Permeability of the Alveolar Membrane
to Solutes
By Aubrey E. Taylor, Ph.D., Arthur C. Guyton, M.D., and
Vernon S. Bishop, Ph.D.
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
• Studies on pulmonary capillary permeability have demonstrated that both water and
low molecular weight solutes pass rapidly
through the pulmonary capillary membrane
into the interstitial spaces, as would be expected from studies on capillary membranes
in other vascular beds. 1 ' 2 However, studies
by Kylstra8 and Swann4 have shown that
transport of solutes from the alveoli into the
pulmonary capillaries is extremely slow, less
than one-hundredth the rate expected on
the assumption that the permeability of the
pulmonary membrane is equal to that measured for capillaries in the isolated hind limb.5
A possible explanation of this difference is
that the alveolar membrane, in contrast to the
pulmonary capillary membrane, represents a
major barrier to the transport of most substances.
The experiments reported in this paper were
designed to study the permeability of the alveolar membrane by analyzing movement of
Na*4, R", glucose, and other test materials
between fluid in the alveoli and fluid circulating through the pulmonary vessels of an isolated, perfused dog's lung.
Methods
Mongrel dogs (6 to 10 kg body wt) were
heparinized (5 mg/kg) and anesthetized with
sodium pentobarbital (30 mg/kg). The lower
left lobe was excised, weighed, and connected
into the system shown schematically in figure 1
as follows:
A. VASCULAR PERFUSION SYSTEM
The lobe was placed in a plastic box (H) to
From the Department of Physiology and Biophysics,
University Medical Center, Jackson, Mississippi.
Supported by research grants-in-aid from the Life
Insurance Medical Research Fund and the Atomic
Energy Commission.
Accepted for publication August 10, 1964.
Circmlttion Reiurch, Vol. XVI, April 196}
prevent drying of the lung. A catheter holder
(G) held three catheters rigidly in place with
proper angulation for catheterizing the pulmonary artery, via catheter (A), the pulmonary vein,
via catheter (C), and the bronchus, via catheter
(B). Perfusion fluid from the pulmonary vein
flowed into reservoir (D). This was then pumped
by a Sigmamotor pump through a bubble trap
(K) into the pulmonary artery. The arterial
pressure was recorded from a sidearm in the
bubble trap using a mercury manometer.
The perfusion pressure measured on the arterial
side of the system was 8 to 10 mm Hg in all experiments reported in this paper.
The perfusion fluid was Tyrode's solution with
5% high molecular weight dextran added, to
provide colloidal osmotic pressure. Enough blood
or plasma was added to the perfusion fluid to
bring the plasma protein concentration slightly
above 0.2 g per cent. In a few early experiments
in which this was not done, large quantities of
fluid transuded into the pulmonary tissues from
the vascular perfusion system. This was in agreement with the work of Landis and Pappenheimer0 on other capillary beds in which they
showed that a small amount of plasma protein
must be present for dextran to exert its full osmotic effect at the capillary membrane. The
buffer THAM (10 meq/liter) was added to the
perfusion system and a gas mixture composed of
95$ Oo and 5% CO2 was bubbled into the reservoir continuously during the experiment. Control
studies indicated that the pH of the perfusion
fluid remained in the normal physiological range
(7.35 to 7.4) during the entire course of the experiment.
The flow was measured at the beginning and
completion of each experiment using a Ludwig
stromuhr; however, the stromuhr was disconnected during experimentation to minimize volume of perfusion fluid. The flow measurement at
the completion of each experiment reported had
not changed more than 10% from its value at the
beginning. In 12 experiments the flow rate was
set at 150 cc/min (5.1-7.0 cc/min/g tissue),
which was a rate found by Ross et al.7 to perfuse
the lobe completely but not to cause pulmonary
edema. In 22 other experiments, the flow was
varied from 3.8 to 12.8 cc/min/g lung tissue.
353
354
TAYLOR, GUYTON, BISHOP
In nine experiments, the lungs were degassed
before filling the alveoli with fluid by allowing
the dog to breath pure O 2 for 40 minutes and
then clamping the trachea until the lungs collapsed because of oxygen absorption into the
blood.
C. METHOD OF SAMPLING
The pulmonary vascular system was perfused
for 15 minutes prior to any measurements. A
known quantity of test substance was then placed
in the venous reservoir, and its disappearance
rate from the system was followed by sampling
the fluid in the reservoir. Samples of M cc volume
were taken and, the contents analyzed as follows:
FIGURE 1
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Apparatus for perfusing the isolated lower left lohe of
the dog's lung. A. Arterial catheter. B. Bronchial
catheter. C. Venous catheter. D. 20 cc venous reservoir. E. 200 cc bronchial reservoir. F. Sigmamotor
pump. G. Catheter holder to insure proper angidation
of catheters during perfusion of the lobe. H. Plastic
box that encloses lobe. 1. Side arm for collapsing the
excised lobe. J. Clamp. K. Bubble trap.
The flow values used in these experiments compare with normal resting control values of 8 to
12 cc/min/g tissue in the intact dog.
The time from excision of the lobe to the start
of the perfusion pump did not exceed 10 minutes in any experiment, and the temperature
of the system was maintained at 20 to
22°C throughout.
B. BRONCHO-ALVEOLAR LAVAGE SYSTEM
The broncho-alveolar system was filled in most
of the experiments in the following manner: The
lobe was initially collapsed with a negative pressure of —5 mm H2O. Then, 150 cc of the solution described above, but with the dextran and
plasma protein omitted, were placed in the plastic
bronchial reservoir (E) shown in figure 1 and
was allowed to ran through tube (B) into the
bronchi and alveoli. It was then ran in and out
of the lung by raising and lowering the bronchial
reservoir until all air bubbles were washed out
of the system. Care was taken to avoid hydrostatic
pressures greater than 11 cm H2O during the
course of filling, for this was shown previously
by Young (unpublished observations) to cause
leakage of fluid out of the alveoli. Immediately
following washout of the bubbles, the bronchial
reservoir was lowered until the fluid level in the
reservoir was at the mean level of the lung. The
bronchial catheter was then clamped, and the
excess fluid in the reservoir was removed. The
volume of excess fluid was recorded so that the
volume remaining in the broncho-alveolar system
could be calculated.
D. ANALYSIS OF SAMPLES
The concentrations of different test substances
in the samples were analyzed using the following
equipment or methods:
Na IJ and K' 2 : gamma ray spectrometer.*
D.,O: Beckman infrared spectrophotometer,
"model 1R-5.
Urea: analyzed by Gentzkow's8 method.
Glucose: analyzed by the method of FolinWu."
Dinitrophenol (DNP): colorimetric.
E. DETERMINATION OF PERMEABILITY COEFFICIENTS
To determine the permeability of the alveolar
membrane the system of figure 1 was treated
as a closed two compartmental system. Compartment 1, called henceforth the vascular compartment, contained all the fluid on the vascular side
of the alveolar membrane, including the interstitial fluid, the fluid in the blood vessels of the
lungs, and all the fluid in the external perfusion
system itself. Compartment 2, called henceforth
the alveolar compartment, contained the fluid
on the alveolar side of the alveolar membrane,
including only that portion of the lavage solution
which actually exchanged solutes through the alveolar membrane.
This treatment of the system in figure 1 as a
two compartmental system was valid only in
case die dispersion of the test substance in each
compartment was very rapid in relation to its
rate of transport through the alveolar membrane.
Thus, it was necessary to assume (a) that the
rate of flow of the vascular perfusion fluid was
rapid enough that transport through the alveolar
membrane was not "flow-limited," (b) that transport through the pulmonary capillary membrane
was very rapid in relation to transport through the
alveolar membrane, and (c) that the rate of
diffusion of the test substance in the alveoli was
very rapid in relation to its transport through
the alveolar membrane so that its concentration
•Nuclear of Chicago, model 132B.
Circulation Reitarcb, Vol. XVI, April 1965
PULMONARY MEMBRANE PERMEABILITY
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
remained evenly distributed throughout the alveolar compartment. It seemed reasonable to
make these assumptions except in the case of
two test substances, D2O and DNP, for the following reasons: (a) The total volume of fluid
in the vascular perfusion circuit flowed around
the circuit on the average once even' 31 seconds, whereas the half-times for diffusion of
all the different test substances between the
vascular and alveolar compartments (except for
D2O and DNP) were in the order of many minutes, as will be shown later in the paper, (b)
Studies on both peripheral and pulmonary capillaries have demonstrated very rapid transport of
fluid and small molecular weight substances
through capillary membranes, (c) The distances
for diffusion in the alveoli are relatively slight in
comparison with velocities of diffusion of the test
substances in water, (d) Fluid was removed from
the broncho-alveolar system at the end of each
experiment, and the concentrations of test substances were plotted as a function of cc removed
as illustrated in figure 2. It will be noted from
this figure that the concentration rose to a plateau
after all the fluid in the dead space of the lungs
had been removed. This indicated that there
was reasonably homogeneous distribution of test
substance among the different alveoli.
We also assumed that the test substances
crossed the alveolar membrane either entirely or
almost entirely by the process of diffusion. In
ten control studies one of the test substances,
Na'-t, was placed in both compartments in the
CC
REMOVED
Total Volume
In Lavag*
Solution
FIGURE 2
Plot of concentration of test substance in fluid removed from broncho-alveolar system as a function of
volume removed. A, represents alveolar fluid that
actively exchanges solutes; At represents the dead
s/tace.
Circulation Ratarcb, Vol. XVI, April 1963
355
same concentration and left in these compartments for two or more hours. In none of the experiments did the concentration in either compartment change significantly, which illustrated
that active transport, if present, was small enough
that it did not affect the results.
To calculate the permeability coefficients for
different substances three types of information
were necessary: (1) surface area of the alveolar
membrane, (2) the volumes of fluid in the two
compartments, and (3) the rates of diffusion of
the different test substances from the vascular
compartment into the alveolar compartment. The
symbols and units used in these calculations are:
V,, = volume of fluid in the blood vessels and
external perfusion system in cc.
V, — volume of fluid in the interstitial spaces of
the lung in cc.
V , = volume of fluid in the vascular compartment in cc.
V^ = volume of fluid in the alveolar compartment in cc.
Vu = total volume of fluid in the bronchoalveolar system in cc.
V,, = volume of fluid in the dead space of the
broncho-alveolar system in cc.
Q = quantity of test substance added to vascular compartment at time zero.
CV)0 = initial concentration in vascular compartment.
C|.-_ ( = concentration in the vascular compartment at time t.
C r „ = concentration in vascular compartment
at infinite time.
CAi t = concentration in the alveolar compartment at time t.
F = permeability coefficient of alveolar membrane for the test substance in cm/sec.
A = surface area of the alveolar membrane in
cm2.
W = weight of lobe in g.
t ] / 2 = half-time of approach of CVj, to Cv> x in
seconds.
Calculation of the Alveolar Membrane Surface
Area (A). The membrane surface area per g lung
tissue was estimated from the values of Roughton10 for membrane surface area in man and the
average weights of human lungs from Miller.11
This yielded a conversion factor of 661.6 cmVg
lung tissue; the alveolar membrane surface area
was then calculated by the following equation:
A = 661.6 W
(1)
Determination of Actively Exchanging Alveolar
Fluid Volume (VjJ. The volume of actively exchanging fluid in the alveoli was determined by
measuring the volume of fluid in the dead space
of the broncho-alveolar system, V,,, and then subtracting this from the total volume of fluid, VD,
356
TAYLOR, GUYTON, BISHOP
initially put in the broncho-alveolar system. To do
this, the broncho-alveolar fluid was drained out
of the lung at the end of the experiment into the
bronchial reservoir a few cc at a time until at least
60S of the fluid had been removed. A typical plot
of concentration as a function of volume removed
is shown in figure 2. Note that the concentration
of the test substance reached a plateau value.
This plateau was then extrapolated along the
abscissa up to the total volume of broncho-alveolar
fluid.
Area A2 then represents the proportion of the
broncho-alveolar fluid that is in the dead space,
and area Ax represents the proportion in the
alveoli that is actively exchanging solutes. VA is
then calculated using the following equation:
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
(2)
Determination of Vascular Compartment Volume (Vy). The volume of fluid in the vascular
compartment was determined by accounting for
all the test substance that was initially placed in
the vascular perfusion system in the following
manner: The alveolar volume VA as determined
above was multiplied by the concentration of
test substance in the alveolar compartment to
determine the quantity of test substance that had
crossed from the vascular compartment into the
alveolar compartment. This quantity was subtracted from the total quantity, Q, of test substance initially placed in the vascular system.
The quantity remaining in the vascular compartment was then divided by its concentration in
the vascular fluid determined at the same time;
this gave the volume of fluid in the vascular
compartment. The final equation for determining
the vascular compartment volume is:
Q-CA,tvA
Cr.t
(3)
Determination of Perfusion System Volume (V,,)
and Interstitial Fluid Volume (V,). The volume of
vascular perfusion fluid, VP, was measured by the
dye dilution method using T-1824 dye and analyzing the samples with a spectrophotometer. The
final volume was corrected for hematocrit, for
there were between 5 and 10 cc of residual blood
in the lobe. In none of the experiments did the
dye (T-1824) enter the alveolar fluid.
The interstitial fluid volume, V, was then calculated by subtracting the perfusion system volume as determined above from the fluid volume
in the vascular compartment according to the
following equation:
V, = VV-VP
(4)
Calculation of Permeability Coefficients. To
determine the actual permeability coefficients for
different substances, the test substance was placed
in the vascular perfusion reservoir, and 0.5 cc
samples were removed periodically for analysis.
The top curve of figure 3 illustrates a typical
decay curve of concentration drawn on semilogarithmic coordinates. This curve approached a
minimal equilibrium value. Cv x which could be
calculated from the final, uniform distribution of
the test substance in the alveolar plus the vascular
volume. Thus,
(5)
Once the experimental points had been plotted
and the final equilibrium value also plotted, the
equilibrium value was subtracted from the experimental points. The differences were then
plotted to give the lower curve in figure 3. Note
that this curve is linear on the semilogarithmic
plot as would be expected from passive distribution in an ideal two compartment system. A convenient form of the equation describing kinetics
40000-1
I30P00-
2QP00-
lopoo5000-
2000100050020 40 60 80 100 120 140 160 ISO
TIME (minutes)
FIGURE 3
Upper curve. Plot of experimentally obtained values
for the disappearance of Na1* from the vascular perfusion fluid. Dashed line represents equilibrium level
that the curve approaches. Lower curve. Plot of difference between upper curve and equilibrium level.
CircmUtiom Rtietrcb, Vol. XVI, April 196}
=> 0 0 CO
ro
O5
CD
10.
PULMONARY MEMBRANE PERMEABILITY
357
of passive distribution in an ideal two compartment system is:
i n CD as
+1
K
1O
= e
rfa
3
CO
3
.49
.71
u ' cm*
o
u
3
cS
CO ©
Ol
CO
8
o
m
O
©
©
s s
M
± .22
Ol
in CD
??
n
u
a
o
32
m o
<O
.53
.96
.56
.49
.99
.55
.42
.43
r
44.4 ± 12.8
= e
L
V7F2
CO
© in in
CD CO i -
CO ©
Ol o
O l 00
in
CO t -
CO
Ol
Ol
00 o
CD CO CO
•Si
25.0 ± 3.3
CO
Ol
21
CD CO ©
Ol
22
22
in
6.4 ± 3.4
o
S
CM O 5
Ol
Ol O l
11
•mined i
in
Q
•2
>
o o
co
s
o
©
t-
o o
o CO
in in © ©
CD CM 00
00 CO 00 00 CO t in
•5
i n i n co i n O l o CO
©
©
in
a*
ss
I—1
as
00 ©
97
89
81
i
92
82
>
83
79
o
5J
g
+1
as
as
r~
86.3 ± 7.1
©
72
1
CD
©
5
(7)
Thus, knowing the volumes of the compartments,
the area of the pulmonary membrane, and the
half-time of approach of the test substance concentration in the vascular compartment to its
equilibrium value, one can calculate the permeability coefficient.
An Alternate Method for Calculating Permeability Coefficient. The permeability coefficient of
the alveolar membrane for Na*4, PNn, was determined 34 times, and the values were then averaged. To determine the permeabilities of other test
substances, the second test substance was placed
in the vascular compartment along with Na1*. The
half-times of disappearance of Na8* and of the
test substance were then determined as described
above. Since the permeability coefficient for a
substance in an ideal system is inversely proportional to the half-time of exponential decay
(equation 7) one can determine the permeability
of the test substance, P, from the ratio of halftimes according to the following equation:
i n p p p q p q q c o q p p
>'
CO CD
o
as
IDJi
tub
(8)
' / ' • •
The value of this method is that neither volumes
nor alveolar membrane area need be measured.
2
inqpqqqqq-fl;
p o p
>
Results
VOLUMES OF THE DIFFERENT FLUID COMPARTMENTS.
AND SURFACE AREA OF THE ALVEOLAR MEMBRANE
olar
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
CO 00 CO CO
>
(6)
Derivations of equation 6 or its analog may be
found in papers by Sheppard,12 Collander et
al.,13 Solomon,14 or Robertson.15 The permeability coefficients, P, may be determined from the
best fit of the experimental data to equation 6 or
more conveniently by solving equation 6 for the
special case when the left hand term is 0.5, i.e.,
for the half-time, t1/2, of approach to equilibrium.
Under these conditions
0.693
I
M
J
«
S
tu
~
oio—i<7>oiooooia)inooco
>
< 3 ) O ! O H H H 0 3 6 H t o l
3
05*
d
z
"5
Circulation Ratrcb,
Vol. XVI, April 196}
Table 1 gives the different fluid volumes
and the alveolar membrane surface area determined in 12 separate experiments for the
purpose of calculating alveolar membrane
permeability coefficients for Na14. These same
measurements were made in 22 additional experiments, and the average results did not
358
TAYLOR, GUYTON, BISHOP
O5
n
§
+1
CD rt CS) CD CO O CO
CD in t- ocid oci i>
e
s
SP
00 O5 CO ^y CO ^^ O5 CD 0 0
+1
in CD i> oo co oo oo d rt
s
*j 60
i »
in
CD
t - oo cr>
c a.
eg
id
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
vary more than 10% from those reported in
table 1.
Note especially the very small quantity of
fluid in the interstitial space (or more properly in the extravascular sodium space). The
value of V, averaged 0.25 ± 0.129 cm8/g
lung tissue. This value is close to the value
of 0.33 cm3/g which we calculated from the
data of Chinard et al.2 on the assumption that
the mass of the left lower lobe is equal to 26%
of the total lung mass as found by Frank.10
Interstitial Fluid Volume Determined in
Three Edemotus Lungs. The interstitial fluid
volume was determined in three lungs that
had been perfused with a dextran solution
containing essentially no plasma protein. These
lungs had become very heavy, weighing almost twice as much as normal in relation to
the dogs' body weights. The interstitial fluid
volumes in these three lungs were 24, 20, and
25 cc per lobe respectively. The largest single
value in table 1 for normal lungs from dogs of
almost identically the same size was 11 cc.
^ 00 CO 1O ^^ 00 00 l-O O^ O5 t"** t"*" ^^ t***
<
a
•§•§
N « £
NA'l PERMEABILITY COEFFICIENT
4
Table 2 gives alveolar membrane Na* per- 5 - a
meability coefficients, P, determined in 34 separate lungs utilizing three slightly different
techniques.
In series A, the flow rate through the lung
was always 150 cc/min. However, because of
S-5
slight differences in lung weights, the flow
CQ
rate/min/g lung tissue varied from 5.0 cc up
to 7.0 cc. The volume and surface area data
of table 1 were obtained from this series of
experiments.
In series B, the flow rate was purposely
o
varied in 13 different experiments between
05
3.8 and 12.8 cc/min/g lung tissue.
In series C, the alveoli of nine separate
lungs were filled with fluid after the O2 degassing procedure described under Methods
instead of the usual vacuum technique used
in most of the experiments.
The permeability coefficients for Naf* determined in the three different series were
7.5 ±2.1 (SD), 7.0 ±1.8, and 7.1 ± 1.9 respectively X 10~7 cm /sec The mean of all
£c£
34 values was 7.2 ± 2.1 X 10~7 cm/sec.
Effect of Variable Flow Rate. In series B of
2: 05
•«
— coinoJcqincocopcqinco p co
^ co"^^ininincDt~t*-ooo>
CM
C^^ C ^ ^^^ ^ ^
^3^ CLO 0^5 ^^^ ^?1 0 0
cdcooJ-q'CDTiHOJdiri
CD
CO C73
+1
in
O
00 ^
t""~
^H CD ^~^ ^ * t—
O
ininco'iocDiocDcoin
s i1
HCICO^WJONCO^
OrcmUtioa Raetrcb, Vol. XVI, April 196)
359
PULMONARY MEMBRANE PERMEABILITY
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
table 2, the perfusion flow rate was changed
over a range of about fourfold in 13 lungs as
described above. The data showed no obvious
correlation between flow rate and measured
value of P, actually demonstrating a slightly
negative correlation coefficient of —0.13.
Effect of Different Alveolar Filling Procedures. Series C of table 2 gives Na2< permeability coefficients determined in nine dogs in
which the alveoli were filled after the oxygen
degassing procedure. The values determined
in this series were not significantly different
from those determined in Series A and Series
B in which alveolar filling was achieved by
the vacuum procedure.
K» PERMEABILITY COEFFICIENT
K-" decay curves in the vascular compartment were determined in the same way as
for the Na'J curves. However, VA and W
were calculated indirectly from values found
in the Naf* studies. VA was considered to be
the same percentage of the total broncho-alveolar volume as was found in the Na*4 studies. Vv was determined by first measuring
perfusion fluid volume and adding to this the
average interstitial fluid volume in proportion to lung weight as that found in the Na**
studies shown in table 1.
The average K" permeability coefficient determined in four experiments was 56.5 ± 3.5
X 10-7 cm/sec.
PERMEABILITY COEFFICIENTS FOR UREA AND GLUCOSE
The permeability coefficients for urea and
glucose were determined by the alternate
method described under Methods, in which
the test substance was added to the perfusing
system along with Na*4. The half-rimes for
disappearance of the urea or glucose and of
the Naf* were determined simultaneously.
From these values, the urea and glucose permeability coefficients were calculated.
In four experiments, the average permeability coefficient for urea was 22.9 ± 9.2 X 10"7
cm/sec.
The permeability coefficient for glucose in
six experiments was 3.14 ± 0.72 X 10~7 cm/sec.
ATTEMPTS TO DETERMINE D,0 AND 1-4 DYNITROPHENOL
(DNP) PERMEABILITY COEFFICIENTS
DoO and DNP disappearance curves from
the vascular compartment were so steep that
it was impossible to take samples rapidly
enough to make accurate measurements. Calculations showed that the rate of transport of
the solutes across the alveolar membrane approached the flow-rates of the solutes in the
vascular perfusion fluid. Because of this flowlimitation, and because of the uncertainties
regarding the relative roles of flow-limitation
and diffusion through the pulmonary capillary
membrane, it was not possible to calculate
the permeability coefficients of the alveolar
membrane for D^O and DNP. However,
from the steepness of the slopes it was
possible to calculate that the minimum permeability coefficient for each of these two
substances was at least 400 X 10~7 cm/sec.
COMPARISON OF PERMEABILITY COEFFICIENTS FOR DIFFERENT SUBSTANCES
Table 3 shows comparative permeability coefficients for six different substances, illustrating that the permeability of the membrane
was about three times as great for urea as for
Na'4, seven to eight times as great for K**,
and at least four hundred times as great for
DL.O and DNP. The results also show that the
TABLE 3
Relative Permeability Ratios for Na2i, K**, Urea, Glucose, 1-4 Dinitrophenol, and
Obtained as Explained in the Text
Substance
No. of
experiments
Permeability
coefficients.
cm/sec x 107
D..O
DNP
K*«
Urea
Na'J
Glucose
5
8
6
4
28
6
>400
>400
58.5 ±3.5
22.9 ± 9.2
7.2 ±2.1
3.1 ±0.7
CircmlttioM Rnurcb, Vol. XVI, April 196}
D,0
Relative
permeabilities,
P»/PN."
>55.6
>55.6
7.9
3.2
1.0
.43
TAYLOR, GUYTON, BISHOP
360
permeability for glucose was only about onehalf to one-third that for sodium.
EXPERIMENTS IN WHICH THE ALVEOLAR
TAGGED INITIALLY
FLUID WAS
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
In 20 additional experiments the test substance was placed in the alveolar side of the
system, and either buildup curves in the vascular perfusion system or decay curves in the
lavage system were studied. The average
permeability coefficients for Na lj , K if , and
urea studied in this way agreed in general
with the permeabilities listed in table 3,
though the results were more variable and
less reliable because these experiments were
performed in early studies in which the characteristics of the system itself were being investigated.
Discussion
The data presented in this paper indicate
that the permeability of the overall pulmonary
membrane is determined mainly by the permeability of the alveolar membrane and that
this has permeability characteristics essentially the same as those of other cellular membranes but quite different from those of a simple capillary membrane. The data supporting
this conclusion are:
First, the Na*1 permeabilities listed in table
2 are very near to those obtained by others
for other cellular membranes. For instance,
Davis17 has collected permeability coefficients from several investigators and gives
values in the order of 10"8 cm/sec for Na**
transport across erythrocyte membranes, frog
skin, and muscle cell membranes. The values
obtained in this work are in the order of 10~7
cm/sec, which compare favorably with those
for the other cellular membranes. The difference, possibly, results, at least partially, from
underestimation of the pulmonary membrane
surface area; recent electronmicrographic studies have shown that the alveolar lining cells
have many small infoldings and outfoldings
that resemble villi.18 This could easily increase
the surface area of the pulmonary membrane
and reduce the calculated Na permeability
several fold.
Second, D^O crossed the membrane at a
rate too rapid for us to measure in our system. Water movement through cellular membranes is usually so rapid that it is measurable only by special techniques.
Third, DNP also passed through the membrane at a rate too rapid to be measured. This
is an indication that the membrane is very
permeable to at least this one fat-soluble substance, which is also a well known characteristic of cell membranes.
Fourth, the relative permeability ratios obtained in table 3 show that the pulmonary
membrane is selectively permeable to different solutes, as is also true of other cellular
membranes. For instance, the permeability ratio for K4* and Na*1 found in these studies,
7.6 to 1, approached the ratio of 10 to 1 obtained for several other cellular membranes.10
Also, the ratios for Na*^ to glucose and Nafi
to urea agreed well with those found for
other cellular membranes.-"
On the other hand, if the pulmonary membrane had the properties of a capillary membrane, the permeabilities should have been
several orders of magnitude greater than those
actually found, and Naft should have penetrated the membrane almost equally as rapidly-1 as urea or YLia.
Thus, the data support the thesis that the
overall pulmonary membrane acts as if it
were a cellular membrane and not a capillary
membrane. This conclusion was also indicated by the studies of Kylstra3 and Swann and
Spafford4 who studied transport of fluid and
solutes through the pulmonary membrane in
drowning animals. Also, Salisbury et al.,7 in
attempts to use the lung as a dialyzing membrane, showed that urea crosses the pulmonary membrane very slowly, and Quails et
al.22 showed that Na1;i in fluid droplets placed
in the lungs is absorbed only after many minutes.
Two facts indicate that the system was not
flow limited except for the case of D^O and
DNP: 1) The flow rates of the vascular perfusion fluid averaged slightly greater than
150 cc/min. The products of the alveolar surface area times the permeability coefficients of
the test substances were: glucose, 0.32 cm 3 /
Circulttiom Rtsircb, Vol. XVI, April 196S
361
PULMONARY MEMBRANE PERMEABILITY
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
min; Na**, 0.75 cm3/min; urea, 2.31 cms/min;
and K*', 5.7 cms/min. Thus, the ratios of flow
to these products were respectively 469 to
1, 200 to 1, 65 to 1, and 26 to 1. Even the
minimum value, 26 to 1, which was the value
for K", was high enough that flow-limitation
should not have played a measurable role in
the results. 2) The calculated values of the
permeability coefficients for Na" under varying flow conditions that covered a fourfold
range showed no correlation with flow as one
would expect in aflow-limitedsystem.
It is possible that movement of K |f , urea,
and glucose into the cells of the pulmonary
tissues might have caused additional dilution
of these test substances and thereby caused
permeability coefficients that are too high.
However, the average mass of the lung tissue minus the measured mass of interstitial
fluid was only 18 g so that the maximum
volume of cells is small enough that this factor probably played little role in the results.
Finally, there remains the possibility, or
even probability, that active transport occurs
to some extent through the alveolar membrane. However, two aspects of these studies
indicated that active transport did not occur
in sufficient magnitude to affect the results
significantly: First, control studies in which
equal concentrations of Na*4 were effected in
the two compartments showed no evident
transport of the sodium. Second, the decay
curves of all the test substances were single
exponentials, which would be expected in the
case of diffusion but unexpected in most instances of active transport. Yet, it should also
be noted that the temperature of the perfusion
fluids in these studies was only 20 to 22°C,
which was low enough to minimize active
transport in relation to diffusion. Therefore,
there is still a possibility that significant active transport can occur through the alveolar
membrane under appropriate conditions.
cients were measured for diffusion of several substances across the alveolar membrane.
The permeability coefficients of the pulmonary membrane for Ku, urea, Na*-1, glucose,
D,.O, and dinitrophenal(DNP) were 56.5 ±
3.5, 22.9 ±9.2, 7.5 ±2.1, 3.1 ±0.7, 400, and
400 X 10~7 cm/sec, respectively. The effect of
varying the flow rate on the permeability coefficient of Na Ji was investigated, and the data
failed to show any significant correlation between the flow limits of 3.8 to 12. 8 cm/min/g
lung tissue. The effect of two different procedures for filling die alveoli with fluid on the
permeability coefficients was also investigated
and no difference could be discerned in the
results. The data support the thesis that the
alveolar membrane has permeability characteristics similar to those of the usual cell
membrane. The interstitial fluid volume of
the lung (extravascular sodium space) was
measured and yielded a value in normal lungs
of 0.250 ± 0.129 cc/g lung tissiue. In three
edematous lungs, this space averaged three
times the normal value.
References
1. GUYTON, A. C , AND LINDSEV, A. W.: Effect of
elevated left atrial pressure and decreased
plasma protein concentration on the development of pulmonary edema. Circulation Res.
7: 649-657, 1959.
2.
Pulmonary extravascular water volumes from
transit time and slope data. J. Appl. Physiol.
17: 179-183, 1962.
3. KYLSTRA, J. A.: Simplified technique of lavage of
the lung. Acta Physiol. Pharmacol. Neerl. 9:
225-239, 1960.
4. SWAN,V, H. C , AND SPAFFORD, N. R.: Body salt
and water changes during fresh and sea water
drowning. Texas Rept. Biol. Med. 9: 356-382,
1951.
5. PAPPENHEIMER, J. R.: Passage of molecules
through capillary walls. Physiol. Rev. 33: 387423, 1953.
6.
Circul.tio* Rtsurcb, Vol. XVI, April 196}
LANDIS, E. M., AND PAPPENHEIMER, J. R.:
Ex-
change of substances through capillary walls.
In Handbook of Physiology, sec. 2, Circulation,
vol. 2, chapt 29, Washington, D.C., Am.
Physiol. Soc., 1963, pp. 961-1034.
Summary
The lower lobe of the left lung of 54 dogs
was isolated and perfused with a dextran-Tyrode's solution. The alveoli were filled with
Tyrode's solution, and permeability coeffi-
CHINABD, F. P., ENNS, T., AND NOLAN, M. F.:
7.
ROSS, C , RlEBEN, P. A., AND SALISBURY, P. F . :
Urea permeability of the alveolar membrane;
hemodynamic effects of liquid in the alveolar
spaces. Am. J. Physiol. J98: 1029-1031, 1960.
362
TAYLOR, GUYTON, BISHOP
8. GENTZKOW, C. ] . : An accurate method for de-
termination of blood urea nitrogen by direct
Nesslerization. J. Bio. Chem. 143: 531-544,
1942.
9. HAWK,
P. B., OSER,
B. L., AND SUMMERSON,
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
W. H.: Practical Physiological Chemistry, ed.
13. New York, McGraw-Hill Book Company,
Inc., 1947, pp. 568-570.
10. ROUCHTON, F. J. \V.: The average time spent
by the blood in the human lung capillary and
its relation to the rates of CO uptake and
elimination in man. Am. J. Physiol. 143: 621633, 1945.
11. MILLER, W. S.: The Lung. Springfield, Charles
C Thomas, 1937.
12. SHEPPARD, C. W.: Basic Principles of the Tracer Method, ed. 1, New York, John Wiley
and Sons, 1962.
13.
COLLANDER, R., AND BARLUND, H.: Permeabili-
tiitsstudien an chara Ceratophylla. Acta Bot.
Fenn. 11: 1-114, 1933.
14.
SOLOMON,
A.
K.:
The
kinetics
of
biological
processes: Special problems connected with
the use of tracers. Advan. Biol. Med. Phys.
3: 65-97, 1953.
15.
ROBERTSON, J. S.: Mathematical treatment of up-
take and release of indicator substances in relation to flow analysis in tissues and organs.
In Handbook of Physiology, sec. 2, Circulation, vol. 1, Washington, D.C. Am. Physiol.
Soc. 1963, pp. 617-644.
16. FRANK, N. R.: A comparison of static volumepressure relation of excised pulmonary lobes
of dogs. J. Appl. Physiol. 8: 274-277, 1952.
17.
DAVES, J. T.: The mechanism of diffusion
of
ions across a phase boundary and thin cell
walls. J. Phys. Colloid Chem. 54: 185-204,
1950.
18.
KARRER,
H. E.:
The
ultrastmcture
of
mouse
lung. J. Biophys. Biochem. Cytol. 2: 241-251,
1956.
19.
HARRIS, E. J., AND BURN, G. P.: The transfer of
Na and K ions between muscle and the surrounding medium. Trans. Faraday Soc. 45:
508-528, 1959.
20.
CUYTON, A. C : Textbook of Physiology, ed. 2.
Philadelphia, W. B. Saunders Company, 1961,
p. 37.
21. RENKIN, E. M.: Transport of K4- from blood to
tissue in isolated mammalan skeletal muscles.
Am. J. Physiol. 197: 1205-1210, 1959.
22.
QUALLS, G., CURTIS, H. J., AND MENEELY, G. R.:
Rate of uptake of fluid from lung measured
with radioisotopes; comparison of rapid rate
for water with slower rate for saline. Am. J.
Physiol. 172: 221-225, 1953.
Circulation Rutmrcb, Vol. XVI, April 196}
Permeability of the Alveolar Membrane to Solutes
AUBREY E. TAYLOR, ARTHUR C. GUYTON and VERNON S. BISHOP
Downloaded from http://circres.ahajournals.org/ by guest on June 17, 2017
Circ Res. 1965;16:353-362
doi: 10.1161/01.RES.16.4.353
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1965 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/16/4/353
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/