,
* Divisions of Infectious Diseases and Allergy and Inflammation, Department of Medicine, and Charles A. Dana Research Institute and Harvard-Thorndike Laboratory, Beth Israel Deaconess Medical Center; and
Division of Rheumatology, Immunology and Allergy, Brigham and Women’s Hospital,
Harvard Medical School, Boston, Massachusetts
Correspondence: Dr. Anne Nicholson-Weller, M.D., Dana 617, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215. E-mail: anichols{at}caregroup.harvard.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
d/ß2. Surprisingly,
blocking anti-LFA-1 mAb had no effect on superoxide production.
Consistent with this observation, immobilized, purified ICAM-2, a
specific counter-receptor for LFA-1, did not support the
adhesion-dependent phase ofsuperoxide production. Thus, PMN treated
with C5a used signals via CR3, P150/95, and
d/ß2, but not LFA-1, to support superoxide production.
LFA-1 has been shown by others to mediate most of the adhesion
necessary for transendothelial migration in vivo. The
inability of LFA-1 ligation to stimulate superoxide production may be
an important means of preventing blood-vessel damage when PMN migrate
across the endothelium.
Key Words: human • complement • free radicals • inflammatory mediators
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-)-triggered
O2- production, the ß2 integrin
provides an essential second signal for superoxide production
[7
, 40
]. Similarly, ß2
integrins are required for PMN-mediated phagocytosis also
[8
, 9
].
To determine whether ß2 integrins play a role in PMN
O2- production triggered by a soluble
complement agonist, we studied the effects of blocking mAbs to
ß2 integrins on C5a-triggered
O2- production. In our assay, the C5a was
added to PMN in suspension, and the cells were allowed to settle by
gravity onto the surface of microtiter wells that had been coated with
specific proteins. Kinetic assays revealed two phases of
O2- production by these C5a-stimulated PMN.
The first was characterized by a high rate of
O2- production of short duration that was
independent of ß2 integrins. The second phase, which
began at about 10 min as the cells settled on the surface of the
microtiter well, featured a lower rate of O2-
production, but production was sustained over 30 min. This phase was
dependent entirely on selective ß2 integrin signaling,
namely CR3, P150,95, and probably d/ß2,
but not lymphocyte function-associated antigen-1 (LFA-1). The second
phase was independent of further G protein-coupled signal transduction,
because the cells were refractory to additional C5a. The
ß2 integrin-dependent phase accounted for >50% of the
total O2- produced by the cells. We conclude
that ß2 integrin-mediated signaling is essential for
sustained C5a-triggered O2- production by
adherent PMN. Finally, we propose a model where the selective inability
of LFA-1 to mediate this signal when the PMN are stimulated with a
chemotactic agent, in contrast to a stimulus such as C1q where LFA-1 is
able to support adhesion-dependent O2-
production, provides a mechanism by which PMN can adhere to and cross
the endothelium without damaging it by sustained release of
O2-.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Antibodies and preparation of F(ab')2 fragments
Sources of mAbs are noted in Table I
[10
11
12
13
14
15
16
].
|
View this table: [in a new window] |
Table 1. Sources of Monoclonal Antibodies
|
R engagement, which is known to modify
O2- production [17
].
Human proteins
C1q was isolated from fresh human serum by fractional euglobulin
precipitation and gel filtration chromatography, as described
[18
]. C1q purity was analyzed by SDS-PAGE, and the
functional activity of the C1q was assessed by hemolytic assay
[19
]. Recombinant human C5a was purchased from Sigma.
Recombinant soluble intercellular adhesion molecule-1 (ICAM-1) was
kindly provided by Dr. Robert Rothlein, Boehringer Ingelheim, Pearl
River, NY. ICAM-2 was purified from human platelets and was homogeneous
as judged by SDS-PAGE with silver staining and by functional assays
(unpublished results).
Isolation of human PMN
Venous blood (20–25 ml) was collected from normal volunteers in
a syringe containing 2.0 ml acidified citrate (pH 5.2) and 8.0 ml
dextran (6% Dextran 70 in 0.9% NaCl; Kendall McGaw Laboratories,
Irvine, CA). After the erythrocytes had sedimented for 1 h at room
temperature, the leukocyte-rich supernatant was removed, added to a new
tube, underlaid with 15 ml Ficoll-Paque (Pharmacia Biotech, Piscataway,
NJ), and centrifuged for 25 min at 1000 g at 4°C. The
cells were maintained at 4°C during all subsequent steps. PMN (bottom
pellet) were resuspended in 1 ml HBSS=, and contaminating
erythrocytes were lysed with hypotonic saline. Finally, 95–98% pure
PMN were resuspended in HBSS= supplemented with 0.1%
ovalbumin. Cells were used within 2–3 h of venipuncture. Just before
the start of the assay, the cells were resuspended in
HBSS++ and warmed to 37°C.
Microplate assay of O2- production
C1q, ICAM-1, or ICAM-2 was added to 96-well plates (Immulon-2HB,
Dynex Technologies, Chantilly, VA), was allowed to bind at 37°C for
1 h or at room temperature for 2–3 h, and then the wells were
washed twice with HBSS=. All wells, whether or not specific
ligands had been added previously, were blocked with gelatin
HBSS= (0.5% gelatin in HBSS=) for 1 h.
Finally, the wells were washed twice in HBSS= before
starting the assay. All reagents were equilibrated to 37°C. A typical
reaction assay was 150 µl HBSS++ containing
80 µM cytochrome c, 4 x 105 PMN, and the reaction
was started by the addition of C5a (33 nM). Immediately after the assay
was started, the microplate was placed in the reading chamber, and
absorbance was read at 550 nm in the kinetic mode with intermittent
agitation at 37°C. O2- generation was
calculated from the superoxide dismutase-inhibitable reduction of
ferricytochrome c in a microplate assay [20
] using a
Thermomax kinetic microplate reader (Molecular Devices, Sunnyvale, CA).
In all experiments, fMLP (1.0 µM) and PMA (100 nM) were included as
positive controls (triplicates or more). Calculations were based on an
extinction coefficient of 20 AU/µmol • cm
[20
]. In more than 10 experiments, we used parallel
assays in the presence of superoxide dismutase, and the color change
was inhibited greater than 95%, an indication that nearly all of the
cytochrome c reduction was a result of
O2- production.
Adhesion assay
Tip-plate adhesion assays for PMN adherence to purified ICAMs
were performed as described [21
] on plastic petri dishes
(Nunc Lab Tek, Fisher Scientific, Medford, MA).
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (19K): [in a new window] |
Figure 1. a) Immobilized C1q stimulated O2- production
by PMN. Microtiter wells were precoated with C1q (100 µg/ml) in HBSS,
washed, and blocked, and PMN in HBSS++
(4x105/well) were added to start the reaction. The rate of
cytochrome c reduction (mOD/min after each 5 min interval) was
recorded, and O2- production was calculated as
described in Materials and Methods. C1q stimulated
O2- production (•—•); unstimulated PMN
were added to wells that were blocked with gelatin HBSS=
( —). The mean ± SE, n = 4–6,
was plotted. Results are representative of 10 experiments. b) Soluble
C5a stimulated O2- production by PMN.
Microtiter wells were precoated with 0.5% gelatin and washed with
HBSS. PMN and C5a (33 nM) were then added simultaneously to start the
reaction and O2- assay as described above. The
"Buffer" negative control was unstimulated cells in gelatin-blocked
wells (—). The mean ± SE, n =
5–6, was plotted. Results are representative of 10 experiments.
|
 View larger version (20K): [in a new window] |
Figure 2. Modulation of the C5a response by immobilized ICAM-1: Microtiter wells
were precoated with recombinant ICAM-1 (3 µg/ml) for 2 h at
37°C and blocked with 0.5% gelatin. The assay was performed
otherwise as in Fig. 1
. The mean ± SE,
n = 4–6, was plotted. ICAM-1 alone in any
concentration used did not stimulate O2-
production; whereas the combination of C5a plus ICAM-1 stimulated a
sustained adhesion-dependent phase of O2-
production. The "Buffer" negative control was unstimulated cells in
gelatin-blocked wells (—). This experiment was repeated twice
with similar results.
|
d/ß2 (CD11d/CD18). LFA-1, CR3, and P150/95
are the recognized counter-receptors for ICAM-1 [13
,
21
, 25
]. Preincubation of PMN with
F(ab')2 fragments of the anti-CD18-blocking mAb, TS1/18,
inhibited almost completely the second phase of
O2- production but had no effect on the first
phase (Fig. 3
). Similar results were obtained with another blocking
F(ab')2, anti-CD18 mAb, IB4 (unpublished results). There
was no inhibition seen with a nonblocking, anti-CD18 mAb
F(ab')2 preparation of CBR-LFA 1/7 (Fig. 3)
.
 View larger version (19K): [in a new window] |
Figure 3. Pretreatment of PMN with blocking anti-ß2 integrin-CD18
F(ab')2 fragments inhibited the adhesion-dependent phase of
O2- production that was stimulated by C5a plus
ICAM-1. Microtiter wells were coated with recombinant ICAM-1 (3.0
µg/ml) or gelatin HBSS=. PMN were pretreated with 10
µg/ml F(ab')2 fragments of blocking (TS1/18) or
nonblocking (CBR LFA-1/7) mAbs, or buffer for 15 min at room
temperature. Subsequently, the cells were challenged with C5a (33 nM)
and O2- production assayed. Unstimulated,
buffer-treated PMN were added to wells that were blocked with gelatin
HBSS= (—). The mean values ± SE,
n = 4–6, were plotted from a representative
experiment. In a repeat experiment, IB4, another blocking anti-CD18
mAb, inhibited 95%, and TS1/18 inhibited 91%; CBR LFA-1/7 did not
inhibit O2- production.
|
 View larger version (19K): [in a new window] |
Figure 4. The effect of anti-CD11b F(ab')2 fragments on
C5a/ICAM-1-stimulated, second-phase O2-
production in PMN. Microtiter wells were coated with recombinant ICAM-1
(3.0 µg/ml) or gelatin HBSS=. PMN were pretreated with 10
µg/ml F(ab')2 fragments of blocking anti-CD11b CBR M1/29
( —), or LM2/1 ( —), or nonblocking anti-CD11b (CBR M1/20)
mAbs. Subsequently, the cells were challenged with C5a (33 nM), and
O2- production was assayed. Unstimulated,
buffer-treated PMN were added to wells containing immobilized ICAM-1
( —) or were blocked with gelatin HBSS= (—).
Both of the blocking mAbs inhibited O2-
production by 50%, and the nonblocking mAb yielded results that were
similar to the non-mAb-treated group (unpublished results). The mean
values ± SE, n = 5–6, were plotted
from a single experiment. This experiment was repeated with similar
results.
|
 View larger version (27K): [in a new window] |
Figure 5. a) Effect of anti-CD11a F(ab')2 fragments on
C5a/ICAM-1-stimulated O2-. Microtiter wells
were precoated with recombinant ICAM-1 (3 µg/ml) for 2 h at
37°C and blocked with 0.5% gelatin. PMN were pretreated with 10
µg/ml F(ab')2 fragments of blocking (TS1/22), or
nonblocking anti-CD11a (TS2/4), mAbs or buffer. Subsequently, the cells
were challenged with C5a (33 nM) and O2-
production assayed. Unstimulated, buffer-treated PMN were added to
wells containing immobilized ICAM-1 (—) or blocked with gelatin
HBSS= (—). The blocking and nonblocking mAbs did not
inhibit the rate of O2- production
significantly. The mean values ± SE,
n = 5–6, were plotted from a single experiment. This
experiment was repeated twice with similar results. b) ICAM-1, but not
ICAM-2, supported the adhesion-dependent phase of
O2- production in C5a-treated PMN. Microtiter
wells were precoated with recombinant ICAM-1 (3 µg/ml), or a dilution
of purified platelet ICAM-2 that yielded an equivalent number of active
binding, for 2 h at 37°C. Subsequently, the wells were blocked
by 0.5% gelatin. These conditions yielded 450–500 ICAM
molecules/mm2 plastic surface (see Materials and Methods).
The cells were stimulated with 33 nM C5a, and the
O2- production was measured. The mean ±
SE, n = 4–6, was plotted. PMN stimulated
with C5a + ICAM-1 (•—•) but not C5a + ICAM-2 (—)
supported O2-. When the PMN were unstimulated,
ICAM-1 ( —) or ICAM-2 (—) in any concentration used did
not stimulate O2- production. This experiment
was repeated with similar results.
|
 View larger version (20K): [in a new window] |
Figure 6. Purified PMN adhere to ICAM-2 in an LFA-1-dependent manner. ICAM-1 and
ICAM-2 were coated on demarcated spots on 35 mm plastic petri dishes
for 2 h at room temperature. The plates were blocked and washed
with HBSS++ containing 10 mM HEPES, pH 7.3, and
0.5% heat-inactivated human serum albumin (HSA). C5a-treated PMN were
added (1.5x106 total in 1.5 ml), and the plates were left
undisturbed for 4 min at room temperature. Nonadherent cells were
removed, and the plates were washed with a transfer pipet. Adherent
cells were counted in three locations for each spot using a calibrated
ocular grid. Results are shown as mean ± SD from a
representative experiment.
|
d/ß2 on PMN binding
to ICAM-1. First, the presence of blocking anti-CD18 mAb fully
inhibited O2- production by PMN, although
anti-CR3 mAb inhibited it by only 50%, and anti-LFA-1 didn’t inhibit
at all. Second, there was specific binding of PMN to immobilized ICAM-1
even when PMN were treated with the combination of blocking mAbs to
LFA-1 and CR3 (Fig. 6) . Consequently, we tested whether a blocking
anti-CD11c mAb was able to inhibit the adhesion-dependent phase of
superoxide production and found partial inhibition when the C5a-treated
PMN were assayed in wells coated with ICAM-1. Although we lacked the
reagents to test directly for the participation of
d/ß2, a mixture of blocking anti-CD11a, anti-CD11b,
and anti-CD11c was unable to inhibit as completely as anti-CD18.
Because there was variability between donors, results from PMN isolated
from two different donors are shown (Fig. 7a
and b
). Thus, these data suggest that all four
ß2 integrins may bind ICAM-1, and all but LFA-1 may
mediate the adhesion-dependent signal for O2-
production by C5a-treated PMN.
 View larger version (30K): [in a new window] |
Figure 7. The effect of anti-CD11c alone and in combination with anti-CD11a and
anti-CD11b on C5a/ICAM-1-stimulated, second-phase
O2- production in PMN, representing
experiments performed with cells from two different donors. Microtiter
wells were coated with recombinant ICAM-1 (3.0 µg/ml) or gelatin
HBSS=. PMN were pretreated with 10 µg/ml
F(ab')2 fragments of blocking anti-CD11c (3.9) or a
combination of 10 µg/ml each of anti-CD11c (3.9) + anti-CD11a
(TS1/22) + anti-CD11b mAb (CBR M1/29); or blocking anti-CD18
(TS1/18) mAb (—) or buffer (—). Subsequently, each group
of cells was stimulated with C5a (33 nM). Unstimulated, buffer-treated
PMN were added to wells blocked with gelatin HBSS=
(—). The mean values ± SE, n =
5–6, were plotted from a single experiment. This experiment was
repeated with similar results. In both donors, anti-CD11c inhibited a
fraction of the C5a-stimulated second phase of superoxide. Of note, the
combination of blocking anti-CD11a, anti-CD11b, and anti-CD11c still
did not block as efficiently as anti-CD18.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two phases of oxygen radical formation have been noted previously in PMN stimulated with C5a and analyzed in a chemiluminescence assay [22 ]. In this assay, the oxidation of luminol, which can be taken up in the cell, is followed in a scintillation counter [27 ]. In an experiment where the exogenous luminol was washed away from the luminol-loaded cells, the subsequent addition of C5a stimulated only the second peak of chemiluminescence. This was interpreted to mean that the first peak of chemiluminescence was detecting oxygen radicals released outside the cell, whereas the second, prolonged phase of oxygen-radical formation was being generated within the cell [22 ]. In our studies, the delay in the onset of the C5a-induced, second adhesion-dependent phase of O2- production may be a result of the time necessary for the cells to settle onto the ICAM-1. This feature of our experimental assay allowed the separate analysis of the two distinct phases of O2- production. The delayed and prolonged C5a-stimulated, adhesion-dependent O2- production has not been well-studied. However, when PMN are stimulated with fMLP, which, like C5a, uses a G protein-coupled receptor, ß2 integrins are activated and subsequently recruit members of the Src family tyrosine kinases, hck and c-fgr (reviewed in [28 ]).
The specificity of the blocking antibodies used in our experiments was rigorously controlled. We used exclusively F(ab')2 fragments of predominantly isotype-matched antibodies (Table 1 ) and found that only adhesion-blocking mAb inhibited O2- production. mAbs that bound to CD18 at a site away from the ligand-binding site and do not block ß2 integrin-mediated adhesion ("nonblocking" mAb) did not inhibit O2- production (Fig. 3 4 5a) . This indicates that the anti-ß2 mAbs that inhibited PMN O2- generation in vitro must have blocked a ligand-receptor interaction, as opposed to mediating a dominant negative signal for O2- generation because of cross-linking of the integrin by the bivalent F(ab')2 fragments.
The observation that blocking anti-CD18 mAb completely inhibited the
adhesion phase of C5a-stimulated O2-
production, although anti-CR3 mAb afforded only partial inhibition
(Fig. 4)
, and anti-LFA-1 gave no inhibition, strongly suggested that
additional ß2 integrins must be involved. Consistent with
this finding, there was significant PMN adhesion to ICAM-1 despite
treatment of the cells with the combination of anti-CR3 +
anti-LFA-1 (Fig. 6) . PMN express two other ß2 integrins,
namely P150,95 and d/ß2, and
P150,95 has been shown to bind ICAM-1 [13
]
and thus was the leading candidate. Blocking anti-CD11c directly
demonstrated a role for P150,95, but the incomplete
inhibition provided by a mixture of anti-CD11a, anti-CD11b, and
anti-CD11c—compared with the complete inhibition seen with
anti-CD18—inferred a role for d/ß2 (Fig. 7a
and 7b)
.
The integrin d/ß2, when assessed as a recombinant
protein expressed on Chinese hamster ovary cells (CHO), bound to wells
coated with ICAM-1-Fc reportedly twice as well as to plastic coated
with bovine serum albumin (BSA) and better than the control CHO
transfectants [29
]. Hence, although the binding of
d/ß2 to ICAM-1 is much less robust than to ICAM-3, it
is possible that all four ß2 integrins may recognize
ICAM-1. Furthermore, it is possible that immobilized ICAM-1 triggers a
ligand-induced binding site on d/ß2 expressed on PMN,
as it does on lymphocyte LFA-1 [30
], and the
ligand-activated d/ß2 may then bind well to ICAM-1.
This has not been tested to our knowledge. Alternatively, although only
ICAM-1 was immobilized in the wells, it is possible that the PMN bind
to adjacent cells via ICAM-3 and d/ß2. The PMN are
observed to be touching each other in the periphery of the microtiter
wells, and PMN express high levels of ICAM-3. This seems less likely
because ICAM-3 binding to d/ß2 should occur in the
absence of immobilized ICAM-1; however, adhesion-dependent production
of O2- does not occur under these conditions.
An additional possibility is that LFA-1 binding to ICAM-1 may direct
the upregulation of CR3 and P150,95 from intracellular
stores, as suggested previously [14
], and these
("unblocked") integrins may mediate adherence and signaling.
The ability of C5a-treated PMN to use LFA-1 for adherence but not for the production of superoxide was unexpected. This was shown first by the inability of blocking LFA-1 mAb to affect O2- production (Fig. 5a) . Confirmation that LFA-1 was not involved came from the inability of C5a-stimulated PMN to generate an adhesion-dependent phase of O2- production when ICAM-2 was immobilized on the plastic instead of ICAM-1(Fig. 5b) . Although some controversy exists [31 ], most studies indicate that ICAM-2 binds only LFA-1 among the ß2 integrins [21, 26, 32; and unpublished results]. Our results are also consistent with ICAM-2 serving as the specific counter-receptor for LFA-1 (Fig. 6) .
During the isolation procedure, PMN became sufficiently activated for
CR3 and LFA-1 to mediate adhesion (unpublished data). However, this
activation alone was not sufficient to induce
O2- production (Fig. 1b)
; an additional signal
from C5a in our experiments was necessary. Data from other laboratories
indicate that C5a stimulation should enhance LFA-1 avidity
[32
]. In contrast to the results using C5a as the
agonist, when CR3 and LFA-1 are activated by immobilized C1q, both
integrins participate in signaling for O2-
production, confirming that the LFA-1 downstream signaling pathway for
O2- production does exist in PMN
[40
]. Thus, the most likely explanation is that,
although LFA-1 is activated by C5a, there is a selective inability of
LFA-1 to signal the C5a-treated PMN for the second phase of
O2- production. This may be the first example
of PMN selectively using a downstream signaling pathway from activated
CR3, P150/95, and d/ß2 but not LFA-1. This
observation parallels the activation-induced upregulation from PMN
internal stores of CR3, P150/95, and d/2
but not LFA-1. There is precedent for the selective activation of
ß2 integrins for adhesion in monocytes
[33
] and eosinophils [34
], and in PMN,
there are data that indicate CR3 can be activated by a
phosphotidyinositol (PI) 3-kinase-dependent or PI 3-kinase-independent
pathway [35
]. These findings in aggregate suggest that
the cellular interactions of CR3, P150/95, and
d/ß2 on PMN for "inside-out" and "outside-in"
signaling are fundamentally different from those of LFA-1.
There are important functional implications for the differential
activation of integrins by C5a. In responding to a tissue infection,
PMN must contact and move through the endothelium without damaging the
endothelial cells with released O2-. The
C5a-induced surface expression of P-selectin [36
] on the
endothelial cell has the potential to block CR3-ICAM-1 binding-induced
O2- production [37
]. Although
C5a-stimulated PMN would activate LFA-1 and CR3 for transendothelial
migration, the higher density of ICAM-2 compared with ICAM-1 on the
initially resting endothelial cells [38
] might
selectively engage LFA-1, which would not support
O2- production. This hypothesis is supported
by recent in vivo studies, showing that in response to a
local TNF- stimulus, the LFA-1 knockout mouse and the CD18 knockout
mouse show comparable abnormally low levels of PMN in the tissue,
whereas the CR3 knockout mouse had higher than normal numbers of
migrating PMN [39
]. Thus, LFA-1 is the dominant
ß2 integrin involved in the adhesion necessary for
transendothelial migration. In summary, in addition to the previously
described inhibitory role for P-selectin, our description of the
ability of PMN to use LFA-1 for transendothelial migration, but not for
O2- production, is a second mechanism that may
prevent PMN-mediated endothelial damage when these cells migrate across
the vessel wall.
Received December 9, 1999; revised May 30, 2000; accepted May 31, 2000.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
dß2, binds preferentially to ICAM-3 Immunity 3,683-690[Medline] subunit cytoplasmic domains Mol. Biol. Cell 10,861-873This article has been cited by other articles:
 |
|
 |
|
  N. Dumont, P. Bouchard, and J. Frenette Neutrophil-induced skeletal muscle damage: a calculated and controlled response following hindlimb unloading and reloading Am J Physiol Regulatory Integrative Comp Physiol, December 1, 2008; 295(6): R1831 - R1838. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Hardev Ramandeep Singh Girn, S. Ahilathirunayagam, A. I. D. Mavor, and S. Homer-Vanniasinkam Reperfusion Syndrome: Cellular Mechanisms of Microvascular Dysfunction and Potential Therapeutic Strategies Vascular and Endovascular Surgery, September 1, 2007; 41(4): 277 - 293. [Abstract] [PDF]
|
|
|
|
 |
|
 L. K. Cherry, K. S. C. Weber, and L. B. Klickstein A Dominant Jurkat T Cell Mutation That Inhibits LFA-1-Mediated Cell Adhesion Is Associated with Increased Cell Growth J. Immunol., December 1, 2001; 167(11): 6171 - 6179. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
