The contribution of bone marrow-derived cells to angiogenesis and lymphangiogenesis in murine models of carcinogenesis

Cancer is a class of diseases in which a group of cells display uncontrolled growth (division beyond
the normal limits), invasion (intrusion on and destruction of adjacent tissues), and sometimes
metastasis (spread to other locations in the body via the circulatory system). In addition to tumor
cell-intrinsic genetic and epigenetic alterations, the tumor stroma, i.e. endothelial cells, pericytes,
fibroblasts and a diverse immune cell infiltrate, might substantially contribute to tumor progression,
metastatic potential and resistance to therapy.
I therefore investigated the influence of immune cells on the growth of tumors in the
Rip1Tag2 mouse insulinoma model of multistage carcinogenesis. I detected a strong infiltration of
myeloid cells, i.e. macrophages and granulocytes, into insulinomas. Functional experiments in vivo
revealed that depletion of macrophages in tumors led to reduced angiogenesis but did not affect
tumor growth.
During the characterization of the immune cell contribution to tumor growth in the Rip1Tag2
tumor model, I detected bone marrow-derived cells at unexpected sites. In particular, when I
analyzed the spatial contribution of GFP-tagged bone marrow cells in tumors of lymphangiogenic
Rip1Tag2;RipVEGF-C mice, I detected bone marrow-derived cells in lymphatic endothelium
surrounding the tumors.
Detailed analysis of the integrated GFP+-cells revealed the expression of a complete set of
markers that are characteristic for lymphatic endothelial cells, the cell surface proteins LYVE-1 and
Podoplanin, as well as the homeo-box transcription factor Prox-1. Depending on the analysis
technique applied, either confocal microscopy followed by 3D reconstitution or flow cytometry,
between 3 and 9% of lymphatic endothelial cells in tumors are derived from the bone marrow.
These studies were expanded to a second tumor model, the subcutaneous growth of TRAMP-C1
prostate cancer cells in syngenic mice, which confirmed the findings made in Rip1Tag2;
RipVEGF-C mice, and allowed to further substantiate the suggested ontogeny of the integrated,
bone marrow-derived cells.
Cell sorting and genetic lineage tracing experiments indicated that the bone marrow-derived
tumor lymphatic endothelial cells were at least partially derived from the myeloid lineage. Tumor
mice were adoptively transferred with labeled myeloid (progenitor) cells, and subsequent
integration of these cells into tumor lymphatic endothelium was detected. Cre/Lox technology
resulting in myeloid-specific marker gene expression was employed to come to similar conclusions
in a pure genetic experimental system without bone marrow cell-transfer or irradiation.
In a loss-of-function approach, macrophages were pharmacologically depleted in
Rip1Tag2;RipVEGF-C mice. Peritumoral lymphatic vessel coverage was found to be reduced in 2
macrophage-depleted mice as compared to control mice. Expression level analysis of the
lymphangiogenic factors VEGF-C and VEGF-D by tumor-infiltrating macrophages indicated that
their contribution to lymphangiogenesis by supplying growth factors is negligible and that the
reduced lymphangiogenesis might indeed come from the reduced availability of macrophages as
building blocks of lymphatic endothelia.
The same plasticity of myeloid cells I detected in vivo was also observed in vitro, where
bone marrow-derived macrophages start forming tube like structures and also start expressing
lymphatic endothelial markers, when cultured under pro-inflammatory and endothelial specific
conditions.
In conclusion, this data indicates a myeloid origin of cells that trans-differentiate into
lymphatic endothelial cells in an inflammatory tumor environment.
The increasing use of non-invasive imaging technologies prompted us to evaluate an
approach resulting in bioluminescent pancreatic insulinoma, principally an improved Rip1Tag2
tumor model of multistage pancreatic β-cell carcinogenesis. I therefore constructed a bicistronic
expression cassette in which SV40 early region is followed by an internal ribosomal entry site and
a firefly luciferase coding sequence, under the transcriptional control of the Rat insulin promoter 1.
Transgenic expression in mice resulted in β-cell carcinogenesis that could be monitored noninvasively
by in vivo bioluminescence. Numerous tumors of different malignancy stages can be
detected in individual mice, indicating that this model recapitulates multistage carcinogenesis. In
addition, in this mouse strain called RL-1 (RipTag-IRES-Luciferase line 1), due to the very stringent expression exclusively in the β-cells of Langerhans islets, we could determine micro-metastasis in
liver via luciferase expression of metastatic cells. This mouse line will be of value to study antitumoral
therapeutic approaches in real-time, as well as to define roles for tumor-promoting as well
as metastasis-related genes when crossed to other transgenic or gene-targeted mice.