Embryonic stem cell tumor model reveals role of vascular endothelial receptor tyrosine phosphatase in regulating Tie2 pathway in tumor angiogenesis

Abstract

Inhibiting angiogenesis has become an effective approach for treating cancer and other diseases. However, our understanding of signaling pathways in tumor angiogenesis has been limited by the embryonic lethality of many gene knockouts. To overcome this limitation, we used the plasticity of embryonic stem (ES) cells to develop a unique approach to study tumor angiogenesis. Murine ES cells can be readily manipulated genetically; in addition, ES cells implanted subcutaneously in mice develop into tumors that contain a variety of cell types (teratomas). We show that ES cells differentiate into bona fide endothelial cells within the teratoma, and that these ES-derived endothelial cells form part of the functional tumor vasculature. Using this powerful and flexible system, the Angiopoietin/Tie2 system is shown to have a key role in the regulation of tumor vessel size. Endothelial differentiation in the ES teratoma model allows gene-targeting methods to be used in the study of tumor angiogenesis.

Fig. 1.

ES cell-derived blood vessel networks in teratomas. (A) Whole-mount LacZ staining seen at low and high magnification (Upper and Lower, respectively) of teratomas grown from ES cells in which the β-galactosidase gene was introduced into the locus of cell-type specific genes, including Nebulin (muscle cell specific, a and e), Keratin-5 (complex epithelial cells, b and f), Lyve-1 (lymphatic endothelial and other cells, c and g), and VEGF-R2 (vascular endothelial cells, d and h). (B) Whole-mount LacZ staining seen at low and high magnification (Upper and Lower, respectively) of teratomas grown from ES cells in which the β-galactosidase gene was introduced into other endothelial cell-specific genes Dll4 (a and d), angiopoietin-2 (b and e), and endoglin (c and f). (C) FACS analysis of teratomas derived from either WT or Rosa-eGFP ES cells. Cells expressing both CD31/Pecam and eGFP were detected in Rosa-eGFP tumors (b, circled) (5.43% of the total cell population), whereas no such population was detected in WT tumors (a) or in Rosa-eGFP tumors stained with isotype control antibody (c) or no antibody (d).

Fig. 2.

Teratoma vasculature consists of both ES cell-derived and host-derived vessels. (A) Immunohistochemistry of teratomas in which the β-gal gene was targeted into VEGF-R2 locus. Adjacent sections of the tumor were immunostained for Pecam/CD31 (a) or β-gal (b). Areas of interest framed in (a) and (b) shown at higher magnification (magnification 10×) (c–h). (c–e) Pecam/CD31 staining; (f–h) adjacent sections stained for β-gal. (c and f) A network of vascular structures derived from the ES cells (immunostained for both Pecam and β-gal) is shown, whereas (d) and (g) show a network of vessels in the same tumor derived from the host (Pecam⁺ but β-gal⁻). (e and h) Boundary of host-derived vessels on the right hand side and ES-derived vessels on the left hand side. (B) Comparison of ES cell-derived blood vessels from VEGF-R2 het and VEGF-R2 null (KO) ES tumors. Whole-mount view of tumors from VEGF-R2 het (a) shows regions of LacZ staining (ES-derived vessels), whereas tumors from VEGF-R2 KO (b) have very few such regions. Area density (c) of LacZ staining from VEGF-R2 KO tumors is less than 10% of that from (control) VEGF-R2 het tumors.

Fig. 3.

Deletion of VE-PTP results in tumor blood vessel enlargement. (A) Immunohistochemistry of teratomas in which the β-gal gene was targeted into either one (het) or both (KO) alleles of the VE-PTP locus. Tumor sections were immunostained for both β-gal (black-purple) and CD31/Pecam (brown). (a and d) Small-diameter vessels derived from VE-PTP het ES cells (low and high magnification views, 2.5× and 10×, respectively). (b) A low magnification view of vessels in a VE-PTP KO ES tumor, in which the black vessels (β-gal⁺) are derived from the ES cells and the brown vessels (Pecam⁺ only) are derived from the host. (c, e, and f). Enlarged views from (b), showing vessels of different origin. Note the enlarged diameter of the VE-PTP KO vessels (black) compared to the adjacent normal host vessels (brown). (B) Pharmacologic inhibition of angiopoietin-Tie signaling reduces tumor vessel diameter. Tumor sections from ES cells in which the β-gal gene was targeted into either one (het) or both (KO) alleles of the VE-PTP loci were immunostained for both β-gal (black-purple) and CD31/Pecam (brown). (magnification 10×) (a) Small-diameter vessels derived from VE-PTP het ES cells treated with control protein (hFc). (b) Enlarged vessels in a VE-PTP het ES cell tumor treated systemically with recombinant angiopoietin-1 (Ang-1). (c) Small diameter vessels in a VE-PTP het ES cell tumor treated systemically with an inhibitor of angiopoietins. (d) Large-diameter vessels derived from VE-PTP KO ES cells treated with control protein (hFc). (e) Further enlargement of the vessels in a VE-PTP KO ES cells tumor treated systemically with recombinant angiopoietin-1 (Ang-1). (f) Small-diameter vessels in a VE-PTP KO ES cell tumor treated systemically with an inhibitor of angiopoietins.

Fig. 4.

VE-PTP regulates angiopoietin-induced vessel diameter and Tie2 phosphorylation. (A) Distribution of blood vessel diameter (ES-derived blood vessels only) in VE-PTP het and VE-PTP KO tumors treated with control protein (hFc), recombinant angiopoietin-1, or angiopoietin inhhibitor. Tumor-bearing mice were systemically treated with hFc (a and d), angiopoietin-1 (b and e), or angiopoietin inhibitor (c and f). For each individual graph, the y axis shows frequency (percentage) of ES-derived vessels with diameter within a certain range as indicated on the x axis (vessel diameter bins, in micrometers). Mean values of vessel diameter from each group are also shown (Ave). The distributions are from 1,150 to 1,600 individual vessel measurements from four tumors per group (275–400 measurements per tumor). The averages shown in the panels are the means of the four individual mean diameters per group. (B) Western blot showing Tie2 phosphorylation following exogenous angiopoietin-1 treatment of teratomas. Mice bearing VE-PTP het and VE-PTP KO ES tumors were treated with either hFc or Ang1 for 4 h. Total tumor protein lysates were immunoprecipitated with Tie2 antibody and probed with either anti-Tie2 antibody or anti-phosphotyrosine antibody, as indicated. Each lane shows individual tumor (four to five tumors per group). Increased Tie2 phosphorylation was detected in VE-PTP KO ES tumors either with or without Ang1 treatment. (C) Western blot showing effect on Tie2 phosphorylation of pharmacologic inhibition of angiopoietins in teratomas. Mice bearing VE-PTP KO ES tumors were treated with hFc or angiopoietin inhibitor for 12 h. Total tumor protein lysates were immunoprecipitated with Tie2 antibody and probed with either anti-Tie2 antibody or anti-phosphotyrosine antibody as indicated. Tie2 phosphorylation was reduced in VE-PTP KO ES tumors treated with angiopoietin inhibitor. Equal amounts of total protein loaded in each lane. The amount of total Tie2 was somewhat reduced in the tumors treated with angiopoietin inhibitor, which may be because of a reduced overall vascularity in these tumors.