VEGF family members : modulators of tumor angiogenesis and lymphangiogenesis

Members of the vascular endothelial growth factor (VEGF) family and their receptors
(VEGFR) play an essential role in the development and maintenance of the blood and
lymphatic vasculature. To date, five VEGFs have been identified in the mammalian genome,
VEGF-A, -B, -C, -D, and placental growth factor (PlGF), which display distinct binding
affinities for VEGFR-1, -2, and -3. In addition to their central function in physiological
angiogenesis and lymphangiogenesis, VEGFs and VEGFRs are upregulated during
carcinogenesis and are involved in the remodeling of the tumoral blood and lymphatic
vasculature. By activating VEGFR-1 and –2, which are both expressed on blood endothelial
cells, VEGF-A promotes the formation of new tumoral blood vessels and thereby accelerates
tumor growth. In contrast, upregulation of VEGF-C, a ligand for lymphatic endothelial
VEGFR-3 as well as for VEGFR-2, induces the formation of tumor-associated lymphatic
vessels and thus promotes the passive metastatic dissemination of tumor cells to regional
lymph nodes. Much less is known about the functional consequences of tumor-expressed
VEGF-B and PlGF, two selective ligands for VEGFR-1, as well as VEGF-D, the second
VEGFR-3- and -2-binding lymphangiogenic VEGF family member. Also, the biological
effects of selective VEGFR-1, -2 or -3 signaling on tumor angiogenesis and tumor growth as
well as tumor lymphangiogenesis and metastasis are incompletely studied. Only recently, the
identification of VEGF-E, a selective ligand for VEGFR-2, as well as the generation of
VEGF-C156S, a specific ligand for VEGFR-3, has enabled the study of the distinct roles of
these receptors.
To investigate the function of lymphangiogenic VEGF-D under physiological
conditions, I analyzed transgenic mice, in which expression of VEGF-D is specifically
targeted to β-cells of pancreatic islets of Langerhans (Rip1VEGF-D mice). In these mice,
expression of VEGF-D induces the formation of large lymphatic lacunae surrounding most
islets. A few of these lymphatic vessels may be dysfunctional, which causes intra-lymphatic
accumulations of immune cells. Moreover, lymphatic lacunae often contain erythrocytes,
which may result from blood-lymphatic vessel shunts found in the vicinity of some islets.
However, the fact that erythrocytes are drained to regional lymph nodes demonstrates the
draining capacity of the de novo formed lymphatic vessels. To address the impact of VEGF-D
on tumorigenesis and metastasis, I crossed Rip1VEGF-D with Rip1Tag2 mice, a wellcharacterized
transgenic model of poorly metastatic multistage β-cell carcinogenesis.
Tumoral expression of VEGF-D in Rip1Tag2 mice promotes the growth of peri-tumoral
lymphatic vessels that frequently contain leucocyte clusters and hemorrhages. Concomitantly,
these double-transgenic mice exhibit a high incidence of regional lymph node and distant
lung metastases. Since expression of VEGF-D does not significantly affect the invasiveness
of tumors and all metastases are well differentiated, these data indicate that VEGF-D
promotes lymphogenous metastasis by upregulating tumor-associated lymphangiogenesis.
Interestingly, the presence of VEGF-D significantly represses tumor angiogenesis and tumor
growth, yet the mechanisms of this inhibition are thus far uncharacterized. Notably, syngenic
and allogenic subcutaneous transplantation of VEGF-D-producing Rip1Tag2 tumor cell lines
results in the formation of tumors exhibiting a dense intra-tumoral lymphatic network but
lacking peri-tumoral lymphatic vessels. In these transplanted tumors, no immune cell clusters
or hemorrhages are formed in tumor-associated lymphatic vessels and tumor angiogenesis is
unaffected by the expression of VEGF-D. These results demonstrate that the tumor
microenvironment critically modulates VEGF-D-elicited effects. It has been recently shown
that transgenic expression of VEGF-C during Rip1Tag2 tumorigenesis promotes metastasis to
regional lymph nodes but not to the lungs by inducing peri-tumoral lymphangiogenesis.
Tumor-associated lymphatic vessels of these mice neither contain immune cell accumulations
nor hemorrhages, and tumor angiogenesis and tumor growth are not affected by the
production of VEGF-C. Thus, by employing the Rip1Tag2 tumor model, I was able to
identify not only similarities but also significant differences between VEGF-D and –C
function.
Since VEGF-C and –D can bind both VEGFR-3 and –2, it is not fully established
whether selective activation of VEGFR-3 is sufficient to induce tumoral lymphangiogenesis
and to promote lymphogenous metastasis. Therefore, I established transgenic mice expressing
VEGF-C156S in the endocrine pancreas and crossed these mice with Rip1Tag2 animals. The
analysis of single and double transgenic mice revealed that VEGF-C156S phenocopies
VEGF-C in all investigated aspects. These results indicate that VEGFR-3 may be the
predominant receptor mediating VEGF-C-elicited effects in Rip1Tag2 mice and that selective
activation of VEGFR-3 is sufficient to promote tumor-associated lymphangiogenesis and
metastasis. Hence, VEGFR-3 might represent a valuable target for future anti-metastatic
strategies.
To further understand the specific roles of VEGFR-1 and –2 signaling in physiological
angiogenesis as well as in tumorigenesis, I established transgenic mouselines, which express
the VEGFR-1-specific ligands VEGF-B167 and PlGF-1 as well as the selective VEGFR-2
ligand VEGF-ED1701 in β-cells of pancreatic islets (Rip1VEGF-B167, Rip1PlGF-1, and
Rip1VEGF-ED1701 mice). These single transgenic mice were analyzed with regard to islet
blood vessel morphology and density. In a second set of experiments, I crossed singletransgenic
animals with Rip1Tag2 mice. These double-transgenic mice expressing either
VEGF-B167, PlGF-1 or VEGF-ED1701 in tumor cells, were analyzed for changes in tumor
angiogenesis, tumor growth, and tumor progression. The preliminary data provide evidence
that β-cell-specific upregulation of VEGF-B167 does not critically affect physiological
angiogenesis of single-transgenic mice but results in a significant increase in the tumor
microvessel density of double-transgenic animals. However, tumor growth and tumor
progression are not promoted by the stimulation of tumor angiogenesis. In contrast,
overexpression of PlGF-1 in single-transgenic mice leads to a prominent dilation of blood
capillaries, which may at least in part be caused by a significant reduction of stabilizing blood
vessel-associated pericytes. Furthermore, tumoral expression of PlGF-1 significantly inhibits
tumor angiogenesis and tumor growth, suggesting that this growth factor might be a natural
inhibitor of pathological angiogenesis. Hence, although binding to the same receptor, VEGFB167
and PlGF-1 elicit opposing effects on the tumor blood vasculature. These results suggest
that the two growth factors induce distinct signaling pathways via VEGFR-1, which might be
considered when designing inhibitors of angiogenesis involving VEGFR-1. Importantly, the
phenotype of VEGF-B167- and PlGF-1- expressing Rip1Tag2 mice is different from the
recently described VEGF-A165 transgenic Rip1Tag2 mice, which exhibited accelerated tumor
growth and early death. The analysis of VEGF-ED1701-expressing mice and effects induced by
selective activation of VEGFR-2 signaling is currently underway.