Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 Apr 24;14(2):R67.
doi: 10.1186/bcr3174.

Early Vascular Deficits Are Correlated With Delayed Mammary Tumorigenesis in the MMTV-PyMT Transgenic Mouse Following Genetic Ablation of the NG2 Proteoglycan

Affiliations
Free PMC article

Early Vascular Deficits Are Correlated With Delayed Mammary Tumorigenesis in the MMTV-PyMT Transgenic Mouse Following Genetic Ablation of the NG2 Proteoglycan

Krissa Gibby et al. Breast Cancer Res. .
Free PMC article

Abstract

Introduction: The neuron-glial antigen 2 (NG2) proteoglycan promotes pericyte recruitment and mediates pericyte interaction with endothelial cells. In the absence of NG2, blood vessel development is negatively impacted in several pathological models. Our goal in this study was to determine the effect of NG2 ablation on the early development and function of blood vessels in mammary tumors in the mammary tumor virus-driven polyoma middle T (MMTV-PyMT) transgenic mouse, and to correlate these vascular changes with alterations in mammary tumor growth.

Methods: Three different tumor paradigms (spontaneous tumors, transplanted tumors, and orthotopic allografts of tumor cell lines) were used to investigate the effects of NG2 ablation on breast cancer progression in the MMTV-PyMT transgenic mouse. In addition to examining effects of NG2 ablation on mammary tumor growth, we also investigated effects on the structure and function of tumor vasculature.

Results: Ablation of NG2 led to reduced early progression of spontaneous, transplanted, and orthotopic allograft mammary tumors. NG2 was not expressed by the mammary tumor cells themselves, but instead was found on three components of the tumor stroma. Microvascular pericytes, myeloid cells, and adipocytes were NG2-positive in both mouse and human mammary tumor stroma. The effect of NG2 on tumor progression therefore must be stromal in nature. Ablation of NG2 had several negative effects on early development of the mammary tumor vasculature. In the absence of NG2, pericyte ensheathment of endothelial cells was reduced, along with reduced pericyte maturation, reduced sprouting of endothelial cells, reduced assembly of the vascular basal lamina, and reduced tumor vessel diameter. These early deficits in vessel structure are accompanied by increased vessel leakiness, increased tumor hypoxia, and decreased tumor growth. NG2 ablation also diminishes the number of tumor-associated and TEK tyrosine kinase endothelial (Tie2) expressing macrophages in mammary tumors, providing another possible mechanism for reducing tumor vascularization and growth.

Conclusions: These results emphasize the importance of NG2 in mediating pericyte/endothelial cell communication that is required for proper vessel maturation and function. In the absence of normal pericyte/endothelial cell interaction, poor vascular function results in diminished early progression of mammary tumors.

Figures

Figure 1
Figure 1
NG2 expression in mouse mammary tumor stroma. Sections of 16-week MMTV-PyMT mammary tumors were immunolabeled to demonstrate NG2 expression by elements of the tumor stroma. (A-D) Sections were labeled for CD31 (blue), desmin (green), and NG2 (red) to demonstrate the co-expression of desmin and NG2 by pericytes in close association with microvascular endothelial cells. D: high magnification of the boxed area in C. Arrowheads indicate pericytes expressing both desmin and NG2. (E-H) Sections were labeled for CD31 (blue), CD11b (green), and NG2 (red) to demonstrate the expression of NG2 by a subpopulation of CD11b-positive macrophages. H: high magnification of the boxed area in G. Arrowheads indicate macrophages positive for both CD11b and NG2. Arrow denotes blood vessel-associated CD11b-positive, NG2-positive macrophage. (I-L) Sections containing both normal and tumor tissue were labeled for CD31 (blue), desmin (green), and NG2 (red) to demonstrate the non-vascular expression of NG2 by adipocytes in the mammary fat pad. L: high magnification of the boxed area in K. Scale bars = 25 μm in D, H, and L. In all other panels, scale bars = 50 μm. MMTV-PyMT, mammary tumor virus-driven polyoma middle T; NG2, nerve-glial antigen 2.
Figure 2
Figure 2
NG2 expression in human mammary tumor stroma. Sections of non-triple negative human ductal adendocarcinoma were immunolabeled to demonstrate NG2 expression by elements of the tumor stroma. (A-C) Sections were labeled for CD31 (A, blue) and NG2 (B, red). C, merged images from boxed area in B, showing close apposition of pericytes and endothelial cells. (D-F) Sections were labeled for αSMA (D, green) and NG2 (E, red). F, merged images from boxed area in E. Arrowheads indicate overlap between NG2 and αSMA labeling. (G-I) Sections were labeled for CD11b (G, green) and NG2 (H, red). I, merged images from boxed area in H. Arrowheads indicate overlap between NG2 and CD11b labeling. (J-L) Sections were labeled for CD31 (J, blue) and NG2 (K, red). L, merged images from boxed area in K). Images shown are representative of six ductal adenocarcinoma specimens that we examined. Scale bars = 25 μm in C, F, I, L. In all other panels, scale bars = 100 μm. Specimens 1, 3, and 8 in Table 1 were used to obtain these images. NG2, nerve-glial antigen 2; αSMA, α smooth muscle actin.
Figure 3
Figure 3
Mammary gland development in wild type and NG2 null mice. Whole mounts of #4 mammary glands from female wild type (WT: A, B, C) and NG2 null (NG2 KO: D, E, F) mice at the ages of four weeks, ten weeks, and 16 days of gestation (Gest 16d) were stained with carmine alum to visualize development of the mammary epithelium [24]. No changes in mammary gland expansion or branching were noted in the NG2 null mouse. LN: lymph node. (G-I). Whole mounts of normal #4 mammary gland from a 12-week old wild type female were immunostained for NG2 (G, red) and CD31 (H, blue). Panel I shows merged images from the boxed area in G and H. NG2 expression is seen on adipocytes (arrowheads) and on pericytes (arrows) that are closely apposed to CD31-positive endothelial cells. Scale bar = 60 μm (G, H) and 30 μm (I). NG2, nerve-glial antigen 2.
Figure 4
Figure 4
Progression of spontaneous mammary tumors in wild type and NG2 null mice. A. Whole mounts of #4 mammary glands from MMTV-PyMT female mice (WT) and NG2 null MMTV-PyMT females mice (NG2 KO) at six, eight, ten, and 12 weeks of age were stained with carmine alum to visualize the developing mammary epithelium and associated MINs. Arrowheads indicate MINs. LN: lymph node. B. MIN progression in wild type and NG2 null MMTV-PyMT mice was quantified by using image analysis to determine the area occupied by MINs at each time point. Large masses in the nipple area were excluded from this analysis. Eight glands were examined at each time point for each genotype. C. At 14, 17, and 20 weeks of age, total tumor burdens were determined in wild type and NG2 null MMTV-PyMT female mice by dissecting and weighing all tumors. Weights are in grams. At 14 weeks, n = 8 mice of each genotype. At 17 weeks, n = 7 mice of each genotype. At 20 weeks, n = 8 mice of each genotype. Dashed lines: average tumor weights. D. Semi-log plots of the data from panel C show that after an initial fast start by tumors in wild type MMTV-PyMT mice, subsequent tumor growth is similar in wild type and NG2 null hosts. * P = 0.03; ** P = 0.005. E-F. Sections of 17-week tumors from wild type (WT, E) and NG2 null (KO, F) MMTV-PyMT female mice were H & E stained to visualize tissue morphology. Tumors in both wild type and NG2 null hosts are multifocal, heterogeneous for cellularity and tissue morphology, and very cystic in nature. G-J. Sections of 10-week (G, H) and 12-week old (I, J) mammary gland from wild type (WT, G, I) and NG2 null (KO, H, J) MMTV-PyMT female mice were H & E stained to visualize MIN development. At both ages, MINs occupy larger areas in wild type than in NG2 null mammary glands. Scale bars = 100 μm in E, F and 200 μm in G-J. MIN, mammary intraepiethelial neoplasis; MMTV-PyMT, mammary tumor virus-driven polyoma middle T; NG2, nerve-glial antigen 2.
Figure 5
Figure 5
Progression of transplanted mammary tumors in wild type and NG2 null mice. A, B. 1 mm3 fragments of wild type MMTV-PyMT tumors were transplanted to mammary fat pad sites in the #4 glands of wild type (wt) and NG2 null (ko) recipients. Each of the recipients (ten wild type and ten NG2 null females) received two transplants, yielding 20 potential tumor sites in each genotype. Trials 1 and 2 utilized four-month (panel A) and two-month old recipients (panel B), respectively. Data are plotted as the % sites with palpable masses as a function of time after transplantation. C, D. Orthotopic allografts of Py8119 (panel C) and Py230 tumor cells (panel D) were prepared by injecting 106 cells into each of four sites in the #2 and #4 mammary fat pads of wild type and NG2 null recipients (six females of each genotype; 24 potential tumor sites for each genotype). Data are plotted as the % sites with palpable masses as a function of time after transplantation. Indicated P values were obtained via Wilcoxon signed rank tests. MMTV-PyMT, mammary tumor virus-driven polyoma middle T; NG2, nerve-glial antigen 2.
Figure 6
Figure 6
Vessel diameter and macrophage phenotype in wild type and NG2 null mice. A-C. Whole mounts of #4 mammary glands from 14-week old wild type (WT) and NG2 null (KO) MMTV-PyMT females were immunostained for CD31 (red) and αSMA (blue) to investigate vascularization of incipient neoplasms. Although not shown in B and C, αSMA staining was used as illustrated in A to restrict analysis of CD31-positive vessels to areas within developing MINs. Thus, we only analyzed tumor-associated vessel segments and not vessels associated with normal areas of the mammary gland. Tumor-associated vessels in wild type mice (B) appear thicker and are more robustly stained for CD31 than vessels in NG2 null mice (C). To quantify vessel size, ten vessel diameters were measured in each of five fields from each of 12 MINs for each genotype to yield the data plotted in the left-hand side of panel D (spontaneous tumors). Sections of CD31-immunolabeled Py8119 fat pad tumors (at day 12) were used to obtain the data shown on the right-hand side of panel D (Py8119 tumors). A total of 20 vessel diameters were determined in each of six sections from each of four 12-day tumors in each genotype. Scale bars = 100 μm in A and 25 μm in B. * P = 0.02; ** P = 0.007. Double immunostaining for CD31 (red) and NG2 (green) was performed on sections of 17-week spontaneous tumors from wild type (E) and NG2 null (F) MMTV-PyMT females and on sections of 12-day Py8119 tumors from wild type (G) and NG2 null (H) female mice. In both types of tumors, NG2 labeling is seen in association with CD31-positive endothelial cells in wild type hosts, but not in NG2 null hosts. Bar in E = 25 μm. TAM abundance (I-K) and TEM abundance (L-N) were studied by flow cytometric analysis of tumor macrophages in wild type and NG2 null recipients transplanted with wild type or NG2 null β-actin/EGFP bone marrow. F4/80, CD11b, CD45, and Gr1 were used for analysis of TAMs, while F4/80, Tie2, CD206, and CD11c were used for analysis of TEMs. Analysis was restricted to EGFP-positive cells. FACS plots I-J show only the F4/80 and CD11b pairing, while plots L-M show only the F4/80 and Tie2 pairing. In addition, only NG2 null recipient data are shown. Numbers in each quadrant of the FACS plots indicate the percentage of the total number of EGFP-positive cells analyzed. Data for wild type recipients appear very similar. Abundance of TAMs and TEMs are indicated in panels K and N, respectively, as percentages of the total number of EGFP-positive cells. Open bars = wild type bone marrow donors. Black bars = NG2 null bone marrow donors. These data are based on the use of all four markers, not just the two shown in the FACS plot examples. All values in panels K and N are derived from FACS analysis of five separate tumors, all from different mice. Three of the eight P values approach, but do not reach, a statistical significance level of 0.05. EGFP, enhanced green fluorescent protein; FACS, fluorescence activated cell sorting; MMTV-PyMT, mammary tumor virus-driven polyoma middle T; NG2, nerve-glial antigen 2; αSMA, α smooth muscle actin.
Figure 7
Figure 7
Structural deficits of vessels in mammary tumors in NG2 null mice. Py8119 fat pad tumors (2-3 mm diameter) in wild type (WT) and NG2 null (NG2 KO) recipients were used to evaluate several types of structural deficits in tumor vessels due to ablation of NG2. A-C. Pericyte coverage of endothelial cells was evaluated in wild type (A) and NG2 null (B) tissue sections immunostained for desmin (red) and CD31 (green). Since desmin and CD31 are on distinct cell types, they cannot overlap in a single optical section. However, because of the intimate interaction between pericytes and endothelial cells, the two labels appear to overlap when viewed in three-dimensional space. Analysis of confocal z-stacks therefore allows quantification of the extent to which desmin appears to overlap with CD31 labeling. Arrowheads in A indicate areas of overlap between desmin and CD31. Arrows in B show vessel segments without pericyte coverage. The extent to which CD31 pixels are overlapped by desmin pixels provides a measure of pericyte ensheathment of endothelial cells (C). Data were collected from eight tumors per genotype, evaluating four sections per tumor. D. Vascular densities are not significantly different in Py8119 tumors in wild type and NG2 null mice, as quantified by counting CD31-positive vessels in a 10,000 μm2 area. Data were collected from eight tumors per genotype, evaluating four sections per tumor. E-G. Pericyte maturation was evaluated via double immunostaining for desmin (green, all pericytes) and αSMA (red, mature pericytes). Mature pericytes express both desmin and αSMA (arrows), while immature pericytes express only desmin (arrowheads). Pericyte maturation is calculated as the % of desmin-positive pericytes that are αSMA-positive (G). Data were collected from four tumors per genotype, evaluating three sections per tumor. H. Endothelial ensheathment by mature pericytes was quantified by double immunostaining for CD31 and αSMA. Because the number of mature pericytes is reduced in tumor vessels in NG2 null hosts, only vessels with αSMA-positive pericytes are included in this analysis. Endothelial investment by mature pericytes is quantified as % overlap of CD31 pixels by αSMA pixels. For each genotype, five selected vessels each were examined in three different sections from each of three tumors. I-K. Basal lamina assembly was evaluated by immunostaining for CD31 (green) and collagen IV (red) in wild type (I) and NG2 null (J) tumor sections. Arrowheads in I indicate areas of collagen IV/CD31 overlap. Arrows in J show vessel segments with poor basal lamina deposition. Confocal Z-stacks were used to determine the percentage of CD31-positive pixels covered by collagen IV pixels (K). Data were collected from six tumors per genotype, evaluating four sections per tumor. L-M. Endothelial cell sprouting was evaluated by double immunostaining for CD31 (green) and VEGFR-3 (red) in wild type (L) and NG2 null (M) tissue sections. Arrowheads in M indicate VEGFR-3 high/CD31-low structures that characterize sprouting endothelial cells. Confocal z-stacks were used to determine the number of sprouting tip cells per 100 mm2 of CD31-positive vessel area (N). Data were collected from four tumors per genotype, evaluating three sections per tumor. Scale bars = 20 μm (A, B), 90 μm (E, F), 30 μm (I, J), and 12 μm (L, M). * P = 0.02; ** P = 0.006. NG2, nerve-glial antigen 2; αSMA, α-smooth muscle actin; VEGFR-3, vascular endothelial growth factor receptor-3.
Figure 8
Figure 8
Deficits in vessel function in mammary tumors in NG2 null mice. A-C. Leakage from Py8119 tumor vessels was evaluated using intravenously-administered FITC-dextran. Tissue sections were immunolabeled for CD31 (red) to allow image analysis of FITC-dextran (green) located external to tumor vessels in wild type (A) versus NG2 null (B) tissues. Arrowheads in merged images A and B show sites with extravascular FITC-dextran. Confocal z-stacks were used to quantify the percentage of FITC pixels external to CD31-positive vessel walls (C). Data were collected from six tumors per genotype, evaluating four sections per tumor. D-H. Tumor hypoxia was evaluated in mice injected intravenously with pimonidazole hypoxia probe. Immunostaining of wild type (D, F) and NG2 null (E, G) tissue sections for CD31 (red, D, E) and pimonidazole (green, F, G) allowed visualization of hypoxic areas relative to tumor vasculature. Image analysis was used to determine the extent of hypoxia as a percentage of total tumor area (H). Data were collected from six tumors per genotype, evaluating four sections per tumor. Scale bars = 100 μm. * P = 0.05; ** P = 0.003. FITC, fluorescein isothiocyanate; NG2, nerve-glial antigen 2.
Figure 9
Figure 9
Increased VEGF expression in mammary tumors in NG2 null mice. A, B. Because hypoxia induces VEGF expression, we used double immunostaining for VEGF (green) and CD31 (red) to localize VEGF expression relative to tumor blood vessels in wild type (WT, A) and NG2 null (NG2 KO, B) hosts. Some VEGF is associated with CD31-positive tumor blood vessels and some VEGF is dispersed in the tumor tissue. Evaluating overlap of VEGF and CD31 pixels allowed quantification of these two different pools of VEGF. Total VEGF is increased in tumors from NG2 null hosts (C). This VEGF increase in NG2 null tumors is due to non-vascular VEGF dispersed in the tumor tissue (E), rather than to VEGF closely associated with vessels (D). Data were collected from four tumors per genotype, evaluating four sections per tumor. F-H. Triple immunostaining for VEGF, CD31, and pimonidazole hypoxia probe in NG2 null tumors shows that non-vascular VEGF (arrows) is localized to hypoxic areas lacking blood vessels. Scale bars = 60 μm (A, B) and 20 μm (F-H). NG2, nerve-glial antigen 2; VEGF, vascular endothelial growth factor.

Similar articles

See all similar articles

Cited by 29 articles

See all "Cited by" articles

References

    1. Bissell MJ, Radisky DC, Rizki A, Weaver VM, Petersen OW. The organizing principle: microenvironmental influences in the normal and malignant breast. Differentiation. 2002;70:537–546. - PMC - PubMed
    1. Roskelley CD, Bissell MJ. The dominance of the microenvironment in breast and ovarian cancer. Semin Cancer Biol. 2002;12:97–104. - PMC - PubMed
    1. Wiseman BS, Werb Z. Stromal effects on mammary gland development and breast cancer. Science. 2002;296:1046–1049. - PMC - PubMed
    1. Lin EY, Jones JG, Li P, Zhu L, Whitney KD, Muller WJ, Pollard JW. Progression to malignancy in the polyoma middle T oncoprotein mouse breast cancer model provides a reliable model for human diseases. Am J Pathol. 2003;163:2113–2126. - PMC - PubMed
    1. Maglione JE, McGoldrick ET, Young LJ, Namba R, Gregg JP, Liu L, Moghanaki D, Ellies LG, Borowsky AD, Cardiff RD, MacLeod CL. Polyomavirus middle T-induced mammary intraepithelial neoplasia outgrowths: single origin, divergent evolution, and multiple outcomes. Mol Cancer Ther. 2004;3:941–953. - PubMed

Publication types

MeSH terms

Feedback