Science 
Copyright © 1996 by the American Association for the Advancement of Science

----------------------------------------------
Volume 273(5272)             12 July 1996             pp 252-255
----------------------------------------------

Platelet-Mediated Lymphocyte Delivery to High Endothelial Venules
[Reports]
Diacovo, Thomas G.*;  Puri, Kamal D.*;  Warnock, R. Aaron;  Springer, Timothy
A.;  von Andrian, Ulrich H.**
T. G. Diacovo, Center for Blood Research and Department of Cardiology, Harvard
Medical School, 200 Longwood Avenue, Boston, MA 02115, USA, and Division of
Newborn Medicine, Department of Pediatrics, Tufts University School of Medicine,
Boston, MA 02111, USA.
K. D. Puri, T. A. Springer, U. H. von Andrian, Center for Blood Research and
Department of Pathology, Harvard Medical School, 200 Longwood Avenue, Boston, MA
02115, USA.
R. A. Warnock, Department of Pathology, Stanford University, Stanford, CA 94305,
USA.
*These authors contributed equally to this work.
**To whom correspondence should be addressed at the Center for Blood Research,
200 Longwood Avenue, Boston, MA 02115, USA. E-mail: uva@cbr.med.harvard.edu

----------------------------------------------

Outline

Abstract
REFERENCES AND NOTES

Graphics

Figure 1
Figure 2
Figure 3
Figure 4

----------------------------------------------

Abstract

Circulating lymphocytes gain access to lymph nodes owing to their ability to
initiate rolling along specialized high endothelial venules (HEVs).One mechanism
of rolling involves L-selectin binding to peripheral node addressin (PNAd) on
HEVs. Activated platelets are shown to bind to circulating lymphocytes and to
mediate rolling in HEVs, in vivo, through another molecule, P-selectin, which
also interacts with PNAd. In vitro, activated platelets enhanced tethering of
lymphocytes to PNAd and sustained lymphocyte rolling, even in the absence of
functional L-selectin. Thus, a platelet pathway operating through P-selectin
provides a second mechanism for lymphocyte delivery to HEVs.

----------------------------------------------

Lymphocytes continuously migrate from the blood to peripheral lymph nodes (PLNs)
and other lymphoid organs. This ``homing'' event is mediated by sequential
engagement of tissue-specific adhesion and activation pathways [1]. Homing to
PLNs, for instance, is dependent on the lymphocyte homing receptor L-selectin
[2], which interacts with a ligand on HEVs that is defined by the monoclonal
antibody (mAb) MECA-79 [3] in mouse and human. Both MECA-79 and recombinant
L-selectin immunoglobulin chimeric protein precipitate sulfated and sialylated
glycoproteins that are collectively known as the peripheral node addressin
(PNAd) [4,5]. Affinity-purified PNAd glycoproteins mediate L-selectin-dependent
lymphocyte rolling in vitro [6,7]. Except for its physiologic expression in HEVs
of PLNs, venular PNAd expression is also induced in several chronic inflammatory
disease states [8]. Some of these pathologic conditions, such as inflammatory
bowel disease, are also associated with an increase in activated platelets both
in the circulation and at vascular sites of inflammation [9]. Because activated
platelets interact with both leukocytes and endothelial cells in vitro, we
examined whether activated platelets in the bloodstream can alter lymphocyte
behavior in HEVs.

We used an in vivo model to study lymphocytes in murine PLN HEVs by intravital
microscopy [10]. A mAb to L-selectin, Mel-14 [2], inhibited rolling of
fluorescently labeled white blood cells (WBCs) by 80 to 90 % [Figure 1 and
[11]], confirming the important role of this molecule in lymphocyte homing. A
subsequent injection of resting human platelets [12] into the arterial
bloodstream did not detectably alter WBC behavior in HEVs [13]. In contrast,
platelet activation before injection resulted in a marked reappearance of
endogenous rolling WBCs despite the continued presence of mAb to L-selectin.
These findings indicate a previously uncharacterized platelet-mediated mechanism
for WBC adhesion to HEVs. Because the phenomenon was apparently L-selectin-independent,
our observations further suggest that activated platelets may alter the
composition of lymphocyte populations that are recruited to PLNs and other sites
containing HEVs.

----------------------------------------------
Figure 1. Activated platelets mediate rolling of anti-L-selectin-treated mouse
leukocytes in HEVs. Blood-borne nucleated cells (lymphocytes and granulocytes)
were fluorescently labeled by iv injection of rhodamine 6-G and visualized in
HEVs of a subiliac LN by fluorescence microscopy. Treatment of mice with mAb
Mel-14 (100 micro gram) significantly reduced the WBC rolling fraction (the
percentage of rolling cells in the total flux). Injection of thrombin receptor
activating peptide (TRAP)-stimulated human platelets markedly increased
tethering and rolling of endogenous WBCs. Blockade of human P-selectin function
by mAb WAPS 12.2 abolished this effect of activated platelets. Values shown are
the mean plus minus SD of nine venules in three animals.
----------------------------------------------

To define the molecular basis for platelet-dependent WBC rolling, we initially
investigated the role of P-selectin [14]. This molecule is expressed on the
surface of activated, but not resting, platelets and endothelial cells and
supports leukocyte rolling and accumulation in acutely inflamed nonlymphoid
sites [15]. It also mediates leukocyte interactions with activated platelets
that are in circulation or bound to thrombogenic surfaces [16]. When activated
platelets were first incubated with mAb WAPS 12.2 directed against human, but
not murine, P-selectin [12], injection of treated cells did not induce WBC
rolling in anti-L-selectin-treated mice. Thus, P-selectin expressed on
platelets, but not on the animals' endothelial cells, was responsible for WBC
rolling.

To identify the target cells and ligands for platelet P-selectin, we labeled
human platelets fluorescently to directly observe their intravascular behavior.
Resting platelets rarely bound to HEVs or other cells in the bloodstream [13].
Activated platelets, in contrast, frequently interacted with both WBCs and the
venular lining. Numerous rolling WBCs were detected with one or more brightly
fluorescent platelets attached to their surface. Treatment of animals with mAb
Mel-14 blocked rolling of WBCs free of surface-bound platelets. In contrast, the
rosettes of WBCs and platelets persisted to roll in a cartwheeling fashion,
suggesting an indirect mode of WBC adhesion to HEVs through bridging platelets.
This mechanism is supported by the finding that a large number of activated
platelets could bind directly to endothelial cells without WBC association.
These interactions were characterized by slow rolling, occasionally followed by
stationary adhesion Figure 2, A, C, and D), and were nearly abolished by mAb
WAPS 12.2 (inhibition of 80 plus minus 9%, mean plus minus SD), indicating that
platelet P-selectin was responsible. In contrast, endothelial P-selectin,
implicated in platelet rolling in other vascular beds [17], was not required;
activated human platelets rolled in HEVs of P-selectin-deficient mice [13].
Furthermore, platelet rolling mediated by platelet P-selectin in HEVs was
associated with the formation of stronger or a larger number of bonds (or both)
than platelet rolling on endothelial P-selectin because the mean rolling
velocities measured for the former were reduced by 80% compared with those
reported for the latter [17]. To show that platelet rolling in HEVs was not the
result of P-selectin interactions with ligands on endothelium-bound leukocytes
such as P-selectin glycoprotein ligand 1 [PSGL-1; [18,19]], we first treated
some animals for 30 min with mAb Mel-14 to block leukocyte binding to HEVs.
P-selectin-dependent platelet interactions persisted in the presence of
anti-L-selectin and were equivalent to those seen in untreated animals (rolling
fraction of 73 plus minus 8% versus 74 plus minus 18%, respectively; mean plus
minus SD). Treatment of mice with mAb MECA-79, which reduces L-selectin-dependent
homing of lymphocytes [3], also reduced adhesion of activated platelets by
approximate 70%, suggesting that L- and P-selectin can share the same endothelial
ligand, or ligands, on HEVs.

----------------------------------------------
Figure 2. Human platelets and L1-2 transfectants expressing human P-selectin
tether and roll in HEVs of PLNs. (A) Resting or thrombin-stimulated, fluorescently
labeled human platelets were injected into the arterial bloodstream of a mouse
and visualized for 3 to 5 min during their passage through HEVs. Platelets that
interacted with HEVs (Proll) and platelets that did not display detectable
interactions (Pfree) during the observation period were counted. Platelets that
were apparently associated with WBCs were not included. Rolling fractions were
calculated as Proll/(Proll plus Pfree) times 100%. A total of 12 venules were
analyzed in four animals. Values shown are the means plus minus SD. (B)
Fluorescently labeled L1-2P-selectin, but not vector control transfectants
(L1-2vector), roll in HEVs. Transfectant rolling was reduced by treatment of
animals with mAb MECA-79 to PNAd (mean plus minus SD of three to four venules in
each of three animals). (C) (Top) Composite image of digitized intravital
fluorescence micrographs showing the range of observable behaviors of activated
platelets in a branched HEV. The intravascular position of a noninteracting
platelet is shown in 0.1-s intervals (arrows) in the lower branch of the venule.
In contrast, a rolling platelet (arrowheads) in the upper branch displays a much
slower rate of motion in the blood-stream despite comparable wall shear rates
(upper, 216 sminus 1; lower, 248 sminus 1) in both branches. In addition, a
platelet is shown (asterisk) that initially rolled in the lower branch, but had
subsequently arrested. (Bottom) The same cell 120 s later as well as several
others that had subsequently accumulated. (D) Typical velocity profiles of
resting (broken line) and activated platelet (solid line) populations in HEVs
(mean venular diameters of 22.4 and 31.9 micro meter, respectively). Velocity
profiles for activated and resting platelets were determined from separate
experiments by PC-based off-line video analysis [29] of rolling and freely
flowing cells. The velocities of consecutive platelets, either resting (n equals
89) or activated (n equals 58) populations, were determined by measuring the
distance traveled between two or more successive video frames. Assuming a
parabolic flow profile, the highest cell velocity, Vmax, was used to calculate
the mean blood flow velocity (VBlood). VBlood was 597 micro meter /s for resting
platelets and 1014 micro meter /s for activated platelets. The calculated wall
shear stress (assuming a blood viscosity of 0.025 poise) was 5.4 and 6.8 dyne
centered dot cmminus 2, respectively. For comparison, individual platelet
velocities were normalized to the centerline velocity, VCL, which was assigned a
value of 1, and the frequency of cells within a particular velocity class was
determined. Mean rolling velocity for interacting platelets was 33 plus minus 29
micro meter /s (range, 7.8 to 155 micro meter /s).
----------------------------------------------

Because platelets express other adhesion molecules, we examined whether
expression of cellular P-selectin alone was sufficient to promote rolling in
HEVs. L1-2 lymphoma cells transfected with human P-selectin [L1-2P-selectin;
[20]], but not vector control transfectants (L1-2vector), displayed slow rolling
interactions in the identical segments of the lymph node (LN) microvasculature
as platelets Figure 2B. As shown for activated platelets, mAb MECA-79 also
reduced adhesion of L1-2 (P-selectin) (inhibition of 59 plus minus 5%, mean plus
minus SD), suggesting that P-selectin:PNAd interactions per se are sufficient
for tethering and rolling in vivo.

To show directly that human PNAd contains a P-selectin ligand, we evaluated
platelet and L1-2P-selectin binding to affinity-purified PNAd from human tonsils
in vitro. Both activated platelets and L1-2P-selectin tethered to PNAd at a wall
shear stress of 1.6 dyne centered dot cmminus 2 Figure 3A. Once attached, more
than 90% of the tethered platelets rolled continuously. Sticking without
displacement (more than 30 s) or skipping rolling motions (repeated detachment
and reattachment) occurred rarely (5.6 and 4.1%, respectively). Spontaneous
detachment of rolling platelets was not observed. Interactions required Ca2 plus
and were completely inhibited by first treating platelets or L1-2 cells with
antibody to P-selectin (mAb WAPS 12.2), but not with isotype-matched control
antibody to CD31 (mAb PECAM 1.3) (anti-CD31). As observed in vivo, incubation of
PNAd with mAb MECA-79 also reduced P-selectin-mediated rolling Figure 3B. This
antibody recognizes sialomucin-like glycoproteins that are decorated with its
carbohydrate epitope. Members of the sialomucin family mediate L-selectin-dependent
tethering of peripheral blood lymphocytes to purified PNAd [21,22] and are
required for leukocyte adhesion to P-selectin [18,19]. Consistent with these
earlier findings, sialylated, O-glycosylated structures were also essential for
P-selectin:PNAd interactions; O-glycoprotease or neuraminidase treatment of PNAd
abrogated activated platelet attachment in flow.

----------------------------------------------
Figure 3. Purified components of PNAd from human tonsils support in vitro
tethering and rolling of platelets and L1-2P-selectin cells. (A) Platelets (1
times 10 (7) per milliliter) or L1-2 transfectants (1 times 106 per milliliter)
were infused at 1.6 dyne centered dot cmminus 2 over polystyrene plates that
were coated with affinity-purified human tonsil PNAd as described [21]. The
number of adherent cells per field of view (0.69 mm2) was quantitated by video
analysis. C-type lectin dependence of interactions was confirmed by infusing
assay medium containing 5 mM EDTA between consecutive data sets, which resulted
in the release of all bound cells. Antibody inhibition studies were done in
triplicate by preincubating platelets and L1-2 cells with mAb (20 micro gram
/ml; n equals 3). (B) Tethering of thrombin-stimulated platelets is blocked by
treatment of PNAd with O-glycoprotease, neuraminidase, or mAb MECA-79, but not
mAbs to CD34 (shear stress of 1.6 dyne centered dot cmminus 2). Interacting
cells were counted for 3 min within the identical field of view. Antibody
inhibition studies were done by incubating cells or the PNAd substrate with
saturating concentrations of mAb MECA-79 (20 micro gram /ml) or anti-CD34 mAbs
(mixture of 10 micro gram /ml each of mAb 547, 563, 581, and My-10) for 15 min
before each experiment. For some experiments, a pre-analyzed field of PNAd was
incubated with Vibrio cholerae neuraminidase (0.1 U/ml; Calbiochem) in 50 mM
NaH2 PO4 (pH 6.0), 2 mM CaCl2, or O-sialoglycoprotease (0.2 mg/ml; Pasteurella
hemolytica, Cedarlane Lab) in assay medium for 1 hour at room temperature. The
substrate was subsequently washed with assay medium, and activated platelets
were perfused over the identical field of view. Values shown are the means plus
minus SD of three duplicate experiments. (C) Distinct PNAd glycoprotein
constituents support tethering and rolling of platelets in flow. CD34 was
purified from human tonsil PNAd [21], immobilized on plastic (200 ng/ml), and
tested for the ability to support platelet tethering and rolling in flow (1.6
dyne centered dot cmminus 2, n equals 3). Undepleted material and the PNAd
fraction that was depleted of CD34 (approximate 88% depletion) were used for
comparison. Site density and protein concentration of this glycoprotein ligand
was determined by saturation binding of iodinated mAb and capture enzyme-linked
immunosorbent assay, respectively.
----------------------------------------------

To date, PSGL-1, a sialomucin expressed on leukocytes [18,19], is the only known
human P-selectin ligand. It is unclear whether PSGL-1 is also expressed on
endothelium and, in particular, whether PSGL-1 is a component of PNAd, which
consists of several glycoproteins (65 to 200 kD). To show that a glycoprotein
component of PNAd distinct from PSGL-1 can support P-selectin-mediated adhesion,
we examined CD34, a major constituent of murine and human PNAd [21,22]. Purified
CD34 from human tonsil PNAd supported P-selectin-mediated tethering and rolling
as observed for the entire mixture of PNAd glycoproteins Figure 3C. However,
CD34 is not a prerequisite scaffold for carbohydrate presentation to P-selectin,
because its depletion from PNAd [21] resulted in only a 20% decrease in
tethering of platelets compared with nondepleted material under identical
biophysical conditions. This result suggests that other constituents of PNAd
also present P-selectin ligand (or ligands), consistent with the concept of
parallel contribution of multiple PNAd components to L-selectin binding [4,5].

These experiments support an important role for P-selectin in adhesive
interactions of activated platelets with HEVs. They also suggest a mechanism for
platelet-dependent lymphocyte recruitment to HEVs in which activated platelets
bound to PNAd can capture circulating lymphocytes through high-density
expression of P-selectin. As many as 60% of peripheral blood T lymphocytes bear
functional ligands for P-selectin [23]. Thus, it is likely that platelet
P-selectin and endothelial PNAd act in synergy in supporting lymphocyte
recruitment to HEVs through simultaneous interactions with PSGL-1 and L-selectin,
respectively. In vitro evidence supports this hypothesis: low-density binding of
activated platelets to purified PNAd resulted in a threefold increase in the
rolling flux of lymphocytes as compared with lymphocyte binding in the absence
of platelets Figure 4. Furthermore, incubation of lymphocytes with mAb Dreg 200
to L-selectin (20 micro gram /ml) [24] abolished cell attachment to PNAd in the
absence of platelets. In contrast, in the presence of surface bound platelets,
mAb Dreg 200 attenuated lymphocyte attachment only to the level found with
lymphocytes with functional L-selectin. These experiments also demonstrate that
rolling was exclusively mediated by platelets attached to the lymphocyte
surface; antibodies to either platelet P-selectin or lymphocyte PSGL-1 [19]
completely blocked lymphocyte rolling.

----------------------------------------------
Figure 4. Role of activated platelets in lymphocyte rolling on purified PNAd.
Peripheral blood lymphocytes (PBLs, approximate 90% T cells after plastic
adherence) were infused (1 times 106 per milliliter) at a shear stress of 2.1
dyne centered dot cmminus 2 before (first infusion) or after (second infusion)
activated platelets were allowed to accumulate on the PNAd substrate (final
density of 280 plus minus 28 platelets per square millimeter, corresponding to
approximate 0.1% of the PNAd surface area). Three minutes later, the rolling
flux of PBLs in the identical field of view (0.12 mm2) was determined as
described [30]. Some PBLs were first incubated with mAb Dreg 200 (20 micro gram
/ml) and cell attachment to PNAd evaluated in the absence and presence of
surface bound platelets. Simultaneous engagement of L- and P-selectin on
lymphocytes and platelets, respectively, resulted in a threefold increase in PBL
rolling. Blockade of L-selectin function abolished PBL rolling in the absence of
platelets. In contrast, the low number of adherent platelets maintained rolling
of PBLs even without a contribution by L-selectin at levels that were comparable
to rolling of untreated PBLs in the absence of platelets. The requirement for
PSGL-1:P-selectin interactions in platelet-mediated recruitment of mAb Dreg
200-treated T lymphocytes is shown by the inhibitory effects of P-selectin mAb
WAPS 12.2 and the PSGL-1 function blocking mAb PL1 [19], but not mAb PL2
(nonblocking), on T cell adhesion. Error bars represent the SD of three
experiments performed in triplicate.
----------------------------------------------

P-selectin on platelet monolayers that cover thrombogenic surfaces can recruit
circulating leukocytes in vitro and in vivo [16]. However, firmly adherent
platelet monolayers are rarely observed on intact endothelial cells, even in
chronic inflammation. Our findings suggest a more dynamic interplay between
activated platelets and vascular endothelium, characterized by reversible, but
continuous, interactions with much smaller amounts of platelet deposition. Such
interactions may not be apparent in histologic studies of chronic inflammation.
The observation that P- and L-selectin share PNAd as a common endothelial ligand
suggests a unifying mechanism for platelet and lymphocyte recruitment to
lymphoid organs and inflamed tissues where PNAd is expressed. Although it is not
surprising that carbohydrate moieties such as PNAd can bind to more than one
selectin [25], this pathway may allow activated platelets to support a constant
influx of WBCs into chronically inflamed lesions. This may provide an important
synergistic mechanism to enhance the recruitment of pathophysiologically
relevant lymphocyte subsets such as memory cells that express little or no
L-selectin.

Delivery of WBCs by way of activated platelets may not be limited to venules
that express PNAd. P-selectin-mediated platelet adhesion to endothelium may also
occur in acutely inflamed nonlymphoid vascular beds where immunologically
distinct L-selectin ligands have been observed [26]. Moreover, recent in vivo
studies have shown that exposure to atherogenic stimuli such as oxidized low
density lipoprotein or cigarette smoke induces rapid P-selectin-dependent
aggregation and accumulation of leukocytes and platelets in arterioles and
arteries [27]. In light of our present results, it is reasonable to speculate
that activated platelets may have the capacity to deliver leukocytes to vascular
beds such as arteries that may not express selectins or selectin ligands but do
have receptors for other platelet adhesion molecules [28].

REFERENCES AND NOTES

1. T. A. Springer, Cell 76, 301 (1994); [Medline Link] [BIOSIS Previews Link] E.
C. Butcher and L. J. Picker, Science 272, 60 (1996). [Fulltext Link] [Medline
Link] [BIOSIS Previews Link] 

2. W. M. Gallatin, I. L. Weissman, E. C. Butcher, Nature 304, 30 (1983).
[Medline Link] 

3. P. R. Streeter, B. T. N. Rouse, E. C. Butcher, J. Cell Biol. 107, 1853
(1988). [Medline Link] 

4. E. L. Berg, M. K. Robinson, R. A. Warnock, E. C. Butcher, ibid. 114, 343
(1991). [Medline Link] 

5. Y. Imai, M. S. Singer, C. Fennie, L. A. Lasky, S. D. Rosen. ibid. 113, 1213
(1991). 

6. E. L. Berg, L. M. McEvoy, C. Berlin, R. F. Bargatze, E. C. Butcher, Nature
366, 695 (1993). [Medline Link] [BIOSIS Previews Link] 

7. M. B. Lawrence, E. L. Berg, E. C. Butcher, T. A. Springer, Eur. J. Immunol.
25, 1025 (1995). [Medline Link] [BIOSIS Previews Link] 

8. S. A. Michie, P. R. Streeter, P. A. Bolt, E. C. Butcher, L. J. Picker, Am. J.
Pathol. 143, 1688 (1993); [Medline Link] [BIOSIS Previews Link] J. P. Turunen et
al., Eur. J. Immunol. 24, 1130 (1994). [Medline Link] [BIOSIS Previews Link] 

9. C. E. Collins, M. R. Cahill, A. C. Newland, D. S. Rampton, Gastroenterology
106, 840 (1994). [Medline Link] [BIOSIS Previews Link] 

10. ``Mice were anesthetized by intraperitoneal injection (0.2 ml) of a mixture
of ketamine (5 mg/ml) and xylazine (1 mg/ml), and catheters were inserted into
the right jugular vein (for injection of anesthesia and mAb), the left carotid
artery (for blood pressure monitoring), and the right femoral artery (for
retrograde injection of fluorescently labeled cells). Subsequently, the left
subiliac LN was microsurgically prepared for intravital microscopy as described
(U. H. von Andrian, Microcirculation, in press). The preparation was transferred
to a custom-built intravital microscope (IV-500; Micron Instruments, San Diego,
CA). For microscopic observation of endogenous microvascular WBC behavior, a
bolus injection of sterile Ringer's lactate (10 ml per kilogram of body weight)
containing the fluorescent dye rhodamine 6G (2 mg/ml; Molecular Probes) was
given intravenously. Rhodamine 6G stains nuclei and mitochondria in living
cells, thereby permitting visualization of interacting and freely flowing
leukocytes in LN HEVs by fluorescent stroboscopic epi-illumination through a 20
times Zeiss objective. Video scenes were recorded through a low-lag SIT camera
(Dage MTI) and stored on Hi 8 videotape for off-line analysis of rolling
fractions, rolling velocities, and microvascular hemodynamics as described [U.
H. von Andrian et al., Am. J. Physiol. 263, H1034 (1992) [Medline Link]]. After
a 3- to 5-min control recording of HEVs, some animals were treated with mAb
Mel-14 [100 micro gram per mouse, intravenously (iv)], and leukocyte behavior in
the same vascular bed was recorded 15 min later.'' 

11. R. A. Warnock, E. C. Butcher, U. H. von Andrian, unpublished data. 

12. ``Human platelets were purified from the blood of healthy donors by standard
procedures. Aliquots of platelets were stimulated with human thrombin (Sigma;
0.5 U/ml for 5 min) or TRAP (Bachem; 25 micro meter for 5 min) in the presence
of 2 mM Arg-Gl-Asp-Ser (RGDS) (Peninsula Labs) to minimize aggregation. Excess
thrombin was neutralized with hirudin (Sigma; 0.5 U/ml, 5 min, 37 degrees C)
before injection into the animal. Activation status of platelets was confirmed
by flow cytometry of P-selectin expression. Aliquots of activated platelets were
incubated with mAb WAPS 12.2 (20 micro gram /ml) [M. A. Jutila et al., J.
Immunol. 153, 3917 (1994) [Medline Link] [BIOSIS Previews Link]] or control mAb
PECAM-1 [F. Liau et al., J. Exp. Med. 182, 1337 (1995)]. In experiments
requiring direct visualization of circulating platelets, cells were fluorescently
labeled with 2 prime,7 prime-bis-(2-carboxyethyl)-5 (and-6) carboxyfluorescein
(BCECF; Molecular Probes) as described [T. G. Diacovo, A. R. de Fougerolles, D.
F. Bainton, T. A. Springer, J. Clin. Invest. 94, 1243 (1994)]. Fluorescent
platelets were injected through the femoral artery catheter into the arterial
bloodstream supplying the node. In some experiments (n equals 3), mAb MECA-79
(100 micro gram per mouse, iv) was injected after assessment of rolling in
untreated HEVs. After allowing mAb binding to PNAd for 10 min, freshly activated
platelets or transfectants were again injected to assess the effects of mAb
treatment.'' 

13. T. G. Diacovo and U. H. von Andrian, data not shown. 

14. G. I. Johnston, R. G. Cook, R. P. McEver, Cell 56, 1033 (1989). [Medline
Link] 

15. T. N. Mayadas, R. C. Johnson, H. Rayburn, R. O. Hynes, D. D. Wagner, ibid.
74, 541 (1993). [Medline Link] [BIOSIS Previews Link] 

16. K. L. Moore and L. F. Thompson, Biochem. Biophys. Res. Commun. 186, 173
(1992); [Medline Link] S. M. Buttrum, R. Hatton, G. B. Nash, Blood 82, 1165
(1993); [Medline Link] [BIOSIS Previews Link] T. Palabrica et al., Nature 359,
848 (1992); [Medline Link] [BIOSIS Previews Link] T. G. Diacovo, S. J. Roth, J.
Buccola, D. F. Bainton, T. A. Springer, Blood 88, 146 (1996). [Medline Link]
[BIOSIS Previews Link] 

17. P. S. Frenette, R. C. Johnson, R. O. Hynes, D. D. Wagner, Proc. Natl. Acad.
Sci. U.S.A. 92, 7450 (1995). [Medline Link] [BIOSIS Previews Link] 

18. D. Sako et al., Cell 75, 1179 (1993). [Medline Link] [BIOSIS Previews Link] 

19. K. L. Moore et al., J. Cell Biol. 128, 661 (1995). [Medline Link] [BIOSIS
Previews Link] 

20. ``A cDNA encoding full length P-selectin was isolated by polymerase chain
reaction from the human megakaryocyte cell line HEL 92. 17 (American Type
Culture Collection). Sequence fidelity was verified by restriction analysis and
sequencing [14]. The P-selectin cDNA was subcloned into pMRB101 to stably
transfect murine L1-2 pre-B cells as described [4,30]. High P-selectin-expressing
subclones were grown in RPMI 1640 containing 10% fetal calf serum and standard
supplements. For in vivo experiments, transfectants were labeled with BCECF,
resuspended in RPMI (1 times 107 per milliliter), and injected into mice as
described [12].'' 

21. K. D. Puri, E. B. Finger, G. Gaudemack, T. A. Springer, J. Cell Biol. 131,
261 (1995). [Medline Link] [BIOSIS Previews Link] 

22. S. Baumhueter et al., Science 262, 436 (1993); [Fulltext Link] [Medline
Link] L. A. Lasky et al., Cell 69, 927 (1992). [Medline Link] 

23. T. Diacovo et al., J. Exp. Med. 183, 1193 (1996). [Medline Link] [BIOSIS
Previews Link] 

24. T. K. Kishimoto, M. A. Jutila, E. C. Butcher, Proc. Natl. Acad. Sci. U.S.A.
87, 2244 (1990). [Medline Link] 

25. C. Foxall et al., J. Cell Biol. 117, 895 (1992); [Medline Link] R. E. Mebius
and S. R. Watson, J. Immunol. 151, 3252 (1993). [Medline Link] [BIOSIS Previews
Link] 

26. U. H. von Andrian et al., Blood 82, 182 (1993); K. E. Norgard-Sumnicht, N.
M. Varki, A. Varki, Science 261, 480 (1993); [Fulltext Link] [Medline Link] K.
Ley, T. F. Tedder, G. S. Kansas, Blood 82, 1632 (1993). [Medline Link] [BIOSIS
Previews Link] 

27. H.-A. Lehr et al., Lab. Invest. 71, 380 (1994); [Medline Link] [BIOSIS
Previews Link] H.-A. Lehr, B. Frei, K.-E. Arfors, Proc. Natl. Acad. Sci. U.S.A.
91, 7688 (1994). [Medline Link] [BIOSIS Previews Link] 

28. B. Savage, E. Saldivar, Z. M. Ruggeri, Cell 84 (1996). 

29. A. R. Pries, Int. J. Microcirc. Clin. Exp. 7, 327 (1988); [Medline Link] K.
Ley and P. Gaehtgens, Circ. Res. 69, 1034 (1991). [Medline Link] 

30. U. H. von Andrian, S. R. Hasslen, R. D. Nelson, S. L. Erlandsen, E. C.
Butcher, Cell 82, 1 (1995). 

31. ``We thank B. Fors, E. Finger, and G. Cheng for technical assistance; E.
Butcher, P. Newman, D. Wagner, G. Gaudernack, and R. McEver for contributing
reagents; and K. Gauvreau for aiding in statistical analysis. Supported by
National Institutes of Health grants HL48675 and HL54936.''

----------------------------------------------
Accession Number: 00007529-199607120-00017
----------------------------------------------