the axis of the wall segment
|
(6)
|
and
|
(7)
|
Eqs. 5-7 appear in the APPENDIX as
Eqs. A15-A17. Combining these equations with Eqs.
1-4 gives the constriction of the duct
|
(8)
|
its effect on lung recoil
|
(9)
|
the constriction of the airway
|
(10)
|
the effects of airway constriction on lung recoil
|
(11)
|
and the effects of axial tension on lung recoil
|
(12)
|
Pleural model.
In the guinea pig, there is abundant -smooth muscle actin in the
pleura. To estimate its maximal effect on net recoil, we consider a
spherical pleura with contractile elements uniformly distributed with
regard to location and orientation. As shown in the
APPENDIX, pleural tension is the same as is given by
Eq. 5 with = 1, and the effect on lung recoil,
then, is the same as that given by Eq. 9.
Properties of Passively Cycled, Activated Contractile Element
Like the elastic elements, the contractile element is
predictable, although it differs in its substantial plasticity. When activated smooth muscle is cycled, the stress-strain relationship departs dramatically from the classical isometric stress-strain curve
depending on the timing and sequence of forces imposed on it. For
example, when trachealis muscle is activated maximally at a constant
length in vitro, stress rises to a plateau over several minutes. Such
plateau values, obtained at different lengths, generate the familiar
maximum isometric contraction curve (MICC; Fig.
2, solid curve). If the muscle is
forcibly stretched after reaching the plateau tension, stress generally
yields along this curve. If it is then shortened rapidly, stress falls
sharply away from the MICC. If it is cycled with a frequency and
amplitude typical of tidal breathing, it forms a stable, narrow,
stress-strain loop that peaks near the MICC but that is much stiffer
than the MICC (examples in Fig. 2A, dashed and solid loops)
(11, 39, 42). Such loops can be generated anywhere along
the length axis, creating an infinite array of parallel, narrow, stiff
loops.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 2.
Effect of activated contractile elements on the elastic
structure of the parenchyma during tidal breathing. A:
typical plastoelastic behavior of the maximally activated contractile
element. Length is presented as a fraction of that at total lung
capacity (TLC). Nominal stress is referred to the cross-section of the
smooth muscle at functional residual capacity (FRC). Solid curve is
that for maximal isometric contraction. Dashed and solid loops are for
maximally activated smooth muscle undergoing repeated, imposed length
cycles of differing amplitude at typical breathing frequencies. Note
that, for a given cycling pattern, loops are narrow and stable (i.e.,
elastic) and are stiff. Nonetheless, different loops form along the
maximal isometric contraction curve (MICC) depending on the length to
which the contractile element is stretched at the peak of the cycle
(i.e., plastic). B: superposition of properties of the
elastic structure for volumes at intervals of 0.1 of lung fractional
volume relative to TLC (fVLTLC) and
for a range of contractile element stress (ce). Radius
is presented as a fraction of that at relaxed TLC and is equivalent to
contractile element length. In the absence of contractile stress,
radius varies as the cube root of volume, whereas, with increasing
contractile stress at fixed volume, elastic modeling (Eq. 8
or 10) predicts narrowing (solid line). Trajectories of the
nadirs of contractile element strain/stress loops are similar to those
in A (dashed line). End-inspiratory ( ) and
end-expiratory ( ) points for tidal breaths between 0.3 and 0.4, 0.5 and 0.6, 0.7 and 0.8, and 0.9 and 1.0 fVLTLC are identified by the
intersections of the contractile curves with the isovolume elastic
curves as described in the text.
|
|
End-Inspiratory and End-Expiratory Points
For parallel linkage, the lengths follow the cube root of lung
volume, the stresses follow the elastic and plastoelastic descriptions, and the recoil is simply the sum of those stresses applied to the
appropriate areas. For serial linkage, compatible solutions are
required for the elastic and plastoelastic components. In tidal
breathing, the elastic structure lies at end-inspiration on its linear
function for that volume (Eqs. 8 or 10), and the contractile element lies on the MICC. Their intersection identifies the
end-inspiratory point. At end-expiration, the elastic structure lies on
the linear function for that volume, and the contractile element lies
on a curve that comes down from the particular end-inspiratory point of
the tidal breath, following the general trajectory of nadirs of such
loops (11). Their intersection identifies the end-expiratory point. Applying Eqs. 9, 11, and 12, we obtained the effects on lung recoil. From these data,
we can calculate the specific compliances and fractional changes of compliance.
Incorporation of Specific Anatomic and Functional Data
Compatible solutions were found by numerical analysis on a
spread sheet (Excel) using Eqs. 8-12 and curves fitted
to the data for volume/pressure, MICC, and the general trajectory of
nadirs of repeated loops of activated smooth muscle.
Elastic structure.
The isovolume elastic solutions for radius of the airway, vessel or
duct, are obtained as
(fVLTLC)
(1 + u/a1). The bulk modulus of the lung,
, is derived from a generalized lung pressure-volume curve
(37) and human pressure-volume data (12), and
the shear modulus, µ, is derived from lung distortion properties
(22) (See APPENDIX). We take fractional FRC
(FRC/TLC) as 0.5, and the volumetric densities of the spherical
alveolar duct and of the cylindrical airway within their ventilatory
units, , as 
and 1/50,
respectively. The areal density of the contractile element in each
location, Ace/AL, comes
from Table 2.
Contractile elements.
The MICC is adapted from a generalized curve (Fig. 1, Ref.
47) with maximum stress set at 2.88 × 106 dynes cm2 at lo
(15), and with lo of the relaxed
muscle set at FRC. We assume that the length of the relaxed muscle
varies as VL. A generalized trajectory of
end-expiratory points was generated from the end points of a series of
dynamic loops (Fig. 1 of Ref. 11), with length scaled in
proportion to the length at the end-inspiratory point.
Orientation and closure.
The orientation of the contractile elements in the airways is nearly
circumferential (23), thus favoring circumferential over
axial tension. We modeled the airway with a single (77° to the
axis of the airway) because the observed distribution is narrow. We did
not have information regarding the distribution of in the blood
vessels, but predictions made with the assumption of uniform
distribution of (not shown) are not very different. When the airway
closes, i.e., when the outside radius falls to ~0.48 of its initial
radius (26), any excess of 
Tcirc over that
needed for closure has no additional effect on the surrounding parenchyma. We therefore limit the effects of Tcirc on
recoil (Eq. 11) to that threshold. Nonetheless, the full
Tax can still be transmitted along the closed airway tree.
| |
RESULTS |
Anatomic Observations
Location.
-Smooth muscle actin appears in the walls of all conducting airways,
in the walls of all blood vessels larger than capillaries, and in two
of four sites of particular mechanical significance, namely the septal
ends and bends, where elastin- and collagen-rich cables are located
(34). It is present in about three-quarters of the ends
and half of the bends (Fig. 3, Table
1). No -smooth muscle actin was
detected in the septa away from the ends and bends, nor in the
three-way septal junctions. -Smooth muscle actin is present in the
visceral pleura of the guinea pig, but not in that of the rat.
View larger version (109K):
[in this window]
[in a new window]
|
Fig. 3.
Light micrograph of rat lung stained by reaction with
-smooth muscle actin antibody, visualized using the immunoperoxidase
method. Arrows indicate staining at a typical septal "end"
(E) and "bend" (B).
|
|
Quantities.
The amounts of -smooth muscle actin and tissue in the various sites
are given in Table 2 and are of roughly
similar magnitude in the two species, with the exception of the rat
pleura. The rat has slightly less -smooth muscle actin in
the alveolar duct than in either the airways or blood vessels, whereas
the guinea pig has more. The rat has more -smooth muscle actin in
its airways and blood vessels than does the guinea pig, but less in the
ducts. In addition, the guinea pig has substantial -smooth muscle
actin in the pleura, almost as much as in the parenchyma.
The areas of the individual bundles of -smooth muscle actin at cable
sites as seen on section are mostly 1 µm2 or less (Fig.
4). The variation of areas is only partly
due to tangential sectioning-random sectioning of uniform linear
bundles would produce fewer large areas than we saw. It is our
impression from the transmission light micrographs that heavier clumps
of -smooth muscle actin are present in clusters of septal ends
associated with larger alveolar ducts. In an attempt to confirm this
impression, we looked for visual evidence of a gradient of heavy
staining from hilum to pleura but were unable to draw a conclusion.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 4.
Frequency distribution of the areas of patches of
-smooth muscle actin seen in septal end structures. Most are <1
µm2.
|
|
Fine distribution and concentrations.
At the cellular level, the rat shows distinct differences between the
smooth muscle of the airways and blood vessels and the -smooth
muscle actin-containing cells in the cable sites of the alveolar
parenchyma. As shown in Fig.
5A, the -smooth muscle actin-containing cells of blood vessels and bronchioles have the well-known morphologic features of smooth muscle: their outline is
unbranched, compatible with a fusiform shape; their cytoplasm is
nearly filled with electron-dense filamentous material that binds the
gold particles; and caveolae are numerous along the cell periphery. The
basal lamina in these preparations is not consistently visible, owing
to low contrast, but an unfilled gap of several tenths of a micrometer
is present between the periphery of neighboring cells or between cells
and interstitial fibrils. By contrast, the immunogold-tagged cells in
the septal ends (Fig. 5B) are represented by profiles of
fine cell processes that sometimes branch. The gold particles are bound
to packets of electron-dense, filamentous material that occupy a
relatively small portion of the profile, the remainder being electron
lucent with occasional mitochondria or sacs of endoplasmic reticulum.
Caveolae are few, and interstitial fibrils abut the cell periphery,
indicating the absence of an intervening basal lamina. These cells
indeed resemble the myofibroblasts in other portions of the alveolar
wall save for the actin isoform expressed. On the other hand, within
the portions of the cells that do contain -smooth muscle actin, over regions of similar size to the confocal pixels, the concentrations of
-smooth muscle actin (immunogold particles per overlay square) are
similar among the ends (1.98) and the smooth muscles of the airways
(1.84) and blood vessels (2.02). So, although these cells seem to be
myofibroblasts, they appear to have hefty contractile machinery
comparable to smooth muscle.
View larger version (98K):
[in this window]
[in a new window]
|
Fig. 5.
Electron micrographs of rat parenchyma, contrasting the
cells containing -smooth muscle actin (immunogold-stained) in the
airways (A) and in the septal end structures (B).
See text.
|
|
Predictions From Modeling
The end-inspiratory and end-expiratory radii and contractile
element stresses during tidal breathing are shown in Fig.
6, A and B. When
relaxed, radii of ducts, airways, and vessels vary simply as the cube
root of lung volume, but with activation, the structures are narrower.
The degree of narrowing 1) during a given tidal breath is
greater at end-inspiration than at end-expiration by up to 150%,
reflecting the greater contractile stress at end-inspiration; 2) is generally greater in tidal breaths taken at lower lung
volumes, reflecting less opposing tension in the elastic structures;
3) is greater if the density of -smooth muscle actin is
greater, e.g., the ducts of the guinea pig compared with those of the
rat; and 4) is greater for the airways than for the ducts,
reflecting the higher local concentration of -smooth muscle actin.
Maximally activated airways are predicted to close at mid-range lung
volumes. Contractile stresses are much lower at end-expiration than at end-inspiration, reflecting the characteristic dynamic stiffness of the
activated contractile element. Lung recoil (Fig. 6C) is increased modestly, e.g., rat ductal constriction increases recoil by
<1 cmH2O, although this reaches 20% of relaxed recoil at
lowest lung volumes. At high lung volumes, the airways have greater
effects on recoil than do the ducts. Most of this (about
three-quarters, data not shown) is due to circumferential tensions. At
lower lung volumes, where the airways close, the contribution of
circumferential tension falls off, while that of axial tension
increases, reversing the ratio. Specific compliance (Fig. 6,
D and E) is reduced by up to 20-30% with
constriction in any one site alone. The effects of pleural activation
(guinea pig) on recoil (not shown) were somewhat less than that of
ductal activation, but its effects on specific compliance were ~50%
greater (Fig. 6E).
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 6.
Predicted effects of activation of contractile elements on the
radius of various embedded structures (A), on stress in the
contractile elements (B), on lung recoil (C), on
specific parenchymal compliance (D), and on the fractional
change of parenchymal compliance (E) during tidal breathing
at different levels of inflation. The outer radius of the duct or
airway is given relative to its radius at relaxed TLC. Volume is given
as fraction of relaxed TLC. Heavy solid lines, relaxed contractile
elements; dotted lines and , activated rat ducts; thin
solid lines and  , rat airways; dot-dashed lines and
, guinea pig duct; dashed lines and  ,
guinea pig airways;  , guinea pig pleura. See
text.
|
|
| |
DISCUSSION |
The main findings of this study are 1) that the amounts
of contractile material do not vary widely between the alveolar ducts, airways, and blood vessels of the lung parenchyma, 2) that
estimated effects on lung recoil are small, and 3)
that there is the potential for modest but physiologically significant
control of regional compliance.
Methodology
We chose -smooth muscle actin as a marker of contractile
capability, as have others (8). Although the most
plentiful isoforms of actin in lung are nonsmooth muscle
(24), such isoforms are generally involved in processes
(e.g., cell migration, phagocytosis, secretion, and translocation of
organelles within the cytoplasm) that do not involve deformation of the
tissue external to the actin-containing cell. By contrast, the smooth
muscle isoforms are found in cells such as fibroblasts during the
contraction of wounds, the myoepithelial cells of gland ducts, and
smooth muscle itself, i.e., in sites or circumstances where cells put tension on the surrounding tissue. In addition, the force generated by
smooth muscle cells is ~10-fold that generated by contraction of
cells such as endothelium and chick embryo fibroblasts with nonmuscle
actin (20). We recognize that, in limiting ourselves to
the -smooth muscle isoform, we may have underestimated the intrinsic
contractile capability of the alveolar parenchyma. We are not likely,
however, to have overestimated it.
We needed images identifying -smooth muscle actin-containing pixels
in anatomical context to quantify contractile elements in the several
sites. We had tried several approaches before settling on the described
techniques. We failed to obtain reliable immunostaining of -smooth
muscle actin for light microscopy with a number of plastic embedding
materials, combined on occasion with etching and removal techniques.
Paraffin-embedding provided excellent immunostaining, but DAB-HRP stain
deposits were neither well localized nor opaque enough to allow the
video microscopic images to be segmented cleanly into -smooth muscle
actin, tissue, and air regions. We then tried combinations of
immunofluorescent staining of -smooth muscle actin and absorptive or
fluorescent staining of tissue, but the thickness of the sections
(7-12 µm) meant that portions of the section above and below the
plane of focus contributed to both the -smooth muscle actin and
tissue signals, making their separation unreliable. This problem was
partly overcome by the use of confocal laser scanning microscopy to
look at optical sections that are a fraction of the thickness of the
section itself (<1 µm out of 7 µm or more). Nonetheless, even with
emission spectra as well separated as those of BODIPY-SE and Cy-3,
there was enough cross talk between the two detected signals that
separating them quantitatively was insecure. This final limitation was
largely overcome by the use of the novel IMS technique (5)
in which two lasers are used, each tuned to a suitable excitation
wavelength of one of the dyes. The intensities of the two exciting
wavelengths are modulated at different frequencies in the megahertz
region. As a result, the intensities of the light from the
different dyes have different frequencies and can be distinguished by
lock-in amplifiers, each tuned to one of the exciting frequencies,
rejecting in each case the out-of-frequency signal contaminating that
channel. This approach gave us two superimposable images, one showing
the -smooth muscle actin-containing pixels and the other showing lung tissue for context.
The images were further clarified by morphology-based image processing.
On the 620-nm image (-smooth muscle actin), there were individual
pixels that appeared to be out of anatomical context. There were two
reasons for this. First, tissue pixels are subject to a certain random
noise. This is photon quantum noise with a broad frequency spectrum and
random signal level proportional to the square root of the signal level
of the rejected (525 nm) signal. It randomly increases or decreases the
intensity of all tissue pixels, whether -smooth muscle actin
containing or not, creating enough of an overlap of intensity that
simple intensity thresholding of the 620-nm channel does not clear it
of the noise from the 525-nm channel. Because the noise is random,
however, the severely affected pixels tend to occur in isolation, e.g., isolated low-intensity pixels within typical patches of -smooth muscle actin and isolated high-intensity pixels in regions of non--smooth muscle actin-containing tissue. Second, there were many
pixels of equivocal intensity at the borders of patches of -smooth
muscle actin, where many pixels lie partly over -smooth muscle
actin-containing tissue and partly over non--smooth muscle actin-containing tissue. These two features (isolation and edges) lend
themselves to morphology-based image processing.
Contractile Cells
-Smooth muscle actin appears not only in classical smooth
muscle cells, but also, in the rat, in myofibroblast-like cells (Fig.
5). One of us (C. Kuhn III) has previously found similar cells in the
septal ends of mice and hamsters. These cells have concentrations of
-smooth muscle actin at the pixel scale that are comparable to those
in smooth muscle. They lie in the same site (alveolar ducts) as do
classical smooth muscle cells in the guinea pig, human, and other
species (Fig. 3). It seems likely, then, that this cell in this
location has a primary contractile function.
Mechanical Linkage-Anatomic Basis
The effect of activation of contractile elements on lung recoil
depends on how the contractile and elastic structures of the lung are
linked (see MODEL). The outer boundaries of the airway and
blood vessel trees deform nearly exactly with the pleural surface,
because the relative volume of lung that might be mechanically in
series with the trees is a subpleural rim several hundred micrometers in depth, and this is negligible compared with the volume of lung encompassed by the trees. The inner portions of the trees most likely
remain shape stable. To the extent that the trees are tensed, branching
structures, supported in space by terminal branches, the angles at each
branch point are determined by the ratios of axial tensions in the
branches and will not change if the axial tensions are changed
proportionally during change of lung volume or change in tone.
Furthermore, the alveolar parenchyma, within which the tree is
embedded, is itself inherently shape stable. Indeed, airways in situ
generally change length as the cube root of lung volume, which is
completely compatible with shape stability. We conclude that the trees
most likely provide a pervasive network that is generally shape stable
either with lung volume change or with uniform changes of axial tension
in its branches, that the linkage of its axial elements with the lung
parenchyma is mechanically in parallel, and that lung recoil is
directly changed by changes of axial tensions.
The lengths of the pleural contractile elements (guinea pig only)
appear to be tied to lung volume, independent of the elastic elements
of the lung, and are thus also in a parallel linkage.
Circumferential elements of the airways, blood vessels, and alveolar
ducts, on the other hand, appear anatomically to be linked in series
with the elastic structures of the lung. They are situated where they
can shorten the elastic elements of the airway or blood vessel at fixed
lung volume, drawing in radially on the surrounding alveolar septa,
expanding their surfaces and deforming their elastic structures. In
particular, the alveolar ductal contractile elements accompany most of
the connective tissue "cables" (Table 1), and we presume that their
mechanical action is similar. These cables are linear condensations of
elastin-rich connective tissue that are positioned to support the free
(end) and tented (bend) septal borders by their tensions and curvatures
(34, 36) and are not found in the alveolar septa away from
the cabled borders or in the three-way septal junctions, where the
configurations are simply compatible with locally uniform septal
tensions (33). The contractile elements of the duct, then,
are positioned to both relieve the elastic elements of the duct and
tense the surrounding alveolar network. Presumably this is the
mechanism of the dramatic radial constriction of alveolar ducts
reported by Nadel et al. (31) in cat lung fixed after
barium sulfate embolism. Unfortunately, to date there have been no
systematic measurements of the degree and circumstances of such
deformation in situ. The tensions of the activated contractile elements
in this series linkage goes in part to distort the elastic network and
in part to increase the lung recoil. Figure 6, A and
C, shows quantitative predictions of narrowing and change of recoil.
Importance of Incorporating the Dynamic Contractile Properties and
the Series Linkage in the Model
The predictions of the model would be very different if we were to
ignore the dynamic properties of activated contractile elements and
assume instead that the stresses remain high throughout the breathing
cycle. In particular, if the stresses in activated rat alveolar ductal
elements were to lie on the MICC at end-expiration as well as at
end-inspiration, parenchymal compliance would actually increase
slightly with activation of the contractile elements (Fig.
7, curve b), an unexpected
result that relates to the volume-related terms in Eqs. 9 and 11. Reduction of parenchymal compliance, then, is
largely due to the substantial stiffness shown by the activated contractile element when it is stretched and released at typical breathing frequencies.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 7.
Comparison of three models. a: Parallel
(dotted-dashed line, ×). b: Constant stress (dashed line,
+). c: Serial model (dotted line, ). Axes as
in Fig. 6D. The parallel linkage model overestimates the
effect on parenchymal compliance (compare curves a and
c). Dynamic stiffness of the contractile element accounts
for virtually all of the predicted reduction in parenchymal compliance
(compare curves b and c).
|
|
The predictions of the model are also different for serial and parallel
linkages. For example, a model in which the duct is linked in parallel
with the surrounding parenchyma (Fig. 7, curve a) predicts
greater effects on compliance than our model in which the linkage is
serial (Fig. 7, curve c). There are two reasons for this.
First, in a serial linkage, the relatively stiff activated contractile
element has less
-Actin: disposition, quantities, and estimated effects
on lung recoil and compliance
TOP
ABSTRACT
INTRODUCTION
