From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
Opposite Sorting of Tissue Factor in Human Umbilical Vein Endothelial
Cells and Madin-Darby Canine Kidney Epithelial Cells
By Eric Camerer, Serena Pringle, Anne Helen Skartlien, Merete Wiiger, Kristian Prydz, Anne-Brit Kolst0,
and Hans Prydz
Tissue factor (TFI is a 48-kD transmembrane glycoprotein
that triggers the extrinsic pathway of blood coagulation by
interacting with the plasma coagulation factorVI1 (WII).TF
is also a true receptor in that a cellular signal isgenerated
when activated Wll (Wlla) binds to TF. For both of these
functions, the cellular surface distribution of TF
important,
is
since Wll is primarily available on the apical side of vascular
endothelialcellsandon
the basolateralside of epithelial
cells liningthe internal and externalsurfaces. We showthat
in endothelial cells, TF(both antigen and procoagulant activity) is sortedto the apical surface, whereasin wild-type and
stably transfectedMadin-Darbycanine kidneyepithelialcells
(MDCK). whichform tight junctions and express TFconstitutively, TF antigen is on
the basolateral surface.No significant
clotting activity is detectable on this surface. Truncated TF
(cytoplasmic tail residues 246 to 263 deleted) is sorted as
wild-type in MDCK cells.
0 1996 by The American Societyof Hematology.
T
from Nycomed (Oslo, Norway). The serine protease chromogenic
substrate S-2288 was from Chromogenix (Molndal, Sweden). The
neutralizing murine antihuman monoclonal antibodies werepurchased from American Diagnostica Inc (Greenwich, CT) (4504and
4507) or were produced and kindly donated by Dr S . Carson (htfl).13
FITC-conjugated rabbit anti-mouse Ig F(ab’), was from DAKO
(Glostrup, Denmark). The TF ELISA kit was from American Diagnostica. Iz5I, PD-IO columns, streptavidin-biotin-HRP, and’ECL reagents were from Amersham (Amersham, UK). Sulpho-NHS-biotin
and Iodo-beads were from Pierce (Rockford, IL). Sheep anti-mouse
Ig-coated magnetic beads were from Dynal (Oslo, Norway). The
mammalian expression vector pcDNA3 was from Invitrogen (San
Diego, CA). Soluene and scintillation fluid were fromPackard
(Groningen, The Netherlands). Human TF cDNA was kindly provided by Dr J.H. Morrissey.
ISSUE FACTOR (TF) is a 48-kDtransmembrane glycoprotein that triggers the extrinsic pathway of blood
coagulation by interaction with plasma coagulation factor
VI1 (FVII).“4 The membrane-bound complex of TF and activated FVII (FVIIa) activates zymogen factors IX and X,
which in turn initiate a series of reactions culminating in the
deposition of fibrin. TF, which is not normally expressed in
tissues that are in direct contact with circulating blood,5” is
thought to form a protective lining around tissues and blood
vessels exposed to the plasma factors only in the event of
tissue damage. However, in monocytes/macrophages* and
endothelial cells: TF synthesis can be induced by endotoxins, inflammatory cytokines, and a number of other agents
and conditions, both physiologic and nonphysiologic.”
To understand the (patho)physiologic role of inducible
endothelial TF expression both as a procoagulant trigger and
as a receptor,’ it is important to know whether TF is exposed
to circulating bloodand thus to its ligand, WIT. Mature
human TF consists of 263 amino acids organized into three
domains: a 219-amino acid N-terminal extracellular domain
organized into two modules, a 23-amino acid transmembrane domain, and a 21 -amino acid cytoplasmic domain.”,’*
To investigate if any sorting signals were present in the TF
cytoplasmic domain, we prepared a construct coding for a
cytoplasmic tail-truncated TF molecule. We report here
studies on the polar distribution of wild-type TF antigen in
human umbilical vein endothelial cells (HUVEC) and of
wild-type and truncated TF antigen and activity in MadinDarby canine kidney epithelial cells (MDCK).
MATERiALS AND METHODS
Materials
Cell Culture media and supplements were obtained from BioWhittaker (Walkerville, MD), except fetal calf serum (FCS), which
was from Biological Industries (Kibbutz Beth Haemek, Israel). Tissue culture plastics (including Anapore and Transwell membrane
inserts) was obtained from Nunc (Roskilde, Denmark) and Costar
(Cambridge, MA). TF was isolated from human brain essentially as
described by Hjort.’ Purified factor X (FX), bovine serum albumin
(BSA), Hoechst 33258, HEPES, propidium iodide, Brefeldin A
(BFA), and Geneticin (G418) were obtained from Sigma (St Louis,
MO). Recombinant human FVIIa, interleukin-10, heparin, and antiTFPI were kindly donated by Professor U. Hedner, Dr M. Ezban,
and Dr T. Jmgensen, Novo-Nordisk (Bagsvcrd, Denmark). Annexin
V purified from human placenta was from Alexis Corp (Laufelfinger,
Switzerland). Dr W.L. van Heerde, Institut de Pathologie Cellulaire,
HBpital de Bicstre (Pans, France), kindly donated the polyclonal
antibody to Annexin V. The FXa chromogenic substrate, FXa-l, was
Blood, Vol 88, No 4
(August 15). 1996: pp
1339-1349
Vector Construction
To see if the cytoplasmic domain of the TF molecule contained
any transport signal, we constructed hTFI.245, a recombinant TF
molecule lacking the cytoplasmic domain, with the exception of the
three membrane-proximal amino acids for membrane anchoring by
cysteine-palmitatektearate thioester binding.I4
The hTF,.245construct was designed using an overlap extension
method of splicing two otherwise noncompatible DNA fragments.
The primers used were p1 GACAGCCAACAATTCAGAGTTIT
(forward), p2 AGTGCTTCCTTTAACACTTGTGTAGAGATATAGCCAGGA (reverse), p3 TCTACACAAGTGITAAAGGAAGCACTGTTGGAGC (forward), and p4 CTAGAAATGCACCCAATTTCCT (reverse), all given 5’-3’. Two internal primers, p2 and
p3, both pointing away from the sequence to be deleted, were placed
just outside the sequence coding for the TF cytoplasmic domain.
Each had a 13-bp tail (underlined) complementary to the sequence
on the other side of the region to be deleted. Individual amplification
with pl + p2 and p3 + p4, isolation from excess primer, pooling,
denaturation, annealing of the 26-bp overlap created by the two 13bp overhangs, and a final fill-in polymerization yielded a fragment
From the Biotechnology Centre of Oslo and the Department of
Biochemistry, University of Oslo, Oslo, Norway.
Submitted January 29, 1996; accepted April l I , 1996.
Supported by grants from The Research Council of Norway and
the Norwegian Council of Cardiovascular Diseases.
Address reprint requests to Hans Prydz, MD, PhD, The Biotechnology Centre of Oslo, Gaustadallhen 21, N-0371 Oslo, Norway.
The publication costs of this article weredefrayed in part by page
charge payment. This article must therefore be hereby marked
“advertisement” in accordance with 18 U.S.C. section I734 solely to
indicate this fact.
0 1996 by The American Society of Hematology.
0006-4971/96/8804-0022$3.00/0
1339
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
1340
coding for the region up to cysteine 245 dxectly followed by a stop
codon and the 3‘ untranslated region of the cDNA. This fragment
was amplified with primers 1 + 4, digested with NcoI and HpaI,
and subcloned into the NcoI and HpaI sites of wild-type human TF.
The entire PCR-amplified sequence and the splice junctions were
verified by restriction digests and sequencing.
This truncated TF construct (hTF1.245)and wild-type human TF
cDNA (hTF,.263)were cloned into the mammalian expression vector
pcDNA3. This vector contains the human cytomegalovirus promoter
for high-level expression and a neomycin gene for selection of stable
transfectants.
Transfection
All transfections were made using standard calcium phosphate
coprecipitation procedures with glycerol shock. Clones with stably
integrated constructs were selected with 500 pg/mL (3418. Selection
of transfectants for total TF expression was made by mixing cell
homogenates (50 pL) with human plasma (50 pL) and 30-mmoVL
Ca” (50.pL) and recording clotting time as previously described.”
Cell Preparation and Culture
MDCK strain 1 and COS-l cells were maintained in Dulbecco’s
modified Eagle’s medium (DMEM) supplemented with 5% inactivated fetal calf serum (FCS), L-glutamine, and antibiotics. Stably
transfected cells were maintained in the above medium supplemented
with 100 pglmL G418. For polarity studies, MDCK cells were plated
either at low density (70,000 cells/cm2 membrane) to obtain tight
junctions 4 to 5 days after plating or at high density (300,000 cells/
cm2 membrane) to obtain tight junctions 2 to 3 days after plating,
in tissue culture-treated filter chambers 1.2 or 2.45 cm in diameter
and with 0.45 or 3 pm pore size (Transwell; Costar). The plating
density did not affect TF distribution. The medium was changed
every day, and the experiments were performed 1 to 2 days after
the formation of tight junctions. MDCK cells of high passage number
were not used, due to their reduced ability to form tight junctions.
Tight junction formation was assessed by resistance measurements
across the membrane using a Millipore Ohmmeter. A steep increase
in electrical resistance accompanied the formation of junctional complexes. Leakage controls were also made using radioactive inulin.
There was no detectable leakage of FXa. Leakage of ‘2sI-labeled
antibody was less than 0.5% after 1.5 hours’ incubation on ice.
Where indicated, cells were incubated with 5 pg/mL BFA in
growth medium for 6 or 24 hours and during the subsequent experiment.
HUVEC were isolated and cultured, and TF synthesis was induced
with 1 0 0 U/mL IL-10 as previously described.16 The cells were
grown on transparent permeable membranes to allow ligbt-microscopic observation. Leakage of 1251-labeled
antibody across the tight
cell layers was, on average, 2% over a 1.5-hour incubation period
on ice. Other tests to confirm the integrity of the tight monolayer
were performed as recently described.I6 DNA level was measured
by Hoechst 33258 fluorometry. Immunofluorescence on cover slips
was performed using subconfluent cell layers.
Determination of TF ActiviQ
Total TF activity in cell homogenates. Total TF activity in frozen, thawed, and homogenized cells was measured using a standard
clotting assay with human plasma and 30 mmol/L CaCl2.I5TF activity was determined from a standard curve made with a human brain
TF preparation’that clotted human plasma in 14 to 15 seconds.
This was taken to represent 100 U TF (UTF)/mL. One unit TF
corresponded to 1.5 ng TF as determined by the IMUBIND TF
ELISA kit. The values are expressed as UTF/mg cell protein.”
Surface TF activity determined by the Xa chromogenic substrate
assay was related to total TF activity determined by the clotting
CAMERER ET AL
assay by reference to a common standard curve made with the human
brain TF preparation.
FX activating activity of TFNIIa onthecell s u ~ a c e . The method
used was based on the activation of FX by the TF-VIIa-Ca2+membrane complex. The reaction between FXa and its chromogenic peptide substrate was monitored by the increase in OD4o5over 20 minutes. The cells were washed carefully thrice and incubated with BSA
( 5 mg/mL) in veronal-buffered saline (VBS), pH 7.4, for 30 minutes
at 37°C with shaking at 100 rpm, in the presence or absence of a
neutralizing monoclonal antibody (American Diagnostica anti-TF
antibody no. 4507) as appropriate. This antibody neutralized human
TF, but didnot alter the endogenous canine TF activity inuntransfected MDCK cells. CaC12was then added to a final concentration of 4 mmol/L. The compartments to be tested for TF expression
were supplied with FVIIa (10 U/mL) and FX (0.5 U/mL), and the
culture trays were again incubated with shalung at 100 rpmand
37°C. At 5 , 10, 15, and 20 minutes, samples of 20 p L were taken
from each compartment and added to a microtiter plate containing
150 pL VBS, pH 7.4, with 10 mmol/L EDTA per well to stop the
reaction. After thorough mixing, 20 pL (80 nmol) FXa-l substrate
was added to each sample. The ODas was monitored in a Titertek
Multiscan at 10-minute intervals for 60 minutes, and AODm5was
calculated. A linear increase in OD,os over this period confirmed
sufficient amounts of substrate. FVIIa and FX were also titrated to
give linear curves within the time span of the assay. Controls with
human brain TF demonstrated that the system was linear with increasing TF concentration under the established conditions. The first
incubation (with cells, FVIIa, and FX) was standardized to I O minutes, andthe second (with chromogenic substrate) to 20 minutes.
The linearity of both reactions was controlled in each experiment.
Under these conditions, a AODa5 value of 1.O corresponded to 0.7 I
UTF ( I .I ng TF) andthe three standard curves (TF antigen, TF
activity, and FX activating activity) were colinear.
Amidolytic activity of TFNIIa onthe cell surface. The cell surface activity of the TFNIIa complex toward the peptide substrate
S-2288 was determined essentially as described by Le et al.’* Cell
monolayers were incubated with 20 nmol/L VIIa in HEPES-buffered
saline ([HBS] 136 mmol/L NaCl, 5 mmol/L CaCI,, 5 mmollL KC1,
1.2 mmol/L MgCIZ, 11 mmol/L Bacto-dextrose, 10 mmol/L HEPES,
and 1 % BSA, pH 7.4) for 1 hour at37°Cto saturate TF specific
binding. Following washing, the chromogenic substrate S-2288 was
added to a final concentration of 0.5 mmoliL in HBS. After incubation at 37°C for 1 hour, 200-pL aliquots were transferred to a 96well microtiter plate and absorbance at 405 nm was measured. Controls were cells not receiving VIIa and cells preincubated with a
neutralizing anti-hTF monoclonal antibody (4504). The values are
expressed as AOD,,/hour.mg cell protein.” Comparisons of this
peptidolytic activity of TF/rFVIIa complexes on the cell surface to
complexes formed using isolated brain TF/rFVIIa showed that with
otherwise equal TF activities and antigen levels, isolated human
brain TF demonstrated threefold to fourfold lower specific activity
toward the chromogenic substrate. The cause remains unexplained.
Immunoj3uorescence Microscopy
For staining of TF on IL-10-induced HUVEC, cells on cover
slips were fixed in absolute ethanol, incubated for 1 hour at 37°C
with primary antibody (4 pg/mL htfl in PBS with 1% BSA) and
for l hour at 37°C with the rabbit antimouse FITC-conjugated antibody (25 pg/mL in PBS). Propidium iodide (100 ng/mL) was added
during the last 10 minutes of the last incubation. Cells were rinsed
four to five times in PBS between each step. A final rinse was made
in PBS, pH 8. Controls of unperturbed HUVEC were included, with
no evidence of nonspecific binding.
For detection of TFPI-1 in wild-type MDCK cells, confluent cells
were rinsed three times in serum-free medium at 37°C fixed in 4%
paraformaldehyde at 37°C for 30 minutes, and incubated with a
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
1341
SORTING OF TISSUE FACTOR
polyclonal goat antiserum to human TFPI-l at a 1:500 dilution in
PBS with 0.1% BSA on either the apical or basolateral side for 1
hourat37°C. The cells were rinsed in PBS at 37°C.and FITCconjugated antigoat antiserum at 1:50 in PBSlBSA was applied for
30 to 45 minutes at 37°C. Staining with an irrelevant polyclonal
antibody (anti-von Willebrand factor) was used as a control. Other
controls were filters without cells incubated with medium and filters
with cells stained with the second antibody only.
To detect endogenous or binding of exogenous Annexin V, the
following procedures were used at 4°C. Transfected or untransfected
MDCK cells and HUVEC (with or without induction of TF synthesis) were first washed thrice in HEPES buffer with 5 mmoVL CaC12
to conserve bound Annexin V or with HEPES buffer containing 1
mmol/L EDTA to remove it. (HEPES buffer consists of 25 mmoU
L HEPES, 135 mmol/L NaCI, 5 mmoVL KCI, 0.01% wt/vol glucose,
and 0.3% wt/vol BSA, pH 7.4.) The cells were then incubated with
Annexin V or FITC-conjugated Annexin V both at a final concentration of 142 nmol/L, or a polyclonal rabbit antibody to Annexin V
as appropriate, for 1 hour at 4°C. Inthe latter case, a FITC-conjugated
goat antirabbit antiserum was used for detection. An irrelevant polyclonal antibody (raised in rabbit against human von Willebrand factor) that did not stain MDCK cells served as control.
'2'I-Labeling and Cell Surface Binding of FVIIa and AntihTF Monoclonal Antibodies
Factor VIIa and two different murine monoclonal antibodies to
TF (htfl or American Diagnostica no. 4504) were labeled with 0.5
mCi 'zsI/lOO pg protein using Iodo-beads, obtaining specific activities of 2 to 4 X lo6 c p d p g . Free iodine was removed by running
labeled proteins through a PD 10 column.
Cells were washed in ice-cold Tris-buffered saline, pH 7.5, containing l mmol/L calcium chloride and0.5 mmol/L magnesium
chloride (TBS+). The amount of monoclonal antibody necessary to
saturate the TF-specific binding sites was titrated in preliminary
experiments. The cells were then incubated with labeled monoclonal
antibody (2.5 pg/mL), saturating the TF-specific binding in TBS+
(RPM1 for HUVEC) + 0.5% BSA for 1.5 hours on ice. The cells
were washed extensively in TBS+, and the filters were cut out
and immersed in Soluene before addition of scintillation fluid and
counting in a Packard Tri-Carb Liquid Scintillation Analyzer. The
number of binding sites were determined by relating the filter count
(cpm) (background subtracted) to the specific activity of the antibody
or FVIIa (cpdmolecule) after correction for isotope decay.
Nonspecific binding of FVIIa to the cell surfaces was controlled
in experiments where 'ZSI-FVIIawas allowed to bind in thepresence
or absence of a 50-fold molar excess of cold FVIIa. The difference
in total 12sI-FVIIabound was taken to represent the specific binding.
The surface to total antigen ratio in MDCK cells was determined
by comparison of apical and basolateral antibody binding to intact
monolayers on Transwells to binding to monolayers on Transwells
opened by 30 minutes' exposure to 100% ethanol on ice. Binding
was as above, except that 5 pg/mL rather than 2.5 pg/mL '251-labeled
antibody was used.
Cell Surface Biotinylation
Biotinylation of apical and basolateral surfaces was performed
essentially as described by Sargiacomo et al.'9 All steps of the procedure prior to sample boiling were performed on ice. Cells were
washedwith PBS, pH7.4, containing 1 mmoVL calcium and 0.5
mmoVL magnesium (PBS+). They were then incubated twice with
0.5 mg/mL freshly made sulpho-NHS-biotin in PBS+ for 20 minutes, in the apical or basolateral compartment as appropriate. The
cells were washed, and excess free biotin was removed by incubating
for 10 minutes in PBS+ with 50 mmol/L m C 1 (pH 7.1). After
extensive washing, the filters were cut out, and cells were lysed in
1 mL lysis buffer (150 mmoVLNaC1, 20 mmol/L Tris, pH 8, 10
mmol/L EDTA, 1% Triton X-100,0.2% BSA, 1 mmol/L PMSF, 10
pg/mL aprotinin, 10 pg/mL leupeptin, 10 pg/mL antipain, 10 pgl
mL chymostatin, and 10 pg/mL phosphoramidon) for 1 to 2 hours.
Samples were precleared with Dynabeads (M-280 coated with sheep
antimouse Ig) for 1 to 2 hours. Beads were discarded and the lysates
incubated for 2 hours with an anti-hTF monoclonal antibody (htfl).
The TF/antibody complexes were then immunoprecipitated for 2
hours with Dynabeads, washed thoroughly in lysis buffer, boiled in
a reducing sample buffer, andrun on 12% SDS-PAGE beside a
biotinylated molecular weight standard. Proteins were then blotted
onto Immobilon-P nitrocellulose (Millipore). Nonspecific binding
was blocked with 3%dry milk, and biotinylated proteins were visualized by binding of streptavidin-biotin-HRP and exposure toEnhanced ChemiLuminescence reagents (ECL; Amersham). The relative intensities of the bands were determined by scanning different
exposures in a Shimadzu dual-wavelength flying-spot scanner.
RESULTS
Using the Transwell culture system described, we have
recently dernonstratedl6 that TF activity expressed in HUVEC induced with E-lp is highly polarized to the apical
(luminal) surface of the cells. To obtain this apical expression, the endothelial cells must be grown beyond confluence
on a permeable support, to a state where junctional complexes are formed and reduced intercellular passage of macromolecules is evident. For comparison, we also used
MDCK strain 1, an epithelial cell line that forms stable and
well-characterized tight junctions. Transfecting these cells
with human TF constructs allowed studies of the sorting of
both canine and human TF.
TF Activity in Transient Assays and Stable Clones
Constructs coding for wild-type (hTF.263)and truncated
(hW,.245)human TF were transiently transfected into a monkeykidney fibroblast-like cell line (COS-l) that does not
normally express TF. In total cell homogenates (Fig 1A) the
truncated clone yielded about 40% higher activity than the
wild-type clone, confirming that the cytoplasmic domain of
TF is not needed for formation of an active complex with
FVII, Ca2', and phospholipids.20
MDCK strain 1 cells were transfected with the same constructs and selected for stable integration with G418. TF
activity was tested by coagulation of homogenized cells with
human plasma and related to total cell protein (Fig 1B).
MDCK cells expressed significant background levels of canine W, which reacted with human FVII. The clone
transfected with the hTF1-245
construct expressed more total
TF activity than that transfected with the hnI.263 construct.
When these MDCK clones were grown to confluence in
normal wells and tested for surface activity using a chromogenic factor Xa-dependent peptide assay, the ratios between
the surface and the total activity of each clone (tested by the
clotting assay) were similar (Fig 1B and C), indicating that
the truncated TF reached the cell surface as efficiently as
the wild-type.
In agreement with studies in other cell lines,18,21,22
we
found that the activity of homogenized cells (Fig 1B) was
20- to 30-fold higher than that of intact cells (Fig 1C).
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
CAMERER ET AL
1342
150
150
100
100
50
50
0
0
I
hTFl-263
I
hTF1-245
B
-c 2 5
$20
25
20
15
15
10
10
E
=
5
5
0
0
hlF1-263
hTF1-245
MDCK
C
250
m
150
100
50
..
8
0
0
hTF1-2&3
hTF1-245
MDCK
junctions were confirmed by measurements of transepithelial
electrical resistance, the TF activity available on the surface
of the cell monolayers was tested on either side using the
Xa chromogenic substrate assay.
To separate endogenous canine TF activity from the human TF activity, we first performed experiments where the
cells were preincubated with a neutralizing murine anti-human TF monoclonal antibody. This brought the activity of
the transfected clones down to the level of nontransfected
cells (Fig 2A). The antibody had no effect on the background
canine TF activity. To exclude this background canine TF
activity when measuring human TF activity due to transfection in subsequent experiments, the apical or basolateral activity of cells incubated with human TF neutralizing antibody
was subtracted from the apical or basolateral activity of the
transfected cells a s appropriate. As discussed above, only a
limited fraction of the total cellular TF activity of transfected
MDCK monolayers was available on the surface. The major
part (79% for hTFl.263or 81% for hTFI.245)
of this surface
activity was found to be apical (Fig 2B). The total surface
available activity of the cells transfected with hTFl.2,5relative to that of cells transfected with hTFl.263was lower in
tight layers (Fig 2R) than in subconfluent layers (Fig IC).
This observation was consistent in all experiments and in
both hTFI.245cell clones investigated, andthe gradual decrease of hTFI.2A5
surface activity relative to hTFI.26 surface
activity could be followedfrom subconfluence through to
the formation of tight junctions.
The total surface available wild-type human TF activity of
subconfluent cell layers in normalwells (0.27 U hTFl.26,,/mg
cell protein; Fig IC) was similar to thatof cells in Transwells
(0.24 UhTFl.26,Jmgprotein; Fig 2B), suggesting that surface
TF was equally available to our assay system underboth
culturing conditions. Furthermore, control experiments in
Transwells with 3-pm pore size gave essentially the same
activity distribution as did Transwells with 0.4-p.m pores.
To determine if the apical predominance of TF activity on
MDCK cells was limited to the activation of macromolecular
substrate, we also looked at the direct hydrolysis of the peptide substrate S-2288 by TFNIIa (the amidolytic activity)
on apical and basolateral surfaces. Figure 2C shows the sum
of S-2288 hydrolysis by human and canine TFNlIa on the
apical and on the basolateral surface of MDCK hTFl.26,,cells.
The figure demonstrates that the amidolytic activity of the
complex was greater on the apical surface than on the basolateral, in agreement with what was found for activation of
macromolecular substrate (Fig 2B).
Fig 1. Total and surface TF activity in nonpolarized cells. (A) Total
TF activity in transiently transfected COS-lcells determined by clotting assay of homogenizedcells using humanplasma. Clotting activities were relatedto totalcell protein and normalizedto the activity
of cells transfected with human TF (hTF,.263). The total activity of the
wild-type human TF clone (hTF,.)
was set at 100%. and ranged
from 0.5 t o 6 UTFlmg cell protein depending on the
efficiency of the
transfection. Data are the meanf SD of 5 independent experiments,
each in triplicate. An equal amount of
cells were transfectedwith an
equal amount of DNA for each construct. (B) Total TF activity in stably
transfected MDCK clones determined by clotting assay of homogenized cells using human plasma. Data are the mean f SD of 3 independent experimentswith each clone, each in duplicate. (C) Surface
activity of stable MDCK clones l day after plating. Surface activity
L.ocatior1
TF Sut$lce Antigen in HUVEC and MDCK
of the high-density-plated layers of adherent cells was determined
Cells
by chromogenic FXa substrate test and relatedto totalcell protein.
The total surface activity of the clone transfected with the wild-type
The apical/basolateral distribution of TF antigen on the
human TF construct (hTF1.263)
was set at 100% (0.43 UTFlrng protein,
surface of polarized monolayers of HUVEC (after 6 hours
of which0.27 UTFlmg protein due
is t o human TF). One typical experiment in triplicate (mean t SDI.
of IL-Ip stimulation) and MDCK was determined by binding
of
TF Surfnce Activie
Permeaide Support
of
Stable MDCK Clones Grown on
The stably transfected MDCK clones and nontransfected
MDCK cells were grown on permeable filters, feeding the
cells every day from both sides of the cell layer. When tight
of1'51-labeled anti-hTF antibodies or ligand FVIIa andby
cell surface specific biotinylation followed by irnmunoprecipitation (MDCK cells only).
For binding of iodinated antibodies, intact monolayers
were incubated on ice with 1251-labeled monoclonal murine
anti-hTF antibody (htfl or 4504) to saturation of TF-specific
binding. In preliminary experiments, this was determined to
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SORTING OF TISSUE FACTOR
245
MDCK
hTF1hTF1hTF1hTF1263
263
245
b
80"
i
1343
80
"
60"
"
E
40"
"
z
20
"
I
"
60
40
20
-0
0-
hTFl-245
hTFl-263
I r I
b
1
:::
:::
0.5
-- 0.4
"
0.3 --
0
d
o
-- 0.3
I
I
"
"
apical
basolateral
0.2
0.1
0
Fig 2. Surface TF activity on polar MDCK
cell layers. TF surface
activity of tight layers of MDCK stable clones grown on permeable
membranes.The activity was determined separatelythe
in apical (m)
or balolateral(0) compartmentsusing a chromogenic FXa substrate
test. No leakage across the cell layers was detected. (A) Preincubation of cella with a monoclonal antibodyto TF decreased the surface
advity of the cell lines transfectedwith human TF constructs (hTFla
and hTF,I
to the lavd of activii of the nontransfectedcells (MDCK).
The data are from 1 typical experiment in triplicate after 10 minute
reaction with MlalFX. A total of 6 experiments gave similar results.
(B) hTFl, and hTF,,
represent thewtface TF a c t i v i i (apical + besolateral) caused by the respectbe stably transfectad constructs (total
surfam TF activity-endogenout background of canine TF not neutralized by an anti-hTF monoclonal antibody). Data am the mean f SE of
six independent experiments,
each of 2 to 3 parallels. The total available
surface clctivity (apical+ barolateral)of the clone stably transfected
with hTFI.= was set at 100%(0.24 UTFlmg cell protein). (C) Amidolytic
activii of T F M a complex onthe apical (W) and badateral (U)surface
of the MDCK hTF,,
done determined usingthe peptide substrateS2288. The columns mpresent the sum of canine and human advity in
the done. Backgrounddeavage of S-2288 independent of Vlla was
Data are the mean f SD of 3 to 5 wells.
wbtmctd.
require about 1 hour. Both antibodies gave the same final
result, although the htfl antibody gave higher nonspecific
binding to both HUVEC and MDCK cells. In HUVEC, 78%
? 5% of the surface available antigen was found on the
apical surface (Fig 3A). In MDCK cells, the distribution of
binding sites was the opposite of that in HUVEC: 94% 2 3%
(hTF1.263)
or 92% 2 3% (hTFI.24s)of the surface available TF
antigen was located on the basolateral surface of the MDCK
cells (Fig 3B). Thus, in HUVEC, TF antigen and activity
colocalized to the apical surface of the cell layer. However,
in MDCK cells, the situation was different. More than 90%
of the surface available TF antigen was detected onthe
basolateral side, whereas essentially all of the (very limited)
TF surface activity was observed on the apical side. This is
clearly illustrated in Fig 3C, wherethe apical and basolateral
human TF activity of the MDCK hTFI.263clone (Fig 2B)
was related to the number of apical and basolateral cell
surface antibody binding sites on the same clone (Fig 3B).
The specific activity of the basolateral TF was about 100fold lower thanthat of the apical TF. All these binding
experiments were performed at 4°C. Control experiments at
37°C gave similar results. Comparison of 12sI-labeledantibody binding to intact and ethanol fixed (opened) cells on
Transwells revealed that the surface available human TF
antigen accounted for 70% i. 8% (n = 10) of the total human
TF antigen in the transfected MDCK cells.
Binding of iodinated FVIIa was used as an independent
method for determining surface available TF and also allowed determination of the distribution of canine TF. This
was especially important inview of the discrepancy in
MDCK cells between the very high amount of basolateral
TF antigen (94% of total surface available antigen; Fig 3B)
and the low basolateral procoagulant activity (less than 20%
of total surface available activity; Fig 2B). Nontransfected
cells and cells transfected with hTF1.263 were used.
The binding assays were performed in the presence or absence of a
50-fold molar excess of cold FVIIa. The competable binding
(binding in the absence of cold FVIIa-binding in its presence)
was taken to be the specific binding. These FVIIa binding
data (Fig 3D) demonstrated a distribution of human TF that
was similar to that obtained when antibody binding was
studied, although the basolateral portion was slightly lower
(88% rather than 94% of total surface available). A basolatera1 predominance was found also for the canine TF (84%
of the total surface available).
The total number of antibody binding sites on the cell
surface was essentially the same for wild-type and truncated
human TF inMDCK cells (25.5 5 3.1 and 25.9 ? 4.1 X
lo9 antibody binding sites/cm2 filter, respectively, n = 21).
The same number of total surface binding sites was found
with FVIIa (24.4 2 7.1 X lo9FVIIa binding sites/cm2filter,
n = 9), suggesting that the number of binding sites corresponded to the number of cell surface TF molecules. The
induced endothelial cells had much less TF antigen on their
cell surface (3.4 2 0.6 X lo9 antibody binding sites/cm2
filter, n = lo), partly due to the heterogeneity in response
to IL-1P discussed below and to the lower number of cells
per filter area (DNA content was determined by Hoechst
33258 to be 17 ? 6 pglcm' filter for MDCK cells and 4.7
2 0.8 &cm' filter for HUVEC).
For immunofluorescence studies of subconfluent HUVEC
stimulated by IL-lP to synthesize TF, ethanol fixed on glass
cover slips and costained with propidium iodide and an antihTF antibody showed marked cell-to-cell heterogeneity of
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
CAMERERET
1344
A
apical
basolateral
hTl-E3
hF1-24!5
B
AL
Fig 3. Surface TF antigen on polar HUVEC and MDCK cell layers
determined by binding of iodinated antibody and FVlla (A) IL-1Pinduced and control polarHUVEC layers labeled on ice for 1.5 hours
with "%anti-TF antibodies on apical (m) or basolateral (D) surface.
Binding t o control HUVEC subtracted. The resulting totalsurface TF
binding wasset at 100% (3.42 x lo9antibody bindingsiteslcm' filter).
Data are the mean ? SD of 4 independent experiments, each of 2 t o
3 parallels. (B) Polar cell layers of MDCK hTF,.'- and MDCK hTFl.2rs
labeled with lz5l-anti-TFantibodies on apical (W) or basolateral 10)
surface. Binding t o nontransfected MDCK cells subtracted. Total surface binding t o human TF was set at 100% for each clone (25.5 and
25.9 x lo9 antibody binding siteslcm' filter for MDCK hTFl.2s, and
MDCK hTF,.2.5, respectively). Data are the mean t SE of 8 independent experiments, each of 2 t o 3 parallels. (C)Specific activity toward
macromolecular substrate of apical (m) and basolateral (cl)surface
anti-hTF antibody binding sites on MDCK hTF,.'= Transwells determined by relating resultsin Fig 26 (surface activity) and 38 (surface
binding sites for anti-hTF monoclonal antibody). (D) Binding of '251
FVlla t o apical (W) or basolateral (c11 surface of polar layers of stably
transfected (hTF,.)
and nontransfected MDCK cells. Uncompetable
binding (in the presence of 50-fold molar excess of cold FVllal was
subtracted in both cases. Net binding t o human TF caused by the
stable transfection was obtainedby further subtracting the binding
t o canine TF in wild-type MDCK cells. Total binding (the sum
of specific binding on both sides after subtraction) was taken as 100% in
each case (24.4and 22.9 x lo9 FVlla binding siteslcm' filter t o human
and canine TF, respectively). Data are the mean 2 SD of 3 independent experiments, each in triplicate.
i
untransfected MDCK cells (Fig 5 ) showed a distribution of
biotinylated TF antigen to the basolateral (hTF,.263r94% 5
1%; hTFI.245r
90% t 4%) and apical (hTF1.2n3,6% 2 1%;
hTFI.24s.10% 2 4%) surface domains essentially identical
to that obtained with binding of iodinated monoclonal antibody and FVIIa. Wild-type MDCKgave no recoverable biotinylated TF.
D
TF expression (Fig 4).Regularly, only one third of the cells
responded to IL-10 induction of TF synthesis. Similar observations were made by Drake et aI2' and Kirchhofer et al.24
Stably transfected MDCK cells (hTFI.2hS)
displayed TF positive staining in 1 0 0 % of the cells. Immunofluorescence of
paraformaldehyde-fixedtight layers of IL- ID stimulated HUVEC andstably transfected MDCK cells appeared to support
the apicalhasolateral expression pattern seen using iodinated
antibodies, but staining of cells grown on Transwell filters
gave too much background fluorescence.
Biotinylation of transfected (hTFI.263and hTFI.Z45)and
Fig 4. Heterogenous HUVEC response t o IL-lP. HUVEC were
grown onglass cover slips, induced by IL-lfl to
express TF. ethanolfixed, and costained with mAb htfl/FITC-rabbit-antimouselg F(ab'),
and propidium iodide.
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SORTING OF TISSUEFACTOR
m
1
2
1345
-
l
180 ~
"."._"._"I._"__.."._
3
"
-
140
m116-
97.
u45-
340
BFA24h
rrnmel
hTF1-263
Fig 5. Surface TF antigen on polar MDCK cell layers determined
by biotinylationlimmunoprecipitation. Polar cell layers of MDCK
hTFl.nu, MDCK hTFl.245, and nontransfected MDCK cells surfacelabeled w i t h sulpho-NHS-biotin on apical or basolateral surface,
immunoprecipitated with an anti-hTF antibody and antimouse IgG
magnetic beads, run on12% PAGE, and visualized with streptavidinbiotin-HRP and ECL. One typical of4 experiments. The high-molecular-weight bands in lanes 4 and 6 represent nonspecific binding of
streptavidin-biotin-HRP.Lane 1, MDCK hTFl.m apical; lane 2, MDCK
hTF1.263basolateral; lane 3, MDCKhTF1.246apical; lane 4, MDCK
hTFl.2.5 basolateral; lane 5, MDCK apical; lane 6, MDCK basolateral.
Redistribution of MDCK TF Activity and Antigen on
Treatment With Brefeldin A
In an attempt to induce a shift of the polarity of TF antigen,
polarized MDCK cell layers were treated with Brefeldin A
(BFA). The concentration used ( 5 ,ug/mL) wouldbe expected
to have an effect on the expression of proteins in MDCK
cells under these conditions.".*" By attempting to induce a
shift of the TF expression, wewantedto
investigate the
importance of the local environment for the activity of the
TFlFVIIa complex.
The cells were treated with BFA for 6 and 24 hours.
After 6 hours, only a slight shift in antibody binding and
TF activity toward the apical surface was seen. After 24
hours, total surface binding of the '2'I-labeled monoclonal
anti-hTF antibody increased by about 40%, caused by an
approximately equal increasein binding sites on the apical
and the basolateral side (Fig 6 ) . This resulted in an apical
shift from 6% to 8% to about 20% of the total surface
binding. Cells transfected with the truncated construct
(hTFI.24s)showed a slightly higher increase on the apical
side than cells transfected with the complete TF cDNA
(hTFI-263). Theseobservations were confirmed by direct
surface biotinylation (Fig 7). Measurements
of cell surface
Xa generation showed a 2.3 5 0.7-fold increase in apical
activity compared with untreated cells, but hardly any (1.3
2 0.2-fold) increase in basolateral activity (n = 12). The
corresponding increase in apical antigen (4.0 5 1.2-fold)
was not significantly different from this activity increase.
These results suggest that it is the local environment of
the membrane domainthat determines if the expressed TF
is active. Treatment with BFA led to moderate cellular
shapechanges and reduced theelectricalresistance
by
about 50% without causing detectable leakagethrough the
layer, confirming previous observations.*'
BFA24h
normel
hTFl -245
Fig 6. Distribution of human TF antigen on the surface of polar
MDCK cell layers treated with Brefeldin A for 24 hours determined by
binding of iodinated antibody. Transfected MDCK hTF1.= and MDCK
hTFl.245 were labeled with
antibody on theapical ( W or basolateral
(U)side after incubation for24 hours in the presence or absence of
Brefeldin A (5 pg/mL). Binding t o nontransfected cells was subtracted. Total surface human TF specific antibody bindingt o cells not
treated with BrefeldinA (normal) was taken as 10046 (1009'0 values
are the same as in Fig 38). Data are the mean of 3 independent
experiments, each in triplicate.
Activih of Basolateral TF in MDCK Cells
Basolateral TF in MDCK cells had very low specificactivity (Fig 3C). To study the possible causes, we have looked
for TFPI-I and Annexin V by immunofluorescence microscopy. With anti-TFPI-l an almost uniform fluorescence was
seen on the basolateral side; all the controls were negative.
The fluorescence pattern was compatible with diffuse cellular or matrix staining. On the apical side all cells were unstained except for a few cells growing in dumps or piles.
These cells expressed TF and were apparently not polarized.
Thus, this antibody to human TFPI-I recognized also the
canine factor. TFPI-I expression was essentially observed
on the basolateral side. Treatment of the cells with heparin
did not change the fluorescence pattern. Neither heparin ( I 0
kD.
1
2
3
4
S
6
m116-
n-
u45
-
Fig 7. Distribution of human TF antigen on the surface of polar
MDCK cell layers treated with Brefeldin A for 24 hours determined
by biotinylationhnmunoprtxipitation. Polar cell layers of MDCK hTFl-,
MDCK hTF1.245,and nontransfected MDCK cells were treated with 5
p g l m L Brefeldin A for 24 hours, surface-labeled with sulpho-NHSbiotin on apical or basolateral surface, lysed, and immunoprecipitated withanti-hTF antibody and antimouse lgG magnetic beads, run
on 12% PAGE, and visualized with streptavidin-biotin-HRP and ECL.
Lane 1, MDCK hTF,.2m apical; lane 2, MDCK hTF1.nu basolateral; lane
3, MDCK hTFl.245 apical; lane 4, MDCKhTFl.26
basolateral; lane 5,
MDCK apical; lane 6, MDCK basolateral.
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
CAMERER ET AL
1346
U/mL for 6 hours) nor a neutralizing polyclonal antibody to
human TFPI-I significantly altered the activity of TF on the
basolateral side.
Polyclonal antibodies to Annexin V did not disclose any
endogenous Annexin V bound to the MDCK surface. However, purified human Annexin V bound to HUVEC and was
easilydetectedusing
thesame polyclonalantibodies. No
binding was observed to MDCK cells, neither on the apical
nor on the basolateral side, indicating a low content of negatively charged phospholipidsin the outerleaflet of the plasma
membrane.
DISCUSSION
an increased prothrombotic state of the endothelium without
the requirement for endothelialdisruption. This occurs in
concert withnovelsynthesis
of leukocyteadhesionmolecules (E-selectinj' and vascular cell adhesion molecule- 1 '?)
and with a decrease in the cell surface fibrinolytic activity
(upregulation of luminally expressed type 1 plasminogen
activator inhibitor," downregulation of cell surface thrombom ~ d u l i n ~and
~ " tissue-type
~
plasminogen activator productionjx).All these factors would contribute to theinitiation and
stabilization of fibrin deposition on the perturbed endothelial
surface, eg, as part of a disseminated intravascular coagulation process. Whether these in vitrofindings reflect what
actually takes place in vivo is still in the dark, as expression
of TF in perturbed vascular endothelium in vivo (except in
spleen) has been hardto demonstrate.23~'y~J" These
difficulties
may be related to thelow level andheterogeneity of T F
expression in perturbedendothelialcells,
or possibly to
masking of the antigen. Even very low levels of T F expression in vivo could still be of pathological significance, eg,
by disturbance of the plasma VIIdVlI ratio, once the T F is
presented directly to flowing blood as our findings suggest.
The binding of FVII to the extracytoplasmic part of TF
is an obligatory step in its triggering of blood coagulation.'
Recent studies in our laboratory' have shown that TF has a
dual functionin that FVII binding also elicits
an intracellular
signal in the form of Ca'' spikes. TF is constitutively expressed in many cell types that are not normally in contact
with flowing blood, such as epithelial cells lining the body
cavities. In two cell types, monocytes/macrophages and endothelial cells, which are both in contact with flowing blood,
TF I s Expressed Mainly on the Basolateral Surface of'
TF synthesis can be induced.839In vivo, epithelial cells and
endothelial cells form
tight junctions or junctional complexes MDCK Cells
In contrast to endothelial cells, epithelial cells would be
that seal off an apical surface area from rest
the of the surface
expected to interactwithFVIIreachingthe
cell from the
(basolateralj with regard to lateral diffusion in the cell membasolateral side. Studies of human TF antigen on the surface
braneandthus
allowthedefinitionofthese
twosurface
oftransfected MDCKcellsshowed
that 88% to 94% of
domains. In vitro experimentsto definecharacteristics of
surface TF indeed was localized to the basolateral side of
these domains must consequently be performed under condithe cells. The endogenous canineTF in the MDCK cells was
tions where the formation of such junctional complexes is
distributed in the same way; ie, both the human and canine
possible, and only under such conditions can the polarity of
TF in MDCK cells showed the opposite pattern of what was
TF expression be studied.
foundfor endothelial cells. The difference in localization
suggests that the sorting machinery either looks at different
TF Is Expressed Mainly on the Luminal Surface of IL-10
sorting signals in MDCK and endothelial cells or reads the
Induced Endothelial Cells
samesignals to adifferentconclusion.
The strictly polar
By growing cells on permeablefilters and carefully moniexpression in both cell types suggests active sorting in both
toring the formation of junctional complexes, we have recases rather than active in oneand default in theother,
cently shown that in endothelial cellsinduced by interleukinalthough thebasolateral dominance in MDCK isreduced
I D , the fraction of total TF activity ( -25%j2x that becomes
when the results are normalized per unit surface area constiavailablefor triggering of coagulationon intact cells is
tuting the two domains. Normalized for the ratio of apical
mainly onthe apical surface.16 In the presentreport, we
to basolateral surface areas (-1:3.8j:'
there is still a greater
confirmthisdistribution
by measurement of binding of a
than two-fold basolateral preference for TF antigen expreslabeled monoclonal antibody to the cell surface domains.
sion.Thisis less compatible with passivedelivery to the
Previous report^^*,^^ have suggesteddifferent localizations
apical and basolateral membrane domains. Prolonged treatof TF in endothelial cells at variance with our results and
ment of the MDCK cells with Brefeldin A caused an apparwiththose of others. As far as we can judge,
neitherof
ent redistribution of TF antigen as well as activity toward
these groups have grown the endothelial cellson permeable
the apical surface. Whether this is caused by altered sorting
supports (except for EM studies by Ryan et al"), nor have
of newly synthesized T F or (less likely) represents a true
theycultured the cells for the time we
find necessary to
redistribution of basolateral or intracellular TF is not known.
establish junctionalcomplexes.The
differences arethus
It is interesting to notethat essentially allof the TF appearing
most likely due to difference in experimental designs. The
on the apical surface becomes fully active. This local enviinadequacy of such setups for the formation of junctional
ronment thus allows functional expression of an increased
complexes in endothelial cells has also
been suggested by
number of TF molecules.
Schleef et al.'"
If thepolarized expression patterndescribed herealso
Discordant Expression of TF Antigen and Activity on the
holds true for the in vivo situation, induced TF is uniquely
Surface of MDCK Cells
positioned for interacting with FVIIa on the luminal surface
Expression of TF antigen of thebasolateralsurface
of
of the endothelial cell. An important role is suggested for
MDCK wasdetected by antibodybinding, FVIIa binding
endothelial TF in the inflammatory response, contributing to
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SORTING OF TISSUE FACTOR
1347
and by biotinylatiodimmunoprecipitation butwasnotaccompanied by detectable procoagulant activity. This was not
a temperature-dependent phenomenon, as the same binding
distribution was obtained at 37"C, nor was it due to limited
access of activity assay components through the permeable
filters, as the same activity distribution was observed with
cells grown on large pore (3.0 pm) membranes. It was clear
from our binding studies that FVIIa, a molecule similar to
FX also used in the Xa chromogenic substrate assay, was
able to reach andbindtothe
basolateral TF. The limited
activity of the basolateral TF was not restricted to activation
of macromolecular substrate, as a similar lack of activity
was found when looking at direct hydrolysis of a peptide
substrate (S-2288) by the TFNIIa complex.
It thus appears that there was a large fraction of TFNIIa
complexes on the basolateral surface of the MDCK cells
which was not active and a much smaller number of complexes on the apical surface which was highly active.
Several different not mutually exclusive explanations are
possible. Restricted access of substrates to bind with TF is
unlikely as discussed above. The presence of one or more
inhibitors is possible. Candidates might be TFPI-1,2,42TFPI' :2 antithrombin III,M Annexin V4' or ~phingosine.~~
The
latter two are unlikely; Annexin V was not detected and
maynot be able to inhibit the TFNIIa complex in intact
s u r f a ~ e s . ~Sphingosine
~,~'
acts by inhibiting binding of VIIa
to TF,& which we have found to be unhampered in MDCK
cells. Immunofluorescence studies of polar MDCK cell layers clearly showed basolateral localization of TFPI-1. This
inhibitor is normally expected to require FXa for binding to
the TFNIIa complex, whereas we demonstrate lack of activity of the TFNIIa complex also in the absence of FX.
Another possibility might be the lack of phosphatidylserine (PS) in the outer leaflet of the plasma membrane. PS has
previously been shown to be crucial for TF
In
1992, Le et a]'' suggested the existence of two classes of
TFNIIa complexes on the surface of OC-2008 cells. Both
fractions had the same binding affinity, both were active
upon cell lysis, butonly the minor fraction supported the
activation of macromolecular substrate on the cell surface.
To explain this, Le et a1 suggested that the recruitment of
an additional surface component, essential for the catalytic
activity of the TFNIIa complex, was necessary. Cell lysis
would supply more of this limiting compound, suggested to
be PS. This would be consistent with the absence of binding
of Annexin V, but the limited difference in gross composition of apical and basolateral plasma membrane domains"
does not support this hypothesis, although a very special
distribution of PS between the outer and inner plasmamembrane leaflet" is not excluded.
One should also consider the fact that the increase in
apical TF antigen after treatment with BFA was coupled to
an increase in activity of the same scale. This would be
compatible with a limiting amount of an additional surface
241
Fig 8. Amino acid sequence alignment of the TF
cytoplasmic domain in4species. Conserved residues
are boxed. Numbering as in human TF. Arrow indicates position of truncation in hTFI.2S.
component required for activity only if the amount of this
component was also affected by BFA treatment or by polarization of the cells.
Removal of the TF Cytoplasmic Tail Does Not Alter the
Distribution of TF in MDCK Cells
Until recently, all of the proteins studied which localize
to the basolateral domain of MDCK cells, have had active
sorting signals within their cytoplasmic domains. Two types
oforting motifs have been revealed, one type dependent on
a tyrosine residue and the other dependent on a double leucine (or other hydrophobic doublets).53When these basolateral determinants are deleted, the proteins are distinctly missorted to the apical surface of the ~ e l l s Cytoplasmic
. ~ ~ ~ ~ ~
domain-independent sorting to the basolateral surface of polarized MDCK cells was recently demonstrated for the (Y2Aadrenergic receptor.60
The TF cytoplasmic domain does not contain any known
signal for endocytosis or sorting. Amino acid sequence alignment of the human, murine, bovine andrabbit TF cytoplasmic domains shows a high degree of conservation comprising a serinekhreonine phosphorylation site6' and several
other amino acid residues (Fig 8). As was the case for the
azA-adrenergicreceptor,m deletion of the cytoplasmic domain of TF had no obvious effect on surface distribution of
the protein; the truncated mutant showed the same basolateral phenotype as wild-type TF. We have not excluded the
possibility thatthe three remaining membrane proximal
amino acids of the cytoplasmic tail are of importance, as
palmitoylation signals have in some cases been shown to
have an effect onendocytosis.62.6' The cytoplasmic tail is
not important for sorting of TF, but it may have other roles,
such as in the modulation of TF half life on the cell surface
or in mediating the intracellular signal generated by TF on
binding of its ligand FVIIa.' Whythe activity of clones
stably transfected with truncated T F C D N A , . was
~ ~ ~in fact
greater than wild-type transfected clones is not known. Paborsky et a1 made the same observation*' and reported that
there was no difference in specific activity between the two.
In our studies, both constructs contained the 3' untranslated
region of TF cDNA; therefore, it is not likely to be due to
different mRNA stability.
Sorting of TF in MDCK cells is also unusual in the sense
that although TF does not contain any known basolateral
sorting or endocytosis signals, it still sorts to the basolateral
domain of the cells. This, in combination with the lack of
cytoplasmic tail involvement, raises the possibility of a new
class ofyetunidentifiedbasolateral
sorting signals (if the
sorting is indeed active). Together with what was seen for
the a,,-adrenergic receptor, this suggests that basolateral
sorting in MDCK cells is not always regulated by a cytoplasmic determinant intimately linked to an endocytosis signal.
Prior to this study, there are few studies of sorting of the
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
CAMERER ET AL
1348
same protein in both endothelial and epithelial cells. GPIanchorshave been shownto directproteins totheapical
domain inboth polarizedendothelial cells64 and MDCK
cells.h5The transferrin receptor is localized on the basolateral
surface of MDCK cells,"6 and on the apical surface of the
high resistance endothelium of the blood-brain barrier,67 but
not the normal low-resistance vascular endothelium.hR Sortnormal lowing of TF to opposite domains in epithelium and
resistance vascular endothelium remains without precedent.
ACKNOWLEDGMENT
We are grateful for the gifts of monoclonal TF antibody from Dr
S. Carson (University of Nebraska Medical Center), TF cDNA from
Dr J. Momssey (Oklahoma Medical Research Foundation), polyclonal Annexin V antibody from Dr W. van Heerde (HBpital de
Bicetre), and recombinant FVIIa, interleukin-10, and TFPI antibodies from Prof U. Hedner and colleagues (Novo-Nordisk), and to A.
Simonsen for helpful advice.
REFERENCES
I . RQttingen J-A, Enden T, Camerer E, Iversen J-G, Prydz H:
Binding of human factor VIIa to tissue factor induces cytosolic Ca2'
signals in 582 cells, transfected COS-l cells, MDCK cells and in
human endothelial cells induced to synthesize tissue factor. J Biol
Chem 270:4650, 1995
2. Hjort PF: Intermediate reactions in the coagulation ofblood
with tissue thromboplastin. Scand J Clin Lab Invest [Suppl] 9:27:1,
1957
3. Bj~rklidE, Storm E, Prydz H: The protein component of human brain thromboplastin. Biochem Biophys Res Commun 55:969,
1973
4. Nemerson Y: Tissue factor and haemostasis. Blood 71:1, 1988
5. Astrup T: Assay and content of tissue thromboplastin in different organs. Thromb Diath Haemorr 14:401, 1965
6. Drake TA, Momssey JH, Edgington TS: Selective cellular
expression of tissue factor in humantissues-Implications for disorders of haemostasis and thrombosis. Am J Pathol 134:1087, 1989
7. Flossel C, Luther T, Muller M, Albrecht S, Kasper M: Immunohistochemical detection of tissue factor on paraffin sections of
routinely fixed human tissue. Histochemistry 101:449, 1994
8. Rivers RPA, Hathaway WE, Weston WL: The endotoxin-induced coagulant activity ofhuman
monocytes. Br J Haematol
30:31 I , 1975
9. Johnsen UHL, Lyberg T, Galdal KS, Prydz H: Platelets stimulate thromboplastin synthesis inhuman endothelial cells. Thromb
Haemost 49:69, 1983
IO. Camerer E, Kolstfl A-B, Prydz H: Cell biology of tissue factor, the principal initiator of blood coagulation. Thromb Res 8 l : l ,
I996
1 I . Ruf W, Edgington TS: Structural biology of tissue factor, the
initiator of thrombogenesis in vivo. FASEB J 8:385, 1994
12. Harlos K, Martin DA, O'Brien DP, Jones EY, Stuart DL
Polikarpov I, Miller A, Tuddenham EGD, Boys CWG: Crystal structure ofthe extracellular region of human tissue factor. Nature
370:662, 1994
13. Carson SD, Ross SE, Bach R, Guha A: An inhibitory antibody
against human tissue factor. Blood 70:490, 1987
14. Bach R, Koningsberg WH, Nemerson Y: Human tissue factor
contains thioester-linked palmitate and stearate on the cytoplasmic
half-cystine. Biochemistry 27:4227, 1988
15. Hvatum M, Prydz H: Studies on tissue thromboplastin. Solubilization with sodium deoxycholate. Biochim Biophys Acta 130:92,
1966
16. Narahara N, Enden T, Wiiger M, Prydz H: Polar expression of
tissue factor in human umbilical vein endothelial cells. Arterioscler
Thromb l4:l, 1994
17. Markwell MAK, Haas SM, Bieber LL, Tolbert NE: A modification of the Lowry procedure to simplify protein determination
in membrane and lipoprotein samples. Anal Biochem 87:206, 1978
18. Le DT, Rapaport SI, Rao LVM: Relations between factor
VIIa binding and expression of factor VIIdtissue factor catalytic
activity on cell surfaces. J Biol Chem 267:15447, 1992
19. Sargiacomo M, Lisanti M, Graeve L, Le Bivic A, RodriguezBoulan E: Integral and peripheral protein composition of the apical
and basolateral membrane domains in MDCK cells. J Membr Biol
107:277, 1989
20. Paborsky LR, Caras IW, Fisher U, Gorman CM: Lipid association, but not the transmembrane domain, is required for tissue
factor activity: Substitution of the transmembrane domain with a
phosphatidylinositol anchor. J Biol Chem 266:21911, 1991
21. Drake TA, Ruf W, Momssey JH, Edgington TS: Functional
tissue factor is entirely cell surface expressed on lipopolysaccharidestimulated human blood monocytes and a constitutively tissue factor-producing neoplastic cell line. J Cell Biol 109:389, 1989
22. Ryan J, Brett J, Tijburg P,Bach RR, KisielW, Stem D:
Tumor necrosis factor-induced endothelial tissue factor is associated
with subendothelial matrix vesicles but is notexpressed on the apical
surface. Blood 80:966, 1992
23. Drake TA, Cheng J, Chang A, Taylor FB Jr: Expression of
tissue factor, thrombomodulin and E-selectin in baboons with lethal
Escherichia coli sepsis. Am J Pathol 142:1458, 1993
24. Kirchhofer D, Tschopp TB, Hadviry P, Baumgartner HR:
Endothelial cells stimulated with tumor necrosis factor-a express
varying amounts of tissue factor resulting in inhomogenous fibrin
deposition in a native blood flow system: Effects of thrombin inhibitors. J Clin Invest 93:2073, 1994
25. Hunziker W, Whitney JA, Mellman I: Selective inhibition of
transcytosis by Brefeldin A in MDCK cells. Cell 67:617, 1991
26. Wagner M, Rajasekaran AK, Hanzel DK, Mayor S, Rodriguez-Boulan E: Brefeldin A causes structural and functional alterations of the trans-Golgi network of MDCK cells. J Cell Sci 107:933,
1994
27. PrydzK,Hansen SH, Sandvig K, van Deurs B: Effects of
Brefeldin A on endocytosis, transcytosis and transport to the Golgi
complex in polarized MDCK cells. J Cell Biol 119:259, 1992
28. Andoh K, Pettersen KS, Filion-Myklebust C, Prydz H: Observations on the cell biology of tissue factor in endothelial cells.
Thromb Haemost 63:298, 1990
29. Mulder AB, Hegge-Paping KSM, Magielse CPE, Blom NR,
Smit JW, van der Meer J, Ruud Halie M, Bom VJJ: Tumor necrosis
factor a-induced endothelial tissue factor is located on the cell
surface rather than in the subendothelial matrix. Blood 5: 1559, 1994
30. Schleef RS, Podor TJ, Dunne E, Mimuro J, Loskutoff DJ:
The majority of type 1 plasminogen activator inhibitor associated
with cultured human endothelial cells is located under the cells and
is accessible to solution-phase tissue-type plasminogen activator. J
Cell Biol 110:155, 1990
31. Bevilacqua MP, Pober JS, Wheeler ME, Cotran RS, Gimbrone MA Jr: Interleukin 1 activation of vascular endothelium. Effects on procoagulant activity and leukocyte adhesion. Am J Pathol
121:393, 1985
32. Carlos TM, Schwartz BR, Kovach NL, Yee E, Rosa M, Osborn L, Chi-Rosso G, Newman B, Lobb R, Rosa M: Vascular cell
adhesion molecule-l mediates lymphocyte adherence to cytokineactivated cultured human endothelial cells. Blood 76:965, 1990
33. Schleef RS, Loskutoff DJ, Podor TJ: Immunoelectron microscopic localization of type 1 plasminogen activator inhibitor on the
surface of activated endothelial cells. J Cell Biol 113:1413, 1991
34. Nawroth PP, Handley DA, Esmon CT, Stem DM: Interleukin
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
SORTING OF TISSUE FACTOR
1 induces endothelial cell procoagulant while suppressing cell-surface anticoagulant activity. Proc Natl Acad Sci USA 83:3460, 1986
35. Moore KL, Andreoli SP, Esmon NL, Esmon CT, Bang NU:
Endotoxin enhances tissue factor and suppresses thrombomodulin
expression ofhuman vascular endothelium in vitro. J Clin Invest
79: 124, 1987
36. Moore KL, Esmon CT, Esmon N L Tumor necrosis factor
leads to the internalization and degradation of thrombomodulin from
the surface of bovine aortic endothelial cells in culture. Blood
73:159, 1989
37. Bartha K, Brisson C, Archipoff G, de la Salle C, Lanza F,
Cazenave J, Beretz A: Thrombin regulates tissue factor and thrombomodulin mRNA levels and activities in human saphenous vein endothelial cells by distinct mechanisms. J Biol Chem 268:421, 1993
38. Schleef RS, Bevilacqua MP, Sawdey M, Gimbrone MA Jr,
Loskutoff DJ: Cytokine activation of vascular endothelium. J Biol
Chem 2635797, 1988
39. Wilcox JN, Hasko JA, Deisher TA, Bunting S : Identification
of cells expressing tissue factor mRNA by in situ hybridization after
in vivo administration of endotoxin to rats. Thromb Haemost 62:348,
1989 (abstr)
40. Mackman N, Sawdey MS, Keeton MR. Lokustoff DJ:Murine
tissue factor gene expression in vivo: Tissue and cell specificity and
regulation by lipopolysaccharide. Am J Pathol 143:76, 1993
41. Parton RG, Prydz K, Bomsel M, Simons K, Griffiths G: Meeting of the apical and basolateral endocytic pathways of the MadinDarby canine kidney cell in late endosomes. J Cell Biol 109:3259,
1989
42. Broze GJ, Warren LA, Novotny WF, Higuchi DA, Girard
JJ, Miletich JP: The lipoprotein-associated coagulation inhibitor
that inhibits the factor VII-tissue factor complex also inhibits factor Xa: Insight into its possible mechanisms of action. Blood
71:355, 1988
43. Sprecher CA, Kisiel W, Mathewes S, Foster DC: Molecular
cloning, expression, and partial characterization of a second human
tissue-factor-pathway inhibitor. Roc Natl Acad Sci USA 91:3353,
1994
44. Rao LVM, Rapaport SI, Hoang AD: Binding of factor VIIa
to tissue factor permits rapid antithrombin IIYheparin inhibition of
factor VIIa. Blood 81:2600, 1993
45. Kondo S , Noguchi M, Funakoshi T, Fujikawa K, Kisiel W:
Inhibition of human factor VIIa-tissue factor activity by placental
anticoagulant protein. Thromb Res 48:449, 1987
46. Conkling PR, Patton KL, Hannun YA, Greenberg CS, Weinberg JB: Sphingosine inhibits monocyte tissue factor-initiated coagulation by altering factor VII binding. J Biol Chem 264:18440, 1989
47. Rao LVM, Tait JF, Hoang AD: Binding of annexin V to a
human ovarian carcinoma cell line (OC-2008). Contrasting effects
on cell surface factor VIIdtissue factor activity and prothrombinase
activity. Thromb Res 67:517, 1992
48. Almus FE, Rao LVM, Rapaport SI: Regulation of factor VIId
tissue factor activity in an umbilical vein model. Arterioscler Thromb
13:105, 1993
49. Bj0rklid E, Storm E Purification and some properties of the
protein component of tissue thromboplastin from human brain. Biochem J 165:89, 1977
50. Bach R, Gentry R, Nemerson Y: Factor VI1 binding to tissue
factor in reconstituted phospholipid vesicles-Induction of co-operativity by phosphatidylserine. Biochemistry 25:4007, 1986
51. van Meer G, Simons K: Viruses budding from either the
1349
apical or the basolateral plasma membrane domain of MDCK cells
have unique phospholipid compositions. EMBO J 1:847, 1982
52. van Meer G, Simons K: The function of tight junctions in
maintaining differences in lipid composition between the apical and
the basolateral cell surface domains ofMDCK cells. EMBO J
5:1455, 1986
53. Matter K, Yamamoto EM, Mellman I: Structural requirements
and sequence motifs for polarized sorting and endocytosis of LDL
and Fc receptors in MDCK cells. J Cell Biol 126:991, 1994
54. Casanova JE, Apodaca G, Mostov KE: An autonomous signal
for basolateral sorting in the cytoplasmic domain of the polymeric
immunoglobulin receptor. Cell 66:65, 1991
55. Hunziker W, Harter C, Matter K, Mellman I: Basolateral
sorting in MDCK cells requires a distinct cytoplasmic domain determinant. Cell 66:907, 1991
56. Naeme SJ, Isacke CM: The cytoplasmic tail of CD44 is required for basolateral localization in epithelial MDCK cells but does
not mediate association with the detergent-insoluble cytoskeleton in
fibroblasts. J Cell Biol 121:1299, 1993
57. Geffen I, Fuhrer C, Leitinger B, Weiss M, Huggel K, Griffiths
G , Spiess M: Related signals for endocytosis and basolateral sorting
of the asialglycoprotein receptor. J Biol Chem 268:20772, 1993
58. Kundu A, Nayak DP: Analysis ofthe signals for polarized
transport of the influenza virus (A/WSN/33) neuraminidase and human transferrin receptor, type I1 transmembrane proteins. J Virol
68:1812, 1994
59. Haass C, Koo EH, Capell A, Teplow DB, Selkoe DJ: Polarized sorting of P-amyloid precursor protein and its proteolytic products in MDCK cells is regulated by two independent signals. J Cell
Biol128:537, 1995
60. Keefer JR, Kennedy ME, Limbird LE: Unique structural features important for stabilization versus polarization of the qA-adrenergic receptor on the basolateral membrane of Madin-Darby canine
kidney cells. J Biol Chem 269:16425, 1994
61. Zioncheck TF, Roy S, Vehar GA: The cytoplasmic domain
of tissue factor is phosphorylated by a protein kinase C-dependent
mechanism. J Biol Chem 267:3561, 1992
62. Alvarez E, Girones N, Davis RJ: Inhibition of the receptormediated endocytosis of differic transferrin is associated with the
covalent modification of the transferrin receptor with palmitic acid.
J Biol Chem 265:16644, 1990
63. KawateN, Menon Kh4J: Palmitoylation of luteinizing hormonehuman choriogonadotropin receptors in transfected cells: Abolition of palmitoylation by mutation of cys-621 and cys-622 residues
in the cytoplasmic tail increases ligand-induced internalization of
the receptor. J Biol Chem 269:3065 1, 1994
64. Lee SW, Kahn ML, Dichek DA: Expression of an anchored
urokinase in the apical endothelial cell membrane: Preservation of
enzymatic activity and enhancement of cell surface plasminogen
activator. J Biol Chem 267:13020, 1992
65. Lisanti MP, Caras IW, Davitz MA, Rodriguez-Boulan E: A
glycophospholipid membrane anchor acts as an apical targeting signal in polarized epithelial cells. J Cell Biol 109:2145, 1989
66. Fuller SD, Simons K: Transfemn receptor polarity and recycling accuracy in “tight” and “leaky” strains of Madin-Darby canine kidney cells. J Cell Biol 103:1767, 1986
67. Roberts RL, Fine RE, Sandra A: Receptor-mediated endocytosis of transferrin at the blood-brain barrier. J Cell Sci 521, 1993
68. Jefferies WA, Brandon MR, Hunt SV, Williams AF, Gatter
KC, Mason DY: Transfemn receptor on endothelium of brain capillaries. Nature 312:162, 1984
From www.bloodjournal.org by guest on June 16, 2017. For personal use only.
1996 88: 1339-1349
Opposite sorting of tissue factor in human umbilical vein endothelial
cells and Madin-Darby canine kidney epithelial cells
E Camerer, S Pringle, AH Skartlien, M Wiiger, K Prydz, AB Kolsto and H Prydz
Updated information and services can be found at:
http://www.bloodjournal.org/content/88/4/1339.full.html
Articles on similar topics can be found in the following Blood collections
Information about reproducing this article in parts or in its entirety may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests
Information about ordering reprints may be found online at:
http://www.bloodjournal.org/site/misc/rights.xhtml#reprints
Information about subscriptions and ASH membership may be found online at:
http://www.bloodjournal.org/site/subscriptions/index.xhtml
Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American
Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.
Copyright 2011 by The American Society of Hematology; all rights reserved.