| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Department of Biological Sciences, Wayne State University, Detroit, Michigan 48202, USA; and
Department of Ophthalmology, University of Michigan, Ann Arbor, Michigan, USA
1Correspondence: Department of Biological Sciences, Wayne State University, Detroit, MI 48202, USA. E-mail: hpetty{at}biology.biosci.wayne.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: transmigration • metalloproteinase • HT1080 cells • T-47D cells • gelatin-nylon mesh matrix
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Invasiveness requires several distinct cellular
functions involved in tumor cell motility, including adhesion,
detachment, and pericellular proteolysis. Metastatic cells display
increased production of proteolytic enzymes including lysosomal
hydrolyses, collagenases, and plasminogen activator (2)
.
The expression of certain cell surface protease receptors is also
increased on metastatic cells (3)
. Thus, the nature of
matrix proteolysis during tumor cell migration is central to
understanding the invasive cell phenotype (4
5
6
7
8)
.
A variety of methods have been developed to explore tumor cell
invasiveness (9)
. In vitro assays of tumor cell
invasiveness include penetration of the urinary bladder wall (10
, 11)
, human amnion (12
, 13)
, chorioallantoic
membrane (11
, 14)
, blood vessels (11)
, and
lens capsule (15)
. Reconstituted basement membrane
components have also been used to assay invasive properties of tumor
cells. Such assays typically involve a layer of Matrigel atop a porous
filter, which is then placed in a Boyden chamber. The invasive
potential of tumor cells placed in the Boyden chamber can be
quantitated by counting tumor cells across the filter
(16)
. However, these techniques do not allow for direct
optical microscopy of a sample nor do they allow for the detection of
biochemical reactions occurring during invasion. We developed a novel
3-dimensional model based on a supported gelatin layer that more
closely resembles the physiological environment and permits
visualization of cell migration pathways and biochemical reactions via
optical microscopy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. The T-47D and T47 A2–1 cell lines were the generous gifts of Dr. Robert Todd (Division of Hematology and Oncology, University of Michigan School of Medicine, Ann Arbor, Mich.). They were grown at 37°C in RPMI1640 (Life Technology) supplemented with 10% FBS (Summit), 1% PSA (Life Technology), and 0.2 IU/ml bovine insulin (Sigma Chemical Co, St. Louis, Mo.). Cells were passaged as described above.
DiI labeling
Labeling tumor cells by DiI
(1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine; Molecular
Probes, Eugene, Oreg.) was performed according to the method of
Kuriyama et al. (18)
with minor modifications. In brief,
DiI was prepared as a 2.0 mg/ml solution in DMSO. The DiI solution was
routinely sonicated before application to cells. Cells were detached by
treatment with trypsin-EDTA, suspended with Hank’s balanced salt
solution (HBSS) at a concentration of 5 x
105 cells/ml, and stained by incubating with DiI
(final concentration is 20 µg/ml) for 1 h at 37°C with gentle
agitation at 15 min intervals. After staining, the cells were washed
twice by centrifugation and seeded onto the surfaces of the
gelatin-nylon mesh.
Nylon membrane preparation
Nylon mesh (125 µm pores, 80% open) cut in 12 mm squares was
rinsed in 75% ethanol for 5 min. A 6% to 10% gelatin solution was
prepared by boiling gelatin in HBSS. To avoid denaturation of the DQTM
collagen, the solution was allowed to cool to 
42°C before the
addition of this compound. Ten microliters of cell attachment factor
(Cell Systems Corporation, Kirkland, Wash.) was placed on a coverslip
for 15 min at 42°C in an incubator. After drying, a sterile nylon
mesh was placed on the coverslip. Fifty microliters of 6 to 10%
gelatin solution containing 10% fetal bovine serum (FBS) and
Bodipy-BSA (at a final concentration of 1 mg/ml) or DQTM collagen (at a
final concentration of 1 mg/ml) were placed on the nylon mesh and
allowed to harden in a refrigerator at 4°C for 10 min. The
gelatin-nylon mesh matrix on the coverslip was then placed at room
temperature.
Transmigration of tumor cells
HT1080 cells were detached from culture flasks by treatment with
trypsin-EDTA. After detachment, the HT1080 cells were resuspended in
culture medium supplemented with 10% FBS for 10 min at room
temperature to inactivate any remaining trypsin activity. After
centrifugation at 1000 rpm for 5 min, HT1080 cells were seeded onto the
surfaces of the gelatin-nylon mesh matrices in 50 µl of 6% gelatin
solution containing 10% FBS. They were then incubated for 2 to 4 h at 37°C in a 5% CO2 incubator. The
gelatin-nylon mesh matrices and coverslips were then mounted on glass
slides for microscopic observation.
Inhibition of the collagenase activity
To inhibit metalloproteinase activity, 1,10 phenanthroline
(Eluka Chemical Corp., Ronkonkoma, N.Y.) was used (19)
.
During preparation of the gel-nylon mesh matrix, 1,10 phenanthroline
was incorporated into the matrix at a final concentration of 10 mM.
Control gels contained only ethanol, the solvent used for 1,10
phenanthroline. HT1080 cells were seeded and incubated as described
above. After 2 h of incubation at 37°C, we counted the number of
cells with associated pericellular fluorescence in four high-power
microscopic fields.
Inhibition of urokinase-type plasminogen activator
To inhibit urokinase-type plasminogen activator, PAI-I
(plasminogen activator inhibitor I; American Diagnostica Corporation,
Joplin, Mo.) was used. During preparation of the gel-nylon mesh matrix,
PAI-I was incorporated into the matrix at a final concentration of 0.05
mg/ml. Control gels contained only HBSS, the solvent used for PAI-I.
HT1080 cells were seeded and incubated as described above. After 2 h of incubation at 37°C, we counted the number of cells with
pericellular fluorescence in four high-power microscopic fields.
Fluorescence Microscopy and Image Reconstruction
To visualize proteolytic activity, Bodipy-BSA
(4,4-difluoro-5,7-dimethyl-4-bora-3a,4a diaza-5-indacene-3-propionic
acid-conjugated BSA; Molecular Probes) and DQTM collagen (type IV from
human placenta, fluorescein conjugate; Molecular Probes cat. no.
D-12052) were used. Bodipy-BSA becomes fluorescent upon exposure to
several proteases, especially urokinase-type plasminogen activator
(20)
. DQTM collagen becomes fluorescent upon exposure to
collagenase activity. Bodipy-BSA or DQTM collagen was incorporated into
the gelatin-nylon mesh matrix layers at a concentration of 1 mg/ml, as
described above.
An axiovert inverted fluorescence microscope (IM-135) with HBO-100 mercury illumination (Carl Zeiss Inc., New York, N.Y.) interfaced to a Dell 410 workstation (Round Rock, Tex.) via a Scion SG-7 video card (Vay-Tek, Fairfield, Iowa) was used. The fluorescence images were collected by an intensified, cooled charge-coupled device camera (model XC-77; Hamamatsu Photonics, Bridgewater, N.J.). Images were processed with ScionImage software and stored as TIFF files. The fluorescence of Bodipy-BSA and DQTM collagen fluorescence was detected using a 485DF22 nm and 530DF30 nm filter combination and a 510 long-pass dichroic mirror (Omega Optical, Brattleboro, Vt.).
For image reconstruction, images from 50 to 70 optical sections were taken from a sample. To remove out-of focus fluorescence, each image was deconvoluted using MicroTome software (Vay-Tek). The 3-dimensional rendering was performed with VoxBlast software (Vay-Tek). Computations were performed on a Dell Precision Workstation. Images were printed using an Epson Stylus-Pro printer.
Imaging spectrophotometry
Fluorescence emission properties of the matrices were studied
using imaging spectrophotometry. Experiments were performed as
described previously (21)
. Briefly, a Zeiss IM-135
axiovert microscope was fiber optically coupled to the input side of an
Acton-150 (Acton, Mass.) imaging spectrophotometer. The exit side was
connected to an intensifier, which was in turn attached to a
Peltier-cooled I-MAX-512 camera (Princeton Instruments, Inc., Trenton,
N.J.). The collection of spectra or images was controlled by a
high-speed Princeton ST-133 interface and a Stanford Res. Systems
(Sunnyvale, Calif.) DG-535 delay gate generator. This equipment was
interfaced to a Dell 410 workstation running Winspec/32 version 2.3.2.5
software (Princeton Instruments, Inc.) to manage image acquisition,
quantitate fluorescence levels, and analyze data. Cells were
illuminated using an optical filter at 485DF22 nm and a 510lp dichroic
mirror.
Statistical analysis
For comparisons, Student’s t tests were performed.
Differences were considered significant at P<0.01 level.
Quantitative data are given as the mean ±
SD.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In
the present study, we microscopically examine the invasiveness of
HT1080 cells into 3-dimensional nylon mesh-supported matrices. To
visualize proteolysis, Bodipy-BSA and fluorescein conjugated collagen
(DQTM collagen; Molecular Probes), which become highly fluorescent on
proteolysis, were used.
We first characterized the fluorescence properties of nylon
mesh-supported matrices containing 1 mg/ml Bodipy-BSA or DQTM collagen.
Figure 1a
shows the fluorescence emission spectra of Bodipy-BSA-doped
matrices. The dim emission spectrum shows the inherent background
fluorescence of the matrix. The companion spectrum of Fig. 1a
shows the emission spectrum in the presence of an
adherent HT1080 cell (15 min at 37°C). Clearly, the total amount of
fluorescence emission is greatly increased in the presence of these
cells. To more specifically image collagenase activity, we used
matrices containing DQTM collagen in the presence and absence of
adherent tumor cells (Fig. 1b
). Again, there was a
substantial increase in fluorescence intensity, although not as great
as that seen for Bodipy-BSA. In both cases the maximal fluorescence
intensity was observed in the region of 530–550 nm, thus indicating
the fluorescein optical filters were appropriate for imaging.
|
We next sought to test the ability of certain inhibitors to effect the
generation of pericellular fluorescence on matrices containing
Bodipy-BSA or DQTM collagen. Cells were placed on Bodipy-BSA or DQTM
collagen-doped matrices, then incubated for 2 h at 37°C. The
number of fluorescent sites corresponding to sites of cell adherence
and pericellular proteolysis were then counted. The number of
fluorescent pericellular zones per high-power microscopic field for the
probe-doped matrices in the presence and absence of protease inhibitors
is shown in Fig. 2
. Phenanthioline, a broad-spectrum metalloproteinase inhibitor,
substantially reduced proteolytic disruption of matrices containing
DQTM collagen, but not those containing Bodipy-BSA (Fig. 2a
, b
, respectively). Inasmuch as type IV collagen is cleaved by
metalloproteinases (22)
, the ability of phenanthioline to
inhibit DQTM collagen fluorescence enhancement was expected. Similar
experiments were conducted with the urokinase inhibitor PAI-1 (Fig. 2c
, d
). PAI-1 substantially inhibited the proteolysis of
Bodipy-BSA-containing matrices and DQTM collagen matrices. The
inhibition of Bodipy-BSA cleavage by PAI-1 was expected based on prior
experiments (20)
. The inhibitory effect of PAI-1 on
collagen disruption may be partially accounted for by the fact that
MMP-2 and MMP-9, which cleave type IV collagen (22)
, are
activated by urokinase-type plasminogen activator (23)
.
Another potential contributing factor is the ability of plasmin to
degrade type IV collagen (24)
. Therefore, these matrices
demonstrate their anticipated functional sensitivities to proteolytic
disruption.
|
Having established the fluorescence spectroscopic properties and
functional sensitivities of these matrices to protease inhibitors, we
next analyzed proteolytic detection methods in 3 dimensions using the
nylon mesh-supported matrices. Matrices were prepared using multiple
gel layers supported by a nylon mesh skeleton, as described above.
HT1080 tumor cells were seeded onto these matrices and incubated for
2 h at 37°C. This incubation time allowed cells to enter and
migrate through this model matrix. We found that the migratory pathways
of the cells in the matrix could be visualized as a region of
proteolytic destruction in the gel matrices. This simultaneously
revealed the migration pathway in the gel and the nature of proteolytic
action on gel during the tumor cell invasion. Figure 3
shows proteolytic trails in these extracellular gel matrices. Matrices
loaded with 1 mg/ml DQTM collagen or Bodipy-BSA are shown in Fig. 3a
, b
, respectively. The fluorescence emission of hydrolyzed
DQTM collagen or Bodipy-BSA was photographed at multiple focal planes.
The stacks of multiple images were deconvoluted by MicroTome software,
then reconstructed by VoxBlast software (see Material and Methods).
These images revealed tracks of proteolytic activity, i.e., tumor cell
migration, through the gel matrices. In general, these 3-dimensional
reconstructed images show that the proteolytic signal is most intense
near cells and progressively decreases toward the point of cell entry
into the matrix. The reduction in relative signal intensity and the
increase in signal diameter are likely due to the diffusion of the
small, cleaved fluorescent peptides from the pericellular region. Cell
migration during image acquisition may also cause some blurring of the
image. Figure 3
also shows selected 2-dimensional slices through the
reconstructed image. We observed donut-shaped regions of fluorescence
(Fig. 3b
, intermediate panel on right). Since the cell had
already passed this point and was lower in the matrix, the dark area in
the center likely corresponds to the wound caused in the matrix by cell
migration. Later (further up the migratory trail), the fluorescence
became more diffuse. Similar properties were noted with matrices
containing DQTM collagen (Fig. 3a
). Thus, protein
degradation in general and collagen disruption in particular are
observed during tumor cell invasion of these matrices. Moreover, this
provides a means to visualize the pathways followed by cells during
invasion.
|
To illustrate how larger numbers of cells act during the invasion of
these model matrices, experiments were repeated using different numbers
of cells added to the top of the matrix. Figure 4
shows several independent repetitions of these experiments. Proteolysis
is common in all transmigration events. The proteolysis of DQTM
collagen and Bodipy-BSA can both be observed (Fig. 4a
, b
,
respectively). Most migration pathways were found to be relatively
straight. The lengths of the proteolytic destruction zones were
generally 150 to 200 µm. In 6% gel matrices, the pathways of cell
migration generally were funnel shaped due to diffusion of fluorescent
peptide fragments.
|
To confirm the role of pericellular proteolysis in the migration of
tumor cells across this model matrix, we determined the fraction of
migrated cells in the presence and absence of protease inhibitors.
HT1080 cells were placed atop a matrix and incubated for 4 h at
37°C in a humidified incubator. The number of cells reaching the
bottom of the matrix was then assessed. The number of cells reaching
the bottom of the matrix divided by the number of input cells is shown
in Fig. 5
. Under the conditions used in these experiments, approximately one-half
of the cells reached the bottom of the matrix. The fraction of cells
reaching the bottom of the matrix was not significantly affected by the
inclusion of 10 mM 1, 10-phenanthroline (Fig. 5b
). However,
the fraction of HT1080 cells undergoing transmigration was dramatically
reduced by the inclusion of PAI-1 (Fig. 5a
). These
differences could not be explained by changes in cell viability. Cell
viability was assessed in the presence and absence of PAI-1 and 1,
10-phenanthroline; in all cases, cell viability was >96%.
Occasionally, multiple cells migrated in the same region of the gel or
as a cluster using pathways made by previously migrated tumor cells. In
some cases, cells would stop for short periods in the gel, resulting in
inhomogeneities in fluorescence intensity. Nonetheless, cell migration
through the gel was characterized by linear pathways rich in
proteolysis.
|
Cell–cell cooperation may participate in the invasive program of tumor
cells under certain conditions (25
26
27)
. Specifically,
uPAR-positive tumor or stromal cells may participate in the
invasiveness of uPAR-negative tumor cells. We therefore used our
in vitro 3-dimensional invasion imaging method with a
heterogeneous cell population. We tested the hypothesis that
uPAR-expressing cells affect the invasive behavior of uPAR-negative
cells. To perform this experiment, we used the T-47D cell line, which
is less invasive than the HT1080 cell line, and a derivative of the
T-47D cell line, T-47D A2–1, which has been transfected with a
uPAR-containing vector. The T-47D A2–1 transfectant expresses uPAR and
demonstrates greatly enhanced proteolytic activity in comparison to the
parental T-47D line (see below). The use of the T-47D and T-47D A2–1
cell lines avoids the introduction of other potentially confounding
variables with cells from different sources. The T-47D cells were
labeled with the fluorescent probe DiI, as described in Materials and
Methods. A single cell suspension of 2 x
104 T-47D A2–1 cells and 2 x
104 DiI-labeled T-47D cells was prepared. These
cells were suspended in 50 µl of HBSS containing 6% gelatin and
layered atop a gel matrix containing DQTM collagen. After 30 min at
37°C, samples were placed on a 37°C microscope stage. Forty to 50
image planes were collected for the DiI probe (T-47D localization, red)
and another precisely matched image stack was collected at a separate
wavelength for DQTM collagen (proteolytic action, blue). DQTM collagen
fluorescence emission is restricted to the location of, or trail of,
T-47D A2–1 cells. Although sensitive imaging techniques have revealed
some pericellular proteolytic activity near T-47D cells, the intensity
of this fluorescence is too low to be detected in these experiments.
Figure 6
shows representative examples of mixed tumor cell invasion experiments.
Figure 6A
shows an example of a T-47D cell following a path
produced by the T-47D A2–1 cell. The DiI emission profile in this
figure is significantly larger than the size of a cell. Since these
cells are migrating and time is required to capture an image and move
the stage to the next plane, some blurring of the images is expected.
Figure 6B
shows results similar to those of Fig. 6A
except that the T-47D cell is migrating at an angle
toward the pathway created by the T-47D A2–1 cell. In other examples
(Fig. 6C
), the pathways followed by T-47D cells were not
colinear with T-47D A2–1 cells. Quantitative analyses of cell
migration are shown in Fig. 7
. This figure shows the migration ratio of T-47D and T-47D A2–1 cells
in the presence and absence of the other cell type. As expected,
migration of the uPAR-expressing T-47D 2A-1 cell line was substantially
greater than the wild-type uPAR-negative T-47D cells when examined
separately. However, when the mixed cell experiment was performed, the
migration of the T-47D 2A-1 cells was not affected whereas the
migration of the DiI-labeled T-47D cells became indistinguishable from
their uPAR-expressing counterparts. One potential mechanism
contributing to this is matrix disruption provided by the uPAR-positive
cells (Fig. 6A
).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3
4
5
6
7
8
9
, 19)
. However, the relative importance of these two
types of enzymatic activities apparently varies among tumor cell types
(2
3
4
5
6
7
8
9
, 19)
. The present study focused on pericellular
proteolysis-dependent events occurring during tumor cell invasion of a
model matrix. The present study is of considerable interest from both technological and physiological viewpoints. Using an open mesh nylon grid for support, multiple gel layers were deposited on the mesh. By adding a small amount of gel to the mesh at a time, multiple layers could be built up into a thicker matrix-like assembling plywood. The open mesh also affords the ability to use both transmitted light and epifluorescence microscopy to image events in a sample. This is a substantial advantage over invasion assays that rely on cell migration through artificial pores. Another advantage of these nylon mesh-supported matrices is that most of the surface area is available for transmigration to the bottom of the matrix. In contrast, only a small fraction of a filter’s surface area is available in the form of pores to permit migration to the bottom of the matrix. In many experiments using the filter assay, cells must ‘squeeze’ through the pores in the plastic filters, an action that has no clear counterpart in cell physiology. It is not clear whether the pore assays are measuring the kinetics of transmigration or the kinetics of finding an available pore. Moreover, these gel matrices can also be tailored to address specific experimental hypotheses. In this study, for example, we included substrates (Bodipy-BSA and DQTM collagen) within the matrix that become fluorescent upon proteolytic cleavage, thus revealing the location of functioning enzymes. When these assays were conducted for extended periods of time, the pathways followed by cells during this model transmigration process could be imaged. Our approach is particularly valuable in studying transmigration since bright-field microscopy is unrevealing under these circumstances because the refractive index of cells matches that of the extracellular matrix. Thus, we have developed the technology suitable for addressing several critical issues surrounding the invasive ability of cancer cells.
Proteolysis of the extracellular matrix is believed to be necessary for
the invasive and migratory capacity of cells. The proteinases thought
to be largely responsible for the degradation of the extracellular
matrix are the matrix metalloproteinases and plasminogen activators
(2
3
4
5
6
7
8
9
, 19)
. The detection of these activities in
situ relies on the generation of fluorescent peptides after
protein cleavage. To confirm the relevance of the Bodipy-BSA and DQTM
collagen probes, we tested their ability to disrupt matrices in the
presence of the inhibitors PAI-1 and 1, 10-phenanthroline. PAI-1 and 1,
10-phenanthroline substantially inhibited the release of fluorescent
peptides from Bodipy-BSA and DQTM collagen, respectively. Thus, these
probes are appropriate for the detection of these enzymatic activities.
The present study demonstrates proteolysis and collagen disruption
during cell migration across these model matrices. Although much work
has suggested a role of collagen disruption in cell migration through
complex matrices, our work provides direct evidence for the
cell-mediated disruption of collagen during tumor cell invasion.
Several previous studies have traced the pathways of leukocyte movement
in 3-dimensional gels (28
, 29)
. These studies are very
time consuming because they necessarily followed just one cell over a
long period of time. By detecting the enzymatic activities associated
with transmigration, we were able to image the migratory pathways of
many cells. By collecting image stacks from different focal planes in a
gel, we can collect data from multiple cells from a single experiment.
We also confirmed the relevance of particular enzyme activities imaged
in the present study with in vitro migration in these model
matrices. PAI-1 and 1, 10-phenanthroline were incorporated into the
matrices at high concentrations before their solidification. PAI-1, but
not 1, 10-phenanthroline, was found to substantially inhibit the
ability of tumor cells to cross the matrix. This is consistent with
earlier studies indicating that urokinase-type plasminogen activator
plays an important role in tumor cell invasiveness (e.g., ref
30
). Since there is some variability in the relative
importance of urokinase and matrix metalloproteinases among different
types of cancer cells (2
3
4
5
6
7
8
9
, 19)
, this result should not
be interpreted to mean that metalloproteinases are unimportant to tumor
cell invasion in general.
Previous studies have shown that uPAR-positive cells express an
invasive phenotype (e.g., refs 8
9
10
11
, 30
). uPAR-negative
cells can also be invasive. One mechanism that may account for the
invasiveness of uPAR-negative cells is cross-talk among uPAR-positive
and uPAR-negative cell types. For example, uPAR-negative tumor cells
may collaborate with uPAR-positive stromal cells during invasion (e.g.,
ref 25
). Having developed the technology to observe
proteolytic trails in matrices, we examined the ability of
uPAR-positive cells to facilitate the invasive behavior of
uPAR-negative cells. To perform this test, we used uPAR-negative T-47D
cells and T-47D transfectants expressing uPAR. We found that the
presence of uPAR-positive cells enhanced the ability of uPAR-negative
cells to migrate across the matrix. The microscopy studies of Fig. 6
suggest that the ability of uPAR-negative cells to follow uPAR-positive
cells across a matrix can partially account for this. Thus, matrix
disruption by uPAR-positive cells could facilitate a ‘follow the
leader’ type mechanism during matrix invasion by uPAR-negative cells.
The 3-dimensional model of cell transmigration and extracellular
proteolysis described above should be generally applicable in studies
of invasion and transmigration. For example, the extracellular matrix
components could be tailored to reflect specific microenvironments.
Other cell types could be added to the matrix itself during formation
and to the top or bottom of the nylon mesh-supported matrix after
formation. Moreover, it should be possible to monitor several
biochemical reactions in addition to the proteolytic reactions noted
above. For example, it is known that tumor cells can produce superoxide
anions (31)
. Such oxidants could be detected in this
tissue-like environment as well. Thus, we expect that this approach
will become a valuable tool in cancer biology.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 16, 2000.
Revision received October 3, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
RIIIb to CR3 proximity on neutrophils. J. Immunol. 164,953-958This article has been cited by other articles:
|
|
 |
|
  M. Anghelina, P. Krishnan, L. Moldovan, and N. I. Moldovan Monocytes/Macrophages Cooperate with Progenitor Cells during Neovascularization and Tissue Repair: Conversion of Cell Columns into Fibrovascular Bundles Am. J. Pathol., February 1, 2006; 168(2): 529 - 541. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. M. Frederiks and O. R.F. Mook Metabolic Mapping of Proteinase Activity with Emphasis on In Situ Zymography of Gelatinases: Review and Protocols J. Histochem. Cytochem., June 1, 2004; 52(6): 711 - 722. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. V. Artym, A. L. Kindzelskii, W.-T. Chen, and H. R. Petty Molecular proximity of seprase and the urokinase-type plasminogen activator receptor on malignant melanoma cell membranes: dependence on {beta}1 integrins and the cytoskeleton Carcinogenesis, October 1, 2002; 23(10): 1593 - 1602. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
