Inflammatory breast cancer (IBC) is clinically characterized by rapid tumor enlargement with skin erythema. The disease tends to invade locally and metastasize systemically and the histopathologic features are consistent with an extremely aggressive phenotype.1 Despite recent advances in strategies for multimodal treatment of patients with IBC, this disease is still considered incurable and there is a great need for further improvement of treatment strategies to combat this breast cancer variant.2 With most solid tumors, growth, proliferation and metastasis are thought to be dependent on angiogenesis, a process in which vessels grow, in part, by incorporating circulating endothelial cells (ECs) and endothelial precursor cells (EPCs).3, 4, 5, 6, 7, 8, 9, 10 Studies have shown that soluble Flt-1 (sFlt-1) can inhibit vascular endothelial growth factor (VEGF) activity not only by sequestering VEGF but also apparently via a dominant negative mechanism in which it forms heterodimers with the full-length membrane-spanning receptors Flt-1 and KDR and thereby prevents VEGF-dependent transphosphorylation and activation of these receptors.11, 12, 13, 14, 15 Soluble Tie2 (sTie2), an artificial form of Tie2 (a full-length membrane-spanning receptor), is a heparin-binding protein that complexes with angiopoietin (Ang) with the same affinity and specificity as Tie2.16, 17, 18
In the present study, we transfected tumor cells with genes for sFlt-1 and sTie2, assessed the effects of these genes in association with tumor-infiltrating (TI) ECs/EPCs in IBC xenografts and found that VEGF and Ang pathways are crucial in IBC angiogenesis and metastasis. The present results help to clarify the mechanism of induced migration of ECs/EPCs into IBC nests and the effects on IBC of blocking VEGF and Ang pathways.
MATERIALS AND METHODS
Informed Consent and Institutional Approval
Informed consent was obtained from each patient prior to study participation. Our study was approved by the Japanese Foundation for Cancer Research Committee. All animal protocols were approved by the Animal Use Committee of the National Cancer Center.
WIBC-9 and the established non-IBC xenografts MC-2, MC-5 and MC-18 were maintained as described elsewhere.19
Semiquantitative RT-PCR Analysis
To assay gene expression in IBC and non-IBC tumors, in both tumor cells and host cells, levels of mRNA were determined by semiquantitative RT-PCR. The oligonucleotide primers for RT-PCR were designed to amplify specifically the mRNA of the proteins shown in Table I. After 20, 25 or 30 cycles, the products were stained with ethidium bromide. The relative expression levels were calculated as the ratio of the density of each product to that of the internal control (β-actin).
Immunohistochemistry and FACS (Detection of ECs and EPCs)
To investigate putative recruitment of host ECs and EPCs by tumors, we compared IBC specimens with non-IBC immunohistochemically. Immunohistochemistry of paraffin-embedded resected specimens from patients with IBC or non-IBC tumors was performed using conventional methods, with anti-human (hu) Tie2, anti-huVEGF, anti-huFlt-1 and anti-huCD31 Abs (PharMingen, San Diego, CA). Frozen sections of xenograft samples were embedded in O.C.T. Compound (Mile's Sankyo, Tokyo, Japan) before undergoing the immunoperoxidase procedure. Anti-muVEGF, anti-muCD31, anti-muflk-1 (PharMingen) and anti-huKi-67 (Santa Cruz Biotechnology, Santa Cruz, CA) were used as the first antibodies. Reactions were visualized by streptavidin-biotin (PharMingen) staining.
To investigate migration of ECs or EPCs, we used fluorescence-activated cell sorting (FACS) to compare the human IBC xenograft WIBC-9 with 3 non-IBC xenografts (MC-2, MC-5, MC-18). Mononuclear cells (MNCs) isolated from tumor-infiltrating (TI) cells (n = 5) were subjected to FACS analysis to examine expression of surface markers. (ECs and EPCs are known to have the following pattern of surface markers: CD34+, Flk-1+, Flt-1+, Tie2+, CD31+, CD45−, TER− and Mac-1−.) Resected xenografts ere passed 3 times through a 200 μm gauge stainless steel mesh after being minced. The cells were suspended in a medium containing a 20–60% Percoll gradient (Amersham Pharmacia Biotech, Uppsala, Sweden) and centrifuged at 550g for 20 min at room temperature. The cells in the 30% layer of Percoll were then collected and erythrocytes were removed by treatment with 0.83% ammonium chloride in 10 mM Tris-HCl (pH 7.5).
MNCs isolated from TI cells (n = 5) were subjected to flow cytometric analysis to examine surface expression of Flk-1, Flt-1, Tie-2, CD31, CD34, CD45, TER and Mac-1. Anti-murine (mu) CD31, anti-muCD34, anti-muCD45, anti-muflt-1, anti-muflk-1, anti-mutie-2 and anti-muVE-cadherin (PharMingen) were used as the first Abs. Anti-rat fluorescein isothiocyanate (FITC), anti-rat phycoerythin (PE) and anti-goat FITC (Becton Dickinson, San Jose, CA) were used as the second Abs. Levels of antigen expression were analyzed using a FACS Calibur with CellQuest software (Becton Dickinson, Mountain View, CA).
Preparation of AdsFlt-1 and AdExTek
To examine whether blocking the VEGF and Ang pathways can inhibit the growth of WIBC-9, genes for soluble receptors of these growth factors (sFlt-1 and sTie2) were injected into tumors (WIBC-9 and MC-5). The recombinant adenovirus AdsFlt-1 was constructed from a replication-deficient adenovirus vector carrying the reporter gene β-galactosidase (Ad.Pac β-gal). The reporter gene was replaced with 2.1-kb sFlt-1 cDNA. The method used to prepare AdExTek is described elsewhere.15 The 2 vectors were generated and propagated in monolayer 293 cells (American Type Culture Collection, Rockville, MD) that were maintained using conventional methods.1, 2, 3, 4, 5 They were purified from lysates of infected cells, suspended in storage buffer (20 mM Tris, pH 7.4, containing 10% glycerol) and then immediately portioned and frozen at −80°C. Titers of the viral stocks were determined using standard plaque assays.
In vivo Growth of Vector-Treated WIBC-9 and MC-5 Tumors
WIBC-9 and MC-5 tumors were transplanted into the mammary pad of female BALB/c nude mice. Beginning 3 weeks after transplantation, when the tumors had reached 2.5 mm in diameter, either AdsFlt-1 (5 × 108 pfu), AdExTek (5 × 108 pfu) or Ad.Pacβ-gal (5 × 108 pfu; control) was injected into them twice a week for 2.5 weeks, for a total of 5 injections. Tumor volume, determined using the formula (width2 × length)/2, was measured twice a week after transplantation.
To evaluate the inhibitory effects of AdsFlt-1 and AdExTek on the angiogenesis induced by WIBC-9 cells, an angiogenesis assay was performed using the dorsal air sac method. A Millipore (Bedford, MA) chamber containing a nitrocellulose membrane, which only passes soluble factors, was filled with a 125 mm3 sample of WIBC-9 (treated with 5 × 108 pfu of AdsFlt-1, AdExTek or AdPacβ-gal) in 200 μl of medium and was then injected into the right flank of BALB/c nude mice. The mice were sacrificed on day 4 after injection. The area of neovascularization was measured using an MCID image analyzer and the degree of angiogenesis was determined from the ratio of the area of blood vessels to the area in contact with the chamber.
To examine whether blocking the VEGF or Ang pathway can suppress the number and growth of tumor metastases, 70 days after transplantation of WIBC-9 tumors into the mammary pad of female BALB/c nude mice, the animals were sacrificed and their lungs were removed and fixed in 10% neutral-buffered formaldehyde (Wako, Osaka, Japan) and embedded in paraffin. Then 5 μm thick sections were cut at 200 μm intervals from each paraffin-embedded lung block. Tissue sections were stained with hematoxylin and eosin (H&E) and the number and size of metastatic tumors were determined.
All data are expressed as mean ± SD. StatView computer software (ATMS, Tokyo, Japan) was used to perform statistical analysis of tumor volume, number of lung metastases and ELISA results. Two-sided p < 0.05 was considered to indicate statistical significance.
Expression of Genes for Angiogenic Factors, EGFR and Adhesion Molecules in Resected Specimens from Patients with IBC and Non-IBC Tumors
The results of semiquantitative RT-PCR are shown in Table II. IBC tumor samples contained significantly higher levels of the following mRNAs than non-IBC tumor samples: VEGF, Flt-1, Ang1, Ang2, Tie2, interleukin (IL)-1β, IL-8, ErbB-1, ErbB2 and integrin β3 (p < 0.05). There were no significant differences in the levels of the following mRNAs between IBC samples and non-IBC samples: KDR, VEGF-C, VEGF-D, Flt-4, IL-6 and αv integrin.
Table II. Semiquantitative RT-PCR of IBC and Non-IBC1
VEGF-Flt-1, KDR pathway
(n = 19)
2.3 ± 1.1
0.9 ± 0.3
0.2 ± 0.1
0.6 ± 0.3
0.4 ± 0.2
0.3 ± 0.1
0.9 ± 0.5
1.6 ± 0.8
1.6 ± 0.6
(n = 23)
0.8 ± 0.3
0.4 ± 0.2
0.2 ± 0.1
0.4 ± 0.2
0.4 ± 0.2
0.3 ± 0.1
0.2 ± 0.1
0.5 ± 0.2
0.4 ± 0.2
Relative expression levels were calculated as the density of the products divided by internal control (β-actin). IBC tumor samples significantly contained higher mRNA levels than non-IBC tumors of mRNAs encoding VEGF and its receptor Flt-1, Ang1, 2 and its receptor Tie2, the cytokine IL-1β; 8, the EGFR family ErbB-1, 2 and the integrin 3 subunit. In contrast, IBC did not express substantially higher levels of the VEGF receptor KDR, the VEGF homologs VEGF-C and VEGF-D and their receptor Flt-4 and the αv integrin subunit than non-IBC. For abbreviations, see Abbreviations list.–*P < 0.05.
(n = 19)
1.3 ± 0.4
0.8 ± 0.5
1.2 ± 0.3
1.4 ± 0.3
2.1 ± 0.5
0.6 ± 0.3
1.8 ± 0.5
(n = 23)
0.5 ± 0.3
0.8 ± 0.3
0.6 ± 0.2
0.5 ± 0.4
0.8 ± 0.6
0.4 ± 0.4
0.6 ± 0.4
Immunohistochemical Assay for Tumor-Infiltrating MNCs in Surgically Resected IBC and Non-IBC Specimens
Figure 1 shows typical results for IBC specimens, using H&E staining (Fig. 1a,b), Elastica-vangieson staining (Fig. 1e,f) and immunohistochemical staining with anti-huTie2 (Fig. 1c), anti-huVEGF (Fig. 1d), anti-huFlt-1 (Fig. 1g) and anti-huCD31Abs (Fig. 1h). TI MNCs (brown-colored; positive for all antigens assayed) were clearly visible in tumor-associated stroma of IBC specimens. IBC tumor cells were positive for huTie2 (Fig. 1c) and weakly positive for huFlt-1 (Fig. 1g). The numbers of TI MNCs in surgically resected IBC and non-IBC specimens are shown in Table III. The total number of TI MNCs in IBC specimens was at least 5 times greater than in non-IBC specimens. In particular, the number of putative ECs or EPCs (MNCs positive for all antigens assayed) in IBC specimens was at least 8 times greater than in non-IBC specimens.
Table III. Tumor-infiltrated MNC numbers in Human Breast Cancer Resected Specimens by immunohistochemistry1
Total no. (per specimen)
Total T1 MNC number was 5 times or greater in IBC than in non-IBC. Putative EC and EPC number in particular, which was positive for demonstrated antigens, was 8 times or greater in IBC than in non-IBC. For abbreviations, see Abbreviations footnote.–*, p < 0.05.
(n = 5)
3,641 ± 1,172*
321 ± 138*
371 ± 144
336 ± 155*
327 ± 134*
308 ± 128*
(n = 14)
586 ± 123
24 ± 14
36 ± 22
28 ± 13
21 ± 14
25 ± 24
FACS Analysis of IBC and Non-IBC Xenografts for Tumor-Infiltrating ECs and EPCs
Results of FACS are shown in Table IV. In WIBC-9 specimens, the total number of TI MNCs was markedly elevated. Specifically, the number of ECs/EPCs in WIBC-9 xenografts was at least 20 times greater than in non-IBC xenografts. The ECs/EPCs incorporated acetylated lipoprotein (acLDL) and they were integrated within a HUVEC monolayer in vitro culture on day 5 (Fig. 2).
Table IV. Tumor-infiltrated MNC population in Human Breast Cancer Xenografts by Fluorescence-Activated Cell Sorting (FACS)1
Breast cancer xenograft
Total number (×105)
CD34+ + Flk-1+ (%)
CD34+ + Flt-1+ (%)
CD34+ + Tie2+ (%)
CD34+ + CD31+ (%)
CD34+ + CD45− + TER + MAC-1− (%)
MNCs derived from T1 cells (n = 5) were subjected to FACS analysis to examine surface expression of CD34+ with Flk-1+, Flt-1+, Tie2+, CD31+, CD45−, TER−, and Mac-1− in each xenograft. In WIBC-9, the total numbers of TI MNC population were prominently elevated. Specifically, the EC or EPC population in WIBC-9 was 20 or more times greater than in non-IBC xenografts. For abbreviations, see abbreviations footnote.–*, p < 0.05.
9.1 ± 1.7*
28.3 ± 5.3*
30.4 ± 4.2*
26.4 ± 5.3*
7.3 ± 2.6*
14.4 ± 2.8*
0.9 ± 0.1
1.2 ± 0.4
1.4 ± 0.2
0.9 ± 0.3
0.5 ± 0.2
0.3 ± 0.2
1.3 ± 1.4
1.4 ± 0.2
2.1 ± 0.3
1.2 ± 0.3
0.8 ± 0.4
0.5 ± 0.2
3.1 ± 1.2
3.8 ± 0.6
4.5 ± 0.2
3.7 ± 1.4
2.1 ± 0.8
4.5 ± 2.1
Inhibition of WIBC-9 Growth by Intratumoral Administration of AdsFlt-1 or AdExTek
AdsFlt-1 and AdExTek produced 100% tumor growth inhibition in WIBC-9 and 90% inhibition in MC-5 (Fig. 3a,b). On day 50 after the last administration, tumor regrowth was evident and was accompanied by erythema of the overlying skin.
Inhibition of Lung Metastases by Blocking VEGF and Ang Pathways
The results of our analysis of lung metastases are shown in Figure 3c. Both AdsFlt-1 and AdExTek suppressed lung metastases, as demonstrated by the metastatic ratio and number of metastases. None of the tumors treated with AdsFlt-1 metastasized to the lungs.
Concentrations of Human VEGF, Murine VEGF and Human sFlt-1 in Mouse Sera During Gene Therapy
On day 28 after transplantation, in both WIBC-9 and MC-5 recipients, the serum level of sFlt-1 in the group given AdsFlt-1 was significantly higher than that of the control group (Fig. 3d). Also on day 28, the serum level of huVEGF in the AdsFlt-1 group was not statistically different from that of the other 2 groups, as determined using polyclonal VEGF antibody (Fig. 3e). huVEGF was not detected in serum from any of the MC-5 recipients. On day 28, the serum levels of muVEGF in the groups injected with AdsFlt-1 or AdExTek were significantly lower than those in the control groups (Fig. 3f). On day 58, serum levels of huVEGF and muVEGF were significantly lower in the AdsFlt-1 and AdExTek groups than in the control groups. muVEGF was not detected in serum from any of the MC-5 recipients.
Morphologic Analysis of Vector-Treated WIBC-9
To examine the morphologic features of the treated tumors, the animals were anesthetized and tumors were resected on day 70 (Fig. 4a–i). Figure 4a, d and g shows that tumors treated with the control vector were large (2,916 ± 168 cm3), with a reddish tumor center. These features were similar to those of untreated WIBC-9. Figure 4b, e and h shows that tumors treated with AdsFlt-1 were small (4.5 ± 3.2 cm3), white avascular clusters with visible central necrosis and fibrosis. Figure 4c, f and i shows that tumors treated with AdExTek were small (3.2 ± 2.1 cm3), with prominent central necrosis and fibrosis.
Effects of Blocking VEGF and Ang Pathways on Angiogenesis Induced by WIBC-9
Angiogenesis within the subcutaneous fascia of tumors infected with AdsFlt-1 or AdExTek was compared with that of tumors injected with the control. Angiogenesis was clearly suppressed in the AdsFlt-1 (Fig. 4k) and AdExTek (Fig. 4l) groups, compared with the control group (Fig. 4j). The ratios of the area in contact with the Millipore chamber were 57 ± 5% in the control group, 12 ± 4.8% in the AdsFlt-1 group and 5.1 ± 2.1% in the AdExTek group.
Immunohistochemical Analysis of Vector-Treated WIBC-9
We detected the expression of muCD31 and muflt-1 in the endothelia of the tumors treated with the control vector (Fig. 4m,p) but not in endothelia of tumors treated with AdsFlt-1 (Fig. 4n,q) or AdExTek (Fig. 4o,r). The cell proliferation marker Ki67 was clearly detected in the central area of tumors treated with the control vector (Fig. 4s) but was only faintly detected in tumors treated with AdsFlt-1 (Fig. 4t) or AdExTek (Fig. 4u). Paucity of muCD31 and huKi 67 immunoreactivity coincided with the histologic findings (Fig. 4g–i) of central necrosis and scarcity of neovascularization in tumors treated with AdsFlt-1 or AdExTek.
The present results support 2 main conclusions. The first is that IBC tumors induce angiogenesis and vasculogenesis that involves recruitment of ECs/EPCs. The second conclusion is that VEGF and Ang pathways are crucial for IBC xenograft angiogenesis. Resected specimens of IBC expressed higher levels of mRNA for VEGF, flt-1, Ang-1, Ang-2, Tie2, IL-1β,IL-8, ErbB-1, ErbB-2 and integrin β3 (all of which are associated with vasculogenic phenotype) than non-IBC specimens. The finding that mRNA expression of VEGF-C, VEGF-D and Flt-4 (the common receptor for these 2 VEGF homologs) is not higher in IBC tumors than in non-IBC tumors does not support the postulated lymphangiogenic pathway. To investigate the response of host cells (especially ECs and EPCs) to IBC, the human IBC xenograft WIBC-9 and species-specific primers and Abs (WIBC-9 is of human origin and the host cells in the present study were of mouse origin) were used in our study. In a previous study, higher levels of murine mRNAs encoding VEGF, flt-1, Ang1, Ang 2, Tie1, Tie2 and integrin β3 were detected in WIBC-9 xenografts than in non-IBC xenografts.19 In the present study, WIBC-9 induced host responses including expression of genes encoding murine VEGF, Ang1 and Ang2 and genes encoding the corresponding murine receptors Flt-1 and Tie2.
Thus, WIBC-9 not only produces angiogenic factors but also induces the host to produce them. To examine the response of host cells to IBC tumors in detail, putative ECs/EPCs were isolated from tumor xenografts according to cell surface antigen expression. ECs/EPCs were defined as MNCs that fit one of the following criteria: (i) positive for CD34 and negative for CD45, Mac-1 and TER; or (ii) positive for CD34 and positive for either flk-1, flt-1, tie2 or CD31. Surprisingly, in WIBC-9 specimens, the number of TI ECs/EPCs was 20-fold higher than for non-IBC xenografts.
To explore the therapeutic potential of blocking the VEGF and Ang pathways involved in migration, proliferation and tube formation by putative ECs/EPCs in IBC, adenoviral vectors encoding sFlt-1 and sTie2 (AdsFlt-1 and AdExTek, respectively) were injected into WIBC-9 xenografts. Both AdsFlt-1 and AdExTek inhibited growth, vascularization and metastasis of WIBC-9. Our results show that blocking either the VEGF or Ang pathway with adenovirally induced sFlt-1 or sTie2 can block angiogenesis and completely inhibit tumor growth in WIBC-9 tumors and can also inhibit growth in MC-5 tumors. On day 28, 1 day after the second administration of adenoviral vectors, the serum level of human sFlt-1 was significantly higher in the AdsFlt-1 group than in the control group. However, on day 58, this difference in serum sFlt-1 levels was no longer observed, probably due to the cessation of cell proliferation in the WIBC-9 tumor. The levels of serum murine VEGF were significantly lower in the AdsFlt-1 and AdExTek groups than in the control groups; this may be due to attenuated host response to the tumor. Tumor regrowth with erythema of the overlying skin was evident 50 days after the last administration of AdsFlt-1 or AdExTek, presumably reflecting a significant decrease in expression of the adenovirally introduced genes. AdsFlt-1 and AdExTek almost completely suppressed lung metastases, according to the metastatic ratio and the absolute number of metastatic foci. Tumor-associated angiogenesis was significantly suppressed in WIBC-9 tumors infected with AdsFlt-1 or AdExTek, compared with controls, in vivo.
The results indicate that AdsFlt-1 and AdExTek inhibit sprouting of vessels by 79 and 91%, respectively. This strong inhibition of tumor-induced angiogenesis seems to result in decreased lung metastasis. However, recently, several kinds of cancers, including breast cancers, were found to be associated with increased levels of circulating ECs in the peripheral blood.20 We have found evidence of tumor-induced vasculogenesis in WIBC-9, by counting the ECs and EPCs in peripheral blood and bone marrow (data not shown) and the present marked effects of AdsFlt-1 and AdExTek may be associated with inhibition of tumor-induced vasculogenesis. IL-1β and IL-8 are inflammatory cytokines that are secreted by WIBC-9. Inflammatory cytokines may directly modulate angiogenesis and indirectly affect adhesion molecules and matrix metalloproteinases that interact with macrophages and lymphocytes.21
Although we did not examine the effects of IL-1β and IL-8 on angiogenesis and vasculogenesis, it is clear whether the VEGF and Ang pathways are crucial for tumor growth and lung metastasis in the present models. Recently, there have been reports describing autocrine pathways of angiogenic factors such as VEGF in cancer cells.22 The tumor necrosis that we observed in WIBC-9 tumors infected with AdsFlt-1 and AdExTek may be the result of the blocking of autocrine pathways. The present experiments were preliminary and several issues need to be clarified. However, the present results may represent the first step in the development of a new treatment for human breast cancer, specifically IBC.
We are grateful to Miss M. Takahashi and Mr. T. Morikawa for technical assistance.