Antiangiogenic treatment effects on neuroblastoma xenografts

1. Research Article
Thrombospondin-1 Peptide ABT-510 Combined with Valproic Acid
Is an Effective Antiangiogenesis Strategy in Neuroblastoma
1 1 1 1 1
Qiwei Yang, Yufeng Tian, Shuqing Liu, Rana Zeine, Alexandre Chlenski,
1 3 2
Helen R. Salwen, Jack Henkin, and Susan L. Cohn
1
The Robert H. Lurie Comprehensive Cancer Center and 2Department of Pediatrics, Northwestern University,
Feinberg School of Medicine, Chicago, Illinois and 3Abbott Laboratories, Abbott Park, Illinois
Abstract substantial progress has been made in the treatment of children
with low- and intermediate-risk neuroblastoma with reduced
In the pediatric cancer neuroblastoma, clinically aggressive
therapy approaches. However, more effective therapy is still needed
disease is associated with increased levels of angiogenesis
for children with high-risk disease. Despite intensive multimodality
stimulators and high vascular index. We and others have
treatment, <30% of high-risk patients are cured (1). Similar to other
hypothesized that blocking angiogenesis may be effective
solid tumors, angiogenesis is required for neuroblastoma tumor
treatment for this pediatric malignancy. However, little is
growth and metastasis. The majority of published studies indicate
known about the efficacy of antiangiogenic agents in pediatric
that angiogenesis contributes to the aggressive behavior of
malignancies. Recently, promising results have been reported
neuroblastoma tumors, and a positive correlation between poor
in an adult phase I study of ABT-510, a peptide derivative of
outcome and high levels of factors that stimulate blood vessel
the natural angiogenic inhibitor thrombospondin-1. Histone
growth and vascular index has been reported (3, 4). In addition,
deacetylase inhibitors, such as valproic acid (VPA), have also
preclinical experiments have shown that angiogenesis inhibitors
been shown to have antiangiogenic activity in several cancer
are capable of suppressing neuroblastoma growth, further support-
models. In this study, we evaluated the effects of ABT-510
ing the concept that the blood vessels in neuroblastoma tumors are
and VPA on neuroblastoma tumor growth and angiogenesis.
likely to be clinically relevant therapeutic targets (5, 6).
Although only VPA was capable of blocking the proliferation
Thrombospondin-1 (TSP-1) is a well-known natural inhibitor of
of neuroblastoma cells and inducing neuroblastoma cell
angiogenesis, and down-regulation of TSP-1 plays a critical role in
apoptosis in vitro, treatment with VPA or ABT-510 alone
the angiogenic switch in several tumor types (7–9). In neuroblas-
significantly suppressed the growth of neuroblastoma xeno-
toma, we have previously shown that TSP-1 is epigenetically
grafts established from two different MYCN-amplified cell
silenced in a subset of undifferentiated, advanced-stage tumors and
lines. Combination therapy more effectively inhibited the
cell lines, whereas it is expressed in tumors with morphologic
growth of small neuroblastoma xenografts than single-agent
evidence of neuroblast differentiation (10). The antiangiogenic
treatment, and in animals with large xenografts, total
effect of TSP-1 is mediated in part by the CD36 receptor, which
cessation of tumor growth was achieved with this treatment
triggers a signaling cascade that leads to apoptosis in activated
approach. The microvascular density was significantly re-
endothelial cells via the CD95 death receptor. The inhibitory
duced in the xenografts treated with combination therapy
activity of TSP-1 has been mapped to the properdin type 1 repeats
compared with controls or tumors treated with single agents.
in the NH2-terminal third of the molecule, and a single substitution
In addition, the number of structurally abnormal vessels was
of D-isoleucine for L-isoleucine in one properdin-region heptapep-
reduced, suggesting that these agents may ‘‘normalize’’ the
tide leads to a 1,000-fold increase in the antiangiogenic activity
tumor vasculature. Our results indicate that ABT-510 com-
(11, 12). ABT-510 is a modified analogue of the active TSP-1
bined with VPA may be an effective antiangiogenic treatment
heptapeptide, which has slowed tumor growth in both xenograft
strategy for children with high-risk neuroblastoma. [Cancer
and syngeneic models (13). Administration of ABT-510 to TSP-1
Res 2007;67(4):1716–24]
knockout mice for 5 consecutive days has been shown to reduce by
f5-fold both elevated circulating endothelial cells and endothelial
Introduction precursor cells to near wild-type levels (14). Although no pediatric
The clinical hallmark of neuroblastoma is heterogeneity, with the cancer studies have been done to date, a recently completed phase
likelihood of tumor progression varying widely according to age at I trial of ABT-510 in adults with advanced cancer has shown a
diagnosis and extent of disease (1). In addition to these clinical favorable toxicity profile and a decrease in the median serum basic
factors, biological markers, such as MYCN gene status, tumor cell fibroblast growth factor (bFGF) levels. Stable disease lasting for six
ploidy, and tumor histology, have also been shown to be strongly cycles or more was seen in 6 of the 39 patients enrolled on the
predictive of risk of relapse (2). Modern treatment strategies are phase I study, and additional response data are being collected on
stratified according to these clinical and biological classifiers, and ongoing phase II studies (15).
Valproic acid (VPA), a branched chain fatty acid initially used for
the treatment of epilepsy, has also been shown to have anticancer
Note: Supplementary data for this article are available at Cancer Research Online activity and is currently being tested in clinical trials. VPA inhibits
(http://cancerres.aacrjournals.org/).
Current address for S.L. Cohn: University of Chicago, Chicago, Illinois.
the growth of many types of cancer, including glioma, neuroblas-
Requests for reprints: Susan L. Cohn, Institute for Molecular Pediatric Sciences, toma (16, 17), breast cancer (18), acute myelogenous leukemia
University of Chicago, 5841 Maryland Avenue, MC 4060, Room N114, Chicago, IL (19, 20), thyroid cancer (21), and endometrial carcinoma (22).
60637. Phone: 773-702-2571; Fax: 773-834-1329; E-mail: scohn@peds.bsd.uchicago.edu.
I2007 American Association for Cancer Research. Recently, VPA and IFN-a have been reported to synergistically
doi:10.1158/0008-5472.CAN-06-2595 inhibit neuroblastoma growth, and increased anticancer activity
Cancer Res 2007; 67: (4). February 15, 2007 1716 www.aacrjournals.org

2. Valproic Acid and TSP-1 Inhibit Neuroblastoma Growth
has also been seen when VPA is combined with all-trans retinoic The experiment was done in duplicate. The cell cycle data were analyzed
acid (16, 17). Although the mechanisms responsible for the anti- with an Epics XL-MCL flow cytometer (Beckman Coulter, Miami, FL), with
cancer activity of VPA are not completely understood, like other System II (version 3.0) software (Beckman Coulter). Further analysis of cell
short chain fatty acids, it inhibits histone deacetylase (HDAC) cycle distribution was determined by using ModFit LT (Verity Software
House, Topsham, ME).
activity and plays a role in the regulation of gene expression.
Annexin V/propidium iodide staining. Apoptosis and/or necrosis were
Recent studies have indicated that HDAC inhibitors are also evaluated using two-color flow cytometric analysis with fluorescein-labeled
capable of blocking hypoxia-induced and vascular endothelial Annexin V and propidium iodide (10). Cells were fluorescein labeled using
growth factor–induced angiogenesis (23). methods provided by the manufacturer of the Annexin V-FITC Apoptosis
In this study, we investigated the effects of the antiangiogenic Detection kit (Immunotech, Marseille, France). The experiment was done
agent ABT-510 on neuroblastoma growth and angiogenesis. in duplicate. In brief, cultured neuroblastoma cell lines were treated with
Because the efficacy of angiogenesis inhibitors is greatly improved 10, 100, and 1,000 nmol/L of ABT-510 for 24 h. Cells were washed twice
when they are combined with other anticancer drugs (24, 25), we with PBS and resuspended at 1 Â 106/mL in binding buffer with FITC-
conducted additional experiments with ABT-510 combined with conjugated Annexin V and propidium iodide. Samples were analyzed using
VPA. VPA was selected for the combination studies because it is the Epics XL-MCL flow cytometer.
Endothelial cell migration assay. Endothelial cell migration assays
capable of inhibiting neuroblastoma cell proliferation and tumor
were done in a modified Boyden chamber (NeuroProbe, Gaithersburg, MD)
growth and can block HDAC activity and angiogenesis. In addition, with human microvascular endothelial cells (Cambrex Corp., East
VPA is known to have a low toxicity profile. We found that VPA, Rutherford, NJ) in EBM medium (Cambrex) containing 0.01% bovine serum
but not ABT-510, was capable of blocking the proliferation of albumin as described previously (26, 27). Test substances were assayed with
neuroblastoma cells and inducing neuroblastoma cell apoptosis or without 5 ng/mL bFGF (National Cancer Institute Preclinical Repository,
in vitro. However, treatment with either ABT-510 or VPA alone Bethesda, MD). To generate dose-response curves, the data were normalized
significantly suppressed the growth of neuroblastoma xenografts as percentage of maximum migration using the difference between bFGF/
established from two different MYCN-amplified cell lines. ABT-510 EBM-induced migration and background migration in EBM alone as 100%
combined with VPA more effectively inhibited the growth of small control. The experiment was done in triplicate.
tumors than single-agent therapy, and in animals with large cDNA synthesis and Taqman assay. Total RNA (2 Ag) was used for
reverse transcription using SuperScript III (Invitrogen) and 50 Amol/L
tumors, combination therapy stabilized tumor growth. In addition
oligo(dT)20 in a 20 AL reaction volume according to standard protocols
to blocking angiogenesis, tumor cell apoptosis and inhibited tumor supplied by the manufacturer. cDNA synthesis was done at 50jC for
cell proliferation were seen in the tumors treated with combina- 50 min. Following heat inactivation at 85jC for 5 min and RNase H
tion therapy. Our results, together with the known low toxicity treatment at 37jC for 20 min, 480 AL distilled water was added and 3 AL of
profile of both agents, suggest that VPA combined with ABT-510 diluted cDNA were used for Taqman analysis. Taqman reactions were set
may be an effective, well-tolerated antiangiogenic treatment up containing 1Â Taqman Fast Universal PCR Master Mix (Applied
strategy for children with high-risk neuroblastoma. Biosystems, Foster City, CA), 250 nmol/L FAM-labeled specific probe, and
900 nmol/L forward and reverse primers in a 20 AL reaction. All assays
were carried out in a 96-well format. Real-time fluorescent detection of PCR
Materials and Methods products was done with an Applied Biosystems 7500 Fast Real-Time PCR
Cell lines. The biological and genetic characteristics of the SMS-KCNR System using the following thermocycling conditions: 1 cycle of 50jC for
and NMB MYCN-amplified neuroblastoma cell lines used in this study have 2 min and 95jC for 20 s; 40 cycles of 95jC for 3 s and 60jC for 30 s. The
been previously described (10). Cells were grown at 5% CO2 in RPMI 1640 sequences for the primers and probes for MYCN, TSP-1, SPARC, and
(Invitrogen, Carlsbad, CA) supplemented with 10% heat-inactivated fetal b2-microglobulin were described previously (10, 27, 28). Data were analyzed
bovine serum (FBS), L-glutamine, and antibiotics. using the comparative method (DDC t) to calculate relative quantities of
Peptide. ABT-510 (Abbott Laboratories, Abbott Park, IL) is a capped a nucleic acid sequence. Nontreated SMS-KCNR and NMB were used as the
nonapeptide based on the linear TSP-1 heptapeptide sequence containing a calibrator sample, and h2-microglobulin was used as an endogenous
fragment of the second type I TSP-1 repeat. The sequence of ABT-510 is control detector.
acetyl-sarcosine-glycine-valine-D-alloisoleucine-threonine-norvaline-isoleu- Neuroblastoma xenograft studies. Four- to 6-week-old female
cine-arginine-proline-ethylamide. Lyophilized peptides were dissolved in homozygous athymic nude mice (Harlan, Madison, WI) were inoculated
sterile 5% dextrose and stored at 4jC. s.c. into the right flank with 1 Â 107 SMS-KCNR or 1.25 Â 107 NMB cells.
Proliferation assay. The in vitro proliferation of SMS-KCNR cells was Once tumors were palpable (70 mm3), mice were treated once daily for
determined by using the CellTiter 96 AQ nonradioactive cell proliferation 20 days with either ABT-510 (i.p., 40 mg/kg), VPA (i.p., 400 mg/kg), or ABT-
assay (Promega Corp., Madison, WI). SMS-KCNR cells were seeded into 510 and VPA in combination. In separate experiments, treatment was
96-well plates at a density of 7.5 Â 103 per well. After 24 h, ABT-510 or VPA started after the animals developed larger tumors z380 mm3 in size and
was added to quadruplicate wells at various concentrations in medium continued for 10 days. Tumor volumes were measured twice weekly and
containing either 10% or 2% FBS. The lower concentration of serum was calculated using the following formula: tumor volume = (length  width2) /
used to ensure that effects of the drugs were not counteracted by excessive 2 (29). The Student’s t test was used to compare tumor size in the control
growth factors present in the serum. Following treatment for 24 to 72 h, and treatment groups. Mice were sacrificed after 20 or 10 days of treatment,
MTS reagent was added and cells were further incubated for 3 h. The respectively. Animals were treated according to NIH guidelines for animal
absorbance of the samples was measured using a Synergy HT Microplate care and use, and protocols were approved by the Animal Care and Use
Reader (Bio-Tek Instruments, Winooski, VT). Committee at Northwestern University.
Measurement of cell cycle phase distribution. Cell cycle distribution Quantification of apoptosis. In situ detection of cleaved, apoptotic
was determined by flow cytometric analysis. Briefly, SMS-KCNR cells were DNA fragments was done using the In Situ Cell Death Detection kit (Roche
cultured in RPMI 1640 containing various concentrations of ABT-510 or Diagnostics Corp., Indianapolis, IN) as described previously (30). Sections of
VPA for 24 h. Control cells were cultured in medium lacking ABT-510 or human colon mucosa processed similarly to the xenografts were included in
VPA. Cells were then washed with PBS, fixed in 70% ethanol, and each assay. To check the intra-assay and interassay consistency, the
hypotonically lysed in 1 mL of DNA staining solution [0.05 mg/mL apoptotic bodies in colonic mucosa, stained as the first and last slide of
propidium iodide (Sigma, St. Louis, MO) and 0.1% Triton X-100]. The cells each assay, were assessed. The assays were considered adequate only when
were incubated, while protected from light, at 4jC overnight before analysis. the frequency of apoptosis on the two sections of colonic mucosa was
www.aacrjournals.org 1717 Cancer Res 2007; 67: (4). February 15, 2007

3. Cancer Research
similar. Each terminal deoxynucleotidyl transferase–mediated dUTP nick endothelial cells. Primary antibodies were incubated in a humidity
end labeling (TUNEL) in situ–labeled section was quantified using a chamber overnight at 4jC. The staining for Ki-67 was developed using
Leica (Heidelberg, Germany) DM IRB inverted fluorescence microscope the DAKO Envision+ System-HRP (DakoCytomation) according to the
equipped with Image-Pro Plus version 4.5 software (Media Cybernetics, manufacturer’s instructions. The CD31 goat anti-mouse primary antibody
Silver Spring, MD) at a magnification of Â400. The frequency of labeled was bridged by biotinylated anti-goat IgG (1:200) and detected with
apoptotic cells was obtained by quantifying 10 consecutive fields starting streptavidin (1:400; Vector Laboratories, Burlingame, CA). Sections were
with areas with the highest number of TUNEL-labeled nuclei, avoiding areas counterstained with Gill’s hematoxylin. In addition, the primary antibodies
of necrosis, and expressed as apoptosis per 10 high-power fields (HPF). were omitted in each assay to ensure the specificity. Immunohistochem-
Immunohistochemical studies. Xenograft tissue was fixed in 10% istry for TSP-1 expression was done as described previously (10). Mean
formaldehyde/zinc fixative (Electron Microscopy Sciences, Hatfield, PA) vascular density (MVD) was quantified by measuring pixels in 10
and embedded in paraffin. Serial sections (4 Am thick) were deparaffinized consecutive fields at Â200 magnification.
and rehydrated as described previously (31). Sections of each tumor were Morphologic evaluation of blood vessels. Blood vessel morphology and
stained with H&E, and selected sections were stained with Masson’s tumor histology were evaluated by a pathologist (R.Z.) on H&E-stained
trichrome special staining. Additional sections were used for immunohis- sections. The presence or absence of hemorrhage was evaluated, and tumor
tochemistry. Antigen retrieval was done with 0.01 mol/L citrate buffer blood vessels were classified morphologically into four types. Vessels were
(pH 6.0) for Ki-67 or 1 mmol/L EDTA (pH 8.0) for CD31 antibody, heated counted, and the percentage of each vessel type in the tissue section was
in a boiling steamer for 20 min, and then cooled down to room calculated. Type 1 vessels had disorganized cellular proliferation with poorly
temperature for 20 min. For analysis of proliferation rate, slides were formed lumina and vascular endothelial proliferation (VEP). Type 2 vessels
incubated with Ki-67 (clone MIB-1, 1:200; DakoCytomation, Carpinteria, were structurally more normal, thin walled, and capillary like. Type 3
CA) mouse anti-human monoclonal antibody. CD31 (platelet/endothelial vessels were structurally similar to type 2 vessels but were embedded in
cell adhesion molecule 1, M-20, 1:100; Santa Cruz Biotechnology, Santa fibrocollagenous stroma. Type 4 vessels had concentric layers of cells within
Cruz, CA) polyclonal goat anti-mouse antibody was used to highlight a thickened wall often accompanied by fibrinoid necrosis and vasculitis.
Figure 1. Analysis of SMS-KCNR cell
apoptosis, necrosis, cell cycle, and
endothelial cell migration after treatment
with ABT-510 or VPA. A, flow cytometric
analysis of SMS-KCNR cell apoptosis and
necrosis after treatment with ABT-510 for
1 d. Top left quadrant, necrotic cells; top
right quadrant, late apoptotic cells; bottom
right quadrant, early apoptotic cells; bottom
left quadrant, live cells. B, sub-G1 apoptotic
analysis of SMS-KCNR cells treated with
VPA. C, cell cycle analysis of
control, ABT-510–treated, and VPA-treated
neuroblastoma cells. SMS-KCNR was
treated with either vehicle or various
concentrations of ABT-510 and VPA for
24 h and analyzed by flow cytometry.
D, ABT-510 and VPA inhibited
angiogenesis in vitro .
Cancer Res 2007; 67: (4). February 15, 2007 1718 www.aacrjournals.org

4. Valproic Acid and TSP-1 Inhibit Neuroblastoma Growth
Evan’s blue dye assay. The Evan’s blue dye assay was used to assess
tumor microvascular permeability as described previously (32). Evan’s blue
dye (2%, 100 AL; MP Biomedicals, Aurora, OH) was given via tail vein
injection of tumor-bearing mice and allowed to circulate for 20 min. To
remove remaining intravascular dye, mice were perfused with 25 mL saline
through the left ventricle with a right atrial vent. Tumors were excised, cut
into pieces, and weighed. Formamide was added at a ratio of 1 mL/100 mg
of tissue and let stand for 72 h at room temperature to facilitate the
extraction of Evan’s blue dye. Tissues were then removed and resultant
extract was centrifuged. The levels of Evan’s blue were quantified using a
spectrophotometer at a wavelength of 620 nm. The amount of Evan’s blue
dye remaining in the tumor reflects the degree of intratumoral vessel
permeability. All samples were run in duplicate and compared with those of
standards.
Statistical analysis. The two-tailed Student’s t test was used to compare
the statistical significance of differences between the mean values of
quantification of apoptosis, proliferation, CD31-positive vessels, and
permeability in the neuroblastoma xenografts comparing vehicle-treated
with ABT-510, VPA, and combination-treated groups. ANOVA followed by
Dunnett’s multiple comparison procedure was used to compare the
statistical significance of differences between control and treatment groups
for sub-G1 and cell cycle analysis.
Results
VPA, but not ABT-510, inhibits neuroblastoma cell prolifer-
ation, blocks cell cycle progression, and induces apoptosis. To
investigate the effects of VPA and ABT-510 as single agents on
neuroblastoma cell growth in vitro, SMS-KCNR neuroblastoma cells
were cultured in a range of drug concentrations. As shown in
Supplementary Fig. S1, treatment with ABT-510 did not inhibit the
proliferation of SMS-KCNR neuroblastoma cells even at concen-
trations as high as 1,000 nmol/L (Supplementary Fig. S1A and C).
In contrast, VPA significantly inhibited neuroblastoma cell proli-
feration at concentrations of 5 and 10 mmol/L of VPA when the
cells were cultured in medium containing either 10% or 2% serum
(P < 0.01, respectively; Supplementary Fig. S1B and D). Further-
more, 2.5, 5.0, and 10 mmol/L of VPA induced neuroblastoma cell Figure 2. Growth of neuroblastoma xenografts. SMS-KCNR (A ) or NMB (B )
apoptosis as shown by the increased population of sub-G1 cells mice were treated once the tumors were palpable with vehicle control, VPA
alone, ABT-510 peptide alone, or ABT-510 and VPA in combination.
(P = 0.0014, P = 0.0008, and P < 0.0001, respectively; Fig. 1B), and C, SMS-KCNR xenografts were treated once they grew to z380 mm3 in size with
cell cycle analysis revealed that treatment with VPA at concen- ABT-510, VPA, or combination therapy.
trations of 2.5, 5.0, and 10 mmol/L induced G2 growth arrest
(P = 0.0003, 0.0002, and 0.0004, respectively; Fig. 1C). Neither
apoptosis (Fig. 1A) nor changes in cell cycle (Fig. 1C) were seen antiangiogenic activity of VPA was due, at least in part, to up-
with ABT-510 (P > 0.01). regulation of endogenous inhibitors of angiogenesis in neuroblas-
VPA and ABT-510 inhibit bFGF-induced endothelial cell toma cells, we also evaluated the level of expression of TSP-1 and
migration. To test the antiangiogenic activity of VPA and ABT-510, SPARC by RT-PCR in the SMS-KCNR neuroblastoma cells following
endothelial cell migration assays were done using a range of drug treatment with VPA. We found that treatment with 1 mmol/L VPA
concentrations in the presence or absence of bFGF. bFGF-induced for 24 h resulted in up-regulation of the level of TSP-1 and SPARC
endothelial cell migration was inhibited by ABT-510 and VPA in a expression (Supplementary Fig. S2C and D).
dose-dependent manner at concentrations ranging from 0.1 to 100 VPA and ABT-510 inhibit neuroblastoma xenograft growth.
nmol/L and 10 nmol/L to 1 mmol/L, respectively (Fig. 1D). To examine the antitumor effects of VPA and ABT-510, nude
VPA down-regulates MYCN expression and induces TSP-1 mice with xenografts (f70 mm3 in size) established from two
and SPARC expression in neuroblastoma cells. Because MYCN different MYCN-amplified neuroblastoma cell lines were treated
plays a critical role in the regulation of neuroblastoma cell with each agent alone and in combination. Tumor growth was
proliferation and apoptosis, we examined the levels of MYCN inhibited with single-agent therapy, and enhanced effects were
mRNA expression in neuroblastoma cells following treatment with seen with combination therapy. After 20 days of treatment with
1 mmol/L VPA by quantitative real-time reverse transcription-PCR VPA combined with ABT-510, tumor volume was reduced by
(RT-PCR). As shown in Supplementary Fig. S2A and B, MYCN 91% (P < 0.001; Fig. 2A) in the SMS-KCNR model and by 87% in the
expression was significantly down-regulated in both SMS-KCNR NMB xenografts (P = 0.029; Fig. 2B) compared with controls.
and NMB cell lines within 2 h of treatment. In contrast, MYCN Because the effectiveness of some antiangiogenic strategies is
expression did not significantly change following treatment with inversely related to tumor burden, we repeated our studies in a set
100 nmol/L ABT-510 (data not shown). To determine if the of animals with large xenografts that were z380 mm3 in size. As
www.aacrjournals.org 1719 Cancer Res 2007; 67: (4). February 15, 2007

5. Cancer Research
shown in Fig. 2C, control tumor volumes increased 3.5-fold over the highest number of apoptotic cells was seen in the xenografts
the subsequent 10 days, and tumors in animals treated with ABT- treated with VPA alone or VPA combined with ABT-510 (93 F 13
510 or VPA, respectively, showed 33% and 44% reduced growth and 141 F 39, respectively, P < 0.001).
increases (P < 0.01 and 0.001). However, during this interval, VPA and ABT-510 inhibit neuroblastoma tumor cell prolif-
animals receiving combined treatment had total cessation of tumor eration in vivo. Cell proliferation in the xenografts was evaluated by
volume increase from the start of treatment through the entire Ki-67 expression, a nuclear protein that is preferentially expressed
10-day period. Thus, the combination of ABT-510 and VPA achieved during active phases of the cell cycle. Ki-67–negative cells represent
immediate tumor growth stasis (P < 0.001), whereas the individual quiescent G0 phase cells. A significantly higher percentage of
treatments gave only modest inhibition. neuroblasts in G0 were detected in tumors treated with VPA alone
Cell apoptosis is increased when VPA is combined with ABT- compared with control tumors (50.6 F 4.7 versus 34.2 F 9,
510. To evaluate whether treatment induced apoptosis, TUNEL respectively, P < 0.01; Fig. 3B). The percentage of neuroblasts in
assays were done on the neuroblastoma xenografts treated with G0 in the control tumors and ABT-510–treated tumors was not
ABT-510, VPA, or combination therapy. The tumors treated with significantly different (34.2 F 9 versus 42.4 F 5.3, P = 0.076;
ABT-510 contained significantly more apoptotic cells than the Fig. 3B). However, the percentage of G0 cells was significantly
control tumors (53 F 16 versus 34 F 9, P = 0.015; Fig. 3A). However, increased in the tumors treated with combination therapy compared
Figure 3. A, representative photographs
of TUNEL fluorescent apoptosis in situ
detection assay from tumors treated with
VPA alone, ABT-510 alone, or ABT-510 and
VPA in combination. Magnification, Â400.
Mean apoptosis counts per 10 HPFs in
treated tumors versus controls. B, MIB-1–
immunostained (Ki-67) sections. Proliferating
cells were identified by MIB-1 antibody in
control versus ABT-510, VPA, and
combination-treated groups. Percentage of
G0 (Ki-67–negative cells) per 10 HPFs of the
control and treated tumors. C, MVD in
SMS-KCNR xenografts after treatment with
ABT-510, VPA, or ABT-510 and VPA in
combination. Columns, mean MVD
measured in SMS-KCNR xenografts
following treatment with vehicle, ABT-510,
VPA, or combination therapy; bars, SD.
P < 0.001, vehicle versus VPA; P < 0.001,
vehicle versus ABT-510; P < 0.001, vehicle
versus combination.
Cancer Res 2007; 67: (4). February 15, 2007 1720 www.aacrjournals.org

6. Valproic Acid and TSP-1 Inhibit Neuroblastoma Growth
P = 0.00024; Fig. 3C). The MVD was significantly lower in the
tumors treated with combination therapy (247,805 F 73,062 versus
31,406 F 18,401; Fig. 3C) than in either single-agent group (P <
0.001). We also evaluated TSP-1 expression in the control and
treated SMS-KCNR tumors using immunohistochemistry to deter-
mine if the treatment with either agent alone or in combination
was capable of inducing expression of this angiogenic inhibitor. No
significant up-regulation of TSP-1 was observed in treated tumors.
VPA and ABT-510 treatment decreases hemorrhage and
modifies tumor blood vessel structure. To investigate the effects
of VPA and ABT-510 on tumor growth, treatment was initiated in
animals with tumors measuring f70 mm3. Mice were sacrificed
after 20 days of treatment, and the tumors were histologically
examined. The control NMB neuroblastoma xenografts were grossly
hemorrhagic, and extravasated RBCs were seen microscopically. The
tumor blood vessels were morphologically classified into four types
as described in Materials and Methods. More than half of the vessels
in the control neuroblastoma xenografts were structurally abnormal
with VEP (type 1 vessels; 58 F 9.04%; Fig. 4, Control, left and right;
Table 1A). The remaining vessels were thin walled and capillary like
Figure 4. Representative sections from H&E-stained and Masson’s (type 2; Fig. 4, Control, middle; Table 1A). To investigate if the
trichrome–stained control and treated NMB xenografts. Hemorrhage is moderate structurally abnormal vessels were related to the large size of the
in control tumors, mild in ABT-510–treated group, minimal in VPA-treated group,
and absent following combination treatment. Control tumors show predominantly
control tumors, we repeated these experiments and sacrificed mice
VEP (type 1 vessels; left and right ) and a minority of thin-walled vessels when the tumors were similar in size to the treated tumors (ranging
(type 2; middle ). In contrast, tumors from treatment groups show predominantly from 220 to 726 mm3). Similar to the large tumors, significant
thin-walled vessels that are either capillary-like or arteriolar with hyaline
arteriolosclerosis (middle ) or embedded in fibrocollagenous stroma (type 3 hemorrhage was present in the smaller control tumors, and the
vessels; left and right ). Magnification, Â600. majority of vessels were structurally abnormal and with VEP.
Tumor hemorrhage was decreased in experiments in which
treatment with ABT-510 was started when tumors were f70 mm3
with control tumors or tumors treated with either single agent in size. Significantly fewer type 1 blood vessels were observed
(86.4 F 9.1 versus 34.2 F 9, P < 0.01; 86.4 F 9.1 versus 42.4 F 5.3, (7.85 F 5.35%; P = 0.0028) in these tumors, and the majority of the
P < 0.01; 86.4 F 9.1 versus 50.6 F 4.7, P < 0.01; Fig. 3B). vessels were thin walled (type 2 and 3 vessels; Fig. 4, ABT-510, right,
Treatment with ABT-510 and VPA significantly inhibits middle, and left; Table 1A). Virtually no hemorrhage was seen in
neuroblastoma tumor angiogenesis. To investigate the anti- tumors treated with VPA, and the percentage of blood vessels with
angiogenic effects of ABT-510 and VPA, the vascularity of tumors VEP (type 1 vessels) decreased to 12.69 F 3.5% (P = 0.0039). The
treated early with ABT-510, VPA, or ABT-510 and VPA in majority of vessels in these tumors were thin walled and many were
combination was evaluated. Histologic sections were stained with capillary like (Fig. 4, VPA, right, middle, and left; Table 1A). Some of
CD31 to highlight endothelial cells, and a MVD was quantified. the vessels were embedded in fibrocollagenous stroma, highlighted
Compared with control tumors, a significantly lower MVD was seen by Masson’s trichrome special stain (type 3 vessels), and others
in xenografts following treatment with VPA or ABT-510 (247,805 F were arteriolar with mild hyaline sclerosis (Table 1A). In tumors
73,062 versus 114,597 F 47,568 versus 75,880 F 31,638, respectively, treated with VPA combined with ABT-510, hemorrhage was not
Table 1. Effects of ABT-510 and VPA treatment on neuroblastoma tumor blood vessel morphology
Vehicle (% F SE) ABT-510 (% F SE) VPA (% F SE) ABT-510 + VPA (% F SE)
A. NMB xenografts
Treatment started when tumors were <70 mm3
VEP (type 1) 58.00 F 9.04 7.85 F 5.35 12.69 F 3.53 3.55 F 1.12
Thin-walled vessels 42.00 F 9.04 92.15 F 5.35 87.31 F 5.53 96.45 F 1.12
Thin or capillary like (type 2) 69.05 F 9.91 62.04 F 7.11 68.55 F 12.12 59.95 F 9.83
Embedded in fibrocollagenous stroma (type 3) 30.95 F 9.91 37.96 F 7.11 31.45 F 12.12 40.05 F 9.83
Hyperplastic arteriolosclerosis (type 4) 0 0 0 0
B. SMS-KCNR xenografts
Treatment started when tumors were z380 mm3
VEP (type 1) 100 (network) 14.68 F 3.83 2.09 F 0.72 0
Thin-walled vessels 0 85.32 F 3.85 97.91 F 0.72 88.14 F 10.58
Thin or capillary like (type 2) 0 19.95 F 4.50 21.33 F 4.96 32.23 F 9.89
Embedded in fibrocollagenous stroma (type 3) 0 80.05 F 4.50 78.67 F 4.96 67.77 F 9.89
Hyperplastic arteriolosclerosis (type 4) 0 0 0 11.86 F 10.58
www.aacrjournals.org 1721 Cancer Res 2007; 67: (4). February 15, 2007

7. Cancer Research
VPA-treated tumors, hemorrhage was minimal (Fig. 5, VPA, left and
middle) and the number of the type 1 vessels was only 2.09 F 0.72%
(P < 0.0001; Table 1B). Almost all of the vessels in the VPA-treated
tumors were thin walled (Fig. 5, VPA, middle and right; Table 1B).
Some vessels were capillary like (Fig. 5, VPA, middle), some
had mild hyaline sclerosis (Fig. 5, VPA, middle), and others
were embedded in fibrocollagenous stroma (Fig. 5, VPA, right; Table
1B). Minimal hemorrhage was seen in the tumors treated with both
VPA and ABT-510 (Fig. 5, ABT-510 + VPA, left and middle), and VEP
was absent. The majority of vessels in these tumors were thin
walled and classified as type 2 or 3 vessels (Fig. 5, ABT-510 + VPA,
middle; Table 1B). In three of the seven tumors treated with
combination therapy, there were small numbers of vessels with
thickened walls exhibiting concentric lamellar cellular arrangement
(‘onionskin’ pattern) with marked luminal narrowing, focal fibrosis,
fibrinoid necrosis, and vasculitis (type 4 vessels, hyperplastic
arteriolosclerosis; Fig. 5, ABT-510 + VPA, right; Table 1B).
VPA and ABT-510 treatment decreases vascular permeability.
Tumor blood vessels are functionally abnormal and leaky. To
determine if treatment with VPA and ABT-510 modified vessel
function, we assessed intratumoral vessel permeability using the
Evan’s blue dye assay. In this assay, Evan’s blue dye is given
systemically, mice are then perfused with saline to wash out the
intravascular dye, and the amount of Evan’s blue dye remaining in
the tumor correlates to permeability of the microvasculature. Our
results indicate a decrease in intratumoral vessel permeability in all
treatment groups (Fig. 6). The decrease was modest in the tumors
treated with ABT-510 alone compared with controls and did not
reach statistical significance (P = 0.124; Fig. 6). However, the
reduction in vascular permeability was statistically significant in
tumors treated with either VPA alone or in combination with ABT-
510 compared with controls (P = 0.044 and 0.007, respectively; Fig. 6).
Figure 5. Representative H&E-stained sections of SMS-KCNR xenografts
treated after reaching z380 mm3 in size. Severe hemorrhage is seen in the
control tumors (left and middle ). Hemorrhage is moderate in the tumors treated Discussion
with ABT-510 (left and middle ), minimal in VPA-treated tumors (left and middle ),
and absent following combination treatment (left and middle ). Control tumors More than 30 years ago, Folkman (33) proposed the concept
show extensive VEP (type 1 vessels; middle and right ). Tumors treated with that blocking angiogenesis would result in tumor dormancy.
ABT-510 show a few type 1 vessels (right ) and a majority of thin-walled type 2
and 3 vessels (middle ). In the VPA-treated tumors, most vessels are thin walled,
Clinical trials in adults with cancer have supported this
including capillary like and arteriolar (type 2 vessels; middle ) as well as those hypothesis, and the antiangiogenic agent bevacizumab (Avastin)
embedded in fibrocollagenous stroma (type 3 vessels; right ). Tumors from the
ABT-510 + VPA combination-treated group show thin-walled vessels surrounded
by fibrosis (middle). A subset of these tumors showed small vessel changes with
features of hyperplastic arteriolosclerosis (type 4 vessels; right ). Arrows,
vessels.
seen and the proportion of vessels with VEP was further reduced to
3.55 F 1.12% (P = 0.0007; Table 1A). The vessels in the tumors
treated with combination therapy were thin walled and embedded
in fibrocollagenous stroma (type 3; Fig. 4, ABT-510 + VPA, left and
right) or capillary like (type 2; Fig. 4, ABT-510 + VPA, middle).
We also evaluated the effects of ABT-510 and VPA on the growth
of large established neuroblastoma tumors. Similar to the previous
experiment, significant hemorrhage was seen in the control SMS-
KCNR tumors that were z380 mm3 in size when treatment was
started (Fig. 5, Control, left and middle). These tumors had a
continuous network of structurally abnormal type 1 vessels (Fig. 5,
Control, right, middle, and left; Table 1B). Following treatment with
ABT-510, the tumors were less hemorrhagic than the controls
(Fig. 5, ABT-510, left and middle; Table 1B) and type 1 vessels Figure 6. Evan’s blue dye assay for intratumoral vessel permeability of the
constituted only 14.68 F 3.83% (P < 0.0001) of the total (Fig. 5, ABT- SMS-KCNR xenografts. Quantitative analysis shows reduction of Evan’s blue
extravasation in mice treated with ABT-510, VPA, and ABT-510 and VPA
510, right). The majority of vessels were thin walled, and some were in combination compared with controls. RQ, relative quantification by
embedded in fibrocollagenous stroma (Fig. 5, ABT-510, middle). In spectrophotometry.
Cancer Res 2007; 67: (4). February 15, 2007 1722 www.aacrjournals.org

8. Valproic Acid and TSP-1 Inhibit Neuroblastoma Growth
was recently granted Food and Drug Administration approval for block angiogenesis and that TSA specifically inhibits hypoxia-
use in metastatic colon cancer (34). ABT-510, HDAC inhibitors, induced angiogenesis in the Lewis lung carcinoma model.
and other promising antiangiogenic agents, either alone or in In the experiments done on both small and large neuroblastoma
combination with cytotoxic agents, are currently being tested in xenografts, the combination of ABT-510 and VPA immediately and
adult clinical trials. Although we and others have reported that fully stabilized the size of the tumors for 20 and 10 days,
the growth of common pediatric cancers, such as neuroblastoma respectively, in the NMB (early treatment) and SMS-KCNR (late
and Wilms’ tumor, is also angiogenesis dependent (4, 35), to treatment) models during which time vehicle controls substantially
date, the experience with antiangiogenic agents in childhood increased in size. The MVD was significantly lower in xenografts
malignancies is limited. In this study, we evaluated the effects of treated with combination therapy compared with tumors treated
ABT-510 and VPA on neuroblastoma angiogenesis and tumor with single agents or controls. Furthermore, ABT-510 combined
growth. As expected, both agents blocked bFGF-induced with VPA induced significant apoptosis and inhibited neuroblas-
endothelial cell migration in vitro. In addition to inhibiting toma cell proliferation. Previous studies done in our laboratory
angiogenesis, VPA also inhibited neuroblastoma cell proliferation with TNP-470 showed that this antiangiogenic agent was capable of
and induced neuroblastoma apoptosis. Furthermore, MYCN inhibiting the growth of small neuroblastoma xenografts but did
expression levels were significantly down-regulated by VPA, not affect the rate of growth of large neuroblastoma tumors (5). In
suggesting that this agent is capable of affecting the biology of contrast, we found that the combination of ABT-510 and VPA was
neuroblastoma cells. Similar changes in MYCN expression occur also capable of stabilizing the growth of large xenografts.
in neuroblastoma following treatment with retinoic acid, a drug Hemorrhage was reduced in the neuroblastoma xenografts
that is of clinical benefit in children with high-risk neuroblas- treated with single-agent therapy compared with control tumors,
toma (36). We also found that VPA induced the expression of and virtually no hemorrhage was seen in the tumors treated with
endogenous inhibitors of angiogenesis, including TSP-1 and combination therapy. In addition, striking changes in blood vessel
SPARC. Previous work from our laboratory has shown that morphology were seen following treatment with ABT-510 and
SPARC is a potent inhibitor of angiogenesis in neuroblastoma VPA alone or in combination. Control tumors contained large
(27). In contrast, ABT-510 had no effect on the neuroblastoma numbers of structurally abnormal blood vessels with VEP. VEP is an
cell growth in vitro. As single agents, both ABT-510 and VPA established unfavorable factor in several cancers, including
inhibited neuroblastoma tumor growth in vivo, and enhanced melanoma, breast, prostate carcinomas, and glioblastoma (43, 44),
antitumor effects were seen with combination therapy. and it is one of the criteria for raising the WHO grade of astrocytic
ABT-510 has been extensively studied in cancers that are neoplasms (45). The number of vessels with VEP was significantly
common in adults (15), but to our knowledge, this is the first decreased in the treated tumors, whereas the proportion of thin-
report showing that this agent has antiangiogenic activity in the walled, capillary-like vessels was increased. Some of the thin-walled
pediatric cancer neuroblastoma. TSP-1 and ABT-510 act by inducing vessels in the treated tumors were embedded in fibrocollagenous
endothelial cell apoptosis, and recent studies suggest that TSP-1 stroma, whereas others exhibited hyaline sclerosis, a feature that is
generates CD95L, a ligand for the CD95 death receptor (37). usually seen in hypertensive disease and attributed to endothelial
However, because CD95 expression on the vascular endothelial cell cell injury (46–48). In the large SMS-KCNR tumors treated with
is independent of TSP-1, the efficacy of the drug is limited. In an combination therapy, rare vessels with hyperplastic arterioloscle-
effort to develop a more effective antiangiogenic treatment strategy, rosis were also seen. Using the Evan’s blue dye assay, we were also
ABT-510 has been combined with other agents in preclinical models able to show that treatment with ABT-510 and VPA resulted in
of adult cancer (38). We selected VPA for our combination studies modification of vessel function with decreased permeability of the
for several reasons. First, VPA is a potent inhibitor of angiogenesis. tumor vasculature.
Although its mechanism of action remains unclear, VPA has been Our results indicate that, in addition to reducing the number of
shown to relieve HDAC-dependent transcription repression of blood vessels, the combination of ABT-510 and VPA modifies the
several antiangiogenic proteins (39). Second, VPA directly affects structure and function of the vasculature in neuroblastoma tumors.
the growth of neuroblastoma cells by inhibiting cell cycle Transient ‘‘normalization’’ of tumor vasculature following treatment
progression and inducing differentiation (16). Third, VPA has been with other antiangiogenic agents has recently been reported (25).
shown to have synergistic effects with other drugs that regulate These changes create a vasculature that is more efficient for oxygen
angiogenesis, including IFN-a and retinoic acid (17, 18). Finally, VPA and drug delivery and may thereby lead to chemosensitization (34).
has limited toxicities even during long-term treatment. Thus, in addition to inhibiting neuroblastoma tumor growth directly,
Other HDAC inhibitors have also been shown to inhibit it is possible that ABT-510 and VPA may also enhance the antineu-
neuroblastoma tumor growth. The hybrid polar HDAC inhibitor roblastoma activity of cytotoxic agents. Additional studies evaluat-
m-carboxycinnamic acid bis-hydroxamide (CBHA) induces apo- ing the clinical efficacy of this antiangiogenic treatment strategy
ptosis in neuroblastoma cells in vitro and suppresses the growth in children with high-risk neuroblastoma are warranted.
of neuroblastoma xenografts. The combination of CBHA and all-
trans retinoic acid led to an even greater inhibition of tumor Acknowledgments
growth (40). Jaboin et al. (41) reported that the synthetic
Received 7/14/2006; revised 11/14/2006; accepted 12/14/2006.
benzamide derivative MS-27-275 (MS-275) inhibited [3H]thymidine Grant support: NIH/National Institute of Neurological Disorders and Stroke grant
uptake in a series of pediatric tumor cell lines and suppressed the NS049814; Neuroblastoma Children’s Cancer Society; Friends for Steven Pediatric
Cancer Research Fund; Elise Anderson Neuroblastoma Research Fund; Neuroblastoma
growth of undifferentiated sarcoma, Ewing’s sarcoma, and Kids Fund; Abbott Laboratories; and Robert H. Lurie Comprehensive Cancer Center,
neuroblastoma in vivo. These investigators reported a marked NIH, National Cancer Institute Core grant 5P30CA60553.
decrease in the vascularization of neuroblastoma xenografts The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
following treatment with MS-275. Others (42) have indicated that with 18 U.S.C. Section 1734 solely to indicate this fact.
trichostatin A (TSA) and suberoylanilide hydroxamic acid can We thank Dr. Alfred W. Rademaker for the statistical analyses.
www.aacrjournals.org 1723 Cancer Res 2007; 67: (4). February 15, 2007

9. Cancer Research
References the thrombospondin-1-mimetic angiogenesis inhibitor 33. Folkman J. Tumor angiogenesis: therapeutic implica-
ABT-510 in patients with advanced cancer. J Clin Oncol tions. N Engl J Med 1971;285:1182–6.
1. Cheung NK, Cohn SL, editors. Neuroblastoma. Heidel- 2005;23:5188–97. 34. Kerbel RS. Antiangiogenic therapy: a universal chemo-
berg: Springer-Verlag; 2005. 16. Cinatl J, Cinatl J, Driever PH, et al. Sodium valproate sensitization strategy for cancer? Science 2006;312:1171–5.
2. Brodeur GM. Neuroblastoma: biological insights into a inhibits in vivo growth of human neuroblastoma cells. 35. Abramson LP, Grundy PE, Rademaker AW, et al.
clinical enigma. Nat Rev Cancer 2003;3:203–16. Anticancer Drugs 1997;8:958–63. Increased microvascular density predicts relapse in
3. Eggert A, Ikegaki N, Kwiatkowski J, Zhao H, Brodeur 17. Cinatl J, Kotchetkov R, Blaheta R, Driever PH, Vogel Wilms’ tumor. J Pediatr Surg 2003;38:325–9.
GM, Himelstein BP. High-level expression of angiogenic JU, Cinatl J. Induction of differentiation and suppression 36. Matthay KK, Villablanca JG, Seeger RC, et al.
factors is associated with advanced tumor stage in of malignant phenotype of human neuroblastoma Treatment of high-risk neuroblastoma with intensive
human neuroblastomas. Clin Cancer Res 2000;6:1900–8. BE(2)-C cells by valproic acid: enhancement by combi- chemotherapy, radiotherapy, autologous bone marrow
4. Meitar D, Crawford SE, Rademaker AW, Cohn SL. nation with interferon-a. Int J Oncol 2002;20:97–106. transplantation, and 13-cis -retinoic acid. N Engl J Med
Tumor angiogenesis correlates with metastatic disease, 18. Mongan NP, Gudas LJ. Valproic acid, in combination 1999;341:1165–73.
N-myc amplification, and poor outcome in human with all-trans retinoic acid and 5-aza-2¶-deoxycytidine, 37. Quesada AJ, Nelius T, Yap R, et al. In vivo upregulation
neuroblastoma. J Clin Oncol 1996;14:405–14. restores expression of silenced RARh2 in breast cancer of CD95 and CD95L causes synergistic inhibition of
5. Katzenstein HM, Rademaker AW, Senger C, Salwen H, cells. Mol Cancer Ther 2005;4:477–86. angiogenesis by TSP1 peptide and metronomic doxoru-
Nguyen N, Cohn SL. Effectiveness of the angiogenesis 19. Bug G, Ritter M, Wassmann B, et al. Clinical trial of bicin treatment. Cell Death Differ 2005;12:649–58.
inhibitor TNP-470 in reducing the growth of human valproic acid and all-trans retinoic acid in patients with 38. Yap R, Veliceasa D, Emmenegger U, et al. Metronomic
neuroblastoma in nude mice inversely correlates with poor-risk acute myeloid leukemia. Cancer 2005;104: low-dose chemotherapy boosts CD95-dependent anti-
tumor burden. Clin Cancer Res 1999;5:4273–8. 2717–25. angiogenic effect of the thrombospondin peptide ABT-
6. Shusterman S, Grupp SA, Barr R, Carpentieri D, Zhao 20. Tang R, Faussat AM, Majdak P, et al. Valproic acid 510: a complementation antiangiogenic strategy. Clin
H, Maris JM. The angiogenesis inhibitor TNP-470 inhibits proliferation and induces apoptosis in acute Cancer Res 2005;11:6678–85.
effectively inhibits human neuroblastoma xenograft myeloid leukemia cells expressing P-gp and MRP1. 39. Blaheta RA, Michaelis M, Driever PH, Cinatl J.
growth, especially in the setting of subclinical disease. Leukemia 2004;18:1246–51. Evolving anticancer drug valproic acid: insights into
Clin Cancer Res 2001;7:977–84. 21. Catalano MG, Fortunati N, Pugliese M, et al. Valproic the mechanism and clinical studies. Med Res Rev 2005;
7. Bornstein P. Diversity of function is inherent in acid induces apoptosis and cell cycle arrest in poorly 25:383–97.
matricellular proteins: an appraisal of thrombospondin differentiated thyroid cancer cells. J Clin Endocrinol 40. Coffey DC, Kutko MC, Glick RD, et al. The histone
1. J Cell Biol 1995;130:503–6. Metab 2005;90:1383–9. deacetylase inhibitor, CBHA, inhibits growth of human
8. Grossfeld GD, Ginsberg DA, Stein JP, et al. Thrombo- 22. Takai N, Desmond JC, Kumagai T, et al. Histone dea- neuroblastoma xenografts in vivo, alone and synergistically
spondin-1 expression in bladder cancer: association cetylase inhibitors have a profound antigrowth activity in with all-trans retinoic acid. Cancer Res 2001;61:3591–4.
with p53 alterations, tumor angiogenesis, and tumor endometrial cancer cells. Clin Cancer Res 2004;10:1141–9. 41. Jaboin J, Wild J, Hamidi H, et al. MS-27-275, an
progression. J Natl Cancer Inst 1997;89:219–27. 23. Michaelis M, Michaelis UR, Fleming I, et al. Valproic inhibitor of histone deacetylase, has marked in vitro and
9. Zabrenetzky V, Harris CC, Steeg PS, Roberts DD. acid inhibits angiogenesis in vitro and in vivo . Mol in vivo antitumor activity against pediatric solid tumors.
Expression of the extracellular matrix molecule throm- Pharmacol 2004;65:520–7. Cancer Res 2002;62:6108–15.
bospondin inversely correlates with malignant progres- 24. Adams GP, Weiner LM. Monoclonal antibody therapy 42. Deroanne CF, Bonjean K, Servotte S, et al. Histone
sion in melanoma, lung, and breast carcinoma cell lines. of cancer. Nat Biotechnol 2005;23:1147–57. deacetylases inhibitors as anti-angiogenic agents alter-
Int J Cancer 1994;59:191–5. 25. Jain RK. Normalization of tumor vasculature: an ing vascular endothelial growth factor signaling. Onco-
10. Yang QW, Liu S, Tian Y, et al. Methylation-associated emerging concept in antiangiogenic therapy. Science gene 2002;21:427–36.
silencing of the thrombospondin-1 gene in human 2005;307:58–62. 43. Cavenee WK, Furnari FB, Nagane M, et al. Diffusely
26. Huang D, Rutkowski JL, Brodeur GM, et al. Schwann infiltrating astrocytomas. In: Kleihues P, Cavenee WK,
neuroblastoma. Cancer Res 2003;63:6299–310.
cell-conditioned medium inhibits angiogenesis. Cancer editors. Pathology and genetics of tumors of the nervous
11. Dawson DW, Volpert OV, Pearce SFA, et al. Three
Res 2000;60:5966–71. system. WHO classification of tumors. Lyon (France):
distinct D-amino acid substitutions confer potent
27. Chlenski A, Liu S, Crawford SE, et al. SPARC is a key IARC Press; 2000. p. 9–54.
antiangiogenic activity on an inactive peptide derived
Schwannian-derived inhibitor controlling neuroblasto- 44. Birner P, Piribauer M, Fischer I, et al. Vascular
from a thrombospondin-1 type 1 repeat. Mol Pharmacol patterns in glioblastoma influence clinical outcome and
ma tumor angiogenesis. Cancer Res 2002;62:7357–63.
1999;55:332–8. 28. Manohar CF, Short ML, Nguyen A, et al. HuD, a neuronal- associate with variable expression of angiogenic pro-
12. Huang HH, Campbell SC, Bedford DF, et al.
specific RNA-binding protein, increases the in vivo stability teins: evidence for distinct angiogenic subtypes. Brain
Peroxisome proliferator-activated receptor g ligands of MYCN RNA. J Biol Chem 2002;277:1967–73. Pathol 2003;13:133–43.
improve the antitumor efficacy of thrombospondin 29. Chen SY, Yang AG, Chen JD, et al. Potent antitumour 45. Straume O, Chappuis PO, Salvesen HB, et al.
peptide ABT510. Mol Cancer Res 2004;2:541–50. activity of a new class of tumour-specific killer cells. Prognostic importance of glomeruloid microvascular
13. Haviv F, Bradley MF, Kalvin DM, et al. Thrombo- Nature 1997;385:78–80. proliferation indicates an aggressive angiogenic pheno-
spondin-1 mimetic peptide inhibitors of angiogenesis 30. Liu SQ, Tian YF, Chlenski A, et al. Cross-talk type in human cancers. Cancer Res 2002;62:6808–11.
and tumor growth: design, synthesis, and optimization between Schwann cells and neuroblasts influences the 46. Mandal AK, Frohlich ED, Nordquist J. An ultrastruc-
of pharmacokinetics and biological activities. J Med biology of neuroblastoma xenografts. Am J Pathol 2005; tural analysis of renal arterioles in benign and malignant
Chem 2005;48:2838–46. 166:891–900. essential hypertensions. Ann Clin Lab Sci 1977;7:158–68.
14. Shaked Y, Bertolini F, Man S, et al. Genetic 31. Yang QW, Liu S, Tian Y, et al. Methylation-associated 47. Kashgarian M. Pathology of small blood vessel disease
heterogeneity of the vasculogenic phenotype parallels silencing of the heat shock protein 47 gene in human in hypertension. Am J Kidney Dis 1985;5:A104–10.
angiogenesis: implications for cellular surrogate marker neuroblastoma. Cancer Res 2004;64:4531–8. 48. Spector S, Ooshima A, Iwatsuki K, Fuller G, Cardinale
analysis of antiangiogenesis. Cancer Cell 2005;7:101–11. 32. Crone C. Permeability of capillaries in various organs G, Udenfriend S. Increased vascular collagen biosynthe-
15. Hoekstra R, De Vos FYFL, Eskens FALM, et al. Phase I as determined by use of indicator diffusion method. sis by hypertension and reversal by antihypertensive
safety, pharmacokinetic, and pharmacodynamic study of Acta Physiol Scand 1963;58:292–305. drugs. Blood Vessels 1978;15:176–82.
Cancer Res 2007; 67: (4). February 15, 2007 1724 www.aacrjournals.org