243
Effect of Reduced Energy Metabolism
and Reperfusion on the Permeability
and Morphology of the Capillaries
of an Isolated Rete Mirabile
Eugenio A. Rasio, Moise Bendayan, and Carl A. Goresky
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
The effects of reduction in energy metabolism were explored in the eel rete mirabile, an organ
composed predominantly of capillaries. In vitro experiments showed that glycolysis is the major
pathway of energy production in this capillary tissue, and that iodoacetate, KCN, and low POj
in combination markedly reduce its ATP generation. When in situ energy generation was
inhibited by this combination during countercurrent perfusion of the arterial and venous
capillaries of the rete, an approximate doubling of the intercapillary barrier permeability for
human [ lu I]albumin, [l4C]sucrose, and n Na was found. Structural damage was evident, but the
intercellular junctions remained intact. The effect of cessation of flow for 30 minutes, followed
by reperfusion, was then explored. Stasis alone altered the structure, chiefly of the venous
capillary endothelium, but not the permeability of the intercapillary barrier. Stasis with a
hypoxic medium containing the inhibitors of energy generation, followed by reperfusion with
oxygenated control medium, resulted in a progressive breakdown of the intercapillary barrier,
with a threefold to fourfold increase in solute (labeled albumin, sucrose, and sodium)
permeability, evolving during early reperfusion, but no change for labeled water permeability.
Morphologically, the endothelial cells, especially those in venous capillaries, showed substantial
damage; they appeared vacuolated, their cytoplasm was extracted, and cytoplasmic and
membrane debris were found in the lumen; intercellular junctions remained intact. Local
pericyte detachment with interstitial edema also appeared. Thus, stasis and reperfusion
amplified the effects of reduction in energy generation and hypoxia on both permeability and
morphological change. (Circulation Research 1989;64:243-254)
W
hen ATP concentrations are reduced by
hypoxia or ischemia, numerous cell functions are impaired.1 The sensitivity of
blood endothelial capillary cells to low levels of
ATP with respect to permeability is not well documented. Conflicting results have been reported:
while many investigators have not been able to
measure any action of hypoxia on capillary permeFrom The Department of Medicine, Notrc-Dame Hospital and
the Department of Anatomy, Universitd de Montreal; and the
McGill University Medical Clinic in the Montreal General Hospital, Montreal, Quebec, Canada.
Supported by grants from the Medical Research Council of
Canada, the Quebec Heart Foundation, and the Fast Foundation. M.B. is a Scientist and C.G. is a Career Investigator of the
Medical Research Council of Canada.
Address for correspondence: Dr. Carl A. Goresky, University
Medical Clinic, Montreal General Hospital, 1650 Cedar Avenue,
Montreal, Quebec H3G 1A4, Canada.
Address for reprints: Dr. Eugenio Rasio, 9E Pavillion Mailloux, Hopital Notre Dame, 1560 Sherbrooke E., Montreal,
Quebec, H2L 4MI, Canada.
Received November 4, 1987; accepted July 15, 1988.
ability, others have reported moderate to striking
effects. There are many plausible explanations for
this conflicting evidence. One potential major cause
is the difficulty of methods for measuring transcapillary transport in whole organs; associated with
this, there is also the problem of accounting for the
vascular, Theological, and metabolic changes that
occur during hypoxia, which make data interpretation difficult. The effects of hypoxia on capillary
permeability have been frequently probed with macromolecules, mostly albumin; but recent studies
indicate that, at least in the blood-brain barrier, the
permeability to small molecules and to ions can also
be increased by hypoxia. 23 The time at which the
effects of hypoxia on capillary permeability have
been observed varies from a few minutes to hours
or days. Finally, there is a wealth of data establishing the morphological,4 biological,5 and functional6
heterogeneity of vascular endothelial cells; against
this, interpretations of effects must be made.
The effects of hypoxia on capillary permeability
have generally been explored by reducing the oxy-
244
Circulation Research
Vol 64, No 2, February 1989
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
gen tension in blood or perfusate. Very few investigators have used metabolic inhibitors of energy
production, alone or in combination with low oxygen partial pressure. In hypoxia studies, flow has
been generally normal or higher than normal. In
other types of experiments, hypoxia has been
induced by ischemia of various degrees of intensity
and duration, and permeability has necessarily been
assessed during recirculation through the ischemic
tissue, when flow has been varied from small reflow
to overperfusion.
In the present study, we have used the rete
mirabile of the eel swimbladder7 to investigate the
effects of changes of oxygen supply, cell energy
production, and organ flow stoppage on capillary
permeability. In a preliminary series of in vitro
experiments, we have measured the effects of
hypoxia and metabolic inhibitors on substrate utilization and the equivalent yield of ATP by rete
capillary endothelial cells. We have then used
hypoxia and the metabolic inhibitors in the perfused
rete to induce a maximal reduction of energy production under conditions of either continuous flow
or arrested flow followed by reperfusion, and have
characterized the rete permeability to a set of multiple water soluble tracers. In every set of experiments, we have searched for morphological correlates with changes in capillary function.
Materials and Methods
The rete mirabile was isolated from the swimbladder of the freshwater eel (Anguilla anguilla). The
arterial and venous inputs were catheterized at
opposite poles, and the arterial and venous capillaries of the rete were simultaneously countercurrent
perfused with identical flows of Krebs-Ringer bicarbonate buffer (pH 7.4), containing glucose (5 mM)
and purified bovine albumin (4 g/100 ml) and equilibrated with a gas mixture of 95% O2-5% CO2.
In Vitro Metabolic Studies
The rete, cleared of its blood, was removed from
the swimbladder and dissected into small clusters of
filaments. Approximately 100 mg wet weight of
capillary tissue were incubated for two hours at
37° C in vials containing 2 ml of the same buffer.8
The in vitro effects of metabolic inhibitors and of
lack of oxygen (produced by equilibration with 95%
N2-5% CO2) were assessed in paired experiments
where control and experimental metabolic vials
received equivalent amounts of capillaries from the
same eel. The metabolic inhibitors used were sodium
iodoacetate, which reacts with thiol groups in many
enzymes and blocks glyceraldehyde phosphate dehydrogenase, and potassium cyanide, which suppresses electron flow between the cytochrome oxidase complex and oxygen in the mitochondria. The
medium in the closed vial was acidified at the end of
the incubation, and the evolved gaseous carbon
dioxide was trapped in a scoop filled with phenethylamine. Glucose and lactate concentrations in the
medium were measured by specific enzymatic procedures. Glucose conversion into carbon dioxide
was calculated by dividing the activity absorbed to
phenethylamine by the specific activity of [MC(U)]
glucose in the medium at the start of the incubation.
Permeability Studies
Perfusions7 were carried out at room temperature
(22-25° C), with a constant flow in each direction of
0.5 ml/min, at a constant pressure head of 45 cm
H2O. The weight of the rete averaged 200 mg. The
medium at the arterial input contained various combinations of the following labeled materials: human
[123I]albumin (Frosst: more than 95% of the labeled
iodine was precipitated with 10% trichloroacetic
acid), [14C(U)]sucrose or [MC]urea (both from New
England Nuclear, Boston, Massachusetts; 1-5 mCi/
mmol and 2-10 mCi/mmol, respectively), [3H]
water (New England Nuclear; biological quality,
0.1 mCi/g), and H Na (New England Nuclear; as
sodium chloride, 99.9% radionuclidic purity). The
medium at the venous input did not contain radioactive tracers.
In a first set of experiments, the effects of metabolic inhibitors and reduced oxygen tension were
explored, while the rete was continuously perfused.
Perfusion for 1 hour with control medium equilibrated with 95% O2-5% CO2 was followed by a
1-hour perfusion with the same medium where
glucose was removed and the metabolic inhibitors
iodoacetate 1 mM or potassium cyanide 1 mM were
added at both arterial and venous inputs either
alone or in combination; in the latter instance, the
medium was desaturated to a PO2 of 10±3 mm Hg,
by equilibration with a 95% N2-5% CO2 gas mixture.
In a second set of experiments, perfusion of the
rete was stopped and then resumed. The protocol
was designed to simulate the effects of ischemia,
followed by reperfusion. After 30 minutes of perfusion with the control oxygenated medium, the rete
was perfused for another 10 minutes with either the
same control medium or a medium without glucose,
with iodoacetate 1 mM and potassium cyanide 1
mM, and a POj of 10 mm Hg. The arterial and
venous inputs were then clamped for 30 minutes.
Following this, the rete was reperfused with the
control oxygenated medium for an additional hour.
Samples were collected simultaneously from the
arterial and venous outputs for 5 to 10 minute
periods of time throughout the perfusions. There
were no significant differences in flow during control and experimental perfusion. The radioactivity
of [mI]albumin was measured in a Packard gamma
scintillation spectrometer (Packard Instrument Co.,
Downers Grove, Illinois) on a 10% trichloroacetic
acid precipitate washed with excess potassium
iodide. The radioactivity of the other tracers, in
the protein-free supernatant, was measured with a
gamma scintillation spectrometer or with a Packard liquid scintillation spectrometer. All values
Rasio et al Capillary Effects of Hypoxia and Ischemia
245
TABLE 1. Effect of Inhibitors of Energy Production on Glucose
Utilization by Isolated Rete Capillaries
CONTINUOUS PERFTJSION: CONTROL
Glucose
Lactate
Glucose production conversion ATP
uptake (/unol/g/hr)
toCOj
yield
18.2±2.0
Control medium
!—i
i
t !—I—f
i—1—{—f-t I 1
i
+ KCN(lmM)
95%N r 5%CO 2
+Iodoacetate (1 mM)
+Iodoacetate (1 mM)
+ KCN (1 mM)
95% N r 5 % COj
12
10
30
60
TIME min
90
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
FIGURE 1. Lack of change in the VJAC ratio for each
permeability probe with time. Changes in permeabilitysurface area values are reflected by changes in this ratio.
Note that a logarithmic ordinate scale has been used.
Mean values are given for the data with lines for standard
deviations. Number of experiments, 6. Vo and Ao are the
simultaneous tracer concentrations at the venous and
arterial outputs, respectively. The codes for the curves
are 1 for [12il]albumin, 2 for [I4C]sucrose, 3for22Na, and
4for[3H]water.
were corrected for background, quenching, and
cross-over.
The diffusion capacity (PS) of the rete was calculated when steady state concentrations were
achieved, from the equation PS=(F- Vo/Ao), where
P is the permeability coefficient, S is the surface
area, F is the flow, and Vo and Ao are the simultaneous concentrations at the venous and arterial
outputs.7 Summed recoveries of tracers from arterial and venous outputs were within ±2% of the
arterial input. The surface available for capillary
exchange is equivalent to 1 cm2/mg wet wt, under
normal conditions of hydration.7 In this study,
various degrees of cellular and interstitial space
135±2O
132±15
23±2
30±8
36.6±4.5
0.545±0.047 57
(% control value)
135±16
27±3
125±9
14±7
0
29±3
33±5
0
swelling were induced by reduction in the energy
metabolism of the rete. Thus, the diffusion capacity
(PS) of the rete, rather than its coefficient of permeability (P), was followed as a function of time.
Morphological Studies
Fragments of the rete were prepared at the end of
perfusions, as described previously.9 The tissues
were fixed by immersion in glutaraldehyde fixative
solution, postfixed with osmium tetroxide, and
embedded in Epon. Each eel has two symmetric
retia. The rete that was not used in the experiments
was removed during the initial preparative procedures and similarly processed for control morphological examinations. Thin sections were cut and,
after staining, were examined with an electron
microscope (model 410, Philips Electronic Instru-
P
[123I] Albumin
)
[14C]Sucrose
Group 1
Control medium
+ K.CN (1 mM) no glucose
4.64±2.06
3.09±1.00
Group 2
Control medium
+Iodoacetate 1 mM no glucose
3.49+0.88
5.04±1.09*
14.9±2.5
18.9±4.4
4.55±1.49
10.19±3.01t
14.0±2.3
23.5±4.3t
Group 3
Control medium
+ KCN 1 mM + Iodoacetate
(1 mM) 95% N r 5 % CCh no
glucose
21
Results are expressed as mean±SEM. Number of experiments: 32 for control and eight for each group of metabolic
inhibitors. The control medium is a Krebs-Ringer bicarbonate
buffer with purified bovine albumin 4 g/100 ml and glucose 5 mM.
Incubations were carried out for 2 hours at 37° C. All values for
metabolic inhibitors are significantly different from control values: p<0.01 (Student's t test for paired experiments).
TABLE 2. Effect of Inhibitors of Energy Production on Rete Diffusion Capacity to Water Soluble Tracers
Perfusion
%
86
11
H
Na
43.2±7.6
75.0±17.9t
tl4C]Urea
[3H]Water
98.8±9.7
102.5±8.7
l,805±184
l,903±318
98.6±18.4
107.7±23.4
l,832±26O
1,879±268
2,294±328
2,454±522
Results are expressed as mean±SEM. Each tracer was tested in eight paired experiments in groups 1 and 2 except
for sucrose and urea in group 2, which were tested in four paired experiments. Each tracer was tested in six paired
experiments in group 3. PS^, permeability-surface area product at 25° C. The control medium is a Krebs-Ringer
bicarbonate buffer with purified bovine albumin (4 g/100 ml) and glucose (5 mM).
•Value significantly different from control value: p<0.01 (Student's / test for paired experiments).
tValue significantly different from control value: p<0.05.
246
Circulation Research
Vol 64, No 2, February 1989
CONTINUOUS PERFUSION
Vo/Ao
CONTROL
Q DECREASED ATP GENERATION
i
i
f—}—f
J*
10
10
30
mM is added to the medium or when the oxygen
tension is reduced, glucose uptake and lactate output are stimulated by approximately 30% above
control values, while oxidation of glucose to CO2 is
reduced by an average of 80%. The combined effect
is that ATP continues to be generated at a normal
rate. Iodoacetate, at a concentration of 1 mM, when
used alone or in combination with cyanide and low
oxygen tension, suppresses or strongly reduces
both the anaerobic and aerobic catabolism of glucose, so the total equivalent rate of yield of ATP
drops to 10-20% of the control value.
90
60
TIME mln
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
FIGURE 2. Time course of VJAO after a steady arterial
infusion of tracers beginning at time zero, in a countercurrent perfused rete. Perfusion was initially with an oxygenated control medium. At the end of the first hour, the
perfusion was switched to a deoxygenated control medium
lacking glucose and containing 1 mM KCN and 1 mM
iodoacetate. Vo and Ao are simultaneous tracer concentrations at venous and arterial outputs, respectively. Mean
values are given for the data, with lines for standard
deviations. Number of experiments, 6. * Values significantly higher than averaged control values (p<0.05). 1 =
[ml]albumin, 2=['4C]sucrose, 3=nNa, and4=[3H]water.
ments, Mahwah, New Jersey). Several hundreds of
capillaries were examined from each experiment.
Results
In Vitro Metabolic Studies
Table 1 gives the values of glucose utilization by
isolated rete capillaries and the effect of inhibitors
of energy production. Most glucose taken up from
the medium is released as lactic acid; oxidation
through the Krebs cycle accounts for approximately 3% of the glucose uptake. The total equivalent rate of yield of ATP from glycolysis and oxidation is 57 ^mol/g/hr. When potassium cyanide 1
Permeability Studies
Control experiments, shown in Figure 1, demonstrate that the ratio V o /A o , which reflects the
permeability-surface area product in the presence
of constant flow, remains steady for each tracer
during a baseline 2-hour perfusion. These control
studies indicate that, in the absence of experimental
manipulation, a steady baseline can be expected for
this period.
The results of the effects of inhibitors of energy
production and lack of glucose in the perfusate on
the permeability-surface area values of the rete
capillaries are shown in Table 2. There were no
detectable effects of potassium cyanide on the tracer
concentrations and perfusate flows at the arterial
and venous outputs. Consequently, the diffusion
capacities of the rete for [l25I]albumin, [MC]urea,
and tritiated water were similar during control and
experimental conditions. Addition of iodoacetate to
the perfusate increased the permeability-surface
area value for albumin by approximately 50%
(p<0.01); the effect was established and steady
from the 10th minute onward. The permeabilitysurface product values for [uC]sucrose, [14C]urea,
and tritiated water were not modified by the iodoacetate. When the rete was perfused with a mixture
of potassium cyanide and iodoacetate at low Po2
REPER FUSION
CONTROL
»
CONTROL
1
10°
REPERFUSION
M-l1
«
1
I
t*
10"
o
10"
I
a n
i
10"
t -t—f2
10
30
0
TIME m i n
30
10
30
0
TIME m i n
FIGURE 3. Time course ofVJAo after a steady arterial infusion of tracers, beginning at time zero, in a countercurrent
perfused rete. Mean values are shown with standard deviations. *Values significantly higher than averaged control
values (p<0.05). l=[l2iI]albumin, 2=['4C)'sucrose, 3=22Na, and 4=[3H]water. Data are given for the two protocols
involving flow stoppage. Six perfusions were performed for each protocol. Left panel: Perfusion with oxygenated control
medium, stasis for 30 minutes, and reperfusion with oxygenated control medium. Right panel: perfusion with oxygenated
control medium, stasis for 30 minutes with a low POj medium lacking glucose and containing iodoacetate and KCN, and
reperfusion with oxygenated control medium.
Rasio et al
TABLE 3.
Capillary Effects of Hypoxia and Ischemia
247
Effects of Stagnant Hypoxia and Reperfoskm on Rete Diffusion Capadty for Water Soluble Tracers
[123I]Albumin
Perfusion
PSu (x 10"3 cmVsec)
[l4C]Sucrose
^ a
[3H]Water
15.5±5.7
26.3+10.8
57.5±13.7
49.2±6.7
2,643+254
2.26±0.56
16.0+2.2
46.3±6.2
1,817+211
6.27±0.40*
70.2±14.3t
198±44t
2,264+329
Group 1
2.43±1.40
3.59+1.64
Perfusion with control medium
Stasis for 30 minutes+reperfusion with
control medium
2,673 + 278
Group 2
Perfusion with control medium
Stasis for 30 minutes with metabolic
inhibitors +reperfusion with control
medium
Results are expressed as mean±SEM. Each tracer was tested in six paired experiments. PSa, permeabilitysurface area product at 25° C. The control medium is a Krebs-Ringer bicarbonate buffer with purified bovine
albumin (4 g/100 ml) and glucose (5 mM). The medium with metabolic inhibitors has no glucose, contains KCN (1
mM) and iodoacetate (1 mM) and is equilibrated with 95% N2-59& CO2.
*Value significantly different from control value: p<0.01 (Student's / test for paired experiments).
tValue significantly different from control value: p<0.05.
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
(Figure 2), its diffusion capacity for [125I]albumin,
[14C] sucrose, and H Na increased by approximately
100% (p<0.05), while remaining similar for tritiated
water. For the group of tracers exhibiting the
increased permeability, the change became perceptible within 10 minutes and progressed until plateau
values were reached after 30 minutes.
vc
r?
Figure 3 shows the results of the combined effects
of stasis and reperfusion on the permeabilitysurface area values of the rete for water soluble
tracers. In the left-hand panel, the results pertain to
experiments where theflowof an oxygenated energyrich medium was discontinued and then resumed.
The time course of the tracer output concentrations
*r
f
'•>
ISp
AC
FIGURE 4. Electron photomicrograph oj arterial [AC) and venous (VC) capillaries of a rete perfused under control
conditions. Tight junctions (J). The venous endothelium is thin and fenestrated (f). The interstitial space (IS) has
pericytes (P) and bundles of collagen fibers. Magnification, x30,000.
248
Circulation Research Vol 64, No 2, February 1989
VC
BL
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
-J
M
AC
FIGURE 5. Electron photomicrograph of arterial (AC) and venous (VC) capillaries of a rete perfused under control
conditions. Tight junctions (J). The venous endothelium is thin andfenestrated (f). Endothelium is lined by well-defined
basal laminae (BL). M, mitochondria; P, pericyte. Magnification, xSO.OOO.
from arterial and venous capillaries was not altered
by the procedure; the illustrated Vo/Ao value, which
reflects the permeability-surface area values (Table
3), were steady for each tracer following reperfusion. In the right-hand panel, the results are displayed for the experiments in which both flow and
energy generation were arrested: the inhibitors iodoacetate and KCN were introduced in a low Po^
medium before flow stoppage, and normal conditions of flow and energy were then restored by
reperfusion with an oxygenated energy-rich medium.
The V^Ao concentration ratio values for [I23I]
albumin, [MC]sucrose, and ^ a begin to rise after 5
to 10 minutes following resumption of the perfusion, and attained plateau levels at the 30th minute.
The corresponding arterial output concentrations
decreased and then leveled out, pan passu with the
venous output concentrations. There was no significant change with time in the VJA,, ratio for tritiated
water. The permeability-surface area values, calculated from averaged steady-state tracer concentrations during the initial 30 minutes of control perfusion and the last 30 minutes of reperfusion, showed
a threefold to fourfold increase for [l23I]albumin,
[14C]sucrose, and "Na, while for tritiated water they
were not modified (Table 3).
Morphological Studies
Normal rete capillaries are shown in Figures 4
and 5. The endothelial cells in the arterial capillary
were high, and their cytoplasm contained a welldeveloped vesicular-tubular system. The endothelial cells in the venous capillaries were thin and
fenestrated; their fenestrae were closed by a diaphragm. Mitochondria were present. The intercellular junctions were tight, and the basal laminae,
continuous. Pericytes were abundant and lined the
outer aspects of the endothelial cells. The intercapillary extracellular space was small and filled with
bundles of collagen fibers.
When the rete was perfused with potassium cyanide or iodoacetate and no glucose, the mitochondria in the endothelial cells and pericytes became
swollen and vacuolated, and their cristae were
reduced or even absent (Figure 6). However, the
intercellular junctions were tight and the venous
capillary fenestrations remained closed by their
diaphragms. In the series of experiments in which
the perfusate contained the metabolic inhibitors
iodoacetate and KCN, and no glucose, and input
oxygen tension was kept low but flow was maintained, consistent and widespread lesions were
Rasio et al
vc
Capillary Effects of Hypoxia and Ischemia
249
M
BL
-J
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
AC
FIGURE 6. Electron photomicrograph of capillary tissues at the end of perfusion experiments with metabolic poisons in
which KCN was added to the perfusion medium. The mitochondria (M) are swollen with reduced cristae. The intercellular
junctions (J) as well as the interstitial space are intact. P, pericyte; BL, basal laminae. Magnification, x.40,000.
observed in both arterial and venous endothelia
(Figure 7). The luminal membrane of the endothelial
cells was swollen and formed large vacuoles that
protruded into the lumen. These large vacuoles
were outlined by layers of concentric membranes,
displaying myelin-like profiles; some vacuoles were
eventually released into the lumen. The mitochondria were vacuolated. Intercellular junctions and
fenestrae, however, were intact.
Figures 8 and 9 illustrate the lesions that were
seen when flow was arrested for 30 minutes and
then resumed for 1 hour. The lesions were similar
whether or not the medium stagnating in the rete
was enriched with metabolic inhibitors and deprived
of glucose and most of its oxygen. The endothelial
cells of the venous capillaries were vacuolated and
their cytosol was extracted; both cytoplasmic and
membrane debris were found in the lumen. The
endothelial cells of the arterial capillaries were better
preserved. Without metabolic inhibitors, the interstitial space remained intact (Figure 8). With metabolic
inhibitors, patches of edema appeared in the interstitial space at sites where adjacent pericytes were
detached from their endothelial cells (Figure 9).
Discussion
Studies of respiration and energy metabolism of
blood capillary endothelial cells are very scarce. 810 "
One of us8 has reported that glucose is the major
energy substrate of the capillaries and that the
activity of the glycolytic pathway is very high. In
the present study, we confirm and complete these
data with incubation media of the same composition
as those that we used to perfuse the rete during
permeability experiments. The results apply to the
mixed population of the cells: arterial and venous
endothelial cells, and pericytes. The relative contributions of each cell type to the overall metabolic
measurements is not known. In control media, two
thirds of ATP regeneration was through glycolysis
and one third was through mitochondrial respiration. Cyanide and low oxygen tension reduced
respiration, as expected. However, glycolysis was
stimulated, and as a result, the ATP yield remained
comparable to that obtained with the control oxygenated medium. Iodoacetate (1 mM) significantly
reduced the flow of glucose through glycolysis.and
the Krebs cycle. No additional decrease in the rate
of ATP yield could be elicited by the addition of
250
Circulation Research
Vol 64, No 2, February 1989
AC
AC
-v
\
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
i
•
\
7. Electron photomicrograph of capillary tissues at the end of perfusion experiments with metabolic poisons in
which a mixture of iodoacetate and KCN was added to a medium that had a low Po2 and lacked glucose. Large vacuoles
protrude into the lumen (arrows); the mitochondria (M) are swollen and vacuolated; whereas the intercellular junctions
(J) remain tight. AC, arterial capillaries; VC, venous capillaries. Magnification, x7,000; inset, x25,000.
FIGURE
cyanide and low oxygen tension. Under the conditions of maximally inhibited glucose utilization, the
yield of ATP averaged 6 to 12 /xmol/g/hr, compared
with 57 ^.mol/g/hr in the control medium. As previously shown,8 the rete capillaries contain glycogen at a concentration of 0.4 g/100 g wet wt; this
represents 22 /xmol of glucose equivalent, per gram
of wet weight, which would be sufficient to sustain
ATP regeneration for more than 1 hour in the
absence of extracellular glucose. In considering our
experimental design, we therefore concluded that
the most efficient way to assess the effects of energy
deprivation on the permeability of the rete was to
use a perfusion medium deprived of glucose and
most of its oxygen and enriched with metabolic
inhibitors.
The effects of hypoxia and ischemia on capillary
permeability remain controversial. While many
reports have indicated a lack of effect of hypoxia on
the capillary permeability of isolated organs,12-16
there is also a wealth of data demonstrating the
opposite. 2317 - 21 In our first series of permeability
experiments, the intent was to induce a decrease in
cell oxidative metabolism without a loss of flow. We
observed that the permeability response varied with
the metabolic inhibitors added to perfusate and with
the tracer substances used (Table 2). The lack of
effect of potassium cyanide on the diffusion capacity of the rete to albumin, urea, and water could be
ascribed to an apparently adequate availability of
ATP through the compensatory effects of glycogenolysis (Table 1). In conditions of more stringent
inhibition of energy production such as induced
by iodoacetate alone or in combination with KCN
and decreased oxygen, significant increases of
permeability-surface values for albumin, sucrose,
and sodium were observed, indicating a possible
involvement of ATP in maintaining endothelial cell
permeability. The process regulating the permeability, if any, does not, however, seem to make a large
demand on the ATP supply. The diffusion capacity
of the rete to water was unchanged, as would be
expected if water were to diffuse in large part through
the general surface of the endothelium rather than
through selective pathways used by the other tracers. Electron microscopic examination demon-
Rasio et al
Capillary Effects of Hypoxia and Ischemia
251
AC
VC
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
1
FIGURE 8. Electron photomicrograph of capillary tissues at the end ofperfusion experiments in whichflowwas arrested
for 30 minutes. Stagnant control medium. The endothelial cells of the venous capillaries (VC) show signs of degeneration;
vacuoles protude into the lumen of the capillaries. The interstitial space remains compact. AC, arterial capillaries; J,
intercellular junctions; P, pericytes; f, fenestration. Magnification, x20,000.
strated swelling of the endothelial cells but did not
reveal disruption or widening of the intercellular
junctions (Figure 7); major alterations were detected
in mitochondria, reflecting modifications of the energygenerating machinery, and in endothelial cell membranes that formed vacuoles with myelin-like structures, some of which were shed intraluminally.
Endothelial swelling has also been described in human
muscle cell capillaries as a sequel to prolonged
ischemia.22
The biochemical and physiological factors underlying capillary permeability changes during anoxia
or ischemia or during recirculation of ischemic
areas are only partially known, and their relative
importance is conjectural. One can speculate that
anoxia increases permeability by modifying the
contractility of the cytoskeleton, the shape of the
endothelial cell, and the tightness of the intercellular junction, 2324 through restructure or loss of actin
filaments brought about by low ATP generation.^-^
There is evidence that oxygen-derived free radicals
produced in the mitochondrial electron transport
chain or along the pathways for adenine nucleotide
degradation are toxic27 and are capable of directly
damaging the cell membrane by peroxidation of the
unsaturated fatty acids of its phospholipid component. The localization of xanthine dehydrogenase
and the important content and high turnover of
intracellular adenine nucleotides in the rhicrovascular endothelium of the heart,28 have been thought to
provide a basis for understanding the induction of
lesions by ischemia followed by reperfusion.29
Our second series of experiments was designed to
test the effects of hypoxia and decreased energy
generation during flow stoppage followed by reperfusion on the diffusion capacity of the rete. Stasis
and reperfusion induced extensive damage of the
venous endothelial cells while the arterial endothelial cells were merely injured (Figures 8 and 9).
Striking differences in permeability-surface area
values were observed depending on the composition of the medium bathing the rete capillaries
during the 30-minute period of stasis (Table 3). With
the medium initially well oxygenated and supplemented with glucose and no inhibitors, permeabilitysurface area values for all tracers were similar
252
Circulation Research
Vol 64, No 2, February 1989
vc
J
IS
vttk'}
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
AC
FIGURE 9. Electron photomicrograph of capillary tissues at the end ofperfusion experiments in which flow was arrested
for 30 minutes. Stagnant medium with metabolic poisons. In addition to the damage described above, the interstitial
space (IS) is edematous, and the pericytes (P) are separated from their adjacent endothelial cells. The intercellular
junctions (J) appear normal. AC, arterial capillaries; VC, venous capillaries; f, fenestration. Magnification, X-15,000.
during control and recirculation conditions. The
concurrent important evolution of structural lesions
in the venous capillaries indicates that, as previously suspected, the arterial endothelial cell sheet is
the rate limiting factor in the rete barrier permeability. With a medium containing inhibitors of energy
metabolism, no glucose, and little oxygen during
the period of stasis, permeability-surface area values for albumin, sucrose, and sodium during reperfusion were initially unchanged and then increased
progressively to reach levels three to four times
higher than during the control period. Concurrent
morphological examination detected the presence
of significant patchy interstitial edema, together
with separation of some of the pericytes from
adjacent endothelial cells. There may be a role for
the pericytes in this regulation of capillary permeability. Their cytoplasmic projections contain bundles of filamentous actin and extend to the edges of
intercellular junctions and fenestrations of the endothelium. The speculation is that the width of these
pathways may vary with the degree of tension on
the actin filaments.3031 The initial lack of change in
permeability and subsequent evolution during reperfusion support the idea that a major part of the
alterations observed is due to the events put into
motion with and occurring during the reperfusion.
One possible explanation is that free radical-mediated oxidative injury of the endothelial cells (and,
potentially, pericytes) occurs during reoxygenation
of the hypoxic rete and produces the damage. This
remains to be tested.
It is of interest that in our series of experiments,
whenever solute permeability is increased by
hypoxia, labeled water permeation has not been
modified. During hyperosmolar infusions, we also
observed an increase in solute permeability
in the
absence of change in water permeability.32 We have
also reported that the permeability coefficient for
water in the rete capillary is higher than expected
from the log-log linear relation between permeability and diffusion coefficients for hydrophilic molecules and have inferred that the surface available
for tracer water transfer is larger than that for
solutes. Thus, their respective pathways for diffusion would not be expected to be modified similarly
by experimental maneuvers.
Rasio et al
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
The observation that solute permeability is
increased by hypoxia while the interendothelial
junctions remain intact is intriguing. For large solutes, such as albumin, we have shown in the
rete,32-33 and others have confirmed in the murine
myocardium,34 that there is entry into plasmalemmal vesicles; transport across the endothelium could
take place through these vesicles. For sodium and
sucrose, however, the commonly held view is that
the major path for diffusion is the interendothelial
junction. This space is likely to be filled by a matrix
of fibrous molecules, with associated charge effects.35 Our observation that ionic transport is
reduced in the rete capillaries by comparison with
neutral solutes with similar diffusion coefficients36
is consonant with this idea. Alterations of the
matrix density or charge during hypoxia could
potentially account for changes in small solute
permeability without visible widening of interendothelial junctions.
In conclusion, we have shown that the micro vascular endothelial cells of the rete have an important
energy metabolism mostly accounted for by glycolysis. When ATP generation is inhibited, structural
injury and selective permeability increases are
observed. Stasis with an oxygenated energy-rich
medium damages the venous capillary endothelium,
but the lesions of the arterial capillary endothelium
are not so severe as to alter its permeability. On the
other hand, stasis, together with hypoxia and inhibition of energy generation, results in a progressive
breakdown of the capillary barrier to the passage of
albumin, sucrose, and sodium, which evolves during the reperfusion, while the diffusion capacity for
water remains unchanged.
Acknowledgments
We thank Marie-Paule Dea for her skillful technical assistance and Nancy Firth for typing the
manuscript.
References
1. Lipmann F: Metabolic generation and utilization of phosphate bond energy. Adv Enzymol 1941; 1:99-162
2. Sage JI, van Uitert RL, Duffy TE: Early changes in bloodbrain barrier permeability to small molecules after transient
cerebral ischemia. Stroke 1984;15:46-50
3. Olesen SP: Rapid increase in blood-brain barrier permeability during severe hypoxia and metabolic inhibition. Brain
Res 1986;368:24-29
4. Simionescu M, Simionescu N: Ultrastructure of the microvascular wall: Functional correlations, in Renkin EM, Michel
CC (eds): Handbook of Physiology, Section 2: The Cardiovascular System, Volume IV, Microcirculation. Bethesda,
MD, American Physiological Society, 1984, pp 41-101
5. Shepro D, D'Amore PA: Physiology and biochemistry of the
vascular wall endothelium, in Renkin EM, Michel CC (eds):
Handbook of Physiology, Section 2: The Cardiovascular
System, Volume IV, Microcirculation. Bethesda, MD, American Physiological Society, 1984, pp 103-164
6. Crone C, Levitt DG: Capillary permeability to small solutes,
in Renkin EM, Michel CC (eds): Handbook of Physiology,
Section 2: The Cardiovascular System, Volume IV, Microcirculation. Bethesda, MD, American Physiological Society,
1984, pp 411-466
Capillary Effects of Hypoxia and Ischemia
253
7. Rasio EA, Bendayan M, Goresky CA: Diffusion permeability of an isolated rete mirabile. Circ Res 1977;41:791-798
8. Rasio EA: Glucose metabolism in an isolated blood capillary
preparation. Can J Biochem 1973^1:701-708
9. Bendayan M, Sandborn E, Rasio E: Studies of the capillary
basal lamina. I Ultrastructure of the red body of the eel
swimbladder. Lab Invest 1975;32:757-767
10. Dobrina A, Rossi F: Metabolic properties of freshly isolated
bovine endothelial cells. Biochim Biophys Ada 1983;
762:295-301
11. Kjellstrom BT, Ortenwall P, Risberg B: Comparison of
oxidative metabolism in vitro in endothelial cells from different species and vessels. / Cell Physiol 1987;132:578-580
12. Diana JN, Laughlin MH: Effect of ischemia on capillary
pressure and equivalent pore radius in capillaries of the
isolated dog hind limb. Circ Res 1974;35:77-101
13. Petito CK: Early and late mechanisms of increased vascular
permeability following experimental cerebral infarction. J
Neuropathol Exp Neurol 1979^8:222-234
14. Bland RD, Demling RH, Selinger SL, Staub NC: Effects of
alveolar hypoxia on lung fluid and protein transport in
unanesthetized sheep. Circ Res 1977;40:269-274
15. Perry MA, Shepherd AP, Kvietys PR, Granger DN: Effect
of hypoxia on feline intestinal capillary permeability. Am J
Physiol mS$4S(.Gastrointest Liver Physio[):G272-G216
16. Fjeld NB, Kluge TH, Stokke KT, Skrede S: The effect of
generalized hypoxia upon flow and composition of cardiac
lymph in the dog. Eur J Clin Invest 1976;6:255-259
17. Land is EM, Pappenheimer JR: Exchange of substances
through capillary walls, in Hamilton WF, Dow P (eds):
Handbook of Physiology Section 2: Circulation. Bethesda,
MD, American Physiological Society, 1963, pp 961-1034
18. Granger DM, Sennett M, McEleamey P, Taylor AE: Effect
of local arterial hypotension on cat intestinal capillary permeability. Gastroenterology 1980;79:474-480
19. Persson NH, Eriansson M, Svensjd E, Takolander R,
Bergqvist D: Microvascular permeability changes following
local ischemia to the hamster cheek pouch and its inhibition
by a betaragonist(abstract). Int J Microcirc Clin Exp 1984;
3:488
20. Kinasewitz GT, Groome LJ, Marshall RP, Leslie WK,
Diana JN: Effect of hypoxia on permeability of pulmonary
endothelium of canine visceral pleura. J Appl Physiol 1986;
61:554-560
21. Ullal SR, Kluge TH, Gerbode F: Changes in cardiac lymph
of dogs during and after anoxia. Ann Surg 1972; 175:472-478
22. Gidlof A, Lewis DH, Hammersen F: The effect of prolonged
total ischemia on the ultrastructure of human skeletal muscle
capillaries. Int J Microcirc: Clin Exp 1987;7:67-87
23. Crone C: Modulation of solute permeability in microvascular
endothelium. Fed Proc 1986;45:77-83
24. Shasby DM, Shasby SS, Sullivan JM, Peach MT: Role of
endothelial cytoskeleton in control of endothelial permeability. Circ Res 1982 ;51:657-661
25. Bereiter-Hahn J, Tillmann U, Voth M: Interaction of metabolic inhibitors with actin fibrils. Cell Tissue Res 1984;
238:129-134
26. Tillmann U, Bereiter-Hahn J: Relation of actin fibrils to
energy metabolism of endothelial cells. Cell Tissue Res 1986;
243:579-585
27. Slater TF: Free-radical mechanisms in tissue injury. Biochem J 1984222:1-15
28. Gerlach E, Nees S, Becker BF: The vascular endothelium:
A survey of some newly evolving biochemical and physiological features. Basic Res Cardiol 1985;8O:459-474
29. Schrader J: Mechanisms of ischemic injury in the heart.
Basic Res Cardiol 1985;80(suppl 2): 135-139
30. Sims DE: The pericyte. A review. Tissue Cell 1986;18:153-174
31. Kelly C, D'Amore P, Hechtman HB, Shepro D: Microvascular pericyte contractility in vitro: Comparison with other
cells of the vascular wall. J Cell Biol I987;1O4:483-49O
32. Rasio EA, Bendayan M, Goresky CA: The effects of hypermolarity on the permeability and structure of the capillary of
the isolated rete mirabile of the eel. Circ Res 1981 ;49:661-676
254
Circulation Research
Vol 64, No 2, February 1989
33. Bendayan M: Use of the protein A-gold technique for the
morphological study of vascular permeability. J Hislochem
Cytochem 1980;28:1251-1256
34. Milici AJ, Watrous NE, Stukenbrok H, Palade GE: Transcytosis of albumin in capillary endothelium. J Cell Biol 1987;
105:2603-2612
35. Curry FE: Is the transport of hydrophilic substances across
the capillary wall determined by a network of fibrous molecules? Physiologist 1980,23:90-93
36. Rasio EA, Goresky CA: Passage of ions and dextran molecules across the rete mirabile of the eel. The effects of
charge. Circ Res 1985;56:74-83
KEY WORDS • rete mirabile • capillary function • ischemia
• reperfusion • reduced energy generation • hypoxia
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
Effect of reduced energy metabolism and reperfusion on the permeability and morphology
of the capillaries of an isolated rete mirabile.
E A Rasio, M Bendayan and C A Goresky
Downloaded from http://circres.ahajournals.org/ by guest on July 31, 2017
Circ Res. 1989;64:243-254
doi: 10.1161/01.RES.64.2.243
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1989 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/64/2/243
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/