The role of pinocytic vesicles in the
transport of materials across the
walls of small blood vessels
V. T. Marchesi*
The purpose of this presentation is to focus attention on the pinocytic vesicle system in
endothelial and mesothelial cells and to present some remits of studies on the structure,
function, and biochemical properties of vesicles which might be relevant to the problem of
transport across these cells and their membranes.
S,
1
mall blood vessels can be characterized
by the anatomy of their endothelial lining.
In one type of vessel the endothelium
forms a continuous cell barrier without
any apparent discontinuities or channels
across the cytoplasm. Vessels of this type
are found in cardiac and skeletal muscle,
lung, connective tissue, brain, and retina.
In this paper this type of vessel will be
referred to as "muscle-type" capillaries. A
second type of vessel is found in liver,
kidney, endocrine glands, intestine, choroid and ciliary body, and other sites. The
endothelium which lines capillaries of this
type may have openings or holes across
the cytoplasm of the endothelial cells
(such as are found in liver and glomerular
capillaries); or the cytoplasm of the endothelial cells may be extremely thin in
places which results in an apparent bloodtissue barrier of the thickness of a single
membrane (examples of this type are
From the Department of Pathology, Washington
University, St. Louis, Mo.
Supported by United States Public Health Research Grant No. HE 08293-02 and Training
Grant No. 5T1-GM-897.
'Present address: Department of Cell Biology,
The Rockefeller Institute, New York, N. Y.
found in the intestine, endocrine glands,
and choroid).
The endothelial cells lining the vessels
of the first type (muscle-type) frequently
possess a system of microvesicles—called
pinocytic vesicles—which line the plasma
membranes of these cells. Capillaries in
some tissues such as heart and lung have
large numbers of these vesicles, while those
in the retina and brain have relatively few.
When these vesicles were first described
it was hypothesized that they might serve
as a transport system for the movement of
salts, water, and large molecules across the
cytoplasm of the endothelial cells.1 Since
this suggestion was proposed many studies
have been reported which show that this
vesicle system is associated with the movement of tracer particles under certain experimental conditions.2"4
A vesicle system with similar properties
is found in mesothelial cells which line the
serosal surfaces of the rat mesentery.
These mesothelial cells are more accessible
to experimental manipulation than are the
endothelial cells of capillaries, and they
have been used as a model system for
combined morphologic and biochemical
studies.
1111
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
1112 Marchesi
Investigative Ophthalmology
December 1965
Fig. 1. A segment of capillary endothelium from an isolated rat heart perfused with iron
oxide. The dense iron particles are seen within pinocytic vesicles. Tissue fixed in OsO-i. (From
Jennings, M. A., Marchesi, V. T.3 and Florey, H. W.: Proc. Roy. Soc, Ser. B. 156: 14,
1962.) (xl005O00.)
Fig. 2. Part of. a mesothelial cell lining the serosal surface of the rat mesentery. The surface
membranes form typical pinocytic vesicles similar to those found in capillary endothelium.
Tissue fixed in glutaraldehyde followed by Os(X (x60,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
Pinocytic vesicles 1113
Voltnne 4
Number 6
Observations
Pinocytic vesicles of vascular endoihelium. The evidence that pinocytic vesicles
of muscle-type capillaries serve to transport materials across the endothelial barrier is based on studies of electron-dense
tracer particles with the electron microscope. Tracer particles such as ferritin or
iron oxide were introduced into the circulation of intact animals and, at various
times afterward, tissue was removed and
processed for microscopic study. The electron-dense tracers could be identified in
the bloodstream, and, in time, some particles were also found in the extravascular
tissue spaces. Particles were seen within
pinocytic vesicles of the endothelial cells
but they were not seen in the interspaces
between the endothelial cells. Therefore
it was concluded that the particles which
passed across the endothelial barrier were
transported by way of the vesicle system.2' 3 Further experimental evidence was
provided by similar studies using the isolated, perfused heart preparation.'1 A high
concentration of tracer particles was introduced directly into the arterial circulation of the hearts, and the transport of particles by the pinocytic vesicles was readily
demonstrated (Fig. 1). In this experimental system there was a much greater
movement of tracer particles across the
endothelium than could be achieved in
the intact animal. Even in this case the
particles traversed the endothelium only
by way of the pinocytic vesicles.
A..
Fig. 3. Part of a mesotheh'al cell lining tlie serosal surface of tlie mesentery. Tin's mesentery
was immersed in cold distilled water for two minutes, then fixed immediately in glutaraldehyde followed by OsO*. The arrows point to distorted remnants of pinocytic vesicles. Most
of the vesicles have been pulled out to form linear membranes. (x40,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
7
1114 Marchesi
Pinocytic vesicles of mesoihelium. The
mesothelial cells which line the serosal
surfaces of the mesentery have vesicles
which are the same size (700 to 900 A),
and shape, and they occur in the same distribution as those found in heart capillaries
(Fig. 2). These vesicles are formed by
"unit membranes" which are directly continuous with the linear cell membranes.
Apart from the vesicular configuration the
membranes which make up the vesicles are
indistinguishable from other parts of the
cell membrane when viewed by conventional thin section electron microscopy.
To test the stability and reversibility of
the vesicles and their membranes, strips
of mesentery were incubated in hypotonic
media for varying periods of time. Electron micrographs of mesothelial cells at
various stages of osmotic lysis show that
Investigative Ophthalmology
December 1965
the pinocytic vesicles pull out as the cells
swell, and the vesicle membranes eventually revert to the linear form (Fig. 3).
Transport function of mesothelial vesicles. The same experimental approaches
described earlier for the demonstration
of transport activity by endothelial cells
have been used to show transport activity
in mesothelium.5' 6 Tracer particles introduced into the peritoneal cavities of animals pass across the mesothelium within
pinocytic vesicles.
However, the mesothelium has an additional virtue as an experimental system
in that the transport of tracer materials
can be studied easily under controlled in
vitro conditions. When strips of mesentery
are incubated in vitro in a balanced salt
medium containing a tracer particle such
as ferritin the particles can be identified
Fig. 4. Part of a mesothelial cell lining the serosal surface of the mesentery. This mesentery
was incubated for five minutes in a balanced salt medium containing 0.1 per cent ferritin,
then fixed in glutaraldehyde followed by OsO*. Ferritin molecules have been taken up from
the medium by the mesothelial. cell, transported across the cytoplasm, then released into the
extracellular space. Some ferritin is still within pinocytic vesicles. (x75,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
Volume 4
Number 6
Pinocytic vesicles 1115
free in the connective tissue spaces of the
mesentery within minutes after initial exposure (Fig. 4). By studying preparations
fixed at different intervals after exposure
to the ferritin it was clear that the ferritin
molecules passed across the mesothelial
lining cells by way of the pinocytic vesicles. Ferritin was not seen free in the
cytoplasm of intact mesothelial cells nor
was any detected in the narrow (100 to
150 A) spaces between adjacent mesothelial cells.
Properties of an adenosinetriphosphatase associated with pinocytic vesicles.
Many observers have reported that blood
capillaries of all types have an ATPase
activity localized presumably in their
endothelial linings.8"12 These observations
were based on histochemical techniques
using the light microscope. Recently the
Wachstein-Meisel lead salt method for the
demonstration of ATPase activity13 has
been modified for use with the electron
microscope, and this technique has been
applied to studies of small blood vessels
in different tissues.14"10 One of the most
interesting findings was the localization
of ATPase activity in the pinocytic vesicles in capillaries of the muscle type. The
success of this technique was based on the
use of glutaraldehyde as a fixative17 since
this aldehyde fixative provided preservation of the structure of the tissue which
was adequate for electron microscopic
study while still retaining some of the
enzymatic activity.
Fig. 5 shows the localization of ATPase
activity on the membranes of pinocytic
vesicles of capillary endothelium. The reaction product is localized only on parts
1
*•
Fig. 5. Part of an endothelial cell of a heart capillary demonstrating ATPase activity. The
tissue was fixed in 4 per cent glutaraldehyde for one hour then incubated in a WachsteinMeisel lead salt medium13 for one hour. The tissue was then fixed in O s d . The reaction
product, lead phosphate, appears as a dense precipitate and is confined to the pinocytic
vesicles and to parts of the cell membrane which form vesicles. (From Marchesi, V. T., and
Barrnett, R. J.: j . Cell Biol. 17: 547, 1963, Rockefeller Institute Press.) (x60,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
1116 Marchesi
of the cell membrane which form vesicles;
the linear membrane appears unreactive.
Since this observation was made on tissue
fixed in glutaraldehyde it was not clear
whether the linear membrane was truly
nonreactive or whether the activity on that
portion of the membrane was inhibited by
the fixative.
This problem was resolved by incubating fresh, unfixed mesentery in a modified
histochemical medium for the demonstration of ATPase activity. Figs. 6 and 7 show
clearly that even when the incubation is
performed with unfixed tissue the reaction
product is confined solely to the inner surfaces of the parts of the cell membrane
which form vesicles. This activity was
demonstrated in the vesicles in both fresh-
In vextigative Ophthalmology
December 1965
ly isolated mesentery and in mesentery
which was incubated in vitro for one hour
before the reaction was run.7 This latter
observation suggested that the enzyme or
enzymes responsible for the reaction were
bound to the vesicle membranes. This point
was substantiated when histochemical reactions were carried out on mesothelium
following lysis by osmotic shock. Fig. 7
shows that even when most of the vesicles
were pulled out to form linear membranes
they were still reactive, and the reactive
substances were not dissolved or destroyed
by numerous washes in cold distilled
water. Fig. 7 also shows that the enzymatic
activity is on the membrane itself and is
not dependent upon the vesicular configuration of the membrane.
1
4i
Fig. 6. Part of a mesothelial cell lining the serosal surface of the mesentery. The mesentery
was washed in buffer and incubated for two minutes without prior fixation in a modified
lead salt ATPase medium. The reaction was stopped by fixation in cold glutaraldehyde followed by OsOi. Dense reaction product is localized to the pinocytic vesicles. Areas of the
linear membrane not involved in vesicle formation are not reactive. (x80,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
Volume 4
Number 6
Electron microscopic histochemical studies of glutaraldehyde-fixed tissue were
also earned out on different vascular
beds, and the presence of ATPase activity
within the endothelium correlated with the
vesicle populations of the different vessels.
Thus the endothelial cells lining the capillaries of heart and lung which have
large numbers of vesicles were highly
reactive,14'1!i while the endothelial cells
of capillaries of brain and retina with
few vesicles were essentially unreac*ive.10-1S
An ATPase activity has been associated
with the movements of sodium and potassium across membranes of other cell types
such as brain, kidney, and red cells.10"*3
This activity is stimulated by magnesium,
sodium, and potassium and specifically inhibited by the cardiac glycoside ouabain.
Thus the demonstration of ATPase activity
in the pinocytic vesicles was significant
Pinocytic vesicles 1117
since it raised the possibility that the vesicle membrane might somehow be involved
in the transport of small electrolytes such
as sodium and potassium as well as large
molecules.
In order to determine whether the
ATPase associated with the pinocytic vesicles was also activated by Mg, Na, and
K and inhibited by ouabain, quantitative
assays of ATP hydrolysis were carried out
on strips of osmotically lysed mesentery.7
Since the ATPase activity of the vesicles
remained firmly bound to the mesothelial
membranes after lysis—and they seemed
to be the only parts of the cell showing
activity—it was assumed that. the breakdown of ATP reflected the activity of these
structures.
The results of these experiments showed
that the ATPase of the mesothelial membranes was activated by magnesium alone,
and was specific for nucleotide substrates
Fig. 7. Part of a mesothelial cell from a strip of mesentery which was immersed in cold
distilled water for two minutes, then incubated unfixed in an ATPase reaction medium for
two minutes. Most of the pinocytic vesicles have pulled out to form linear membranes (as
shown in Fig. 3). Reaction product is now localized on these linear membranes. (x603000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
1118 Marchesi
with a pH optimum in the neutral range.
However, there was no indication that the
ATPase activity was stimulated by sodium
and potassium nor was it sensitive to
ouabain. These findings suggested that the
ATPase on the vesicles was not similar
to the Na-K ATPase of other membranes
nor was it due to a nonspecific acid or
alkaline phosphatase.
In order to explore further some of the
chemical properties of this membranebound ATPase and to verify the finding
made on the lysed mesentery strips,
methods were devised for the isolation
of mesothelial cells free from the connective tissue, blood vessels, adipose cells,
and assorted cell types of the whole mesentery. Mesenteries were removed from rats
following perfusion with saline and incubated in a buffered medium with collagenase (Worthington) which separated the
mesothelial cells from the other cell types.
The mesothelial cells were preferentially
Investigative Ophthalmology
December 1965
harvested on a discontinuous sucrose
gradient.7 Fig. 8 shows the appearance of
isolated mesothelial cells following this
treatment. These cells were then disrupted
by ultras onication and a "membrane" fraction was prepared by differential centrifugation (Fig. 9).
Experiments with aliquots of washed
and dialyzed membrane fractions gave
ATP hydrolysis rates of 20 to 40 micro-
Table I. Properties of membrane bound
adenosinetriphosphatase
Activated by Mg2+ or Ca2+ ions
Not stimulated by Na+ or K+, or inhibited by
ouabain
pH optimum in 6-8 range
Active toward nucleotide substrates (ATP, ADP,
AMP)
Not active toward glycerophosphate, hexosediphosphate, or glucose-6-phosphate
Fig. 8. Phase contrast photomicrograph of a preparation of isolated mesothelial cells. These
cells were freed from the connective tissue and other cellular elements of die mesentery by
the separation of a collagenase-treated preparation on a discontinuous sucrose gradient.7 The
mesothelial cells remain attached to one another in sheets, and this provides the basis for
their separation from the otiier cells.
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
Volume 4
Number 6
Pinocytic vesicles 1119
moles of inorganic phosphate released per
milligram of protein per hour. This activity was stimulated by magnesium and calcium independently, and the activity was
unaffected by sodium, potassium, or ouabain. These results are summarized in
Table I.
Discussion
The results presented here show that
the pinocytic vesicles of endothelial and
mesothelial cells are specific adaptations
of the cell membrane which can function
in the transport of large molecules. The
portions of the cell membrane which form
the vesicles appear to have an enzyme sys-
tem (ATPase) not present on the linear
membrane. This ATPase is not a nonspecific acid or alkaline phosphatase nor is it
similar to the sodiiun-potassium activated
ATPase found on other cell membranes.19"
23
Rather, some of the properties of this
ATPase are similar to those of myosin
ATPase of muscle (activation by magnesium and calcium, insensitivity to ouabain2'1).
The system of pinocytic vesicles as a
possible transport mechanism applies to
other epithelial barriers as well as the vascular endothelium and mesothelium described above. One stage in the transport
of tracer particles across the endothelium
I
*
Fig. 9. Membrane fraction prepared from isolated mesothelial cells. Mesothelial cells were
disrupted by ultrasonication and a membrane fraction prepared by differential cenbifugation.
This fraction is made up largely of "unit membrane" structures. (x90,000.)
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
1120 Marchesi
of the cornea has been shown to involve
vesicles,25'2G and recent observations on
the movement of tracer particles across
the endothelial walls of the canal of
Schlemm27 also implicate a vesicle mechanism.
Several important questions and problems relating the pinocytic vesicle system
to transport across epithelial barriers remain to be resolved. There is as yet no
clear indication as to the physiological
significance of the vesicle system in vivo
nor is there any simple way to investigate
this point. Even if the vesicle system does
operate in vivo, there would be no reason
to assume that it is the only route—or
even the principal route—for the transepithelial passage of materials. Thus the
demonstration of a transport fimction by
vesicles does not rule out the possibilities
of movement via intercellular or other
channels.
To assess the possible contribution of
the vesicle system to transport across
endothelial and mesothelial barriers it
might be profitable to find out whether
vesicle formation or function can be modified experimentally and what effect this
would have on vascular or mesothelial
permeability. As one approach to this problem one might explore the hypothesis that
the ATPase associated with vesicle membranes plays a role in the formation of the
vesicle configuration perhaps in the formation and movement of membrane or its
energy utilization analogous in some way
to the ATPase involved in muscle contraction.
I am grateful to Mrs. Beverly Young for technical assistance, to Mrs. Bertha McClure for the
photographic work, and to Dr. S. L. Marchesi for
helpful criticism of the manuscript.
Investigative Ophthalmology
December 1965
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
REFERENCES
1. Palade, G. E.: The fine structure of blood
capillaries, J. Appl. Physics 24: 1424, 1953.
2. Wissig, S. L.: An election microscope study
of the permeability of capillaries in muscle,
Anat. Rec. 130: 467, 1958.
3. Palade, G. E., and Bruns, R. R.: Structure
and function in normal muscle capillaries,
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
16.
in Small blood vessel involvement in diabetes mellitus, Washington, D. C., 1964,
The American Institute of Biological Sciences,
p. 39.
Jennings, M. A., Marchesi, V. T., and Florey,
H. W.: The transport of particles across the
walls of small blood vessels, Proc. Roy. Soc,
Ser. B 156: 14, 1962.
Odor, L.: Uptake and transfer of participate
matter from the peritoneal cavity of the rat,
J. Biophys. & Biochem. Cytol. 2 (Suppl.):
105, 1956.
Staubesand, J.: Zur histophysiologie des
Herzbeutels. II. Electronmikroskopische Untersuchungen uber die Passage von Metallsolen durch mesotheliale Membranen, Ztschr.
Zellforsch. 58: 915, 1963.
Marchesi, V. T.: To be published.
Padykula, H. A., and Herman, E.: The specificity of the histochemical method for
ATPase, J. Histochem. & Cytochem. 3: 170,
1955.
Spater, H. W., Novikoff, A. B., and Masek,
B.: ATPase activity on the cell membranes
of kidney tubule cells, J. Biophys. & Biochem.
Cytol. 4: 765, 1958.
Wachstein, M., and Meisel, E.: The histochemical demonstration of secretory capillaries in the pancreas with the aid of substrate-specific phosphatases, J. Biophys. &
Biochem. Cytol. 6: 119, 1959.
Freiman, D. G., and Kaplan, N.: Studies
on the histochemical differentiation of enzymes hydrolyzing ATP, J. Histochem. &
Cytochem. 8: 159, 1960.
Novikoff, A. B., Drucker, J., Shin, W-Y.,
and Goldfischer, S.: Further studies of the
apparent ATPase activity of cell membranes
in formal-calcium fixed tissues, J. Histochem.
& Cytochem. 9: 434, 1961.
Wachstein, M., and Meisel, E.: Histochemistry of hepatic phosphatases at a physiologic
pH, Am. J. Clin. Path. 27: 13, 1957.
Marchesi, V. T., and Barmett, R. J.: The
demonstration of enzymatic activity in pinocytic vesicles of blood capillaries with the
electron microscope, J. Cell. Biol. 17: 547,
1963.
Marchesi, V. T., and Barmett, R. J.: The
localization of nucleoside phosphatase activity in different types of small blood vessels,
J. Ultrastruct. Res. 10: 103, 1964.
Marchesi, V. T., Sears, M. L., and Barmett,
R. J.: Electron microscopic studies of nucleoside phosphatase activity in blood vessels and glia of the retina, INVEST. OPHTH.
3: 1, 1964.
17. Sabatini, D. D., Bensch, K., and Barmett,
R. J.: Cytochemistry and electron micros-
Volume 4
Number 6
18.
19.
20.
21.
22.
copy. The preservation of cellular ultrastructure and enzymatic activity by aldehyde fixation, J. Cell Biol. 17: 19, 1963.
Torack, R. M., and Barrnett, R. J.: The fine
structural localization of nucleoside phosphatase activity in the blood-brain barrier,
J. Neuropath. & Exper. Neurol. 23: 46, 1964.
Skou, J.: Preparation from mammalian brain
and kidney of the enzyme system involved
in active transport of Na+ and K+, Biochim.
et biophys. acta 58: 314, 1962.
Whittam, R.: The asymmetrical stimulation
of a membrane adenosine triphosphatase in
relation to active cation transport, Biochem.
J. 84: 110, 1962.
Bonting, S. L., Caravaggio, L. L., and Hawkins, N. M.: Studies on sodium-potassiumactivated adenosinetriphosphatase. IV. Correlation with cation transport sensitive to
cardiac glycosides, Arch. Biochem. & Biophys.
98: 413, 1962.
Kinsolving, C. R., Post, R. L., and Beaver,
D. L.: Sodium plus potassium transport
adenosine triphosphatase activity in kidney,
J. Cell. & Comp. Physiol. 62: 85, 1963.
Downloaded From: http://iovs.arvojournals.org/ on 06/19/2017
Pinocytic vesicles 1121
23. Wallach, D. F. H., and Ullrey, D.: Studies
on the surface and cytoplasmic membranes
of Ehrlich ascites carcinoma cells. II. Alkalication-activated adenosine triphosphate hydrolysis in a microsomal membrane fraction,
Biochem. et biophys. acta 88: 620, 1964.
24. Glynn, I. M.: The action of cardiac glycosides on ion movements, Pharmacol. Rev. 16:
381, 1964.
25. Kaye, G. I., and Pappas, G. D.: Studies on
the cornea. I. The fine structure of the rabbit
cornea and the uptake and transport of colloidal particles by the cornea in vivo, J. Cell.
Biol. 12: 457, L962.
26. Kaye, G. I.: Pappas, G. D., Donn, A., and
Mallett, N.: Studies on the cornea. II. The
uptake and transport of colloidal particles by
the living rabbit cornea in vitro, J. Cell Biol.
12: 481, 1962.
27. Fine, B. S.: Observations on the drainage
angle in man and rhesus monkey: A concept of the pathogenesis of chronic simple
glaucoma, INVEST. OPHTH. 3: 609,
1964.