Microsoft word - seeon meeting report 2009.doc

2nd International Kloster Seeon Meeting
“Cellular and Molecular Mechanisms of Tumor Progression and Metastasis”
Hosted by the Nationwide DFG Priority Research Program 1190 “The Tumor-Vessel Interface” with support from the Verein für Wissenschaftliche Fachtagungen in der Biomedizin e.V.
Session 1:
ANGIOGENESIS AND TUMOR PROGRESSION

“Reversing Angiogenesis in solid tumors”

Ruth Ganss (Perth) presented new data regarding regulator of G protein signaling (RGS) 5 as a
master gene responsible for the abnormal morphology of tumor vessels. RGS5 is a marker for
progenitor perivascular cells and regulates vessel stability in tumors. She found enriched expression
of RGS5 in pericytes, whereas after tumor vessel normalization the signal was reduced. Several
angiogenic tumors (astrocytomas, renal cell carcinoma and hepatocellular carcinoma) showed a
correlation of high RGS5 expression and low vessel normalization. Loss of RGS5 altered intratumoral
pericytes, and led to a shift to more mature pericytes within the vascular bed. Importantly, tumors in
RGS5-deficient mice became susceptible to immune effector cell influx after adoptive transfer of in
vitro activated, anti-Tag CD4+ and CD8+ T cells. In addition, RGS5 knockout mice were highly
responsive to therapeutic vaccination. The tumors exhibited increased blood perfusion, reduced tumor
hypoxia as well as reduced angiogenic activity. Ruth Ganss further showed that upon ligand binding to
the G protein-coupled receptor (GPCR), RGS5 increased the rate of GTPα hydrolysis and therefore
negatively regulated the signaling through the GPCR. Taken together, RGS5 regulates pericyte
maturation and possibly the vascular composition in tumors with profound changes in the tumor
environment. She proposed that a combination of vascular remodeling with anti-cancer immune
strategies would hold great promise for the development of more effective anti-cancer therapies.
“The Role of Loxl2 in angiogenesis and tumor progression”
Gera Neufeld (Haifa) investigated the role of lysyl oxidase 2 in tumor angiogenesis. Loxl2 and Loxl3
mRNA expression was found in metastatic and non-metastatic breast cancer derived cell lines. He
described a novel asymmetric 3D in vitro assay consisting of a tumor cell monolayer embedded
between collagen layers. This assay permits a direct comparison of the effects of different tumor
stroma compounds on tumor invasion. Inhibition of Loxl2 inhibited invasion of tumor cells and induced
MET-like changes in several types of tumor cells, whereas Loxl2 gain-of-function resulted in increased
tumor cell migration and invasion into vessels, nerves and muscles. Inhibition of Loxl2 by antibody
treatment resulted in reduced tumor growth of MDA-MB-435 cells and reduced angiogenic factor
expression such as SDF1 and VEGF in tumors. In addition, metastasis formation was also reduced.
He introduced Platelet factor 4 (PF4) as a substrate for Loxl2 that inhibited VEGF function by a
mechanism that does not interfere with VEGF binding to heparin sulfates. PF4 inhibited the enzymatic
activity of Loxl2 and it is itself oxidized by Loxl2. External Loxl2 abrogated PF4 induced inhibition of
bFGF induced proliferation of HUVECs. Inhibition of Loxl2 resulted in the inhibition of bFGF and VEGF
induced proliferation and migration of HUVECs. Furthermore, inhibition of Loxl2 inhibited adhesion
and tube formation on Matrigel and induced morphological changes in endothelial cells. In summary,
Loxl2 plays a role in angiogenesis and tumor progression.
“Proteolytic processing controls pro-angiogenic and pro-tumorigenic effects of Semaphorin
3G”

Simone Kutschera (Heidelberg) presented data about a novel class 3 semaphorin, Semaphorin 3G
(Sema3G) that is selectively expressed by sprouting endothelial cells upon stimulation with VEGF-A or
FGF-2. Full length Sema3G (100 kDa) is a secreted molecule that is cleaved by pro-protein
convertases similar to the other class 3 semaphorins to yield 65 kDa and 35 kDa fragments. Cleavage
enhanced the angiogenic activity of Sema3G in an in vitro 3D HUVEC sprouting assay. Antibody
staining for Sema 3G revealed a peri-endothelial cell location suggesting an additional function on
mural cells. To further unravel the role of Sema 3G on tumor progression, she generated Hek293 cells
expressing different Sema3G constructs (full length wild type [WT-L], processed 65 kDa form [WT-S]
and a full length cleavage-resistant form [MUT]) and implanted these cells subcutaneously into nude
mice. WT-L tumors grew significantly faster compared to tumors growing from mock-transfected, WT-
S or MUT Hek293 cells. Silencing Sema3G in MiaPaCa cells endogenously expressing Sema3G
resulted in reduced tumor growth of MiaPaCa tumors. Taken together, Sema3G is a novel regulator of
vascular function and tumor growth.
“Targeting tumor vessels by therapeutic vaccination”

Anna-Karin Olsson (Uppsala) - content embargoed.
Session 2:
ANGIOGENESIS: EMERGING TARGETS

“miR-132 acts as an angiogenic switch by suppressing endothelial 120RasGAP”

David Cheresh (San Diego) presented the microRNA, miR-132, as a regulator of angiogenic signaling
by targeting p120RasGAP in quiescent endothelium and thereby facilitating angiogenesis. His group
identified miR-132 via a screen of human microRNAs that were upregulated in HUVECs upon growth
factor treatment and in a human ES cell model of vasculogenesis. Expression of miR-132 in HUVECs
resulted in increased proliferation as well as increased tube formation in a 3D assay, whereas down
regulation of miR-132 resulted in decreased proliferation and tube formation, respectively. Injection of
anti miR-132 in a Matrigel plug assay decreased FGF-mediated angiogenesis and intraoccular
injection resulted in disruption of angiogenesis in the developing retina. Furthermore, using multiple
prediction programs for interacting partners, he identified p120RasGAP as a direct target of miR-132.
Expression of miR-132 in HUVECs results in decreased p120RasGAP levels and increased Ras
activity. SiRNA of RasGAP phenocopied miR-132 mediated EC proliferation. RasGAP was associated
with the development of human hemangiomas and was found in hyperproliferative hemangiomas.
Further analysis revealed that p120RasGAP was expressed in normal vessels, whereas it was lost in
tumor vessels (i.e. pancreas and beast). Using an αVβ3 targeting nanoparticle to deliver miR-132 to
the tumor endothelium, David Cheresh showed that tumor angiogenesis and growth was suppressed
in melanoma and breast cancer models. Taken together, he showed that miR-132 and p120RasGAP
function as a unique angiogenic switch to maintain endothelial quiescence.
“Identification of tumor dormancy specific microRNAs”

Nava Almog (Boston) studied the involvement of microRNAs in tumor dormancy. For this purpose,
she used experimental in vivo models of human tumor dormancy. Human tumor cells (glioblastoma,
osteosarcoma, liposarcoma and breast carcinoma) generate in these models small tumors when
injected into immunodeficient mice. Following long latency periods, they spontaneously emerge from
dormancy and initiate rapid growth. Using genome wide transcriptome analysis and advanced
microRNA profiling, she compared dormant tumors with tumors that had switched to rapid growth and
revealed critical players of tumor dormancy. Among the genes upregulated in the dormant state, she
found elevated levels of thrombospondin, angiomotin, EphA5 and Insulin-like growth factor binding
protein 5 (IGFBP5). High levels of EGFR, IGFR and Endothelial Specific Marker (ESM) were observed
in fast growing tumors. Furthermore, she showed that the plasma levels of EphA5 in mice, similar to
human patients, correlated with disease stage. Interestingly, overexpression of miRNAs associated
with tumor dormancy in rapid growing tumors resulted in a significant inhibition of tumor growth. In
conclusion, this study may lead to the detection and treatment of tumors long before they are
symptomatic or their anatomical location is known.
“Using novel mouse genetic approaches to study signaling pathways in tumor angiogenesis”

Gavin Thurston (Tarrytown) presented data regarding the role of Angiopoietin-2 (Ang-2) during tumor
angiogenesis. He used embryonic stem (ES) cell-derived teratomas as a new approach to study
specific gene knockouts within the tumor entity. Teratomas with the LacZ allele under the control of
muscle (nebulin), epithelial (keratin5), or endothelial (VEGFR-2) specific genes showed distinct
patterns and structures. The VEGFR-2-LacZ ES derived cells form networks, which co-express the
vascular specific marker CD31 and are connected to the host vasculature. Expression profiling studies
comparing ES-derived endothelial cells with host derived endothelial cells showed a similar
enrichment of numerous vascular specific genes, both, in teratomas and in host-derived endothelial
cells. In addition, Gavin Thurston introduced VE-PTP null ES tumors exhibiting enlarged blood
vessels. Treatment of these tumors with systemic Ang-1 resulted in further vessel enlargement. VE-
PTP-/- teratomas showed increased Tie2 phosphorylation with or without systemic Ang-1 treatment,
whereas inhibition of Ang-1 and Ang-2 reduced Tie2 phosphorylation accompanied by a reduction of
tumor vessel diameter. Blockade of Ang-2 in Colo205 a colon carcinoma cell line resulted in
decreased tumor vessel diameters. However, no significant difference of tumor vessel numbers was
observed. Blockade of Ang-2 was accompanied by increased expression of FOXO target genes that
are usually suppressed by the Ang-1-Tie2 axis. Based on these findings, Thurston concluded that
Ang-2 acts as a Tie2 agonist in tumor angiogenesis and restores Tie2-AKT signaling.
“Compromised tumor angiogenesis and vessel maturation in mice with broad Junb ablation”
Tobias Nübel (Heidelberg) discussed the function of Junb during angiogenesis. Junb is a subunit of
the Ap-1 transcription factor complex and mediates gene regulation. Complete as well as endothelial-
cell specific ablation of Junb results in similar angiogenic defects with subsequent embryonic lethality.
He showed that Junb acts as a critical regulator of several pro-angiogenic molecules including the
major angiogenic growth factor VEGF-A, the CBF transcription factor subunit Cbfbeta and its target
MMP13. He injected intradermally B16 mouse melanoma cells into JunbΔ/ΔColl(I)α2-Cre mice, in
which Junb was ablated in a variety of cell types, including endothelial cells and fibroblasts. Junb-
ablated mice exhibited reduced tumor growth, reduced relative blood volume and decreased blood
supply to the tumor. However, the density of the microvasculature and the total vessel number was
increased. Furthermore, he observed an altered pattern of pericyte recruitment and maturation of the
tumor vasculature in Junb-deficient hosts. Junb-deficient tumor vessels exhibited higher pericyte
coverage indices whereas coverage with vascular smooth muscle cells extensively found in tumor
vasculature of control animals was almost completely missing in Junb-deficient tumor vessels. In
summary, Junb is a critical factor for vessel size regulation and vessel maturation within tumors.
Session 3:
CANCER MALIGNANCY

“Inhibition of c-Met and VEGFR suppresses tumor invasion and metastasis and prolongs
survival of the RIP-TAG2 mice”
Donald McDonald (San Francisco) investigated the phenomenon that inhibition of vascular
endothelial growth factor (VEGF) signaling can promote invasiveness and metastasis in preclinical
tumor models. The mechanism underlying this consequence of VEGF inhibition is not fully understood
but activation of the hepatocyte growth factor (HGF)/c-Met (HGFR) pathway may be a factor. He
reported that selective inhibition of VEGF by a function-blocking antibody increased c-Met expression
and exaggerated tumor invasiveness and metastasis. Treatment of tumor-bearing RIP-Tag2
transgenic mice with anti-VEGF antibody reduced tumor growth but increased the expression of c-Met
in tumor blood vessels and tumor cells and also increased invasiveness and the number and size of
liver metastases. Also, inhibition of HGF/c-Met signaling and VEGF/VEGFR signaling together by the
multi-targeted receptor tyrosine kinase inhibitor XL 184 (developed by Exelixis, South San Francisco,
CA/Bristol-Myers Squibb, NY) significantly reduced growth and vascularity of RIP-Tag2 tumors and
strikingly reduced tumor invasiveness and metastasis accompanied by increased survival rates of
RIP-Tag2 mice. Taken together, inhibition of c-Met and VEGFR signaling not only has potent
antiangiogenic activity but also reduces tumor growth, invasiveness, and metastasis and improves
host survival.

“Phenotypic switching and the anatomy of tumor progression”
George Vande Woude (Grand Rapids) focused his presentation on the phenotypic switching during
tumorigenesis. During malignant progression, tumor cells switch their phenotype from proliferative to
invasive and then back again to a proliferative phenotype during metastatic growth in their new niche.
The acquisition of an invasive phenotype through epithelial mesenchymal transition (EMT) is a critical
step in tumor progression, as is mesenchymal epithelial transition (MET) for producing proliferating
metastatic lesions. George Vande Woude evaluated the genes responsible for both, EMT and MET.
He showed that a significant proportion of either prostate or ovarian carcinoma cells were epithelial (E)
and were non-responsive to hepatocyte growth factor (HGF). E cells were non-invasive in vitro, but
grew abundantly in soft agar and were highly tumorigenic in athymic nude mice. Beyond the E cells,
there was also a significant subpopulation of cells undergoing extensive branching morphogenesis
(EMT) in 3D Matrigel in response to HGF. These cells were mesenchymal (M) and, when treated with
HGF, were highly invasive in vitro, but poorly tumorigenic in vivo. With prostate carcinoma cells, he
showed data demonstrating that M cells spontaneously emerged from the non-invasive E subclones in
vitro. Furthermore, he showed that EMT occured with loss of E-cadherin expression and gain of c-Met
expression. Interestingly, the β-catenin expression levels did not change with phenotypic switching.
However, β-catenin localization shifted from the cell membrane to nuclear/cytoplasm localization
suggesting a significant role for β-catenin in modulating c-Met expression and in cell type switching. In
summary, George Vande Woude proposed that the regulation of β-catenin location and E-Cadherin
expression influences both EMT and MET.
“Molecular characterization of minimal residual cancer”

The detection of disseminated tumor cells (DTC) in bone marrow and lymph node has been
associated with reduced disease-free or overall survival and consequently these cells are thought to
compromise the metastasis founder cells. Christoph Klein (Regensburg) and his group addressed
the question whether primary tumors may serve as surrogate for the genetics of DTC and metastasis.
However, using comparative genomic hybridization studies and loss of heterozygosity analyses, he
found evidence for the parallel tumor progression model in which DTC were genetically less advanced
than the primary tumor cells and that these cells disseminated very early during tumor progression.
Therefore, the molecular characterization of DTC is essential for developing systemic therapies for
cancer. Furthermore, genome-wide characterization of DTCs may uncover the identity and
characteristics of metastasis founder cells.
“Searching for origins of breast cancer malignancy in mice”

Katrina Podsypanina (New York) examined the role of the initiating oncogenes MYC and KrasD12 in
the metastatic process using a modified experimental metastasis model and a three-dimensional (3D)
culture system. Un-induced mammary cells derived from transgenic mice expressing MYC and KrasD12
in a mammary-specific doxycycline-dependent manner were intravenously injected into immuno-
compromised mice exposed or not exposed to doxycycline. She showed that in the absence of an
active oncogene mammary cells homed to and persisted in the ectopic microenvironment of the lung
and could give rise to tumors upon activation of MYC and KrasD12. These metastatic lesions remained
dependent on the continuous expression of the initiating oncogenes. However, regulated regression of
tumors at ectopic sites left a small number of residual mammary cells that responded to re-induction of
the transgenes and may be responsible for tumor recurrency. In a 3D culture system un-induced
mammary cells developed into polarized acini. Due to the expression of MYC and KrasD12, the acini
enlarged and generated solid, depolarized spheres. The deactivation of MYC and KrasD12 resulted in
regression of the depolarized structures and the mammary cells showed a mammary epithelium
progenitor like phenotype. Moreover, residual mammary cells in the de-induced acini retained the
ability to respond to transgene activation and thus may represent the type of cells that give rise to
recurrent tumors.
Session 4:
INVASION & METASTASIS MEETS IMAGING

“Real-time imaging of tumor cell extravasation at the vascular interface reveals a highly
dynamic process regulated by metastatic programming“

Richard Klemke (San Diego) introduced zebrafish embryos as a model to study metastasis of tumor
cells. He showed that cancer cell extravasation was a highly dynamic and complex process regulated
by pro-metastatic genes that target the cytoskeleton and remodel the endothelium. High resolution
time-lapse imaging of arrested tumor cells revealed that these cells were not passively immobilized in
the lumen as previously believed, but instead displayed dynamic amoeboid-like movement along the
endothelial surface. Surprisingly, cell locomotion could be against or with the blood flow and required
integrin-mediated tumor cell adhesion to the blood vessel wall. Extravasating cells did not damage the
vessel wall causing leakage as previously suggested, but rather induced local vessel remodeling
characterized by altered endothelial cell-cell junctions. Induction of the pro-metastatic gene, twist,
caused a switch from integrin-dependent to an integrin-independent mode of extravasation that
required ROCK-mediated formation of dynamic membrane blebs and protrusions that penetrated the
vascular wall. In summary, Richard Klemke reported that cancer cell extravasation is a highly dynamic
and complex process regulated by pro-metastatic genes that target the cytoskeleton and remodel the
endothelium.
“Brain tumor imaging using a novel peptide targeting malignant brain tumors“

Pirjo Laakkonen (Helsinki) used phage displayed peptide libraries to map disease-specific
differences in the vasculature. Using this technology, she isolated several peptides homing specifically
to the tumor blood vessels, lymphatic vessels and/or to tumor cells. She compared low grade
astrocytoma with malignant brain tumors, where the tumor cells propagated and spred efficiently by
cooperating with existing vessels. Thereby, she identified a 9 amino acid peptide, CooP, which
specifically targeted a subset of malignant brain vasculature. The receptor molecule was expressed in
a grade-dependent manner in human brain tumor specimens. The CooP-peptide, covalently linked
with a drug, was used for tumor therapy in a mouse model of brain tumors. The targeted treatment
induced a prolonged survival and decreased tumor burden in the mice. In addition, a radio labeled
CooP peptide could be used to image brain tumors using the SPECT-CT technology.
“The role of HIF hydoxylases in tumor progression and metastasis“

Ben Wielockx (Dresden) - content embargoed.
“The endogenous TLR4 ligands regulate pre-metastatic soil“
Yoshiru Maru (Tokyo) reported on the identification of ligands involved in the bi-directional signaling
between primary tumor cells and pre-metastatic tissues. Using DNA-array analysis, subcutaneously
injected tumor cells representing an experimental system for the pre-metastatic phase were compared
with tumor cells injected into the tail vein representing the metastatic phase. S100A8 was found to be
induced in the metastatic phase. He showed that VEGF and TNFα from primary tumors stimulated
pulmonary resident cells in lungs to secrete the endogenous TLR4 agonists S100A8 and SAA3. SAA3
caused production of TNFα in the lungs, which made TNF from the primary tumors dispensable for the
stimulation of SAA3 production in the lungs. Both, S100A8 and SAA3 activated the two transcription
factors NF-κB and ATF3. ATF3 induced expression of the angiogenic molecules PAI-1, Eph/ephrin
family of proteins, but not VEGF. Abrogation of the S100A8-SAA3-TLR4 cascade inhibited lung
metastasis. Given the embryonic lethality of S100A8.deficient mice, the endogenous TLR4 agonists
could play essential roles even in physiological settings.
Session 5:

“Molecular dissection of epithelial-mesenchymal transition (EMT)”

Gerhard Christofori (Basel) presented data regarding the role of E-cadherin signaling during EMT.
Loss of E-cadherin function appears to play a central role in triggering full EMT in a number of
epithelial cell types in vitro and in transgenic mouse models of carcinogenesis in vivo. He showed that
the transcription factor distal-less homeobox 2 (Dlx2) was upregulated during EMT by TGFβ signaling.
Dlx2 knockdown cells still underwent EMT. However, these cells died especially after TGFβ induction
indicating that Dlx2 knockdown cells were resistant to TGFβ which induced cell survival. Furthermore,
knockdown of Dlx2 in tumors resulted in reduced tumor growth. Dlx2 induced the transcriptional
repression of TGFβRII, leading to attenuated TGFβ signaling and thus reduced expression of TGFβ
target genes. Simultaneously, Dlx2 promoted cell survival and proliferation through EGFR-mediated
activation of the MAPK and the PI3K signaling pathways and, thus, was critical for primary tumor
growth and metastasis. In conclusion, these results establish a mechanistic link between Dlx2 gene
expression, resistance against TGFβ mediated growth inhibition, and promotion of cell survival and
invasion.
“Tumor invasion and metastasis: EMT and cancer stem cells”

Thomas Brabletz (Freiburg) highlighted in his talk the role of ZEB1 in epithelial mesenchymal
transition (EMT). His concept of malignant progression and metastasis is based on migrating cancer
stem cells (MSCs). Tumor cells at the invasive front of colorectal adenocarcinomas (CRC)
accumulated nuclear β-catenin, underwent EMT and aberrantly expressed EMT-associated trans-
criptional repressors, like ZEB1. ZEB1 is known to be involved in EMT during development. In
addition, ZEB1 repressed target genes involved in basement membrane formation and epithelial cell
polarity. Knockdown of ZEB1 induced MET in tumor cells and thereby inhibited invasion and
metastasis in tumor xenografts. ZEB1 directly suppressed transcription of members of the miRNA-200
family. He showed that miR200c, a MET promoting factor, reduced migration and invasion of tumor
cells. ZEB1 directly bound and inhibited the miR200c promoter. Both, miR200c and miR141 negatively
regulated ZEB1 and TGFβ2. Furthermore, he demonstrated in vitro that knockdown of ZEB1 reduced
the cancer stem cell (CSC) population, outgrowth of potential CSC spheres, and the expression of
stem cell factors, like Bmi1 and Sox2. Furthermore, expression of ZEB1 correlated with Bmi1
expression in human pancreatic cancer. Taken together, Thomas Brabletz proposed that EMT-
associated tumor cells at the invasive front act as “migrating cancer stem cells” which can re-
differentiate and depending on the range of dissemination, give raise to primary carcinoma
metastases.
Session 6:
STEM CELLS

“Wnt/β-catenin in stem and cancer stem cells“

Walter Birchmeier (Berlin) gave a talk highlighting the essential functions of Wnt/β-catenin in stem
cells and cancer stem cells. He used genetic mouse models to study the necessity of Wnt/β-catenin
and Bmp signaling in stem cells of the salivary gland. β-catenin loss-of-function prevented stem cel
differentiation towards the hair lineage and led to less precursor cells and more differentiated cells in
the skin. In contrast to this, β-catenin gain-of-function produced dozens of teeth from one tooth bud.
Wnt/β-catenin controlled the balance between self-renewal and differentiation of precursor cells and
the specification of dorsal neurons in the central nervous system. In addition, β-catenin was required
for DMBA/TPA-induced tumor formation in the skin. He showed that β-catenin signaling was essential
to maintain cancer stem cells in the skin. Combined activation of Wnt/β-catenin with blockage of Bmp
signals produced highly aggressive squamous cell carcinomas. These tumors contained a high
number of CD24+CD29+ stem cells and as few as 500 of these cells induced tumors following
transplantation into NOD/SCID mice. If the Wnt/β-catenin and Bmp pathways were individually
mutated, stem cells showed an increased potential for tissue regeneration but no tumors occurred.
Walter Birchmeier concluded that his work defines a switch in stem cell potency from single to double
mutations.
“A hypoxic niche regulates glioma stem cells”

Till Acker (Giessen) presented data regarding CD133+ tumor stem cells (TSCs). Glioma growth and
recurrence have been shown to depend on CD133+ tumor stem cells. He demonstrated that side
population (SP) cells display TSC characteristics. SP cells were able to self-renew, regenerated SP as
well as non-SP cells within a cellular hierarchy and were highly tumorigenic in vivo. SP cells fulfilled
the criteria of tumor stem cells proven by FACS analysis and they formed tumors with a fivefold
increased probability compared to non-SP cells. The SP is characterized by a distinct gene signature
and he identified 73 genes using a comprehensive transcriptional profiling analysis of SP and non-SP
cells. These signature genes were overexpressed by TSCs in vascular and perinecrotic/hypoxic
niches. Specifically, the hypoxic microenvironment played a key role in the regulation of the TSC
phenotype, through the hypoxia-inducible factor (HIF)-2α and subsequent induction of specific TSC
genes. In conclusion, Till Acker proposed that TSCs are maintained within a hypoxic niche, providing a
functional link between the well-established role of hypoxia in stem cell and tumor biology.
“The vascular wall as a source of stem cells in adult organs”

Bruno Péault (Los Angeles) focused on mesenchymal stem cells (MSC) whose identity, frequency
and location have remained obscure. He suggested that these multi-lineage progenitors originated in
blood vessel walls. He documented anatomic, molecular and developmental relationships between
endothelial cells and myogenic cells within human skeletal muscle. Myoendothelial cells coexpressing
myogenic and endothelial cell markers (CD56, CD34 and CD144) were identified by
immunohistochemistry and flow cytometry. Cultured myoendothelial cells proliferated, retained a
normal karyotype, were not tumorigenic and survived better under oxidative stress than regular
myogenic cells. Clonally derived myoendothelial cells differentiated into myogenic, osteogenic and
chondrogenic cells in culture. He identified in multiple human organs perivascular cells expressing
CD146, NG2 and PDGF-Rβ which did not express hematopoietic, endothelial, and myogenic cell
markers. Pericytes purified from skeletal muscle or non-muscle tissues were myogenic in vivo. Long-
term cultured pericytes retained myogenicity, expressed MSC markers and migrated by chemotaxis.
Furthermore, pericytes reduced myocardial fibrosis after acute myocardial infarction. In conclusion,
Bruno Péault showed that blood vessel walls harbor a reserve of progenitor cells that may be integral
to the origin of MSCs and other related adult stem cells.
Session 7:
BONE MARROW-DERIVED CELLS

“Bone marrow derived progenitor cells in angiogenesis and tumor growth”

Petri Salven (Helsinki) investigated the mobilization of BM-derived cell populations to the circulation.
Therefore, he inoculated VEGF polypeptides, adenoviral vectors expressing VEGF or syngeneic
tumors in wild type mice. The current concept indicates that a significant part of neovascular
endothelial cells (ECs) originates from circulating “precursor” or “progenitor” cells. These cells are first
mobilized from the BM and subsequently differentiate to mature ECs and incorporate into the growing
vasculature. These cells were originally defined as VEGF receptor 2-expressing (VEGFR-2) cells that
are mobilized from the BM by VEGF or by tumors. Using a genetically tagged syngeneic BM stem cell
transplantation model Petri Salven showed that systemic VEGF or tumors did not promote the
mobilization of VEGFR-2+ BM cells to the circulation. GFP+ BM derived cells infiltrated the tumor and
were recruited close to blood vessel wall ECs. However, they did not form parts of the endothelium.
He showed that all BM-derived cells in angiogenic tissue were perivascular. Using endothelial cell
specific genetic reporter systems Petri Salven identified no BM-derived VEGFR-2+ or Tie1+ ECs in
angiogenic neovasculature. He concluded that endothelial differentiation is not a typical in vivo
function of normal BM-derived stem cells and it has to be an extremely rare event if it occurs at all. In
addition, he showed that angiogenesis during tumor growth did not involve or require contribution from
BM-derived circulating progenitors for vascular ECs. Instead, large numbers of perivascular cells
derived from CD117(c-kit)+/Sca-1+/lin-/Thy1.1low hematopoietic stem cells of the BM were recruited
to angiogenic sites during neovascularization or tumor growth. Taken together, while BM-derived cells
may play a significant role in the formation of neovasculature and in cancer growth, their contribution
does not involve differentiation of BM stem cells to vascular endothelial cells.
“Tumor-mediated education of bone marrow-derived cells in tumor angiogenesis and tumor
progression”

Curzio Rüegg (Lausanne) presented the role of bone marrow derived (BMD) cells during tumor
angiogenesis. It is unclear whether angiogenic activity of BMD cells is acquired only upon local
recruitment to the tumor or it may occur before these cells reach the tumor site. Using in vitro and in
vivo studies and correlative analyzes of patient material he characterized the effect of tumor-derived
factors on the generation of angiogenic myelomonocytic cells from hematopoietic progenitor cells. He
observed that breast cancer cells could educate monocytes to develop angiogenic capacity during
differentiation from hematopoietic progenitor cells. This angiogenic education was associated with the
expression of CCR5+ on CD11b+ monocytes in cell culture and with the increased frequency of
circulating CCR5+/CD11b+ monocytes in breast cancer-bearing mice and breast cancer patients.
CD11b+ cells educated by breast cancer cells induced endothelial cell (EC) sprouting in co-culture
experiments. Furthermore, circulating CD11b+ cells that were also Tie2 positive promoted EC
sprouting in breast cancer patients. Curzio Rüegg demonstrated that MMP-9 was highly upregulated
during BMD cell education. In addition, MMP inhibitors blocked the angiogenic activity of educated
BMD cells. He concluded that breast cancer can influence hematopoiesis towards the generation of
angiogenic monocytes and that this provides a new mechanism on how tumor cells hijack normal host
events to their advantage.
Session 8:
LYMPHANGIOGENESIS VERSUS ANGIOGENESIS

“Interfering with growth factor crosstalk for angiogenesis vs. lymphangiogenesis”
Kari Alitalo (Helsinki) investigated the crosstalk between angiogenesis and lymphangiogenesis via
growth factors. Vascular endothelial growth factor (VEGF) stimulates angiogenesis via its two
receptors VEGFR-1 and VEGFR-2, but it has only little lymphangiogenic activity. The VEGFR-3
receptor does not bind VEGF and its expression becomes restricted mainly to lymphatic endothelia
during development. Transgenic mice expressing the VEGFR-3 ligand VEGF-C or VEGF-D showed
evidence of lymphangiogenesis. In addition, VEGF-C knockout mice exhibited defective lymphatic
vessels. The proteolytically processed form of VEGF-C also bound to VEGFR-2 and was angiogenic.
He introduced the molecule Claudin like protein 24 (CLP24) which interacted with VEGFR-2 activating
it and leading to its internalization. CLP24 and VEGFR-2 had the same expression pattern except in
the retinas. In addition, CLP24 expression was increased in lymphatic vessels and had been shown to
be important for lymphatic and vascular development in Xenopus. Furthermore, soluble VEGFR-3
inhibited embryonic and tumor lymphangiogenesis and lymphatic metastasis. He showed the role of
VEGFR-3 signaling in the settings of physiological and pathological angiogenesis. VEGFR-3 blocking
antibodies provided significant inhibition of tumor angiogenesis and growth in several xenograft
models. He found that angiopoietin-1 could stimulate lymphatic sprouting while secretion of
angiopoietin-2 by endothelial cells led to hyperplasia in developing lymph vessels. Kari Alitalo
concluded that receptor trafficking and compartmentalization determine the angiopoietin-specific
signals via cell-cell and –pericellular matrix interactions.
“Vascular Endothelial Growth Factor-C protects prostate cancer cells from oxidative stress by
activation of mTORC-2 and AKT-1”
Michael Muders (Dresden) presented data regarding the protective role of vascular endothelial
growth factor-C (VEGF-C) on reactive oxygen stress induced cell death by activation of mTOR
complex 2 (mTORC-2). Recurrence and subsequent metastatic transformation of cancer developed
from a subset of malignant cells which showed the ability to resist stress and to adopt to a changing
microenvironment. These tumor cells had distinctly different growth factor pathways and anti-apoptotic
responses compared to the vast majority of cancer cells. Both microarray and immunohistochemical
analysis of human prostate cancer tissue samples have shown an increased expression of vascular
endothelial growth factor-C (VEGF-C) in metastatic prostate cancer. He discovered that VEGF-C
acted directly on prostate cancer cells to protect them against oxidative stress. VEGF-C increased the
survival of prostate cancer cells during hydrogen peroxide stress by activation of AKT-1/PKBα. This
activation was mediated by mTORC-2 and was not observed in the absence of oxidative stress.
Finally, the transmembrane non-tyrosine kinase receptor Neuropilin-2 was found to be essential for
the VEGF-C-mediated AKT-1 activation. His findings suggest a novel and distinct function of VEGF-C
in protecting cancer cells from stress-induced cell death, thereby facilitating cancer recurrence and
metastasis. This is distinctly different from the known function of VEGF-C in inducing lymph-
angiogenesis.
“The tumor-lymphatic interface: CCL21 as an important player in invasion and immune
escape”

Melody A. Swartz (Lausanne) proposed a synergistically crosstalk between the chemokine (C-C
motif) ligand 21 (CCL21) and the vascular endothelial growth factor (VEGF)-C which enables tumor
cells to escape the immune surveillance as well as to obtain an increased invasiveness and thereby
promoting metastasis via the lymphatic route. The secretion of VEGF-C by tumor cells, tumor-
associated macrophages and stromal cells resulted in an increased expression of the CCL21 on
lymphatics, which in turn drove CCR7-dependent tumor chemoinvasion towards lymphatics.
Furthermore, CCL21 was also expressed by invasive tumor cells. Several in vivo experiments
indicated the recruitment of lymphoid tissue inducer cells to invasive CCL21+ tumors and the
upregulation of HEV marker expression on intratumoral vessels, lymph node stromal cell markers on
surrounding fibroblasts and M2 macrophage switching. These lymphoid-like changes together with the
tolerogenic cytokine profile of the tumor microenvironment led to the recruitment of naïve CCR7+ T-
cells and their priming to regulatory T-cells. Thus, the tumor escaped the immune response and
spread via lymphatics.
Session 9:
TUMOR MICROENVIRONMENT

“Exploring new strategies to target the pro-angiogenic tumor stroma”
Kristian Pietras (Stockholm) talked about bone morphogenetic protein (BMP) 9 as a new player in
activin receptor-like kinase (ALK) 1 signaling. He used the RIP1-TAG2 insulinoma mouse model. He
found progressive upregulated expression of TGFβ and BMP9 during tumorigenesis, which correlated
with highest ALK1 expression during the angiogenic state. Furthermore, impaired ALK1 signaling
retarded tumor progression and reduced tumor vascular density in RIP1-TAG2; ALK1+/- mice. He
generated RAP-041, an ALK1-Fc fusion protein that neutralized specifically BMP9. RAP-041 inhibited
physiological vessel formation in vitro and in vivo. RAP-041 treatment of RIP1-TAG2 mice resulted in
a therapeutic benefit. Genetic as well as pharmacological targeting of ALK1 led to reduced down-
stream signaling in vivo. To analyze the mechanistic basis of these findings he analyzed the
endothelial cell behavior in proliferation and apoptosis assays in vitro. He demonstrated that TGFβ
and BMP9 acted only in concert to enhance VEGF-A function resulting in increased proliferation rates
and reduced apoptosis. The same TGFβ and BMP9 interaction was found in an ex vivo angiogenic
islet sprouting assay and in an in vivo assay using Matrigel plugs. In both cases the cytokines
stimulated endothelial cell sprouting. Inhibition of ALK1 signaling by phosphorylated SMAD2 affected
prototypical ALK5 downstream target genes. Kristian Pietras further introduced endoglin as a co-
receptor of ALK1, which was expressed by endothelial cells in the RIP1-TAG2 tumors. Endoglin
deficient mice showed improved pericyte coverage of blood vessels and increased formation of liver
micrometastases. In summary, Kristian Pietras concluded that the TGFβ-BMP9-ALK1-endoglin
pathway is important for the angiogenic switch.
“Tenascin-C in the tumor microenvironment triggers oncogenic signaling“

Gertraud Orend (Strassburg) introduced the extracellular matrix molecule tenascin-C (TNC) that is
highly expressed in most cancers and which correlates with bad survival prognosis and tamoxifen
resistance. TNC played a role in promoting tumor cell proliferation, angiogenesis, invasion and
metastasis but the molecular mechanisms are poorly understood. She crossed Rip-Tag mice that
express the SV40 antigen under the control of the insulin promoter with TNC transgenic animals and
analyzed the tumor progression in the pancreas of these mice. The animals showed enhanced lung
micrometastases with insulin positive areas and increased β-catenin expression and nuclear β-catenin
localization. TNC stabilized β-catenin and induced tumor vasculature with an aberrant structure.
Tumors overexpressing TNC exhibited tubular structures in the matrix resembling structures described
in human tumors (e.g. human insulioma, certain colon carcinoma). In addition, TNC stimulated
endothelin receptor type A (EDNRA) expression which maintained cell rounding. FAK, paxilin, RhoA
and Tropomyosin-1 were critical targets of TNC downstream of syndecan-4 and EDNRA. Taken
together, TNC in the tumor microenvironment triggers oncogenic signaling.
“Diverse role of mesenchymal cells in cancer progression and metastasis“

Raghu Kalluri (Boston) talk presented the previously unappreciated presence of fibroblast diversity
and its influence on primary tumor growth and metastasis formation. He showed an exclusive
anatomical distribution patterns in fibroblast lineage cells in mice in which fluorescent reporter genes
were expressed under the control of alpha-SMA, NG2 or FSP1 promoter respectively. The functional
relevance of the above finding was shown using a 4T1 breast cancer model in which 90% of cancer
associated fibroblast were FSP1+ and were alpha-SMA-. Additionally, he demonstrated that in
contrast to bone morrow derived cells, the local fibroblast population formed the premetastatic niche.
To confirm the functional relevance of the diverse fibroblast cells in tumor growth, transgenic mice
expressing thymidine kinase under the control of alpha-SMA, NG2 and FSP promoter were generated
and the cells were selectively ablated by ganglocyclovir administration. In the 4T1 orthotropic breast
cancer model, the depletion of alpha-SMA+ cells resulted in reduced tumor growth but had no effect
on metastasis. Absence of NG2+ cells caused leaky vessels, reduced primary tumor and increased
metastasis due to increased levels of HIF1alpha, p38 and TWIST. Mice deprived of FSP1 cells
exhibited no changes in primary tumor growth but altered metastasis. Finally, he addressed the
contribution of mutations occurring in fibroblast cells in pathophysiologcal conditions. Transgenic mice
harboring mutant KrasG12D gene under the control of alpha-SMA promoter predominantly died
shortly after birth and surviving pups developed transformed breast epithelial cells during the first
month. Taken together, he demonstrated that the diversity of fibroblast cells plays an important role in
primary tumor growth and metastasis development.
Session 10:
TUMOR SIGNALING

“EphrinB reverse signaling contributes to endothelial and mural cell assembly into vascular
structures”

Giovanna Tosato (Bethesda) presented her data about ephrinB and EphB receptor and their
influence on vascular development through bidirectional cell-to-cell signaling. She showed that
ephrinB was a critical mediator of post-natal pericyte-to-endothelial cell assembly into vascular
structures. This function was dependent upon extracellular matrix-supported cell-to-cell contact,
engagement of ephrinB by EphB receptors on cells and Src-dependent phosphorylation of the
intracytoplasmic domain of ephrinB. Phosphorylated ephrinB marked angiogenic blood vessels in the
developing and hypoxic retina, in tumor tissue and at contact points between endothelial cells and
pericytes. Inhibition of ephrinB activity prevented proper assembly of pericytes and endothelial cells
into vascular structures. Thus, ephrinB signaling orchestrates pericyte/endothelial cell assembly,
suggesting that therapeutic targeting of ephrinB may be useful to inhibit angiogenesis when it
contributes to disease.
“EphrinB2 regulates VEGFR trafficking to mediate tip cell filopodial extension during
angiogenesis”

Amparo Acker-Palmer (Frankfurt) demonstrated in her talk that ephrinB2 signaling involving PDZ
interactions regulated tip cell guidance to control angiogenic sprouting and branching in mouse retina
and tumor angiogenesis. In vivo, ephrinB2 PDZ signaling deficient mice exhibited reduced number of
tip cells with less filopodial extensions. In pathological settings, blockade of the ephrinB2 PDZ
signaling pathway decreased tumor vascularization and led to reduced angiogenic sprouting and
branching of tumor vasculature. At cellular level, time-laps microscopy revealed that ephrinB2
signaling induced the extension of filopodia by regulating the internalization of VEGFR-2 in endothelial
cells. Interestingly, activation of ephrinB2 rescued the blockade of VEGFR-2 by soluble Flt1 regarding
tip cell filopodia dynamics both in retinal explants and in cultured endothelial cells. In summary,
Amparo Acker-Palmer showed that ephrinB2 has an essential role in the guidance and function of tip
cell filopodia during development and also during sprouting angiogenesis.
“The bone morphogenic protein antagonist DRM/Gremlin as a novel pro-angiogenic factor”
Marco Presta (Brescia) introduced the protein Drm/gremlin, a member of the Dan family of bone
morphogenetic protein (BMP) antagonists. Drm/Gremlin binds to BMP 2, 4, and 7 on the endothelial
surface. He purificated Drm/gremlin from conditioned medium of FGF2-overexpressing endothelial
cells. Recombinant Drm/gremlin stimulated endothelial cell migration and invasion on fibrin and
collagen gels. Moreover, Drm/gremlin exerted a potent angiogenic activity in the chick embryo chorio-
allantoic membrane and in Zebrafish embryos. Sequence alignment analysis showed sequence
similarities to VEGF-A. Drm/gremlin was also able to bind to VEGFR-2 and induces dimerization in
vitro. This interaction mediated the angiogenic activity of Drm/gremlin in vitro and in vivo by
phosphorylation of intracellular proteins including Paxillin, focal adhesion kinase and ERK1/2.
Drm/gremlin also induced Ang-1 upregulation in endothelial cells which was mediated by the activation
of the transcription factor NF-kB. Marco Presta demonstrated that Drm/gremlin was highly
overexpressed in human tumors (e.g. lung, pancreas, prostate) where it stimulated tumor growth. He
concluded that Drm/gremlin has a novel, previously unrecognized capacity to modulate angiogenesis
in vitro and in vivo by interacting directly with VEGFR-2 on endothelial cells.
“EGFR/Ras/ERK-signaling-dependent production of the chemokine CCL20 in tumor cells
critically contributes to angiogenesis and tumor progression”
Andreas Hippe (Düsseldorf) presented his data about the EGFR/Ras/ERK-signaling regulated
chemokine CCL20. Tumors enhance angiogenesis by upregulating the production of CCL20. In vivo,
the chemokine CCL20 was overexpressed in areas of melanoma, breast cancer, colon cancer and
head and neck squamous cell carcinoma with increased ERK phosphorylation. Using large scale
tissue microarrays he identified increased CCL20 expression in most advanced or aggressive tumors
and the expression of CCL20 also correlated with tumor progression and lymph node metastases. Its
specific corresponding receptor, CCR6, was abundantly expressed on endothelial cells in vitro and in
vivo. Activation of CCR6 signaling in endothelial cells induced migration, in vitro-repair and led to
enhanced vessel formation. In vivo, CCL20 specifically induced increased vascularization of Matrigel
plugs in wildtype mice, which was abrogated in CCR6-deficient mice. In addition, CCL20 expressing
B16F10 melanomas showed significantly decreased tumor growth and vessel density in CCR6-
deficient compared to wildtype mice. Andreas Hippe propagates CCL20 as a novel chemokine-driven
mechanism of tumors to promote angiogenesis and tumor progression.
Session 11:
NOTCH SIGNALING

“Notch signaling – new mechanism and role in resistance to anti-VEGF therapy”

Adrian Harris (Oxford) investigated the role of Dll4 and Avastin on tumor growth. MDA231 cells
transfected with Dll4 exhibited increased tumor volume and led to reduced survival of the mice.
Simultaneous administration of Avastin for one and a half week significantly reduced tumor volume.
He also showed that inhibition of Notch signaling sensitized tumors to anti-VEGF therapy. Avastin did
not affect the vessel morphology, but instead it decreased the number of vessels. Resistance of
tumors to VEGF blockade was due to poor antibody penetration. Upregulation of Dll4 in tumors
resulted in resistance to Sorafenib treatment. He demonstrated that Dll4/Notch signaling affected
perivascular coverage. Dll4 regulated EphB4 and blockade of ephrinB2 re-sensitized tumor response
to anti-VEGF therapy. The second story Adrian Harris presented was the role of exosomes in tumor
angiogenesis. He investigated in U87 glioblastoma cells transfected with Dll4 the production of
exosomes. The exosomes contained beside Dll4 the exosomal markers Rab5, Tsg101 and Lamp1.
Addition of Dll4 containing exosomes to HUVECs inhibited Notch signaling and enhanced branching.
The mechanism was related to removal of Notch1 from the cell surface. He showed that the exosomes
were incorporated into mouse tumor endothelial cells in vivo. This incorporation could be one of the
possible mechanisms regulating blood vessels at a distance from tumors and enhancing vessel
branching. Therefore, Notch signaling either produced from stromal cells, endothelial cells themselves
or cancer cells have a highly significant role in tumor angiogenesis and are likely to modify the
response to many different anti-angiogenic therapies.
“Decoding tumor-host interactions in dormancy: Notch3-mediated regulation of MKP-1
promotes tumor cell survival”
Stefano Indraccolo (Padova) presented data regarding the regulation of the phosphatase MKP-1 by
Notch3. The Notch ligand Dll4, induced by angiogenic factors in endothelial cells, triggers Notch3
activation in neighboring T-ALL leukaemia cells and promotes tumorigenesis. He showed new data
that MKP-1 levels were controlled by Notch3 by protein ubiquitination and stability and not by gene
expression. Notch3 and MKP-1 levels were consistently upregulated in aggressive compared to
dormant tumors. A good correlation between Notch3 ICD and MKP-1 levels was observed in T-ALL
primary samples from patients and in a panel of T-ALL cell lines. Silencing Notch3 by RNA
interference or by γ-secretase treatment or stimulation of Notch3 by the Dll4 ligand had marked effects
on MKP-1 levels in T-ALL cells in vitro. Stefano Indraccolo also showed that MKP-1 was
downregulated by anti-DLL4 therapy. On the one hand Notch3 inhibition lowered MKP-1 levels and on
the other hand Notch3 activation increased MKP-1 levels in T-ALL cells. Attenuation of MKP-1 levels
by shRNA did not affect proliferation, whereas it significantly increased T-ALL cell death and thereby
controlled leukemia outgrowth. In summary, Stefano Indraccolo presented a novel mechanism
downstream of Notch3 by which the direct interplay between endothelial and tumor cells promote
survival of T-ALL cells.
“Two Notch ligands with opposing effects on angiogenesis”

Expression of the ligand Delta-like 4 (Dll4) in tip cells activates Notch receptors in adjacent endothelial
cells and downregulates VEGF receptor expression in these cells. Ralf Adams (Münster)
demonstrated that sprouting was also controlled by a second Notch ligand, Jagged1, which is a potent
pro-angiogenic regulator with the opposite role of Dll4. Sprouting and tip cell numbers were reduced in
endothelial cell-specific Jag1 loss-of-function mice, whereas overexpression of Jagged1 produced the
reverse effect. Notch target genes were upregulated in the Jagged1-deficient endothelium and
blocking of Notch signaling in these mutants restored a wild-type-like response. Ralf Adams also
showed that loss of Jag1 led to downregulation of VEGFR-3. These findings and the spatial expression patterns of Dll4 and Jagged1 suggested that the balance between Notch ligands with opposing roles controlled physiological and perhaps also pathological angiogenesis. In addition, he showed new data regarding the expression of ephrinB2 and EphB4 in the retinal endothelium. Knockdown of ephrinB2 in zebrafish resulted in sprouting defects. He demonstrated that ephrinB2 controlled VEGFR-3 endocytosis and signaling. EphrinB2 is also a regulator of lymphangiogenesis. EphrinB2 expressing lymphatic vessels exhibited altered VEGFR-3 signaling. Taken together, ephrinB2 is a regulator of endothelial cell invasiveness and sprouting. (supported by: Ilse Hofmann, Vijayshankar Sivanandam, Sonja Breuninger, Anja Runge, Soniya Savant, Miriel Teichmann, Anja Weick, Matthias Wieland)

Source: http://www.vwfb.de/Seeon2009/Seeon%20Meeting%20Report%202009.pdf