| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Laboratory of Vascular Biology, The Picower Institute for Medical Research, Manhasset, New York 11030, USA; and
* University of Tübingen, Tübingen, Germany
2Correspondence: The Picower Institute for Medical Research, 350 Community Dr., Manhasset, NY 11030, USA. E-mail: cmetz{at}picower.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: fibroblast • antigen • MMP-9 • wound healing • neovascularization
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Much progress has been made in identifying specific factors that regulate angiogenesis. These molecules include polypeptides, proteins, lipids, nucleotides, and copper. An important challenge is to develop an understanding of how the tissue microenvironment and cell-to-cell interactions result in the production of angiogenic mediators necessary for neovascularization in vivo. Several circulating cell types, including neutrophils, monocytes, macrophages, lymphocytes, and granulocytes, which are recruited to sites of tumor growth, chronic inflammation, or wounds, contribute to the angiogenic process. However, their specific roles in the neovascularization process are only partially understood.
In previous studies we have established that a population of connective
tissue fibroblast-like cells exist in the circulating peripheral blood
(2)
. These cells, termed ‘fibrocytes’, were identified
by their rapid and specific recruitment from the blood to
subcutaneously (s.c.) implanted wound chambers in mice, where they
proliferate. Fibrocytes purified from peripheral blood and grown in
culture express vimentin and collagens I and III. However, these cells
are molecularly distinct from fibroblasts exhibiting a unique profile
of cell surface markers including CD34, CD45, CD80, CD86, and MHC class
II (2
, 3)
. These cells do not express typical surface
antigens present on dendritic cells (CD11a, CD25, CD10, CD38),
macrophages (CD14, CD16), B cells (CD19), fibroblasts, or endothelial
cells (von Willebrand factor) (2)
. Past studies have
revealed the unique cytokine and chemokine profile produced by
fibrocytes (4)
and demonstrated their ability to present
antigen in vitro and in vivo (3
, 5)
. More recent studies
show the critical interaction between the chemokine receptor CCR7,
expressed on fibrocytes, with secondary lymphoid chemokine as a
mechanism by which fibrocytes migrate to wound sites (35)
.
Although circulating fibrocytes have been shown to migrate to early wound sites where angiogenesis occurs, the precise contribution of these cells to wound healing is not known. In the present report, we characterize the production of proangiogenic factors by isolated fibrocytes and reveal the potential for fibrocytes to induce an angiogenic phenotype in microvascular endothelial cells in vitro and promote angiogenesis in vivo. Fibrocytes produce and secrete active matrix metalloproteinase 9 (MMP-9; gelatinase B), which is implicated in the proteolysis of the basement membrane early during the invasion stage of angiogenesis. In addition, cultured fibrocytes constitutively secrete several growth factors (i.e., VEGF, bFGF, PDGF, and hematopoietic factors) that induce endothelial cell migration, proliferation, and/or alignment of endothelial cells into tubular-like structures in vitro. Last, we observed that cultured fibrocytes (and fibrocyte-conditioned media) promote angiogenesis in vivo using the Matrigel implant model. Taken together, these results suggest that fibrocytes comprise a population of cells that may play a functional role in angiogenesis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3
4)
. Total PBMCs were first isolated from
blood by centrifugation over Ficoll-Hypaque (Pharmacia, Uppsala,
Sweden), following the manufacturer’s protocol, and plated at 1 x 107 cells/ml in Dulbecco’s minimal essential
media (DMEM) supplemented with 20% fetal calf serum (FCS; Hyclone,
Logan, UT) and 1% penicillin/streptomycin/glutamine (DMEM 20%FCS,
PSQ). After overnight culture, the nonadherent cells were removed by
gentle aspiration and the cells were cultured for 
10–12 days. Mouse
fibrocytes were purified from the peripheral blood of BALB/c mice
(female, retired breeders), as described (3
4)
. Mice were
killed by CO2 asphyxiation and bled by cardiac
puncture. Mouse blood was mixed with PBS (1/5 volume); 4 ml blood was
layered over 8 ml Ficoll-Hypaque and centrifuged at 400 g
for 30 min at room temperature. The buffy coat was isolated, washed
with PBS and spun at 500 g, then washed with PBS. Mouse
fibrocytes were maintained in DMEM supplemented with 10% FCS,
10% mouse sera (Sigma, St. Louis, MO), and 1% P/S/Q
for 2–3 wk. Typical yields were 0.8–4 x
104 cells/ml of mouse blood. Fibrocytes (mouse
and human) were analyzed to be >95% pure by adherent,
spindle-shaped morphology as well as by flow cytometry for CD34 and
collagen I, as described previously (4)
.
Antibodies, other reagents
Neutralizing antibodies to basic fibroblast growth factor (bFGF;
AF233), vascular endothelial growth factor (VEGF; AF493), IL-8
(MAB209), MCSF (AF216), and GCSF (NA-214) were purchased from R&D
Systems (Minneapolis, MN). The matrix metalloproteinase inhibitor,
GM6001 (Ilomastat) was purchased from Chemicon International (Temecula,
CA).
Preparation of conditioned media
Two- to 3-wk-old fibrocytes (80% confluent), mouse NIH/3T3
fibroblasts (80% confluent), or human dermal fibroblasts (80%
confluent) were washed once with PBS and cultured for 48 h in
either OptiMEM (Life Technologies, Grand Island, NY) or DMEM, as
indicated. Fibrocyte-conditioned media (FcCM) was centrifuged at 5000
rpm for 10 min, filtered through 0.22 µm filters, and stored at
-80°C until use. Conditioned media was concentrated as indicated
using Microcon 10 MWCO filters (Amicon, Beverly, MA).
Zymography gels
Gelatin zymography of FcCM (or control media) was performed to
visualize the gelatinolytic activity (6)
. Briefly, 20 µl
aliquots containing 10–20 µg protein were electrophoresed using
10% SDS-PAGE gels copolymerized with gelatin (Bio-Rad, Hercules, CA)
at 60V (4°C). After electrophoresis, the gels were washed with 2.5%
Triton X-100 to remove the SDS, then incubated in development buffer
(50 mM Tris pH 8.0 alone or 50 mM Tris pH 8.0 containing 10 mM EDTA or
10 mM Ca2+) overnight at 37°C. The lysed
regions indicating gelatinase activity were visualized as clear bands
after Coomassie blue staining/destaining.
RT-PCR
Total cellular RNA was isolated from human fibrocytes by a
modified guanidinium isothiocyanate method (RNAzol, TelTest,
Friendswood, TX). The cDNA was prepared from 1.0 µg of RNA using 0.25
ng of oligo-(dT)12–18 and MMLV reverse
transcriptase following the manufacturer’s protocol (Life
Technologies). Two microliter aliquots of cDNA were amplified using
Supermix (Life Technologies) by PCR in a Perkin Elmer model 9600
thermal cycler using specific primers and conditions as described
previously: aFGF, bFGF, VEGF (7)
, ß-actin, GADPH, IL-8,
IL-1ß, GM-CSF, interferon 
(IFN-), tumor necrosis factor
(TNF-
) (8)
, CTGF (9)
, HGF
(10)
, IGF-1 (11)
, platelet-derived growth
factor A (PDGF-A) (12)
, M-CSF (purchased from Clontech,
Palo Alto, CA), and MMP-1, -2, -9, -14 (13)
. Amplified
fragments of expected sizes were analyzed using a 2% agarose gel and
photographed over UV light.
MMP-9 Western blotting
Fibrocyte-conditioned media (or control media) was concentrated
(fivefold) using a Microcon 10 concentrator and electrophoresed (10
µg/lane) using a 12% SDS-PAGE gel under both reducing and
nonreducing conditions. Proteins were Western blotted using an MMP-9
monoclonal antibody (Calbiochem, Cambridge, MA) according to the
manufacturer’s instructions, using detection by enhanced
chemiluminescent methods (Amersham, Piscataway, NJ). Under nonreducing
conditions, both the latent (92 kDa) and active (83 kDa) forms of MMP-9
are detected, whereas under reducing conditions only the latent form of
the protein (92 kDa) is revealed.
Endothelial cell proliferation assay
HuMVEC (3–5x103/well; passages 5–10)
were plated in 96-well plates in EGM-2 media. After 24 h, the
medium was aspirated, replaced with EGM-2 (positive control), human
FcCM (prepared by incubation of fibrocytes with DMEM 0.2% FCS for
24 h), or control media (DMEM 0.2% FCS) containing 0.5 µCi
[3H]-thymidine/well, and incubated overnight
(n=8 per condition). The cells were harvested onto GF-B
filter plates (Packard, Downers Grove, IL) and
[3H]-thymidine incorporation was measured by
ß-liquid scintillation counting (Packard). For endothelial cell
proliferation experiments to identify specific mitogenic factors
present in FcCM, neutralizing antibodies (or isotype control
antibodies) were included in the proliferation assay between 0.1 and 10
µg/ml (according to the manufacturer’s directions). The data were
reported as the mean % change in proliferation ±
SD when compared with control media and the
Student’s t test was used to determine significance
(P<0.05 was considered significant). Similar results were
obtained using FcCM preparation obtained from cultured fibrocytes from
more than six different donors.
Determination of cytokines and chemokines present in
fibrocyte-conditioned media
ELISA kits for the quantification of the following factors were
purchased from R&D Systems: angiogenin, bFGF, EGF, GMCSF, HGF, IL-1ß,
IL-8, IL-12, IFN-, IFN-, VEGF, M-CSF, and substance P.
Endothelial cell migration assay
Endothelial cell migration was determined using the
ChemoTX® disposable chemotaxis 96-well
microplates (Neuroprobe, Gaithersburg, MD) according to the
manufacturer’s instructions. Human FcCM (prepared in OptiMEM or DMEM
and supplemented with 2%FCS), EGM-2, or basal endothelial cell media
containing 2% FCS and 5 ng/ml VEGF (as a positive control) or OptiMEM
supplemented with 2% FCS (as a negative control) was placed in the
bottom chamber (29 µl/well; n=6 per sample condition). The
chemotaxis microplate was then fitted with a filter (8 µ pore size)
bonded to a metal frame containing 96–6.0 mm diameter test sites and
50 µl containing 1–2 x 104 human
microvascular endothelial cells (p5–7; resuspended in OptiMEM 2% FCS)
was pipetted onto each site. After overnight incubation, cells were
wiped from the top filter and the plate was centrifuged at 500
g. The cells that migrated through the filter were fixed and
stained using Diff-Quik reagents I and II (Baxter, Morton Grove, IL).
Migrated cells were counted by microscopy [four or five random fields
(400x)/sample]. The data were reported as the mean ±
SE and the Student’s t test was used
to determine significance (P<0.05 was considered
significant). To identify the specific chemotactic agents present, FcCM
(or control media) was preincubated specific neutralizing antibodies at
25 µg/ml for 30 min before the addition of the endothelial cells. The
experiment was repeated twice with similar results. In addition, the
endothelial cell migration assays in response to FcCM were assessed
using Costar transwell chambers (8 µM; Costar Corp., Cambridge, MA)
under similar conditions and according to the manufacturer’s
directions. Similar results were observed with four donors.
Endothelial cell morphogenesis assay
Approximately 2 x 104 human
microvascular endothelial cells (passages 4–7) were resuspended in 0.4
ml EGM-2, human FcCM (prepared in Opti-MEM or DMEM) supplemented with
2% FCS or control media (OptiMEM 2% FCS or DMEM 2% FCS) and plated
onto 4-well chamber slides (Nunc, Naperville, IL) precoated with
Matrigel (0.3 ml/well; Collaborative Biomedical Products, Bedford, MA),
as described previously (14
, 15)
. After 3.5 h at
37°C/5%CO2, the media was aspirated, the cells
were fixed and stained with Diff-Quik reagents, and the slides were
examined for endothelial cells alignment at 100x using an Olympus
microscope (Model BX60). Images were obtained using MetaMorph Software
(West Chester, PA). Similar results were observed with four donors.
In vivo angiogenesis assay
In vivo angiogenesis was assayed by the growth of blood vessels
from the s.c. tissue into a Matrigel implant (16
, 17)
containing cultured fibrocytes or concentrated FcCM. Animal protocols
were approved by the North Shore University Hospital Institutional
Animal Care and Use Committee. Mice (BALB/c, retired female breeders,
n=5 per group) were injected along the abdominal midline
into the s.c. tissue with 0.4 ml Matrigel containing heparin (60 U/ml)
(Life Technologies) mixed with fibrocytes
(3–5x105/mouse), NIH/3T3 fibroblast cells (as a
negative control), FcCM (4 µl; 20-fold concentrated), control medium
(4 µl; 20-fold concentrated), or NIH/3T3 fibroblast cell-conditioned
media (4 ml; 20-fold concentrated). As positive and negative controls
for angiogenesis, mice were injected with 0.4 ml Matrigel containing
heparin (60 U/ml) and bovine aFGF (1.25 ng/ml; R&D, Abingdon, Oxon, UK)
(positive control) or heparin alone (60 U/ml) (negative control, Life
Technologies). The implanted gels were harvested 6–7 days after
injection. A small piece of the gel implant (100 µg) was saved for
quantitative hemoglobin (Hb) analysis using the Drabkin’s kit (Sigma).
The data are reported as the mean concentration of Hb (g/dl) ±
SD and the Student’s t test
(two-sample, equal variance) was used to determine significance. The
remainder of the Matrigel implant was fixed with 10% buffered
formalin, embedded in paraffin, sectioned at 3 µm, and stained with
Masson’s Trichrome. To further assess vascularization, equivalent
sections were stained using an anti-CD31 mAb (PharMingen, San
Diego, CA) and analyzed as described previously (18)
.
Matrigel angiogenesis was photographed at 100x using an Olympus BX60
microscope and MetaMorph imaging software; representative sections are
shown. The experiment was repeated once with similar results.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In mouse wounds, MMP-9 is induced within 24 h
after injury (20)
and localized mainly to tissue
fibroblast-like cells and the endothelium (21)
. Based on
these observations, we investigated whether fibrocytes express matrix
metalloproteinases. RT-PCR analysis of human fibrocytes revealed the
presence of MMP-9 (gelatinase B) mRNA (Fig. 1A
) but not MMP-1, MMP-2 (gelatinase A), or MMP-14 (data not
shown). Fibrocytes did not produce MMP-3 (active or latent stromelysin
1), as determined using a FITC-labeled MMP-3 substrate (Chemicon
International) or MMP-10 (active or latent stromelysin 2), as
determined by Western blotting using an MMP-10 specific monoclonal
antibody (Chemicon International) (data not shown). To demonstrate the
presence of gelatinase proteins, conditioned media obtained from
unstimulated human fibrocytes was analyzed by SDS-PAGE gelatin
zymography. As shown in Fig. 1B
(left panel), gelatinolytic
bands with an apparent molecular mass of 92 and 83 kDa were revealed,
suggesting the presence of both latent and active forms of MMP-9,
respectively. Incubation of the gelatin zymogram in EDTA prevented the
appearance of these bands (Fig. 1B
, middle panel) whereas
incubation with calcium enhanced proteolysis of the gelatin, further
suggesting the presence of MMP-9 (Fig. 1B
, right panel). In
contrast, control media lacked gelatinase activity. For comparison,
conditioned media from mouse fibrocytes also contained MMP-9 activity,
whereas human monocytes/macrophages (prepared in the same manner)
contained 10- to 20-fold less MMP-9 activity (data not shown). The
presence of MMP-9 was further confirmed by Western blotting using an
MMP-9 mAb (Fig. 1C
). Under reducing conditions, this
antibody reveals the latent form of MMP-9 (lane 4); under nonreducing
conditions, this antibody detects the latent (92 kDa) and active (83
kDa) forms (lane 2). Control media lacks MMP-9 reactivity (lanes 1 and
3).
|
Fibrocytes secrete factors that promote endothelial cell
proliferation
Endothelial cell proliferation is a critical step in angiogenesis.
Previous studies revealed that mouse fibrocytes secrete bFGF, a potent
angiogenic factor (4)
. Therefore, we tested whether
fibrocyte-conditioned media induced endothelial cell proliferation in
vitro. Fibrocyte-conditioned media promoted microvascular endothelial
cell proliferation by 35–40%, as determined by incorporation of
[3H]-thymidine (Fig. 2
A). This increase in endothelial cell proliferation was
confirmed by counting the cells after incubation with EGM-2 media,
FcCM, or control media using a Coulter counter, as well as by a
nonradioactive method for measuring cell proliferation using
sulforhodamine (22)
(data not shown). Similar proliferation
results were obtained using HuVEC (data not shown). A striking
observation was that serum-free human FcCM was as mitogenic for the
endothelial cells as the commercially available EGM-2 containing FCS,
VEGF, bFGF, aFGF, hydrocortisone, heparin, etc. The mitogenic effect of
FcCM (and EGM-2) on endothelial cell proliferation was completely lost
by boiling FcCM for 10 min before the assay, suggesting the presence of
labile growth factors (Fig. 2A
). Several preparations of
FcCM from multiple donors exhibited similar mitogenic properties (data
not shown). When we examined the effect of mouse FCCM on human
endothelial cell proliferation (due to a lack of mouse primary
endothelial cells), we found it was almost half as active as human FcCM
despite species differences (data not shown). In addition, screening
conditioned media (serum-free) obtained from several primary cell
cultures, including human dermal fibroblasts and numerous tumor cell
lines for endothelial cell mitogenic factors, revealed that these media
were significantly less active than human FcCM (<20% change in
proliferation; data not shown).
|
To identify the potential growth factors produced by fibrocytes
mitogenic for endothelial cells, the fibrocytes and
fibrocyte-conditioned media were screened for several known endothelial
cell mitogenic factors by RT-PCR (mRNA) and ELISA (protein),
respectively. As shown in Table 1
, human fibrocytes express and constitutively secrete several factors
known to induce endothelial cell proliferation, including bFGF, GM-CSF,
IGF-I, IL-1ß, IL-8, PDGF, and VEGF. We also characterized the
expression of proangiogenic factors by mouse fibrocytes by ELISA and
found that they secrete VEGF, GM-CSF, IL-1ß (data not shown; bFGF and
IL-8 were not tested). Because angiogenesis is the result of a balance
of both pro- and antiangiogenic factors, we also screened fibrocytes
and FcCM for representative well-known inhibitors of endothelial cell
proliferation (Table 2
). Fibrocytes do not express IFN- or IFN-, both of which have been
shown to block the angiogenic effects of bFGF, PDGF, and MMP-9
(23
24
25
26
27)
. Nor do they express IL-12 or TSP-1, a potent
macrophage-derived antiangiogenic agent.
|
|
To identify the specific growth factors present in the human FcCM
mitogenic for endothelial cells, we performed proliferation assays in
the presence of neutralizing antibodies. Endothelial cell proliferation
induced by FcCM was reduced by >35% in the presence of anti-bFGF
antibodies (10 µg/ml), >30% in the presence of VEGF antibodies (10
µg/ml), and only 15% with G-CSF antibodies (10 µg/ml) (see Fig. 2B
). Inclusion of all inhibitory antibodies (each at 10
µg/ml) reduced FcCM-mediated endothelial cell proliferation to that
of control media containing 2% serum, but no additional growth
factors. In contrast, endothelial cell proliferation induced by FcCM
was not significantly inhibited by inclusion of isotype control
antibodies, antiangiogenin antibodies, anti-IL-8 antibodies, anti-M-CSF
antibodies, or the addition of a MMP inhibitor, GM6001 or Ilomastat,
which inhibits the enzymatic activity of MMP-9, as well as MMP-1, -2,
-3, and -8 (data not shown). The effect of commercially available media
(EGM-2) on endothelial cell proliferation was significantly inhibited
only by the inclusion of anti-bFGF (40%) or anti-VEGF antibodies
(30%) when present at 10 µg/ml (data not shown).
Fibrocytes secrete factors that promote endothelial cell migration
The identification of potent endothelial cell chemotactic factors
and active MMP-9 in the conditioned media of fibrocytes prompted us to
investigate whether the FcCM could also induce endothelial cell
migration. As shown in Fig. 3
, we observed a two- to threefold increase in endothelial cell
chemotaxis with FcCM prepared using Opti-MEM vs. control media alone
(Opti-MEM). Similarly, endothelial cells migrated fourfold more in
response to VEGF (5 ng/ml). The addition of neutralizing antibody to
M-CSF (25 µg/ml) inhibited FcCM-mediated chemotaxis whereas
anti-IL-8, anti-VEGF, and anti-bFGF had no significant effect on
endothelial cell migration (data not shown). The ability FcCM to
promote endothelial cell migration was confirmed using 8 µ Costar
transwell chambers. VEGF induced the migration of 65 ± 6% of the
endothelial cells to the bottom chamber, whereas FcCM promoted the
migration of 46 ± 4% endothelial cells compared with 12 ±
5% migrated cells in the control media group. A similar induction of
endothelial cell migration was observed using DMEM-based
fibrocyte-conditioned media compared with DMEM alone (data not shown).
|
Fibrocytes secrete factors that enhance endothelial cell
differentiation in vitro
Because fibrocyte-conditioned media was chemotactic for
endothelial cells, we explored whether it could support the alignment
of endothelial cells or the formation of lumen-like structures. As
shown in Fig. 4
, microvascular endothelial cells when plated on Matrigel and incubated
with FcCM aligned to form lumen-like structures and anastomosing tubes
with multicentric junctions (Fig. 4B
) vs. nonconditioned
control media (Fig. 4C
). Similar endothelial cell
morphogenesis occurred in response to incubation with DMEM-based
fibrocyte-conditioned media vs. DMEM alone (data not shown). This
capillary-like network of endothelial cells induced by FcCM appears to
be weaker than that induced by EGM-2, commercially prepared endothelial
cell media (Fig. 4A
). By contrast, no endothelial cell
morphogenesis occurred after the incubation of endothelial cells plated
on Matrigel with human fibroblast-conditioned media (data not shown).
|
Fibrocytes secrete factors that are angiogenic in vivo
Based on our observations that fibrocytes constitutively secrete
factors that induce endothelial cell proliferation and migration and
support tube formation, we next examined whether cultured fibrocytes or
FcCM could promote angiogenesis in vivo. We used an experimental model
of blood vessel formation in vivo. When combined with Matrigel
containing heparin, mouse FcCM (Fig. 5
C) and mouse fibrocytes (Fig. 5E
) both induce
neovascularization when implanted s.c. in mice. Similar
neovascularization of Matrigel implants was observed using FcCM
prepared with Opti-MEM and DMEM (data not shown). Human FcCM also
induced blood vessel formation within the Matrigel plug; however, the
response was less than that of mouse FcCM, probably due to species
differences (data not shown). On the other hand, the addition of
control media (Fig. 5D
) or NIH/3T3 fibroblast cells (Fig. 5F
) did not promote blood vessel formation within
Matrigel implants. Vessel formation within the Matrigel plugs can be
analyzed by measuring the Hb content using Drabkin’s reagent. Hb
content present within the Matrigel plugs containing FcCM or mouse
fibrocytes is significantly greater than that implanted with
conditioned media obtained from mouse NIH/3T3 fibroblasts (data not
shown), nonconditioned control media (Opti-MEM or DMEM), or mouse
NIH/3T3 cells (Fig. 5G
).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
).
Although several mediators of wound healing have been identified, few
studies have focused on the role of specific cell types that mediate
blood vessel formation during early tissue repair.
Fibrocytes are a novel peripheral blood cell type that display a
distinct cell surface phenotype and rapidly appear in wound sites
(2)
. More recent studies have demonstrated the interaction
between CCR7 (expressed on fibrocytes) and SCL necessary for the
trafficking of fibrocytes to the wound site (36)
. Based on
their presence in early wounds and their cytokine, chemokine, and
growth factor profile (2
, 4)
, we have examined the role of
isolated fibrocytes in angiogenesis, a critical step in the wound
healing process. Fibrocytes isolated from peripheral blood and cultured
in vitro are able to promote all aspects of neovascularization
including basement membrane dissolution, endothelial cell
proliferation, migration, and tube formation in vitro and induce
angiogenesis in vivo.
Angiogenesis within a wound bed occurs by formation of blood vessels
from the existing vasculature made up of quiescent endothelial cells
and a few pericytes. These cellular structures are surrounded by a
stable basement membrane that physically separates the capillary from
surrounding stromal tissue. In response to injury or other stimuli,
angiogenesis ensues when endothelial cells dissociate from neighboring
cells and migrate into the wounded tissue area to form new vessels.
Therefore, proteolysis of the basement membrane is a critical component
during the very early stages of angiogenesis. Several degradative
enzymes have been implicated in this process. One family of enzymes,
collectively known as matrix metalloproteinases, has been shown to be
important mediators of angiogenesis and tissue remodelling. MMPs are
secreted as latent proenzymes that are subsequently cleaved and
activated extracellularly. MMP activity is also regulated by tissue
inhibitors of matrix metalloproteinases or TIMPs. Although MMPs are not
expressed by normal human skin, MMP-2 (gelatinase B) and MMP-9
(gelatinase A) are strongly induced within 24 h after wounding
(19
, 21)
. In addition, MMP-9 has been shown to be the main
MMP found in wound fluid, with peak activity between 2 and 4 days
postwounding (21)
. Consistent with these observations, we
have found that fibrocytes, which home to cutaneous wound sites within
1–4 days (2
, 36)
, constitutively express MMP-9 mRNA (Fig. 1A
) and secrete high levels of active MMP-9 (Fig. 1
B, C). MMP-9 dissolves several extracellular
matrix proteins including gelatin, collagens I, IV, V, XIV, as well as
aggrecan, elastin, entactin, and vitronectin (29)
.
Although MMP-9 has not been shown to play a key role in basement
membrane/collagen degradation in vivo, it has been proposed to initiate
and promote angiogenesis. Therefore, fibrocytes might be an important
source of active MMP-9 involved in angiogenesis in the very early
phases of wound healing.
After the degradation of the basement membrane during angiogenesis,
endothelial cells invade the stromal tissue and migrate to form a
leading edge. We report that cultured fibrocytes secrete
chemoattractants for endothelial cells (Fig. 3)
. Surprisingly, only the
addition of M-CSF antibodies inhibited FcCM-mediated endothelial cell
migration. The levels of bFGF and VEGF present in FcCM (100–300 pg/ml)
are below the optimal level for inducing endothelial cell migration.
Although M-CSF is not considered one of the prominent chemoattractant
agents for endothelial cells, studies by Sigounas and co-workers have
demonstrated the role of M-CSF and other hemopoietic growth factors on
endothelial cell migration (30)
. The biological role of
hemopoietic growth factors and expression of their respective receptors
on endothelial cells have not been examined thoroughly. However,
endothelial cells do express several surface antigens in common with
hemopoietic cells and respond to numerous hematopoietic growth factors
including G-CSF, GM-CSF (31)
, IL-3 (32)
, and
M-CSF (30)
.
During neovascularization, as endothelial cells migrate away from
preexisting blood vessels they begin to proliferate at the leading
edge. We report that serum-free conditioned media collected from
unstimulated cultured fibrocytes (FcCM) supports endothelial
proliferation comparable to commercially available endothelial cell
growth media containing serum and several mitogenic factors (Fig. 2A
). Analysis of fibrocytes (mRNA) and fibrocyte culture
supernatants (protein) revealed the presence of numerous proangiogenic
factors including bFGF, VEGF, GM-CSF, IL-1, IL-8, and M-CSF (Table 1)
.
Based on these findings, we performed in vitro endothelial cell
proliferation assays using specific neutralizing antibodies and found
that bFGF, G-CSF, and VEGF were the most active mitogenic factors
present in the FcCM. IL-8, a mitogenic factor for endothelial cells, is
present at very high concentrations (>50 ng/ml) in the FcCM. However,
neutralizing antibodies to IL-8 did not inhibit FcCM-mediated
proliferation. These data are consistent with previous studies showing
that IL-8 is a potent angiogenic factor only in vivo, because cultured
endothelial cells do not express the IL-8 receptor (33)
.
Serum-free FcCM induces lumen formation by endothelial cells when
plated on Matrigel. Fibrocytes produce bFGF, the main factor found to
mediate endothelial cell morphogenesis on Matrigel (34)
,
as well as VEGF and IL-8, minor morphogenic mediators
(34)
. Because angiogenesis is the result of a balance
between pro- and antiangiogenic factors, we also examined fibrocytes
for antiangiogenic factors. We found that fibrocytes did not produce
any of the antiangiogenic factors analyzed, including TSP-1, IFN-,
IFN-, TNF-, and IL-12. Finally, we observed that FcCM could
support endothelial cell tube formation in vitro. We speculate that the
production of collagens I, III, VEGF, and angiogenin by cultured
fibrocytes mediates this effect.
Angiogenesis is a multistep process requiring the coordinated activity
of many different cell types. Several circulating cells have been
postulated to be proangiogenic. Many studies have revealed the role of
activated macrophages in inflammatory and tumor-associated angiogenesis
(reviewed in ref 35
). Activated macrophages are able to
promote all phases of angiogenesis in vitro, but attempts to identify
pro- vs. nonangiogenic macrophage subtypes have been unsuccessful.
Macrophages secrete a tissue-type plasminogen activator and
urokinase-type plasminogen activator that degrade the basement
membrane. They secrete MMP-9 in the latent form whereas fibrocytes
secrete significantly more latent and active MMP-9 (data not shown).
Similar to fibrocytes, macrophages secrete many proangiogenic factors
(bFGF, TGF, GM-CSF, IL-8, VEGF, substance P, IGF-1, and PDGF).
However, macrophages also secrete several antiangiogenic agents not
produced by fibrocytes, including IFN-, IFN-, TSP-1,
TNF-, and IL-6.
Fibrocytes produce factors that support every aspect of angiogenesis:
proteolysis of the basement membrane, endothelial cell migration,
proliferation, and vessel formation in vitro. Therefore, we evaluated
whether isolated fibrocytes or fibrocyte-conditioned media would
promote angiogenesis in vivo. FcCM and fibrocytes produced a brisk
neovasculararization response when we used the Matrigel implant
experimental model (Fig. 5)
. The levels of many angiogenic factors
produced by fibrocytes are relatively low. However, because
angiogenesis in vivo is a complex process requiring multiple factors,
it is difficult to assess the contribution of the individual factors
produced by fibrocytes to the overall process of angiogenesis. Many of
these factors act additively or synergistically to promote
neovascularization, further complicating analysis of the contribution
of specific factors to blood vessel formation. Because of the
appearance of fibrocytes within the wound and their proposed role in
tissue repair, it would be of interest to examine the role of
fibrocytes in wound healing in normal and diabetic animals where wound
healing is compromised, specifically at the step of blood vessel
formation.
We propose that fibrocytes play a role in blood vessel formation during the early stages of wound healing based on their early presence within injured tissue and the ability of cultured fibrocytes to promote an angiogenic phenotype in cultured endothelial cells and support neovascularization in vivo. Further characterization of cultured fibrocytes with angiogenic potential provides a unique approach to examine the direct role of fibrocytes in neovascularization during wound healing. Future studies will examine the role of fibrocytes on angiogenesis during disease processes where aberrant blood vessel formation exacerbates the disease pathogenesis, such as tumor growth, rheumatoid arthritis, and psoriasis.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication January 31, 2001.
Revision received June 15, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C.-H. Wang, C.-D. Huang, H.-C. Lin, K.-Y. Lee, S.-M. Lin, C.-Y. Liu, K.-H. Huang, Y.-S. Ko, K. F. Chung, and H.-P. Kuo Increased Circulating Fibrocytes in Asthma with Chronic Airflow Obstruction Am. J. Respir. Crit. Care Med., September 15, 2008; 178(6): 583 - 591. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A M A. El-Asrar, S Struyf, J Van Damme, and K Geboes Circulating fibrocytes contribute to the myofibroblast population in proliferative vitreoretinopathy epiretinal membranes Br. J. Ophthalmol., May 1, 2008; 92(5): 699 - 704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Spiekerkoetter, C. M. Alvira, Y.-M. Kim, A. Bruneau, K. L. Pricola, L. Wang, N. Ambartsumian, and M. Rabinovitch Reactivation of {gamma}HV68 induces neointimal lesions in pulmonary arteries of S100A4/Mts1-overexpressing mice in association with degradation of elastin Am J Physiol Lung Cell Mol Physiol, February 1, 2008; 294(2): L276 - L289. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Wang, H. Jiao, T. L. Stewart, M. V. H. Lyons, H. A. Shankowsky, P. G. Scott, and E. E. Tredget Accelerated wound healing in leukocyte-specific, protein 1-deficient mouse is associated with increased infiltration of leukocytes and fibrocytes J. Leukoc. Biol., December 1, 2007; 82(6): 1554 - 1563. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. N. Gomperts and R. M. Strieter Fibrocytes in lung disease J. Leukoc. Biol., September 1, 2007; 82(3): 449 - 456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Haurani and P. J. Pagano Adventitial fibroblast reactive oxygen species as autacrine and paracrine mediators of remodeling: Bellwether for vascular disease? Cardiovasc Res, September 1, 2007; 75(4): 679 - 689. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Pilling, N. M. Tucker, and R. H. Gomer Aggregated IgG inhibits the differentiation of human fibrocytes J. Leukoc. Biol., June 1, 2006; 79(6): 1242 - 1251. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. J. Davie, E. V. Gerasimovskaya, S. E. Hofmeister, A. P. Richman, P. L. Jones, J. T. Reeves, and K. R. Stenmark Pulmonary Artery Adventitial Fibroblasts Cooperate with Vasa Vasorum Endothelial Cells to Regulate Vasa Vasorum Neovascularization: A Process Mediated by Hypoxia and Endothelin-1 Am. J. Pathol., June 1, 2006; 168(6): 1793 - 1807. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Frid, J. A. Brunetti, D. L. Burke, T. C. Carpenter, N. J. Davie, J. T. Reeves, M. T. Roedersheimer, N. van Rooijen, and K. R. Stenmark Hypoxia-Induced Pulmonary Vascular Remodeling Requires Recruitment of Circulating Mesenchymal Precursors of a Monocyte/Macrophage Lineage Am. J. Pathol., February 1, 2006; 168(2): 659 - 669. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E S Choi, E M Pierce, C Jakubzick, K J Carpenter, S L Kunkel, H Evanoff, F J Martinez, K R Flaherty, B B Moore, G B Toews, et al. Focal interstitial CC chemokine receptor 7 (CCR7) expression in idiopathic interstitial pneumonia J. Clin. Pathol., January 1, 2006; 59(1): 28 - 39. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. R. Stenmark, N. J. Davie, J. T. Reeves, and M. G. Frid Hypoxia, leukocytes, and the pulmonary circulation J Appl Physiol, February 1, 2005; 98(2): 715 - 721. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Pilling, C. D. Buckley, M. Salmon, and R. H. Gomer Inhibition of Fibrocyte Differentiation by Serum Amyloid P J. Immunol., November 15, 2003; 171(10): 5537 - 5546. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. C. Hendrix, R. E. B. Seftor, E. A. Seftor, L. M. Gruman, L. M. L. Lee, B. J. Nickoloff, L. Miele, D. D. Sheriff, and G. C. Schatteman Transendothelial Function of Human Metastatic Melanoma Cells: Role of the Microenvironment in Cell-Fate Determination Cancer Res., February 1, 2002; 62(3): 665 - 668. |
