Inhibition of VEGF receptors significantly impairs mammary cancer growth by charleshogue


More Info
									                                                                         Oncogene (2004), 1–11
                                                                         & 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00


Inhibition of VEGF receptors significantly impairs mammary cancer
growth in C3(1)/Tag transgenic mice through antiangiogenic and non-
antiangiogenic mechanisms

Jung-Im Huh1,6, Alfonso Calvo1,4,6, Jeffrey Stafford2,5, Mui Cheung2, Rakesh Kumar2,
Deborah Philp3, Hynda K Kleinman3 and Jeffrey E Green*,1
 Laboratory of Cell Regulation and Carcinogenesis, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA;
 Cell Biology Section, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA;
 GlaxoSmithKline, Research Triangle Park, NC, USA

Cancer growth and progression is often critically influ-                  Tumor growth, however, rapidly advanced following
enced by the production of vascular endothelial growth                   cessation of treatment. This is the first demonstration
factor (VEGF), a key mediator of angiogenesis. VEGF                      that a VEGF receptor inhibitor, GW654652, has a strong
produced by tumor cells stimulates endothelial cell growth               inhibitory effect on angiogenesis and tumor progression in
through the binding and activation of the KDR/Flk-1                      a transgenic model of mammary cancer, suggesting that
receptor (VEGFR-2) on endothelial cells. Recently, some                  this is a useful approach for preclinical testing of such
human breast cancer epithelial cells have been shown to                  agents.
express VEGF receptors, suggesting a potential auto-                     Oncogene advance online publication, 13 December 2004;
crine-mediated growth stimulation of a subset of cancers                 doi:10.1038/sj.onc.1208221
by VEGF. We demonstrate that mammary tumors in the
C3(1)/Tag transgenic model express VEGF and VEGF                         Keywords: antiangiogenesis; VEGFR tyrosine kinase
receptors and tumor growth is stimulated by this autocrine               inhibitor; mammary cancer; C3(1)/Tag transgenic mice
mechanism. GW654652, an indazolylpyrimidine, is a
VEGFRs tyrosine kinase inhibitor that dramatically
reduces both angiogenesis and tumor cell growth in this
model, as demonstrated using both in vitro and in vivo
assays. GW654652 significantly decreased cell prolifera-                  Introduction
tion and induced apoptosis in human umbilical vein
endothelial cells and M6 mammary tumor cells derived                     Angiogenesis, the formation of new blood vessels, is an
from C3(1)/Tag (Tag: simian virus 40 T-antigen)                          essential process during growth and progression of solid
transgenic mice. A 75% reduction in VEGF-induced                         tumors. Vascular endothelial growth factor (VEGF) is a
angiogenesis was observed with GW654652 using the                        well-known stimulator of tumor angiogenesis (Folkman,
chick chorioallantoic membrane assay, whereas                            1995; Folkman and D’Amore, 1996). Since tumor
GW654652 produced an approximately 85% reduction                         growth and metastases are dependent on angiogenesis,
in angiogenesis as assessed by the Matrigelt plug assay.                 antiangiogenic therapies have become important strate-
A profound inhibitory effect on tumor growth in the                      gies to inhibit tumor progression. Recently, many
C3(1)/Tag transgenic model of human breast cancer was                    approaches have been tried to develop antiangiogenic
observed with oral administration of GW654652 as                         agents. VEGF-targeted drugs, epidermal growth factor
measured by delayed tumor onset, decreased multiplicity,                 (EGF) receptor tyrosine kinase inhibitors (Ciardiello
reduced tumor volume, and extended animal survival. The                  and Tortora, 2002), MMP inhibitors (Kerbel and
antitumor effects of GW654652 were associated with                       Folkman, 2002), and integrin-targeted drugs (Gutheil
reduced tumor vascularization and no apparent toxicity.                  et al., 2000) have been developed for this purpose.
                                                                            Intriguingly, receptors for VEGF have been identified
                                                                         on subsets of human breast cancers and breast cancer cell
*Correspondence: JE Green, Laboratory of Cell Regulation and             lines, suggesting that a potential VEGF-mediated auto-
Carcinogenesis, National Cancer Institute, NIH, Building 41, Room
C629, 41 Medlars Dr., Bethesda, MD 20892, USA;
                                                                         crine growth effect might exist in such tumors (Price et al.,
E-mail:                                                  2001; Nakopoulou et al., 2002). Receptors for VEGF
 These two authors contributed equally to this work                      have also been described in tumor cells from C3(1)/Tag
 Current address: Division of Oncology, Department of Histology and      transgenic mammary cancers (Calvo et al., 2002b).
Pathology, Center for Applied Medical Research (CIMA), University        Whereas antiangiogenesis indirectly inhibits tumor
of Navarra, Pamplona, Spain
 Current address: Syrrx Inc., San Diego, CA, USA                         growth by impairing blood vessel formation, interference
Received 4 June 2004; revised 20 September 2004; accepted 20 September   of an autocrine growth pathway on tumor cells could
2004                                                                     substantially augment the inhibition of tumor growth.
                           Inhibition of mammary cancer by VEGFRs blockade
                                                               J-I Huh et al
          VEGF, the angiogenic factor most commonly pro-                       course of lesion progression (Maroulakou et al., 1994;
       duced by tumors, binds to two distinct receptor tyrosine                Shibata et al., 1996) developing through low-grade
       kinases (Xie et al., 1999) – the fms-like tyrosine kinase               and high-grade mammary intraepithelial neoplasia
       Flt-1 (VEGFR-1) and the kinase insert domain contain-                   (similar to human ductal carcinoma in situ) to inva-
       ing receptor KDR/Flk-1 (VEGFR-2). These two mem-                        sive carcinoma with some mice developing meta-
       brane receptor tyrosine kinases are predominantly                       stases. VEGF and VEGF receptors, VEGFR-1 and
       expressed on endothelial cells. Flk-1 is essential for the              VEGFR-2, are upregulated during the transition from
       development of hematopoietic and endothelial cells                      preinvasive mammary lesions to invasive carcinomas
       (Shalaby et al., 1995; Takahashi et al., 2001), whereas                 (Calvo et al., 2002a).
       Flt-1 is required more for differentiation and vascular                    In this study, we demonstrate that GW654652 has
       organization (Fong et al., 1995).                                       potent antiangiogenic activity as well as inhibitory
          VEGF and VEGF signaling have been intensively                        activity on VEGF-mediated autocrine growth of the
       studied for developing novel anticancer agents and                      tumor epithelial cells. GW654652 inhibits the growth of
       many approaches have been taken, including small                        HUVEC and M6 cells (mammary tumor cells derived
       molecule tyrosine kinase inhibitors (Itokawa et al., 2002;              from the C3(1)/Tag model) and induces apoptosis. We
       Laird et al., 2002; Lin et al., 2002; Spiekermann et al.,               also show that the compound is a potent inhibitor of
       2002), fusion proteins (Holash et al., 2002), ligand                    angiogenesis in vivo using the chick chorioallantoic
       antibodies (Presta et al., 1997), receptor antibodies (Zhu              membrane (CAM) and Matrigelt plug assays. Most
       et al., 2003), and ribozymes (Parry et al., 1999).                      significantly, GW654652 can dramatically inhibit mam-
       Antibodies directed against VEGF that bind specific                      mary cancer progression in the C3(1)/Tag mouse model
       epitopes can block VEGF interaction with its receptors                  of mammary cancer without apparent signs of toxicity.
       (Presta et al., 1997). Recombinant solubilized receptors                These results suggest that GW654652 or similar
       containing the VEGF interaction domains of VEGFR-                       compounds may have potential use as an antiangiogenic
       1 and VEGFR-2 fused to the Fc portion of immuno-                        agent for human breast cancer and may also function
       globulin G1 bind VEGF and can prevent VEGF signaling                    as an inhibitor of VEGF autocrine-mediated tumor
       through both VEGFR-1 and VEGFR-2 (Zhu et al.,                           growth.
       2003). Antibodies that target VEGFR-1 and VEGFR-2
       also inhibit the VEGF signaling pathway (Zhu et al.,
       2003). Ribozymes are RNA-based enzymes that bind to
       and cleave mRNA molecules in a sequence-specific                         Results
       manner (Parry et al., 1999) resulting in the destruction
       of message and decreased expression of the encoded                      Inhibitory effect of GW654652 on VEGFRs kinase
       protein. Anti-Flt-1 and/or anti-KDR ribozymes reduce                    activities
       mRNA levels specifically, decrease proliferation of                      The structure of GW654652 is depicted in Figure 1.
       vascular cells in culture, and inhibit VEGF-stimulated                  Kinase assays revealed that GW654652 was highly
       angiogenesis.                                                           specific for inhibiting VEGFRs phosphorylation com-
          Endothelial cells, such as human umbilical vein cells                pared to other kinase receptors based upon IC50
       (HUVECs), express high levels of Flk-1/KDR                              determinations (Table 1). The IC50 for VEGFR-2 was
       (VEGFR-2), but low levels of Flt-1 (VEGFR-1), and                       0.0023 mM with concentrations of 1.1-, 5.3-, 130-, 139-,
       are responsive to VEGF for growth signaling. Com-                       2086.9-, 156-, 605-, 5125-, and >8800-fold to produce
       pounds that can interfere with VEGF signaling by                        an IC50 for VEGFR-3, VEGFR-1, FGF-R1, FGF-R3,
       blocking the phosphorylation of Flk-1/KDR may be                        FGF-R4, epidermal growth factor receptor (EGFR),
       important agents to inhibit tumor angiogenesis.                         ErbB-2, and CDK-4, respectively. These results demon-
       GW654652 has been developed as a small molecule                         strate that GW654652 has a high degree of specificity for
       inhibitor of VEGFRs autophosphorylation (Dev et al.,                    inhibiting VEGFRs.
       2003; Kumar et al., 2003). A previous study has shown
       that GW654652 blocked VEGF-induced Flt-1 phos-
       phorylation in a multiple myeloma cell line that
       expresses high-affinity VEGFR-1 but not VEGFR-2                                                                             CH3
       (Podar et al., 2004). Structurally, GW654652 belongs
       to the family of indazolylpyrimidines (Cheng et al.,                                        N
       2003).                                                                                                                      N
          Ultimately, one wants to test the efficacy of an agent
       in a preclinical setting using an appropriate model                                   HN        N        N             N
       system. The validation of genetically engineered mouse
       models for preclinical testing has become an important                     CH3O                          CH3
       area of research. The C3(1)/Tag transgenic mouse
       model of mammary cancer shares important biological
       and histological similarities to human breast cancer
       (Maroulakou et al., 1994). Female mice develop
       mammary gland tumors within a well-defined time                           Figure 1 Chemical structure of GW654652

                                                                                           Inhibition of mammary cancer by VEGFRs blockade
                                                                                           J-I Huh et al
     Table 1 Effect of GW654652 on various receptor kinases                                dramatically inhibited in both HUVEC and M6 cells.
                                       IC50 (mM)a           Fold selectivity vs VEGFR-2b   To show the specificity of GW654652 to VEGF-
                                                                                           mediated growth, the cells were stimulated with either
VEGFR-2                     0.002370.0008                                  —               10 ng/ml VEGF or 0.5 ng/ml basic fibroblast growth
VEGFR-3                     0.002570.0005                                  1.1             factor (bFGF), which are known to act as angiogenic
VEGFR-1                     0.012070.0026                                  5.3
FGFR-1                            0.3                                     130              factors. The inhibitory effect of GW654652 on VEGF
FGFR-3                            0.32                                   139.1             stimulation of HUVECs proliferation was more selec-
FGFR-4                            4.8                                    2086.9            tive for VEGF-induced proliferation than bFGF-in-
SRC                           0.3570.24                                   156              duced proliferation at the same concentration. M6 cells
Eph-B4                        0.4670.03                                   204              were much less responsive to growth stimulation by
c-Fms                         0.5370.33                                   237
Tie2                          0.3270.04                                   144              VEGF, although GW654652 significantly inhibited
EGFR                          1.3670.42                                   605              proliferation. Since M6 cells already express relatively
ErbB2                         11.5372.74                                  5125             high endogenous levels of VEGF, the VEGF-specific
CDK2                             >20                                     >8800             response to GW654652 could not be demonstrated with
CDK4                             >20                                     >8800
                                                                                           the addition of exogenous VEGF. To determine the role
 Values are mean7s.e. bRatio of IC50 obtained with a given kinase and                      of endogenous VEGF in M6 cell proliferation, we
that achieved with VEGFR-2                                                                 performed experiments in which endogenous VEGF was
                                                                                           downregulated by small interfering RNA (siRNA). The
                                                                                           level of VEGF mRNA was significantly reduced 24 h
             a                                         HUVEC
                                                                                           after siRNA transfection in M6 cells (Figure 3a) with a
                                                                 FG F
                 Absorbance (420 nm)

                                       1.5                       PBS
                                                                                                  a siRNA (nM)                             0      20        40
                                       1.0                                                                                    VEGF
                                                                                                  b                  200
                                                                                                      VEGF (pg/ml)

                                       0.0                                                                           150
                                             0   0.001 0.005 0.01 0.05   0.1
                                                     GW654652 (µM)                                                   100                         **
                                                                                                                       50                                   ***
             b                                            M6
                                       1.5                       VEGF                                                         0
                                                                                                                                       0       20      40
                 Absorbance (420 nm)

                                                                 FGF                                                                       siVEGF (nM)
                                                                                                  c                                        ***                     siVEGF(-)
                                                                                                                              100                                  siVEGF(+)
                                       0.5                                                                                                             **
                                                                                                       of proliferation (%)


                                             0   0.001 0.005 0.01 0.05   0.1
                                                    GW654652 (µM)                                                                 50                              **

 Figure 2 Inhibition of cell proliferation by GW654652. HUVECs                                                                    25
 and M6 cells were plated in collagen-coated or uncoated 96-well
 plates in RPMI 1640 containing 1% fetal calf serum (FCS) or
 Dulbecco’s modified Eagle’s medium (DMEM) without serum.                                                                           0
 After a 30 min incubation, these cells were subsequently stimulated                                                                        0      0.01   0.05           0.1
 by 10 ng/ml VEGF or 0.5 ng/ml FGF. HUVECs were incubated for                                                                                     GW654652 (µM)
 72 h or M6 cells were incubated for 24 h. Cell proliferation in
 HUVECs (a) and M6 cells (b) was inhibited at the dose range from                           Figure 3 Autocrine stimulation of proliferation by VEGF on M6
 0.001 to 0.1 mM of GW654652                                                                cells. siRNA experiment against mouse VEGF was performed to
                                                                                            demonstrate the autocrine effect of VEGF in M6 cells. (a)
                                                                                            Expression of VEGF mRNA 24 h after siRNA against mouse
                                                                                            VEGF (simVEGF) transfection at 20 and 40 nM concentrations
                                                                                            was detected by semiquantitative reverse transcription–polymerase
GW654652 inhibits HUVEC and M6 cell proliferation                                           chain reaction (RT–PCR). GAPDH was used as an internal
                                                                                            control. (b) Murine VEGF protein level was measured by ELISA
The effects of GW654652 on HUVECs and M6 cells                                              (**Po0.01, ***Po0.001). (c) A 40 nM portion of siVEGF was
(derived from a C3(1)/Tag transgenic mammary tumor;                                         transfected in M6 cells and after 24 h GW654652 was added at
                                                                                            indicated concentrations. Cells were stimulated by 10 ng/ml
Holzer et al., 2003) were investigated in a proliferation                                   mVEGF after 1 h of GW654652 addition. Cell proliferation assay
assay using concentrations of GW654652 ranging from                                         was performed after GW654652 treatment for 24 h (**Po0.01,
0 to 0.1 mM (Figure 2a and b). Cell growth was                                              ***Po0.001)

                             Inhibition of mammary cancer by VEGFRs blockade
                                                                 J-I Huh et al
       70% decrease in VEGF protein level using 40 nM siRNA                      incubation times were used for each cell line because
       (Figure 3b, **Po0.01, ***Po0.001). Cell proliferation                     of differences in the growth characteristics of the cells.
       of M6 cells 24 h after siRNA transfection was reduced                     FACS analysis demonstrated an accumulation of cells in
       by about 30% compared to control cells. This result                       the sub-2N stage, indicating that apoptosis occurred at
       suggests that the growth of M6 cells is positively                        all concentrations of GW654652 tested. For HUVECs,
       stimulated through a VEGF autocrine loop. M6 cell                         levels of apoptosis increased from 13% in the control
       proliferation was additionally inhibited by the combina-                  samples to 32.0, 48.5, and 47.2% for cells exposed to
       tion of siVEGF and GW654652 (Figure 3c).                                  0.05, 0.1 or 0.5 mM GW654652, respectively (Figure 4a
                                                                                 and d, ***Po0.001). For M6 cells, the levels of
                                                                                 apoptosis increased from 7.9% in the control samples
       GW654652 increases apoptosis in HUVECs and M6 cells
                                                                                 to 14.1, 10.8, and 14.6% at 0.05, 0.1 or 0.5 mM
       To characterize potential mechanisms of growth inhibi-                    GW654652, respectively (Figure 4b and e, **Po0.01,
       tion by GW654652, the induction of apoptosis in                           *Po0.05). To observe the morphological changes in M6
       HUVECs (Figure 4a and d) and M6 cells (Figure 4b,                         cells, cells at 80% confluency were treated for 24 h with
       c, and e) in response to GW654652 was determined.                         the same concentrations of GW654652 used for the
       After 24 h incubation in serum-free medium, HUVECs                        FACS analyses. Morphological changes in M6 cells
       were incubated for 24 h and M6 cells were incubated for                   following GW654652 treatment were observed in a
       6 h with 0.05, 0.1 or 0.5 mM GW654652. Different                          dose-dependent manner. As shown in Figure 4c, M6

           Figure 4 GW654652 induces apoptosis. HUVECs and M6 cells were plated and incubated in RPMI 1640 containing 1% FCS or
           DMEM without serum for 24 h. GW654652 was added at concentrations ranging from 0.05 to 0.5 mM and then HUVECs were
           incubated for 24 h and M6 cells for 6 h at 371C. Cells were stained by propidium iodide (PI) and analysed by FACSCalibur within 24 h.
           Representative results are shown. (a) HUVECs and (b) M6 cells. Apoptotic cells in HUVECs and M6 cells (d and e) were quantified
           using flow cytometric analyses (*Po0.05, **Po0.01, ***Po0.001). (c) To observe the morphological changes in M6 cells, cells were
           seeded and when they were 80% confluent, they were treated with GW654652 at the indicated concentrations for FACS analysis in
           DMEM for 24 h. The morphological changes in M6 cells were dose-dependent

                                                                                Inhibition of mammary cancer by VEGFRs blockade
                                                                                J-I Huh et al
cells treated with increasing concentrations of                                 assay, whereas this was reduced to 16% at the highest
GW654652 became more rounded with detachment of                                 dose of GW654652 (Figure 6a and c). We also evaluated
numerous cells from the culture dish (Figure 4c). To                            the effect of the oral administration of GW654652 on
further confirm the apoptotic response, DNA fragmen-                             the inhibition of angiogenesis using the Matrigel plug
tation and caspase-8 activation were measured. DNA                              assay. In control plugs without VEGF, only a few
laddering was determined after 8 h of treatment with                            scattered vascular cells were observed, whereas VEGF
0.l mM GW654652 (Figure 5a). In order to investigate the                        dramatically enhanced the number of vascular cells
kinetics of apoptosis in response to GW654652 in M6
cells, caspase-8 activation was determined. Caspase-8
activation increased about twofold in response to 0.1 mM
GW654652 (**Po0.01, ***Po0.001). These results                                          a
show that GW654652 increases apoptosis of M6 cells
at least in part through caspase-8 activation.

                                                                                                                 Control                   Control                  GW654652
Inhibition of VEGF-induced angiogenesis by GW654652                                                              (-VEGF)                (+VEGF, 50 ng)               + VEGF
in the CAM and Matrigelt plug assays
The CAM and Matrigel plug assays were performed to
test the inhibitory activity of GW654652 on angiogen-
esis (Figure 6). Three concentrations of GW654652 –
0.1, 0.5, and 1 mM, all delivered in a total volume of 2 ml                                                                                Control
                                                                                                                       Control                                      GW654652
– were tested in the CAM assay. Inhibition of VEGF-                                                                    (-VEGF)          (+VEGF, 50 ng)               + VEGF
induced angiogenesis was demonstrated at the three
concentrations compared to the VEGF-positive con-                                                                      100
trols. A total of 66% of the VEGF control samples
exhibited a positive angiogenic response in the CAM                                                 Percent positive    75              (8 /12)

                                                                                                                              (3 /11)             (3 /10) (3 /11)
                                                                                                                        25                                          (2 /12)
            a                       (hr)   0       2    4        8   12
                                                                                                           t ( GF)







                                                                                                                                        GW654652 (µM)

                                                                                        d                                                           P<0.001

                                                                                             Number of blood vessels

                                                                                                    per field




           b                                                                                                             0
            of caspase-8 activity

                                     2.0                             **   ***                                                                    )

                                                                                                                                                          g       g

                                                                                                                                                        m       m
                Fold changes

                                                                                                                                           VE        30

                                     1.5                    **                                                                 -V                            00
                                                                                                                             (          (+                  1
                                                                                                                          nt         nt
                                                                                                                       Co         Co
                                     1.0                                                                                                       GW654652

                                     0.5                                         Figure 6 Antiangiogenic activity of GW654652 in vivo. (a)
                                                                                 Representative pictures from the CAM assay. At least 10 eggs
                                     0.0                                         per each group were used. (b) Representative pictures from the
                                               0        0.01  0.05   0.1         Matrigel plug assay (magnification: Â 40). Female C57BL/6 mice
                                                       GW654652 (µM)             (8-week-old) were injected subcutaneously with 0.1 ml Matrigel or
                                                                                 Matrigel with VEGF and animals were treated with GW654652 at
 Figure 5 GW654652 induces DNA fragmentation and caspase-8                       10, 30, and 100 mg/kg/day doses (described in Materials and
 activity. (a) Agarose gel (1.5%) electrophoretic detection of                   methods). Four mice were used for each group and the blood
 nucleosomal DNA fragmentation in M6 cells after 2, 4, 8, and                    vessels were counted in one random field from each plug. (c)
 12 h treatment of 5 mM GW654652. (b) GW654652 induced caspase-                  Quantification of angiogenesis in the CAM assay. (d) The number
 8 activity in M6 cells in a dose-dependent manner. Measurements                 of blood vessels per field was calculated in random fields from four
 were in triplicate (**Po0.01, ***Po0.001). Bars, s.d.                           mice/group in the Matrigel plug assay

                                                        Inhibition of mammary cancer by VEGFRs blockade
                                                                                            J-I Huh et al
       within the plugs. GW654652 significantly inhibited                                                    tic (MIN) lesions (Figure 7a). No differences in staining
       VEGF-induced angiogenesis in the plugs at all concen-                                                of VEGF and VEGFR-2 were observed among the three
       trations tested – 10, 30, and 100 mg/kg/day – as seen in                                             groups. Microvessel density (MVD; as detected by anti-
       Figure 6b and d (positive control vs 100 mg/kg/day                                                   CD-31 immunostaining) in mammary tumors from 20-
       GW654652-treated group, Po0.001). These data de-                                                     to 25-week-old mice was reduced by more than 50% in
       monstrate that GW654652 inhibits VEGF-induced                                                        the group that received 100 mg/kg/day of GW654652
       angiogenesis in two in vivo model systems and that it                                                compared to control animals (Figure 7b and c,
       is effective as an oral agent.                                                                       **Po0.01). No discernible differences were found at
                                                                                                            the lower dose of 10 mg/kg/day. These data demonstrate
                                                                                                            that GW654652 is effective in reducing tumor vascu-
       Immunohistochemical analysis of VEGF, VEGFR-2, and                                                   lature in vivo after oral treatment.
       CD-31 in C3(1)/Tag mice
       Mammary glands were collected from 15-week-old mice
       after 3 weeks of daily oral treatment with GW654652.                                                 GW654652 delayed tumor onset, reduced tumor volume
       Both VEGF and VEGFR-2 were detected by immuno-                                                       and multiplicity, and increased survival of female C3(1)/
       histochemistry in the mammary intraepithelial neoplas-                                               Tag transgenic mice
                                                                                                            Tumor onset was moderately, but significantly, delayed
                                                                                                            in mice that received 100 mg/kg/day of GW654652,
             a                                                                                              whereas the group that received 10 mg/kg/day did not
                                                                                                            show a significant difference in tumor onset compared
                                                                                                            to controls (Figure 8a) (P ¼ 0.0069, control vs 100 mg/
                                                                                                            kg/day GW654652-treated group). Tumor volume was
                                                                                                            greatly reduced in mice that received 100 mg/kg/day of
                                                                                                            GW654652 compared to controls (Figure 8b). At 20
                                                                                                            weeks of age, the average tumor volume for the high-
                                                                                                            dose group was 150 mm3, whereas the group of mice
                                                             VEGF                         VEGFR2
                                                                                                            given 10 mg/kg/day of GW654652 and controls showed
                                                                                                            an average tumor volume of 850 mm3. Thus, a decrease
             b                                                                                              of 82% in tumor volume was achieved with GW654652.
                                                                                                            At 23 weeks of age, when all of the control mice were
                                                                                                            dead, the reduction in tumor volume in GW654652-
                                                                                                            treated mice was approximately 50% relative to the
                                                                                                            largest tumors observed at the time of death of the
                                                                                                            untreated mice (**Po0.01, *Po0.05). Tumor number
                                                                                                            was also significantly reduced in mice treated with
                                                           Control               100 mg/kg/day              GW654652 (Figure 8c). At 19 weeks of age, the mean
                                                                                                            number of tumors/mouse was less than 0.5 in mice
             c                                                                                              treated with 100 mg/kg/day of GW654652, whereas the
                 staimed areas/total area

                   Percentage of CD-31

                                                                                                            mean number of tumors in control mice was almost 2.0/
                                            12                                                              mouse. In this case, mice treated with 10 mg/kg/day of
                                                                                                            GW654652 showed intermediate values between con-
                                            8                    **
                                                                                                            trols and mice treated with 100 mg/kg/day of GW654652
                                            4                                                               (**Po0.01, *Po0.05). The survival of the mice was
                                                                                                            significantly increased by 5 weeks in animals treated
                                                                                                            with 100 mg/kg/day of GW654652, with respect to the

                                                                                                            control group (Figure 8d) (P ¼ 0.0022). After cessation



                                                                                                            of treatment with GW654652 at 20 weeks of age, tumors

                                                      GW6546452                                             began to grow rapidly (Figure 8b). At 20 weeks of age,
           Figure 7 Immunohistochemical staining of tumors. Mice (15-                                       when treatment with GW654652 was stopped, all mice
           week-old) were killed and mammary glands were collected for                                      treated with 100 mg/kg/day were alive, whereas only
           immunohistochemical analysis. Mammary tumor samples were                                         50% of mice treated with either phosphate-buffered
           collected from 20- to 25-week-old mice for CD-31 staining. (a)                                   saline (PBS) or 10 mg/kg/day of GW654652 were alive
           Both VEGF and VEGFR-2 were strongly detected by immunos-
           taining in MIN lesions from 15-week-old mice ( Â 10). (b) CD-31
                                                                                                            at the same age. These data demonstrate that
           staining from mammary tumors. MVD in tumor mammary gland                                         GW654652 reduced tumor growth and multiplicity
           detected by anti-CD-31 staining was significantly decreased after                                 resulting in increased survival, which appeared limited
           100 mg/kg/day GW654652 treatment. Representative pictures from                                   by the duration of treatment. No differences in weights
           control and 100 mg/kg/day GW654652-treated group. (c) Percen-                                    were observed between treated and untreated mice and
           tage of CD-31-stained areas vs total area was calculated using
           AnalySISs software. Each value represents mean7s.d. Signifi-                                      no signs of toxicity were observed at the doses tested.
           cantly different from control or 10 mg/kg/day GW654652-treated                                   Histopathologic analyses of livers from all mice revealed
           group vs 100 mg/kg/day GW654652-treated group (**Po0.01)                                         no abnormalities in the treated groups.
                                                                                              Inhibition of mammary cancer by VEGFRs blockade
                                                                                              J-I Huh et al
       a                                                 Control                              kinases, might provide new and complimentary thera-
                                                         10 mg                                pies for cancer (Shawver et al., 2002; Traxler, 2003).
                                              100        100 mg                               Small molecule drugs targeted against different tyrosine
           Tumor incidence

                                                                                              kinases have been intensively developed, including

                                                                                              EGFR (Traxler, 2003), platelet-derived growth factor
                                               50                                             receptor (PDGFR), Kit, and Fms-like tyrosine kinase
                                                                                              (Flt-3) (Laird et al., 2002). Solid tumors cannot grow
                                               25                                             beyond a certain size without inducing the formation of
                                                                                              new blood vessels (Folkman and D’Amore, 1996).
                                                    13 15 17 19 21 23 25 27                   Selective VEGFR kinase inhibitors will suppress tumor
                                                            Age (wk)                          growth by inhibiting tumor vascularization. GW654652
                                                                                              was developed to specifically inhibit VEGFRs phos-
       b                                                                                      phorylation (Kumar et al., 2003; Podar et al., 2004).
           Tumor burden(mm3)

                                             1600        Control
                                                         10 mg                                Several tyrosine kinase inhibitors have been shown to
                                             1200        100 mg
                                                                                              have varying degrees of efficacy in clinical trials for
                                              800                                     *
                                                                                              cancer therapy (Shawver et al., 2002). GW654652 adds
                                                                                              an additional dimension to target VEGFRs, which
                                              400                             *               could be beneficial against many forms of cancer.
                                                                * ** **                          Recent studies have demonstrated that VEGF recep-
                                                     15 16 17 18 19 20 21 22                  tors are found on human breast cancer epithelial cells
                                                             Age (wk)                         (Price et al., 2001; Nakopoulou et al., 2002), which also
                                                                                              produce VEGF, suggesting that tumor cells may
       c                                        4                                             generate an autocrine growth mechanism to their
                       tumors per mouse
                        Mean number of

                                                          10 mg                               advantage. Mammary tumor cells from the C3(1)/Tag
                                                          100 mg                              model also express high levels of VEGFR-2 (Flk-1/
                                                2                                             KDR) and VEGFR-1 (Flt-1) (Calvo et al., 2002b) . In
                                                                                              this study, we demonstrate for the first time that a
                                                1                        **                   selective VEGFRs inhibitor, GW654652, has a pro-
                                                                                              found effect on spontaneous mammary cancer growth in
                                                0               * **
                                                     15 16 17 18 19 20 21 22                  a transgenic mouse model. Our results demonstrate that
                                                             Age (wk)                         this compound targets vascularization as well as growth
                                                                                              of the tumor epithelial cells.
       d                                                                                         We have shown that inhibition of proliferation by
                                             100                                  Control     GW654652 in HUVEC and M6 tumor epithelial cells
                                                                                  10 mg
                     Survival (percentage)

                                                                                  100 mg      occurs in a dose-dependent manner at concentrations
                                              75                                              that can be achieved in vivo. The inhibition of
                                                                                              proliferation in HUVECs by GW654652 was VEGF-
                                              50                                              responsive, but the inhibition of M6 cell proliferation
                                                                                              (previously derived from a C3(1)/Tag transgenic mam-
                                              25                                              mary tumor; Holzer et al., 2003) was not affected by
                                                                                              supplementation with VEGF. This difference in re-
                                                                                              sponse to VEGF between HUVECs and M6 cells is
                                                    16   18   20   22    24       26          likely due to the relatively high levels of VEGF
                                                              Age (wk)                        endogenously produced by M6 cells, leading to a
                                                                                              potential autocrine growth effect. Therefore, additional
 Figure 8 Mammary tumor response to GW654652. The mice were
 treated daily with GW654652 at a dose of 0, 10, and 100 mg/kg by                             VEGF does not significantly increase proliferation of
 oral gavage. The control mice received only PBS. The treatment                               M6 cells, but GW654652 does significantly inhibit
 was started from 12 weeks and stopped at 20 weeks. (a) Tumor                                 proliferation signaled through VEGF.
 incidence. P ¼ 0.0069. (b) Tumor burden. **Po0.01, *Po0.05. (c)                                 Epithelial cells in the normal mammary gland secrete
 Tumor number. **Po0.01, *Po0.05. (d) Percent survival
 (P ¼ 0.0022). (E) Control group (n ¼ 8), (’) 10 mg/kg
                                                                                              very low amounts of VEGF to maintain the integrity of
 GW654652-treated group (n ¼ 10), (m) 100 mg/kg GW654652-                                     blood vessels, but in mammary carcinomas upregulation
 treated group (n ¼ 9). The arrows indicate the cessation of                                  of VEGF results in increased angiogenesis accompanied
 treatment                                                                                    by overexpression of the VEGF receptors VEGFR-1
                                                                                              and VEGFR-2 (Heer et al., 2001). A significant
                                                                                              correlation between high levels of VEGF and increased
Discussion                                                                                    KDR/Flk-1 activity associated with increased Ki-67
                                                                                              expression in hormone-induced mammary cancer
Protein kinases play an important role in signal                                              suggests that VEGF may act as an autocrine growth
transductions that regulate a number of cellular func-                                        factor for human mammary cancer cells in vivo and
tions such as differentiation and proliferation. There-                                       that this autocrine regulatory role may be mediated
fore, targeting of protein kinases, especially tyrosine                                       through VEGFR-2 (Nakopoulou et al., 2002). Our
                           Inhibition of mammary cancer by VEGFRs blockade
                                                               J-I Huh et al
       immunohistochemical results are consistent with an                      growth is caused, in part, by the reduced number of
       autocrine activation of VEGF signaling in the C3(1)/                    tumor vessels in GW654652-treated mice. GW654652
       Tag transgenic mammary cancer model since VEGF                          treatment resulted in significant improvement in all
       and VEGFR-2 are expressed strongly in MIN lesions.                      parameters related to tumor growth in vivo and
          siRNA against mouse VEGF significantly inhibited                      significantly increased animal survival. This effect was
       proliferation of M6 cells, strongly suggesting that                     clear in the animals receiving the highest dose of
       endogenously produced VEGF in M6 cells has a                            GW654652, whereas mice treated with a low dose of
       positive autocrine growth effect on these cells. By                     GW654652 (10 mg/kg/day) had little or no benefit. Low-
       reducing endogenous VEGF using siRNA, the inhibi-                       dose GW654652 was, however, able to inhibit angiogen-
       tory effect of GW654652 was even more pronounced in                     esis in the Matrigel plug assay. Additional doses and
       M6 cells. VEGF has been shown to act as an autocrine                    longer treatment of GW654652 were not tested due to
       factor for the survival of metastatic breast carcinoma                  the limited amount of compound available. Once
       cells in vitro and is associated with breast carcinoma                  treatment was terminated, tumor progression occurred
       invasion (Nakopoulou et al., 2002). This is the first                    quickly in this mouse model, demonstrating that the
       transgenic model shown to recapitulate this important                   interference of VEGF signaling in this model did not
       autocrine growth mechanism identified in human breast                    permanently inhibit tumor growth, but rather provided
       cancer.                                                                 a static growth effect during its period of administra-
          Regarding the mechanism of cell growth inhibition,                   tion. These results are consistent with the hypothesis
       our kinase assays show that GW654652 is highly                          that chronic administration of antiangiogenic drugs will
       specific for VEGFR-2. The IC50 for VEGFR-3                               be required for a maximal, sustained effect, and that
       (0.0025 mM) is also similar to that found for VEGFR-                    these compounds may be most efficacious when given in
       2. In spite of the specificity of GW654652 for VEGFR-2                   combination with drugs that target additional oncogenic
       compared to other receptor kinases, we cannot exclude                   pathways. Ongoing studies will determine whether long-
       the possibility that GW654652 may also exert an                         term administration of GW654652 would continue to
       inhibitory effect through other non-VEGFR-2 path-                       inhibit tumor growth or whether resistance to the
       ways. A recent study demonstrated that GW654652                         compound would develop.
       inhibited the growth and migration of multiple myelo-                      Of significance is the fact that no apparent toxicity
       ma cells, which express VEGFR-1 but not VEGFR-2                         was noted in animals that received this compound
       (Podar et al., 2004). Such a result has also been found                 during the course of these experiments. Livers from all
       for other small molecules targeting VEGF receptors.                     treated and untreated mice were also examined micro-
       For instance, SU5416, SU6668, and SU11248 appear to                     scopically and no histological abnormalities were
       have inhibitory activity on VEGF, FGF, and PDGF                         observed in the treated group. While potential toxic
       receptors, thus, potentially broadening the range of                    effects are not excluded by these studies, further
       molecular targets (Itokawa et al., 2002; Laird et al.,                  investigations involving GW654652 appear warranted
       2002; Spiekermann et al., 2002). Although SU5614 is a                   since the development of some kinase inhibitors has
       potent inhibitor of VEGF-induced endothelial cell                       been terminated because of toxicities in clinical trials.
       sprouting in vitro, the sensitivity of leukemic cells                      Our previous studies have demonstrated that addi-
       toward the growth inhibitory activity of the compound                   tional molecular targets can be used to interfere with
       is dependent upon c-kit, not VEGFR-1/2 expression                       tumor angiogenesis in the C3(1)/Tag mammary cancer
       (Spiekermann et al., 2002).                                             model, including endostatin (Yokoyama et al., 2000;
          Our study demonstrates that GW654652 induces                         Calvo et al., 2002a, b) and 2-methoxyestradiol (et al.,
       apoptosis in both HUVECs and M6 cells followed by                       unpublished observations). The data presented in this
       a cell growth arrest at the G1 phase of the cell cycle. We              study demonstrate that interference of the VEGF
       also show that GW654652 induces DNA fragmentation                       signaling pathway using GW654652 has a significant
       and caspase-8 activity in M6 cells during apoptosis.                    effect on tumor growth inhibition and on prolongation
       Other antiangiogenic factors have also been shown to                    of survival. This work may have important implications
       regulate apoptosis. For example, 2-methoxyestradiol, an                 for translating antiangiogenic therapy to human breast
       endogenous compound with antiangiogenic activity,                       cancer. Future studies will address how combination
       triggers apoptosis in various tumor cell lines (Pribluda                therapies may improve these parameters even further by
       et al., 2000; Qanungo et al., 2002). The cell lines                     targeting different pathways involved in angiogenesis,
       responsive to 2-methoxyestradiol undergo apoptosis                      tumor growth, and survival.
       either by G2–M arrest with Bcl-xL phosphorylation or
       through the accumulation of tetraploid cells in the G1–
       S phase without Bcl-2/Bcl-xl phosphorylation. These
       results suggest that GW654652 can suppress tumor                        Materials and methods
       progression, in part, through the activation of apoptosis
       in tumor cells as well as endothelial cells.                            Reagents
          In the present study, we have demonstrated that                      GW654652 was discovered and synthesized by GlaxoSmithK-
       tumor vascularization is significantly decreased in                      line (Research Triangle Park, NC, USA). For in vitro and in
       animals that received GW654652 as compared to                           vivo assays, a stock solution of 1 mM GW654652 was prepared
       control animals. We conclude that inhibition of tumor                   in dimethylsulfoxide (DMSO).
                                                                   Inhibition of mammary cancer by VEGFRs blockade
                                                                   J-I Huh et al
Kinase assays                                                      Reverse transcription–polymerase chain reaction
Kinase activities for VEGFR-1, VEGFR-2, VEGFR-3,                   RT reactions were performed using the Superscriptt first-
FGFR-1, FGFR-2, FGFR-3, SRC, Eph-B4, c-Fms, Tie-2,                 strand synthesis system for RT–PCR (Invitrogen, Carlsbad,
EGFR, ErbB-2, CDK-2, and CDK-4 were conducted as                   CA, USA) following the manufacturer’s instructions: 1 mg total
previously described (Kumar et al., 2003).                         RNA, 1 ml of 10 mM dNTP mix, and 1 ml of oligo(dT)12–18
                                                                   (0.5 mg/ml) in a total volume of 10 ml at 651C for 5 min,
                                                                   followed by the addition of 2 ml of 10 Â RT buffer, 4 ml of
Cell lines                                                         25 mM MgCl2, 2 ml of 0.1 M DTT, 1 ml of RNaseOUTt (40 U/
                                                                   ml) with incubation at 421C for 2 min. Reverse transcriptase
M6 cells, which were derived from C3(1)/Tag transgenic mice
                                                                   (50 U) was added and incubated at 421C for 50 min. The
mammary gland tumors (Holzer et al., 2003), were cultured in
                                                                   reaction was terminated at 701C for 15 min and then chilled on
DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with
                                                                   ice. A 1 ml portion of RNase H (2 U/ml) was added to each tube
10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA,
                                                                   and incubated for 20 min at 371C.
USA) and penicillin/streptomycin. HUVECs (less than pas-
                                                                       PCR mixture included 1 ml of the cDNA generated from the
sage 7) were obtained from freshly delivered umbilical cords by
                                                                   RT reaction, 1 ml of 10 Â PCR buffer, 1.5 ml of 50 mM MgCl2,
treatment with 0.1% collagenase (Jaffe et al., 1973). Cells were
                                                                   1 ml of 10 mM dNTP, 0.4 ml of Taq polymerase (5 U/ml), 0.5 ml
cultured in RPMI 1640 (Invitrogen, Carlsbad, CA, USA)
                                                                   of 20 mM sense primer, and 0.5 ml of 20 mM antisense primer in
supplemented with 20% bovine calf serum (BCS) (Hyclone,
                                                                   a total volume 50 ml. PCR conditions were as follows: 40 cyc-
Logan, UT, USA), 2 mM glutamine, 0.1 mg/ml heparin
                                                                   les of 941C 5 min/941C 30 s; 551C 30 s; 721C 1 min/721C, 7
(Sigma-Aldrich, St Louis, MO, USA), penicillin/streptomycin,
                                                                   min, 41C. The following primer sequences were used: VEGF
and 0.05 mg/ml endothelial cell growth supplement (Sigma-
                                                                   sense 50 -CTGCTCTCTTGGGTGCACTG-30 and antisense
Aldrich, St Louis, MO, USA) at 371C and 5% CO2.
                                                                   50 -CACCGCCTTGGCTT GTCACA-30 ; GAPDH sense
                                                                   50 -ATCACCATCTTCCAGGAGCGAGAC-30 and antisense
                                                                   50 -ATGCCAGTGAGCTTCCCGTTCAGC-30 .
Cell proliferation assay
HUVECs or M6 cells were seeded at 3 Â 103 cells/well in flat-       ELISA analysis
bottom collagen-coated or uncoated 96-well plates (Beckton
Dickinson, Mansfield, MA, USA), in quadruplicate and                M6 cells (106) were plated in six wells and the next day
allowed to adhere overnight. GW654652 (final concentrations         transfected with GFP or siVEGF. After 24 h transfection, the
from 0 to 0.1 mM), in serum-free medium for M6 cells and 1%        conditioned media were collected for VEGF protein measure-
FCS for HUVECs, was added to each well. After a 30 min             ment using a quantitative sandwich enzyme immunoassay
incubation with GW654652, cells were subsequently stimu-           (Quantikine; R&D Systems, Minneapolis, MN, USA) accord-
lated with 10 ng/ml VEGF (R&D Systems, Minneapolis, MN,            ing to the manufacturer’s instructions.
USA) or 0.5 ng/ml bFGF (R&D Systems, Minneapolis, MN,
USA). M6 cells were incubated for 24 h and HUVECs for 72 h         FACS analysis
at 371C. Proliferation was measured using the WST-1 reagent
(Roche Molecular Biochemicals, Mannheim, Germany). A               After 24 h in serum-free medium, HUVECs were incubated for
10 ml portion of WST reagent was added to each well and the        24 h, and M6 cells were incubated for 6 h with 0.05, 0.1 or
cells were incubated for 4 h at 371C. Absorbance at 405 nm was     0.5 mM GW654652. Cells were trypsinized and resuspended in
measured using an ELISA plate reader. All the experiments          2 ml PBS containing 1 mM EDTA. The cells were collected
were performed at least three times.                               following centrifugation and washed once with PBS. Cells were
                                                                   resuspended in 2 ml ice-cold 80% ethanol and placed on ice for
                                                                   30 min. Cells were washed once with PBS and the cell pellet
siRNA transfection                                                 was resuspended in 1 ml propidium iodide (PI) (Sigma-
                                                                   Aldrich, St Louis, MO, USA) staining buffer (50 mg/ml PI
siRNA against mouse VEGF (siVEGF) was employed from                and 25 mg/ml RNase A; Roche Molecular Biochemicals,
a previous study (Reich et al., 2003). siVEGF target sequence      Mannheim, Germany) in permeability buffer (0.1% Triton
is AACGAUGAAGCCCUGGAGUGC. Sense sequence is                        X-100/0.1 mM EDTA in PBS).
CGAUGAAGCCCUGGAGUGC and antisense sequence is                         The samples were analysed using FACSCalibur (Becton
GCACUCCAGGGCUUCAUCG. Green fluorescent protein                      Dickinson, Mansfield, MA, USA) within 24 h after prepara-
duplex (GFP) was used as a control. The target sequence is         tion and the data were analysed with CellQuest software
GGCTACGTCCAGGAGCGCACC. M6 cells (0.5 Â 105)                        (Becton Dickinson, Mansfield, MA, USA).
were plated in six wells with 2% FBS without antibiotics
the day before transfection. siRNAs were synthesized by
                                                                   DNA fragmentation
Dharmacon Inc. siRNA was transfected using Lipofectamine
2000 (Invitrogen, Carlsbad, CA, USA) following the manu-           M6 cells (1 Â 106) were plated in six wells and grown overnight
facturer’s instructions. Transfected cells were incubated for      until they attached. The medium was changed to serum-free
24 h at a 371C incubator. Conditioned media from each well         medium and 0.1 mM GW654652 was added for 2, 4, 8, and 12 h
were collected to measure the concentration of VEGF protein        and harvested by spinning at 1000 r.p.m. for 10 min. Cell
by ELISA analysis. Total RNA was extracted using Trizol            pellets were washed with PBS, resuspended in lysis buffer
reagent (Invitrogen, Carlsbad, CA, USA) from the same              (10 mM Tris pH 7.5, 1 mM EDTA, 0.2% Triton X-100) and
wells for semiquantitative RT–PCR. GW654652 was added              incubated at room temperature for 15 min. After centrifuga-
at 0.01, 0.05, and 0.1 mM concentrations 24 h after siVEGF         tion, the supernatant was transferred to a new tube and 55 ml
transfection to determine the autocrine effect of endogenous       of 10% SDS, 5 ml of 10 mg/ml RNase A (Roche Molecular
VEGF in M6 cells following stimulation with 10 ng/ml               Biochemicals, Mannheim, Germany), and 5 ml of 20 mg/ml
VEGF on cell proliferation. All the experiments were               proteinase K (Sigma, St Louis, MO, USA) were added
repeated at least twice.                                           and incubated at 371C for 45 min. DNA was extracted
                             Inhibition of mammary cancer by VEGFRs blockade
                                                                 J-I Huh et al
       and resuspended in 20 ml of 100 mg/ml RNase A and incubated               killed at 15 weeks of age and mammary glands were fixed in
       for 15 min at 371C and then run on a 1.5% agarose gel stained             4% paraformaldehyde, embedded in paraffin, and cut for
       with SYBR green I (Molecular Probes Inc., Eugene, OR, USA).               immunohistochemical analyses. For CD-31 immunohisto-
                                                                                 chemistry, tumor samples from 20- to 25-week-old mice were
       Measurement of caspase-8 enzymatic activity                               immediately immersed in optimal cutting temperature (OCT)
                                                                                 compound (Tissue-Tek, Sakura, Japan) and frozen on a
       Caspase-8 activity was measured using a caspase colorimetric              mixture of dry ice and isopentane (Fluka, St Louis, MO,
       protease assay kit (R&D Systems, Minneapolis, MN, USA).                   USA). All mice were treated in accordance with the guidelines
       Briefly, after treatment of cells with 0.01, 0.05, and 0.1 mM              of the Animal Care and Use of Laboratory Animals (NIH
       GW654652 for 8 h, cells were harvested and pelleted. Lysis                Publication No. 86-23, 1985) under an approved animal
       buffer was added to the cell pellets. Resuspended cells were              protocol.
       incubated on ice for 10 min and then centrifuged to remove
       insoluble materials. Cell lysates were incubated with 5 ml of             Immunohistochemistry
       1 mM stock of substrate at 371C for 2 h. The release of 7-
       amino-4-trifluoromethyl coumarin (AFC) was measured with                   Paraffin sections were rehydrated through decreasing concen-
       an ELISA reader at 405 nm. Assay was repeated three times.                trations of ethanol. Slides were immersed in citric acid (pH 6)
                                                                                 and heated in a microwave oven twice for 5 min. Anti-VEGF
       Chick chorioallantoic membrane assay                                      antibody (sc-507, Santa Cruz Biotechnology, Santa Cruz, CA,
                                                                                 USA) was used at a 1 : 200 dilution. Anti-Flk-1 antibody (sc-
       Approximately 4 ml of albumin was removed from 3-day-old                  315, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was
       embryonated eggs (CBT Farms, Chestertown, MD, USA). On                    used at a 1 : 400 dilution.
       day 5, windows were opened in the shell for each egg and                     To detect MVD, cryostat-cut sections were immersed in
       covered with tape to prevent drying, and then further                     citric acid (pH 6) and heated in a microwave oven twice for
       incubated at 371C. On day 7, 2 ml of solution containing                  5 min for antigen retrieval. Slides were stained with anti-CD-31
       50 ng human VEGF (hVEGF) and either 0.1, 0.5, or 1 mM                     antibody (PharMingen, San Diego, CA, USA) at a 1 : 150
       GW654652 was applied to the center of quartered 13-mm-                    dilution. The avidin–biotin complex method (Vectastatin ABC
       diameter plastic coverslips (Nalge Nunc International, Roche-             Elite kit; Vector Laboratories, Burlingame, CA, USA) was
       ster, NY, USA) and air dried. hVEGF (50 ng) served as a                   used to visualize the bound antibody. Immunostaining for CD-
       positive control. DMSO (2.5 ml) served as a negative control.             31 (PECAM) was used for quantification of vascular density of
       Each coverslip was put on the CAM, and the eggs were                      tumors. Images from four fields randomly chosen were
       incubated at 371C for 3 days. Angiogenic response was then                captured from each tumor sample using a Nikon Eclipse
       scored as either positive or negative and photographed (Ponce             E800M microscope. Mammary tumors from five mice at
       et al., 2001).                                                            approximately 20–25 weeks of age in each experimental group
                                                                                 were examined, providing 20 fields per group for analyses.
       Matrigel plug assay                                                       Image capture and analysis were performed with the Analy-
                                                                                 SISs software (Soft Imaging System, Lakewood, CO, USA)
       A 1 ml portion of Matrigelt (Becton Dickinson Labware,                    using the same light and color conditions for all the samples at
       Bedford, MA, USA) alone or mixed with 50 ng mVEGF was                     Â 200. The brown signal from CD-31 immunostaining was
       injected subcutaneously into female 8-week-old C57BL/6 mice.              filtered using the same filtering conditions for all of the
       The same concentration of mVEGF was employed as that                      samples. The area fraction (area occupied by the CD-31
       described in a previous paper (Inoki et al., 2002). The mice              staining divided by reference area) was calculated. Data are
       were treated with GW654652 (0, 10, 30, and 100 mg/kg/day) by              presented as the mean percentage of area occupied by CD-31
       oral gavage for 7 days. On day 7, the mice were killed and the            staining with respect to the total tumor area.
       plugs were harvested, fixed in formalin, and stained with
       Masson trichrome blue. Four mice were used in each group.
       One field from each sample was selected randomly and blood                 Statistical analysis
       cells were counted to evaluate vascularization of the plugs as            A one-way ANOVA was used to compare the mean area
       described previously (Inoki et al., 2002).                                occupied by CD-31 staining between the three experimental
                                                                                 groups and for quantification of blood vessels from the
       Animal studies                                                            Matrigel plug assay. Neuman–Keuls analysis was used as a
                                                                                 post hoc test. The two-sided Student’s t-test was used to
       Heterozygous C3(1)/Tag transgenic female mice (12-week-old)               evaluate whether significant differences existed between the
       in the FVB/N background were used for these studies and have              mean values of the groups analysed for the apoptosis assay.
       been previously described (Maroulakou et al., 1994). The mice             The log-rank test was used to determine differences in tumor
       were divided into three groups: (1) control group that received           onset and survival curves.
       PBS (n ¼ 16); (2) group that received 10 mg/kg/day of
       GW654652 (n ¼ 16); and (3) group that received 100 mg/kg/
       day of GW654652 (n ¼ 16). GW654652 was suspended in
       hydroxypropyl methyl cellulose (HPMC) (Sigma-Aldrich, St                  Abbreviations
       Louis, MO, USA) containing 0.1% Tween 80 (Sigma-Aldrich,                  HUVEC, human umbilical vein endothelial cells; VEGF,
       St Louis, MO, USA) in water as a vehicle. GW654652 was                    vascular endothelial growth factor; bFGF, basic fibroblast
       administered to the mice by oral gavage daily for 8 weeks                 growth factor; Tag, simian virus 40 T-antigen.
       beginning at 12 weeks of age (a stage when MIN lesions have
       developed). The control group of mice received PBS with the               Acknowledgements
       same schedule and route of administration from each group.                We appreciate Mizuko Mamura and Sushil G Rane for their
       Tumor incidence, burden (calculated volume ¼ the largest                  assistance and helpful discussion with regard to FACS
       diameter  (the smallest diameter)2  0.4; Calvo et al.,                  analyses. We are also grateful to Lisa Birely for her technical
       2002b), and number were measured weekly. Five mice were                   assistance with regard to animals.
                                                                Inhibition of mammary cancer by VEGFRs blockade
                                                                J-I Huh et al

Calvo A, Feldman AL, Libutti SK and Green JE. (2002a).          Lin B, Podar K, Gupta D, Tai YT, Li S, Weller E, Hideshima
  Cancer Res., 62, 3934–3938.                                     T, Lentzsch S, Davies F, Li C, Weisberg E, Schlossman RL,
Calvo A, Yokoyama Y, Smith LE, Ali I, Shih SC, Feldman            Richardson PG, Griffin JD, Wood J, Munshi NC and
  AL, Libutti SK, Sundaram R and Green JE. (2002b). Int. J.       Anderson KC. (2002). Cancer Res., 62, 5019–5026.
  Cancer, 101, 224–234.                                         Maroulakou IG, Anver M, Garrett L and Green JE. (1994).
Cheng M, Bollor A, Hinkle K, Davis-Ward R, Harris P, Mook         Proc. Natl. Acad. Sci. USA, 91, 11236–11240.
  RJ, Veal J, Truesdale A, Johnson J, Crosby R, Rudolph S,      Nakopoulou L, Stefanaki K, Panayotopoulou E, Giannopou-
  Knick V, Epperly A, Kumar R, Luttrell D and Stafford J.         lou I, Athanassiadou P, Gakiopoulou-Givalou H and
  (2003). Proc. Am. Assoc. Cancer Res., 45, 40 (abstract).        Louvrou A. (2002). Hum. Pathol., 33, 863–870.
Ciardiello F and Tortora G. (2002). Expert Opin. Investig.      Parry TJ, Cushman C, Gallegos AM, Agrawal AB, Richard-
  Drugs, 11, 755–768.                                             son M, Andrews LE, Maloney L, Mokler VR, Wincott FE
Dev I, Dornsife R, Hopper T, Onori J, Miller C, Harrington L,     and Pavco PA. (1999). Nucleic Acids Res., 27, 2569–2577.
  Dold K, Mulligan R and Kumar R. (2003). Proc. Am. Assoc.      Podar K, Catley LP, Tai YT, Shringarpure R, Carvalho P,
  Cancer Res., 45, 4695 (abstract).                               Hayashi T, Burger R, Schlossman RL, Richardson PG,
Folkman J. (1995). Nat. Med., 1, 27–31.                           Pandite LN, Kumar R, Hideshima T, Chauhan D and
Folkman J and D’Amore PA. (1996). Cell, 87, 1153–1155.            Anderson KC. (2004). Blood, 103, 3474–3479.
Fong GH, Rossant J, Gertsenstein M and Breitman ML.             Ponce ML, Nomizu M and Kleinman HK. (2001). FASEB J.,
  (1995). Nature, 376, 66–70.                                     15, 1389–1397.
Gutheil JC, Campbell TN, Pierce PR, Watkins JD, Huse WD,        Presta LG, Chen H, O’Connor SJ, Chisholm V, Meng YG,
  Bodkin DJ and Cheresh DA. (2000). Clin. Cancer Res., 6,         Krummen L, Winkler M and Ferrara N. (1997). Cancer
  3056–3061.                                                      Res., 57, 4593–4599.
Heer K, Kumar H, Read JR, Fox JN, Monson JR and Kerin           Pribluda VS, Gubish Jr ER, Lavallee TM, Treston A, Swartz
  MJ. (2001). Clin. Cancer Res., 7, 3491–3494.                    GM and Green SJ. (2000). Cancer Metast. Rev., 19, 173–179.
Holash J, Davis S, Papadopoulos N, Croll SD, Ho L, Russell      Price DJ, Miralem T, Jiang S, Steinberg R and Avraham H.
  M, Boland P, Leidich R, Hylton D, Burova E, Ioffe E,            (2001). Cell Growth Differ., 12, 129–135.
  Huang T, Radziejewski C, Bailey K, Fandl JP, Daly T,          Qanungo S, Basu A, Das M and Haldar S. (2002). Oncogene,
  Wiegand SJ, Yancopoulos GD and Rudge JS. (2002). Proc.          21, 4149–4157.
  Natl. Acad. Sci. USA, 99, 11393–11398.                        Reich SJ, Fosnot J, Kuroki A, Tang W, Yang X, Maguire
Holzer RG, MacDougall C, Cortright G, Atwood K, Green             AM, Bennett J and Tolentino MJ. (2003). Mol. Vis., 9,
  JE and Jorcyk CL. (2003). Breast Cancer Res. Treat., 77,        210–216.
  65–76.                                                        Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF,
Inoki I, Shiomi T, Hashimoto G, Enomoto H, Nakamura H,            Breitman ML and Schuh AC. (1995). Nature, 376, 62–66.
  Makino K, Ikeda E, Takata S, Kobayashi K and Okada Y.         Shawver LK, Slamon D and Ullrich A. (2002). Cancer Cell, 1,
  (2002). FASEB J., 16, 219–221.                                  117–123.
Itokawa T, Nokihara H, Nishioka Y, Sone S, Iwamoto Y,           Shibata MA, Maroulakou IG, Jorcyk CL, Gold LG, Ward JM
  Yamada Y, Cherrington J, McMahon G, Shibuya M,                  and Green JE. (1996). Cancer Res., 56, 2998–3003.
  Kuwano M and Ono M. (2002). Mol. Cancer Ther., 1,             Spiekermann K, Faber F, Voswinckel R and Hiddemann W.
  295–302.                                                        (2002). Exp. Hematol., 30, 767–773.
Jaffe EA, Hoyer LW and Nachman RL. (1973). J. Clin.             Takahashi T, Yamaguchi S, Chida K and Shibuya M. (2001).
  Invest., 52, 2757–2764.                                         EMBO J., 20, 2768–2778.
Kerbel R and Folkman J. (2002). Nat. Rev. Cancer, 2,            Traxler P. (2003). Expert Opin. Ther. Targets, 7, 215–234.
  727–739.                                                      Xie B, Tam NN, Tsao SW and Wong YC. (1999). Br. J.
Kumar R, Hopper T, Miller C, Johnson J, Crosby R, Onori J,        Cancer, 81, 1335–1343.
  Mullin R, Truesdale A, Epperly A, Hinkle K, Cheng M,          Yokoyama Y, Green JE, Sukhatme VP and Ramakrishnan S.
  Stafford J and Luttrell D. (2003). Proc. Am. Assoc. Cancer      (2000). Cancer Res., 60, 4362–4365.
  Res., 45, 39 (abstract).                                      Zhu Z, Hattori K, Zhang H, Jimenez X, Ludwig DL, Dias S,
Laird AD, Christensen JG, Li G, Carver J, Smith K, Xin X,         Kussie P, Koo H, Kim HJ, Lu D, Liu M, Tejada R,
  Moss KG, Louie SG, Mendel DB and Cherrington JM.                Friedrich M, Bohlen P, Witte L and Rafii S. (2003).
  (2002). FASEB J., 16, 681–690.                                  Leukemia, 17, 604–611.


To top