| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Dermatology,*
Cutaneous Biology
Research Center, and the Center for Imaging and Pharmaceutical
Research,
Massachusetts General Hospital and
Harvard Medical School, Charlestown, Massachusetts; Schering
AG,
Berlin, Germany; and the Haartman
Institute,
University of Helsinki,
Helsinki, Finland
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Whereas tumor angiogenesis involving the formation of new blood vessels has been studied extensively, the role of lymphatic vessels in tumor pathology and the molecules regulating lymphangiogenesis have remained mostly unknown. Recently, a novel member of the vascular endothelial growth factor (VEGF) family has been identified that is distinguished by its capacity to stimulate lymphangiogenesis.10,11 Vascular endothelial growth factor-C (VEGF-C) stimulated lymphangiogenesis in the avian chorioallantoic membrane assay,12 and transgenic mice overexpressing VEGF-C in the skin were characterized by specific hyperplasia of the lymphatic network without obvious effects on blood vessels.13 In normal adult human tissues, the VEGF-C receptor VEGFR-3 (Flt-4) has been found predominantly expressed by lymphatic endothelium.7,14 However, VEGF-C also binds VEGFR-2 (KDR), that is mainly expressed by activated endothelium of blood vessels, suggesting a potential function of VEGF-C in the induction of angiogenesis.10,15 Indeed, VEGF-C stimulates the proliferation and migration of blood vascular endothelial cells in vitro and has been shown to increase vascular permeability in the Miles assay,10,11,16,17 and to induce angiogenesis in the mouse corneal micropocket assay and in the rabbit ischemic hind limb model.18,19
VEGF-C mRNA expression has been demonstrated in a large number of human tumors, including malignant melanomas.20-32 Expression of its receptor VEGFR-3 has been detected in lymphatic vessels and occasionally also in blood vessels adjacent to cancer cells.7,14,21 Most recently, we have shown that VEGF-C overexpression in experimental breast cancer selectively induced intratumoral lymphangiogenesis, leading to increased tumor metastasis, without obvious effects on angiogenesis.33
In the present study we demonstrate the occurrence of intratumoral lymphangiogenesis in VEGF-C-overexpressing human melanomas transplanted onto nude mice. Moreover, VEGF-C also induced tumor angiogenesis and recruitment of macrophages. Our results indicate that the distinct biological effects of VEGF-C are critically dependent on its proteolytic processing in vivo.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
The human malignant melanoma cell lines MeWo,34,35 WM9, and WM239 were provided by Dr. Robert S. Kerbel (Sunnybrook Health Science Center, Toronto, Canada). Human melanoma cell lines PM-WK, RPM-MC, RPM-EP, and MM-LH36 were obtained from Dr. Randy Byers (Boston University Medical School, Boston, MA). The human melanoma cell lines SK-MEL-5, SK-MEL-25, SK-MEL-28, MelJuso, Colo 38, MML-1, HS695T, and MELKL-2 were obtained from the tumor bank of the German Cancer Research Center in Heidelberg, Germany. All melanoma cell lines were maintained in RPMI 1640 medium with 5% fetal bovine serum; human prostatic adenocarcinoma PC-3 cells (American Type Culture Collection, Rockville, MD) in Ham’s F12 medium with 5% fetal bovine serum. All media were purchased from Life Technologies, Inc., Grand Island, NY.
Cell Transfection and Selection
A human VEGF-C cDNA comprising the complete coding sequence (GenBank accession number X94216)10 was cloned into a pcDNA3.1/Zeo expression vector (Invitrogen, San Diego, CA) that contains a CMV-enhancer promoter and a Zeocin selection cassette. The sequence and the orientation of the VEGF-C gene in the construct were verified by restriction mapping and direct sequencing. Human MeWo malignant melanoma cells were transfected either with the human VEGF-C cDNA cloned into a pcDNA3.1/Zeo vector or with the vector alone using the Superfect transfection reagent (Qiagen, Chatsworth, CA). Transfected cells were selected and maintained in growth medium containing 50 µg/ml Zeocin. Stably transfected cell clones were individually expanded and analyzed for VEGF-C mRNA expression and protein secretion.
RNA Isolation and Northern Analysis
Total cellular RNA was isolated from cultured cells and from tumors using the RNeasy kit (Qiagen). The isolated RNA was subjected to electrophoresis (15 µg per lane) and transferred to Hybond-N+ membranes (Amersham, Arlington Heights, IL). 32P-radiolabeled DNA probes were labeled by the random priming method (Multiprime labeling kit; Amersham, Arlington Heights, IL). The VEGF-C probe used was a fragment containing nucleotides 581 to 1634 of human VEGF-C cDNA. The probe for human VEGF hybridizes with a region of VEGF mRNA common to all known VEGF splice variants.37 A human ß-actin cDNA probe (Clontech, Palo Alto, CA) was used as a control for equal RNA loading. Blots were hybridized at 65°C for 24 hours, washed at high stringency, and exposed to X-OMAT MR film (Kodak, Rochester, NY).
Western Analysis
Cells grown to 80% confluency were lysed with Izuhara buffer (50 mmol/L HEPES, 150 mmol/L NaCl, 1% Triton X-100, 30 mmol/L Na-pyrophosphate, 50 mmol/L Na-fluoride, 1 mmol/L Na-orthovanadate, pH 7.4) containing protease inhibitors (phenylmethylsulfonyl fluoride, 0.02 mol/L; leupeptin, 50 µg/ml; aprotinin, 50 µg/ml). The amount of protein was determined with the BioRad protein assay (BioRad, Hercules, CA), and 15 µg were analyzed on the gel. Conditioned media were obtained from subconfluent cells grown for 60 hours in serum-free media, concentrated 100-fold using Centricon-10 columns (Amicon, Beverly, MA), and 15 µg of protein were loaded on the gel. Tumors were snap-frozen in liquid nitrogen, lysed by homogenizing the tissue (0.5 g) in 2 ml of a buffer containing 2% sodium dodecyl sulfate, 50 mmol/L Tris (pH 7.4), and protease inhibitors as above, and 10 µl were loaded onto the gel. All samples were boiled in denaturing Laemmli sample buffer (BioRad), analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and blotted onto polyvinylidene difluoride membranes (BioRad). Filters were blocked overnight with 5% nonfat milk in phosphate-buffered saline (PBS)/0.1% Tween 20 and were incubated with a rabbit antiserum against human VEGF-C (1:1000).16 After washes, membranes were incubated with horseradish-peroxidase-conjugated anti-rabbit IgG (1:2000; Amersham), washed, and analyzed using the Amersham ECL+ chemiluminescence reagents. Receptor phosphorylation assays were performed as described,33 using antibodies against mouse VEGFR-2 (Santa Cruz Biotechnology, Santa Cruz, California) and phosphotyrosine (PY-20; ICN Biomedicals, Aurora, Ohio).
Tumorigenicity Assay
Stably transfected MeWo cells (2 x 106 in 100 µl of serum-free culture medium) were injected intradermally on each side of the back of 8-week-old female Swiss/c (nu/nu) nude mice (five mice for each clone). Three clones were analyzed for each construct. Tumor growth was measured weekly using a digital caliper. Tumors were harvested when they reached a size of >1200 mm3 or after 6 to 8 weeks, and were embedded in OCT compound and frozen in liquid nitrogen or were fixed in 4% paraformaldehyde/PBS and processed for routine histology. For RNA or protein extractions, tumors were snap-frozen in liquid nitrogen. Two independent experiments with five mice in each group were performed.
Generation of a Polyclonal Anti-Mouse VEGFR-3 Antibody
The 19-amino acid synthetic peptide CYPGKQAERAKWVPERRSQ, corresponding to amino acid residues 265 to 285 in the Ig-like domain 3 of the mouse VEGFR-3 extracellular domain, was used to immunize rabbits with standard techniques. The antisera were affinity purified and specific staining of lymphatic vessels was confirmed in adult mouse tissues (heart, lung, spleen, and liver).
Indirect Immunofluorescence
Cryosections were stained as previously described,38 using antibodies against mouse CD31 (dilution 1/30; Pharmingen, San Diego, CA), LYVE-1 (1/300),33,39 VEGFR-3 (1/50; rabbit polyclonal; 1/30; R&D Systems, Minneapolis, MN), CD11b/Mac-1 (1/200; Pharmingen), or F4/80 (1/200; Serotec, Raleigh, NC). The secondary antibodies, labeled with either Texas Red or fluorescein isothiocyanate (Jackson ImmunoResearch, West Grove, PA) were used at the dilution 1/50. Cell nuclei were counterstained with Hoechst bisbenzimide (Sigma, St. Louis, MO) at 20 µg/ml. Specimens were examined by using a Nikon E-600 microscope (Nikon, Melville, NY).
Computer-Assisted Morphometric Analysis
For analysis of tumor vascularization, immunohistochemical stainings were performed on 6-µm frozen sections of tumor xenografts using an antibody against mouse CD31.40 Tissue sections were viewed using a Nikon E-600 microscope and images were captured with a SPOT digital camera (Diagnostic Instruments, Sterling Heights, MI). Two MeWo/control and two MeWo/VEGF-C clones were analyzed, five tumors each. On each tumor section, three randomly chosen fields were evaluated at x60 magnification. Morphometric analysis was performed using the IPLab software (Scanalytics, Fairfax, VA) to determine vessel density, size distribution, and the relative tumor area covered by vessels.33 For analysis of macrophage densities, sections were stained with CD11b/Mac-1 antibody as described above. Three MeWo/control and three MeWo/VEGF-C clones were analyzed, three tumors each. For each tumor, the total area of the adjacent overlying skin was analyzed, and the relative area of the skin occupied by macrophages was determined using the IPLab software. The unpaired t-test was used for statistical analysis.
Computed Tomography
Computed tomography was performed to determine the clearance rate
of a lymphographic contrast agent from the tumors. Mice were imaged 5
weeks after tumor cell inoculation. Mice bearing MeWo/control tumors
were additionally analyzed when the tumor size equaled that of the
MeWo/VEGF-C tumors at 5 weeks (
1000 mm3).
Three mice carrying two tumors each were analyzed in both groups. After
anesthesia with a mixture of ketamine (100 mg/kg) and xylazine (10
mg/kg), mice were injected with an iodinated lymphographic contrast
agent,41
and imaging was performed by using a helical
computed tomography scanner (TCT900S/xII; Toshiba, Tokyo, Japan). The
contrast agent was administered very slowly by injection into the
center of each tumor (20 µl per tumor) by using a 30-gauge needle
attached to a Hamilton syringe. Mice were imaged before, immediately
after, 6 hours, 24 hours, 48 hours, and 72 hours after the injection of
the contrast agent. Each imaging study consisted of a helical data
acquisition through the tumor and the axillary region for the
assessment of axillary lymph nodes. The imaging parameters were 120-kV
tube voltage, 150-mA tube current, 2-mm slice thickness, and a
high-resolution imaging mode with an 80-mm field of view. Images were
reconstructed onto a 512 x 512 image matrix yielding 0.05
mm3
voxels. Data analysis was performed with the
image analysis program DIP Station (HIPG, Boulder, CO) to obtain the
average signal intensity value for the tumor volume. Signal intensity
values in computed tomography are expressed as Hounsfield units that
change linearly as a function of contrast agent concentration in the
tissue.42
A signal-time curve was created for each tumor
and the rate constant was calculated using a mono-exponential model as
a first approximation. The average clearance rate was then calculated
for each group of animals.
Isolation of Mouse Macrophages, Flow Cytometry, and Chemotaxis Assay
Macrophage accumulation was induced in Swiss/c nu/nu mice by intraperitoneal injection of 1 ml of 3% Brewer thioglycollate medium and macrophages were harvested 5 days after injection by lavage with Hanks’ balanced salt solution.43 For flow cytometry, macrophages were detached from the tissue culture dish by scraping, incubated for 15 minutes at 4°C with antibodies to CD11b or mouse VEGFR-3 and with fluorescein isothiocyanate-labeled corresponding secondary antibodies. Chemotaxis was examined using 24-well Transwell migration chambers (8 µm pore size; Costar). Inserts were coated on the underside with 10 µg/ml of collagen type I (Collagen Corp., Palo Alto, CA) and 4 x 105 cells were added to the upper compartment in 100 µl of serum-free Dulbecco’s modified Eagle’s medium. VEGF-C (1 to 100 ng/ml) or the macrophage chemoattractant N-formyl-met-leu-phe (fMLP; 10-6 mol/L) were added to 600 µl of serum-free media in the lower compartment only. After 3 hours, inserts were fixed and washed as described44 and cell nuclei were stained with propidium iodide. The number of migrated cells was determined by using the IPLab software. For every insert, three fields were evaluated at x120 magnification. All assays were performed in triplicate.
Cell Growth Assays
In vitro proliferation rates of transfected MeWo cell clones were determined by using the BrdU labeling and detection kit (Boehringer Mannheim, Mannheim, Germany). Cells were plated in 96-well plates (5 x 103 cells/well; 8 wells/cell clone) and were allowed to proliferate for 24 hours before addition of BrdU (10 µmol/L) for 6 hours. The absorbance was determined at 405 nm using a microtiter plate reader (Titertek, Huntsville, AL). To assess the effect of VEGF or VEGF-C on MeWo cell proliferation, MeWo/control clone 5 was seeded into a 96-well plate at a density of 2.5 x 103 cells/well, and cells were treated for 5 days with 20 ng/ml of recombinant human VEGF-C16 or VEGF (R&D Systems), six wells each. Cell proliferation was determined by using the BrdU kit. In additional experiment, MeWo/control clones were plated in 12-well plate inserts (1 µm pore size; 1 x 104 cells/insert; four inserts each), and were co-cultured for 7 days with either MeWo/VEGF-C cl. 23 or MeWo/control cl. 5 (1 x 104 cells/well). Total cell nuclei were stained with the Hoechst bisbenzimide DNA stain and cells that had incorporated BrdU (10 µmol/L, 2 hours) were visualized by immunofluorescent staining using an anti-BrdU antibody38 (Pharmingen, San Diego, CA). Inserts were mounted on slides and the percentage of proliferating cells was determined by using the IPLab software. For every insert, three randomly chosen fields were evaluated at x120 magnification. All experiments were repeated twice. The unpaired t-test was used for statistical analysis.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
We first examined the expression of VEGF-C in several human
melanoma cell lines by Northern analysis and found that VEGF-C mRNA was
expressed in most (10 of 15) melanoma cell lines studied, as a single
band of 2.4 kb (Figure 1A)
. In contrast,
normal human melanocytes expressed little or no VEGF-C mRNA in
vitro (data not shown). To examine the role of VEGF-C in
tumorigenesis, we transfected the melanoma cell line MeWo, established
from a lymph node metastasis of a nodular malignant
melanoma,34,35
devoid of VEGF-C expression, to
constitutively overexpress VEGF-C. As determined by Northern analysis,
the parental MeWo cell line and three vector-transfected control clones
(MeWo/control) did not express any detectable amounts of VEGF-C mRNA
in vitro or in vivo (Figure 1B)
. Three
VEGF-C-transfected cell clones (MeWo/VEGF-C) expressed high levels of
VEGF-C mRNA in culture, as well as in tumors that reached the size of
1200 mm3
(Figure 1B)
. To exclude the
possibility of altered production of VEGF or other angiogenic factors
by MeWo/VEGF-C cells, we examined the expression of VEGF, basic
fibroblast growth factor (bFGF), and platelet-derived growth factor
(PDGF)-B in MeWo transfectants by Northern analysis. Parental, control,
and VEGF-C-transfected MeWo cells expressed comparable, low levels of
endogenous human VEGF mRNA in vitro and in vivo
(Figure 1B)
, and little or no bFGF or PDGF-B mRNA (data not shown).
|
. We detected
a strong band of 60 kd in cell lysates of VEGF-C transfectants,
corresponding to VEGF-C precursor,16
and only trace
amounts in control cells. Large amounts of the secreted 31-kd form that
activates VEGFR-316
were selectively found in culture
supernatants of VEGF-C-transfected clones. None of the clones expressed
any detectable levels of the 21-kd form in vitro. In
VEGF-C-overexpressing tumors however, we predominantly detected the
mature 21-kd VEGF-C form, a high-affinity ligand for both VEGFR-2 and
VEGFR-3, demonstrating that complete processing of VEGF-C occurred
in vivo (Figure 1C)
. VEGF-C Induces Tumor Lymphangiogenesis
The VEGF-C-overexpressing MeWo cell lines were injected
intradermally into immunodeficient mice and tumors were harvested 6
weeks after the injection. To visualize tumor-associated lymphatic
vessels we used an antibody against LYVE-1, a recently discovered
specific marker of lymphatic vessels in normal
tissues39,45
and in tumors.33
We detected
lymphatic vessels with mostly compressed lumina in the skin surrounding
control tumors (Figure 2A
, arrowheads).
In contrast, lymphatic vessels were frequently enlarged in the skin
surrounding VEGF-C-overexpressing tumors (Figure 2B
, arrowheads) and
also within the peripheral areas of the tumors (Figure 2B
, arrows).
Importantly, numerous small lymphatic vessels were observed within
VEGF-C-producing tumors (Figure 2, C and D
; arrows), but not in control
tumors (Figure 2A)
, demonstrating that VEGF-C induced tumor
lymphangiogenesis. All LYVE-1-positive lymphatic vessels also expressed
VEGFR-3 (Figure 2, E and F)
. The effects of VEGF-C overexpression on
lymphatic vessels were not associated with a modulation of the
expression of the related lymphangiogenesis factor,
VEGF-D,46,47
because, as determined by Northern analyses,
little or no VEGF-D mRNA was detected in cultured cells and in tumor
extracts of control and VEGF-C-overexpressing cells (data not shown).
|
. The clearance rate was independent of
tumor size. Assessment of axillary lymph nodes confirmed that the
contrast agent had been transported by the lymphatic vessels to the
lymph nodes (data not shown).
|
To analyze intratumoral angiogenesis, tumor sections were stained
with an antibody against mouse CD31.48
In both
experimental groups, vessels were distributed rather uniformly
throughout the tumors (Figure 4, A and B)
. Computer-assisted morphometric analysis revealed significantly
increased vascular densities (Figure 4C)
in 6-week-old tumors derived
from VEGF-C-overexpressing clones (40 ± 12
vessels/mm2) as compared to control clones
(23 ± 7 vessels/mm2; P <
0.0002). Furthermore, the relative tumor area covered by vessels was
significantly increased in MeWo/VEGF-C tumors (Figure 5D)
as compared to control tumors
(2.65 ± 0.9% versus 1.48 ± 0.4%;
P < 0.0001). The distribution of vessel sizes however,
was not significantly different (data not shown). The majority of tumor
vessels were smaller than 800 µm2, with
approximately half of all vessels in both control and
VEGF-C-overexpressing tumors smaller than 400
µm2.
|
|
, which is in accordance with
the observed production of the mature form of VEGF-C in these tumors.
Taken together, these results identify the mature form of VEGF-C as a
potent tumor angiogenesis factor. VEGF-C Stimulates Recruitment of Macrophages
Histological analysis revealed an increased inflammatory response
in the skin surrounding VEGF-C-transfected melanomas, as compared to
control tumors. Using an antibody to CD11b that has broad reactivity
with dermal macrophages,49
we observed a markedly
increased density of macrophages surrounding MeWo/VEGF-C tumors (Figure 5, A and C)
. Immunofluorescent analysis demonstrated that
CD11b-positive macrophages expressed VEGFR-3 (Figure 5, B and D)
. The
identity of VEGFR-3-positive cells as macrophages was further confirmed
by staining with an additional macrophage marker, F4/8050
(Figure 5, E and F)
.
Flow cytometry analysis confirmed the expression of VEGFR-3 on cultured
mouse peritoneal macrophages (Figure 5G)
. Importantly, VEGF-C
stimulated macrophage chemotaxis in vitro in a
dose-dependent manner (Figure 5H)
.
Melanoma Cell Growth in Vivo and in Vitro
After orthotopic injection into immunodeficient mice, MeWo/control
tumors reached a volume of 1200 mm3
within 5
to 6 weeks after injection. Overexpression of VEGF-C did not increase
the tumor growth rate, but rather resulted in growth reduction by 40 to
80% after 6 weeks, as compared to control tumors (Figure 6A)
. The extent of tumor growth
suppression was correlated with the macrophage densities in the
peritumoral areas. In the skin surrounding control tumors, macrophages
occupied 4.0 ± 0.8% of the total skin area examined. In
contrast, tumors derived from MeWo/VEGF-C cl.29 cells that exhibited a
growth suppression of up to 40% at 6 weeks showed an increase of
peritumoral macrophage densities to 12 ± 1.5%. Tumors derived
from MeWo/VEGF-C cl.23 cells that exhibited the most prominent growth
suppression (up to 80%), showed further increase in macrophage
densities to 23 ± 1.9%.
|
. Moreover, the
proliferation of control clones was not significantly altered on
addition of 20 ng/ml recombinant human VEGF or VEGF-C to the culture
medium (Figure 6C)
. Proliferation of control cells was also not altered
by secretion of VEGF-C by MeWo/VEGF-C cells into the culture medium in
a co-culture experiment (Figure 6D)
. In accordance with these findings,
MeWo transfectants did not express any detectable levels of VEGFR-2 or
VEGFR-3, as determined by reverse transcriptase-polymerase chain
reaction, Northern analysis, and in situ hybridization (data
not shown). The rate of spontaneous apoptosis in vitro was
comparable between control and VEGF-C clones (1.3 ± 0.3%
versus 1.1 ± 0.2%). Taken together, these results
demonstrate that VEGF-C did not exert direct effects on the growth of
MeWo cells. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
Our results demonstrate infiltration of VEGF-C-overexpressing tumors with lymphatic vessels, even in the most central tumor areas. It has been suggested previously that lymphatic vessels are absent from tumors.8,52,53 However, most evidence in support of this concept has been indirect and has been based on the absence of detectable perfusion of lymphatic vessels after injection of contrast agent into tumors.9,54,55 Moreover, the direct visualization of lymphatic vessels in tissue sections has been problematic because of the lack of molecular markers that reliably distinguish the lymphatic from the blood vasculature.56 By using a recently derived antibody to the mouse hyaluronan receptor LYVE-1, that is selectively expressed on lymphatic vessels in normal tissues39,45 and in tumors,33 we were able to demonstrate intratumoral lymphatics in experimental melanomas overexpressing VEGF-C. All lymphatic vessels selectively expressed VEGFR-3. These results extend our recent findings that demonstrated, for the first time, the occurrence of intratumoral lymphangiogenesis in an orthotopical model of breast cancer.33
VEGF-C overexpression also resulted in enlargement of peritumoral lymphatic vessels, whereas intratumoral lymphatics did not appear to be enlarged and were found either with an open lumen or collapsed. Under normal circumstances lymphatic capillaries are partially or fully collapsed, because of the lack of smooth muscle coverage and the low pressure within the lymphatic system.57-59 Increased demand for fluid transport in tissue results in widening of their lumina; however, if overdistended, lymphatics become dysfunctional.60 Hence, the functional state of lymphatic vessels cannot be deduced from their morphology, because an open lumen can indicate both vessels with dysfunction as well as normally functioning vessels with increased load. Functional studies of lymphatic vessels in mice have been extremely difficult to perform because of the low spatial resolution and the insufficient signal intensity imposed by most currently used methods. By using computed tomography and a recently developed lymphographic contrast agent with an increased concentration of iodine that results in image enhancement,41 we have established a simple, minimally invasive procedure to measure the efficacy of lymphatic transport in mice. In comparison to control tumors, lymphatic transport from MeWo/VEGF-C tumors to regional lymph nodes was markedly retarded, indicating functional impairment of lymphatic vessels in terms of transport or uptake of fluid or functional overload because of an increase in VEGF-C-induced vascular permeability. The significance of delayed fluid clearance from tumors for the rate of tumor cell metastasis via the lymphatics remains to be established. At present, however, it is unclear how tumor cells are transported within lymphatic vessels, and the mechanisms of fluid and particle uptake into lymphatic vessels are most likely distinct from the mechanisms of tumor cell entry into the lymphatics.
VEGF-C has been reported to increase microvascular permeability in the Miles assay, with a fourfold to fivefold lower potency than VEGF.16 VEGF-C was also 50- to 100-fold less potent than VEGF in inducing proliferation of bovine capillary endothelial cells,10,16 human umbilical vein endothelial cells, or lung microvascular endothelial cells in vitro.11,19 In contrast, VEGF-C promoted proliferation of human dermal microvascular endothelial cells with a similar potency as VEGF,17 suggesting that microvascular endothelial cells derived from the skin blood vessels may be more responsive to VEGF-C than endothelial cells from other tissues. Moreover, VEGF-C acted synergistically with VEGF on in vitro angiogenesis,17,61 suggesting that in VEGF-C-expressing melanomas VEGF-C may act in cooperation with VEGF to increase vascular permeability and angiogenesis. In fact, our results demonstrate that, in addition to its role in lymphangiogenesis, VEGF-C also acted as a tumor angiogenesis factor. We detected increased microvascular densities in VEGF-C-overexpressing melanomas by morphometric analyses of CD31-stained tumor sections. In accordance with these findings, computed tomography with an intravascular contrast agent revealed increased perfusion of MeWo/VEGF-C tumors (data not shown). This was not because of up-regulation of the major angiogenic factors VEGF, bFGF, and PDGF-B, because their expression remained comparable between VEGF-C-overexpressing and control tumors. The stimulatory effect of VEGF-C on tumor angiogenesis seemed to be mediated through VEGFR-2, because we observed marked induction of VEGFR-2 expression and phosphorylation in VEGF-C-overexpressing melanomas, as compared to control tumors.
VEGF-C was originally described as a specific growth factor for lymphatic vessels12,13 but was later also found to induce angiogenesis of blood vessels,18,19 raising a question about the mechanisms that determine the distinct effects of VEGF-C in vivo. As demonstrated in vitro, the secreted 31-kd VEGF-C protein predominantly activates VEGFR-3 whereas the mature, fully processed 21-kd form also activates VEGFR-2, suggesting that the biological functions of VEGF-C may be regulated by differential proteolytic processing.16 The major VEGF-C form produced by MeWo/VEGF-C cells in vitro was the partially processed 31-kd form. In tumors however, the secreted VEGF-C was processed further to the mature 21-kd form, indicating an important role of tumor-host interactions for VEGF-C processing. As predicted by in vitro studies, the mature 21-kd form of VEGF-C stimulated both lymphangiogenesis and angiogenesis in VEGF-C-overexpressing melanomas. In contrast, overexpression of VEGF-C in breast cancer has been recently shown to selectively induce tumor lymphangiogenesis, but not tumor angiogenesis.33 In the breast cancer model, however, the predominant VEGF-C form detected was the 31-kd form that selectively activated VEGFR-3 on the lymphatic vessels. Together, these results demonstrate that the biological effects of VEGF-C in tumors are critically dependent on the in vivo proteolytic processing and provide an explanation for the selective induction of lymphangiogenesis in certain physiological and pathological settings.
Lymphatic vessels, serving as a pathway for trafficking of antigen-presenting cells and macrophages, are important components of the immune system; therefore, activation of the lymphatic system by VEGF-C might have an impact on immune functions. In fact, we detected expression of VEGFR-3 on mouse macrophages in vitro and in vivo by flow cytometry and immunofluorescent analysis, and VEGF-C induced macrophage chemotaxis, indicating that VEGF-C can act as a direct immunomodulator. In agreement with our results, expression of VEGFR-3 has been reported in some hematopoietic and leukemia cells.62,63 Furthermore, we observed increased densities of peritumoral macrophages in the skin surrounding VEGF-C-transfected melanomas. Enhanced host-tumor defense capabilities by increased recruitment of macrophages may explain the observed growth impediment of VEGF-C-overexpressing melanomas, because increased densities of peritumoral macrophages were correlated with the extent of tumor growth suppression. VEGF-C did not exert direct effects on tumor cells, because co-culture of VEGF-C-overexpressing cells with control cells did not result in altered proliferation of control cells, and addition of recombinant VEGF-C to control cells did not affect their growth rate in vitro. MeWo transfectants also did not express any detectable levels of VEGFR-2 or VEGFR-3, as determined by reverse transcriptase-polymerase chain reaction, Northern analysis, and in situ hybridization (data not shown). Our finding that VEGF-C did not increase the growth of melanomas is in agreement with a recent report showing that VEGF-C did not promote breast cancer growth, although it promoted tumor metastasis.33
Taken together, our results provide evidence that the proteolytic processing of VEGF-C in vivo is critical for determining its biological function in tumors. In addition to its role as a tumor lymphangiogenesis factor, our findings identify novel functions of the mature form of VEGF-C in melanoma, acting as a tumor angiogenesis factor and as an immunomodulator.
|
Note Added in Proof |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the Human Frontier Science Program (to M. S.), by the National Institutes of Health/National Cancer Institute (grants CA69184 and CA86410 to M. D.), by the Deutsche Forschungsgemeinschaft (to T. H.), by the Dermatology Foundation (to M. St.), and by the Cutaneous Biology Research Center through the Massachusetts General Hospital/Shiseido Co. Ltd. Agreement (to M. D.).
Accepted for publication May 7, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionNote Added in ProofReferences |
|---|
vß3 integrin, osteopontin, and thrombin. Am J Pathol 1996, 149:293-305[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  L. Fang, V. C. Lee, E. Cha, H. Zhang, and S. T. Hwang CCR7 regulates B16 murine melanoma cell tumorigenesis in skin J. Leukoc. Biol., October 1, 2008; 84(4): 965 - 972. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. P. Padera, A. H. Kuo, T. Hoshida, S. Liao, J. Lobo, K. R. Kozak, D. Fukumura, and R. K. Jain Differential response of primary tumor versus lymphatic metastasis to VEGFR-2 and VEGFR-3 kinase inhibitors cediranib and vandetanib Mol. Cancer Ther., August 1, 2008; 7(8): 2272 - 2279. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Halin, H. Fahrngruber, J. G. Meingassner, G. Bold, A. Littlewood-Evans, A. Stuetz, and M. Detmar Inhibition of Chronic and Acute Skin Inflammation by Treatment with a Vascular Endothelial Growth Factor Receptor Tyrosine Kinase Inhibitor Am. J. Pathol., July 1, 2008; 173(1): 265 - 277. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Bock, L. L. Sinclair, N. S. Bedford, R. E. Jackson, J. H. Lee, and D. K. Trask Modulation of Cellular Invasion by VEGF-C Expression in Squamous Cell Carcinoma of the Head and Neck Arch Otolaryngol Head Neck Surg, April 1, 2008; 134(4): 355 - 362. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Uzarski, M. B. Drelles, S. E. Gibbs, E. L. Ongstad, J. C. Goral, K. K. McKeown, A. M. Raehl, M. A. Roberts, B. Pytowski, M. R. Smith, et al. The resolution of lymphedema by interstitial flow in the mouse tail skin Am J Physiol Heart Circ Physiol, March 1, 2008; 294(3): H1326 - H1334. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Brideau, M. J. Makinen, H. Elamaa, H. Tu, G. Nilsson, K. Alitalo, T. Pihlajaniemi, and R. Heljasvaara Endostatin Overexpression Inhibits Lymphangiogenesis and Lymph Node Metastasis in Mice Cancer Res., December 15, 2007; 67(24): 11528 - 11535. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Halin, N. E. Tobler, B. Vigl, L. F. Brown, and M. Detmar VEGF-A produced by chronically inflamed tissue induces lymphangiogenesis in draining lymph nodes Blood, November 1, 2007; 110(9): 3158 - 3167. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Sundar and T. S. Ganesan Role of Lymphangiogenesis in Cancer J. Clin. Oncol., September 20, 2007; 25(27): 4298 - 4307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Jackowski, M. Janusch, E. Fiedler, W. C. Marsch, E. J. Ulbrich, G. Gaisbauer, J. Dunst, D. Kerjaschki, and P. Helmbold Radiogenic Lymphangiogenesis in the Skin Am. J. Pathol., July 1, 2007; 171(1): 338 - 348. [Abstract] [Full Text] [PDF]
|
|
|
|
|
Hounsfield
units/
