Local irradiation in combination with bevacizumab enhances radiation control of
bone destruction and cancer-induced pain in a model of bone metastases
Pawel Zwolak1, Arkadiusz Z. Dudek1*, Vidya D. Bodempudi1, Julia Nguyen1, Robert P. Hebbel1, Nathan J. Gallus2,
Marna E. Ericson2, Michael J. Goblirsch3and Denis R. Clohisy3
1Department of Hematology, Oncology and Transplantation, University of Minnesota, Minneapolis, MN
2Department of Dermatology, University of Minnesota, Minneapolis, MN
3Department of Orthopedic Surgery, University of Minnesota, Minneapolis, MN
Skeletal metastases are a major source of morbidity for cancer
patients. The purpose of this study was to evaluate the effects of
megavoltage irradiation and antiangiogenic therapy on metastatic
bone cancer. A tumor xenograft model was prepared in C3H/Scid
mice using 4T1 murine breast carcinoma cells. Twenty-eight mice
bearing tumors were treated with either bevacizumab (15 mg/kg),
local megavoltage irradiation (30 Gy in 1 fraction), combination
of bevacizumab and local megavoltage irradiation or physiologic
saline solution (control group). Tumor area, bone destruction, tu-
mor microvessel density, pain-associated behaviors and expression
of substance P were assessed. Combined modality treatment
reduced the frequency of pain-associated behaviors, decreased lev-
els of nociceptive protein expression in the spinal cord, maintained
cortical integrity and decreased the density of microvessels as
compared to single modality treatments. We conclude that con-
current antiangiogenic therapy and localized radiotherapy for the
treatment of bone metastases warrants further evaluation in
human clinical trials.
' 2007 Wiley-Liss, Inc.
Key words: bevacizumab; VEGF; local irradiation; metastatic bone
cancer; pain; osteolysis
Bone is a common site of embedment for metastatic breast,
prostate, lung and kidney cancers.1,2Bone metastases usually
result in osteolysis with major sources of morbidity for patients
including bone pain, pathologic fractures, hypercalcemia and
Single modality therapies for bone metastases are often ineffec-
tive at reducing both tumor mass and pain. Radiation therapy for
this frequently painful disease is able to obtain a 15–18% com-
plete response rate with only 33% of patients able to stop narcotic
pain medications at 3 months.3Angiogenic inhibitors have
emerged as a novel approach for augmenting radiation therapy.
Antiangiogenic agents have been shown to inhibit the growth of a
wide variety of solid tumors in animals by reducing tumor blood
supply,4–6and angiogenesis has been shown to promote the
growth of breast cancer metastases.7There had been concern that
antiangiogenic therapy might counteract the effect of radiation
therapy based on evidence that hypoxic tumors, especially in an
acute hypoxic state, are less likely to respond to radiation ther-
apy.8However, Gorski et al. have reported that antiangiogenic
therapy increases the effect of radiation therapy,9and Tsai et al.
have reported that radiation therapy disrupts angiogenesis by dam-
aging endothelial cells.10Further investigations have found that
blockade of vascular endothelial growth factor (VEGF) induces
transient normalization of tumor vasculature by interfering with
abnormal angiogenesis and, at least temporarily, alleviating tumor
hypoxia.11Therefore, the strategy of combining radiation and anti-
angiogenic therapies appears to be an attractive option.
VEGF levels in tumors are regulated by numerous factors such
as metabolic stress, hypoxia, mechanical stress, cytokines and hor-
mones.12,13It has been proposed that VEGF promotes tumor sur-
vival by increasing tumor vascularization.9,14,15In addition to its
role in angiogenesis, activation of VEGF receptors on osteoclast
precursor cells initiates signaling for osteoclast differentiation,
proliferation and lytic destruction of bone.16–18Therefore, inhibi-
ting VEGF might have the 2-fold effect of preventing bone loss
and limiting tumor growth. In turn, both effects would presumably
reduce tumor-induced bone pain, but the literature is inconclusive
on whether inhibiting angiogenic and osteoclastogenic com-
pounds, such as VEGF, would reduce pain.
Bevacizumab is a monoclonal murine antibody against human
VEGF. It improves survival of patients with metastatic colorectal
cancer and lung cancer when combined with chemotherapy.19,20
In this study, we examined the effects of megavoltage radiation in
combination with bevacizumab on tumor growth, bone resorption
and bone pain in an orthotopic experimental model for bone me-
tastasis generated by 4T1 murine breast cancer cells. The 4T1
model closely represents the local bone tumor conditions in natu-
rally arising metastatic breast cancer.21
Material and methods
Experiments were performed on 6-week-old female C3H/Scid
mice (Jackson Laboratories, Bar Harbor, ME), weighing 18–22 g.
The selection of this murine strain was dictated by our previous
experience with a pain model in bone cancer.22Other murine
strains do not exhibit reproducible behavioral pain assessments in
our experience. In accordance with the National Institutes of
Health guidelines, mice were housed under pathogen-free condi-
tions with a 12-hr alternating light-and-dark cycle and fed auto-
claved food and water ad libitum. All procedures were approved
by the Animal Care and Use Committee at the University of
Secretion of VEGF protein from irradiated 4T1 cells
4T1 breast cancer cells at a density of 50,000 cells per well in
6-well plates (BD, Bedford, MA) were grown to 20% confluence,
and then treated with a single dose of local megavoltage irradia-
tion at 0 and 25 Gy. Cells were then incubated up to 96 hr. The
secretion of VEGF into cultured medium was measured by ELISA
(R&D Systems, Minneapolis, MN) at 24, 48, 72 and 96 hr. VEGF
levels were normalized to the number of cells present in wells,
and calculated after collection of conditioned media.
Bevacizumab binding to murine VEGF
To measure bevacizumab binding to VEGF, supernatant from
4T1 tumor cells (0.5 mg protein/ml) was incubated with and with-
out bevacizumab antibody (0.5 lg/ml) overnight in a rotator at
4?C. Then, 50 ll of proteinA-sepharose (Amersham, Piscataway,
NJ) was added to the samples, and the rotation was continued for
1 hr. The sepharose beads were collected by spinning, and the
*Correspondence to: Department of Hematology, Oncology and Trans-
plantation, University of Minnesota, Mayo Mail Code 480, 420 Delaware
Street SE, Minneapolis, MN 55455, USA. Fax: 1612-625-6919.
Received 26 April 2007; Accepted after revision 20 August 2007
Publishedonline 12October 2007inWileyInterScience (www.interscience.
Int. J. Cancer: 122, 681–688 (2008)
' 2007 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
remaining supernatant was used for quantification of VEGF levels
using ELISA (R&D Systems).
Quantitative endothelial cell invasion assay
The membrane endothelial cell invasion assay was performed
in Matrigel-coated invasion chambers (BD Biosciences, Bedford,
MA). Briefly, resuspended 4T1 cells (5 3 104) were plated in the
lower chamber without bevacizumab (control) or with bevacizu-
mab in increasing concentrations (5–15 lg/ml) and incubated at
37?C in humidified 5% CO2atmosphere. Two triplicates with mu-
rine VEGF (164) (R&D Systems)—which included 1 triplicate
with and 1 without the addition of bevacizumab—were added to
the experiment as positive controls. Forty-eight hours after incuba-
tion, murine endothelial cells were stained with carboxyfluores-
cein diacetate and succinimidyl ester (Vybrant CFDA SE Cell
Tracer Kit; Molecular Probes, Eugene, OR) according to the kit
manual and then added into the upper chamber (3 3 103cells).
Two hours after incubation a fluorescence plate reader (SynergyTM
HT Multi-Detection Microplate Reader; Bio-Tek Instruments,
Winooski, VT) was used to quantify the number of endothelial
cells invading the membrane by measuring the amount of fluores-
cence at 485 nm excitation and 530 nm emission as an indication
of cell number. All assays were run in triplicate.
Intrafemoral implantation of 4T1 cells into C3H/Scid mice
For evaluating the effect of local radiation and assessing pain,
direct local intramedullary injection of tumor cells is a preferred
method over systemic administration of tumor cells. Localized
implantation of tumor cells results in a uniform site of skeletal tu-
mor and allows for corresponding analysis of neuroanatomic pain
markers associated with behavioral pain.21Cultured 4T1 cells
(90% confluence) were harvested after a 3-min exposure to 0.25%
trypsin. Trypsinization was stopped with medium containing 10%
fetal bovine serum, and cells were washed once in medium con-
taining 10% fetal bovine serum and resuspended in Hank’s bal-
anced salt solution (HBSS; Sigma, Deerfield, IL). Only suspen-
sions consisting of single cells with 90% viability were used. The
4T1 injection technique was performed as described previously.23
Briefly, a general anesthetic cocktail of ketamine (97.5 mg/kg)
and xylazine (37.5 mg/kg) was administered intraperitoneally. A
28-gauge boring needle was inserted, via arthrotomy, into the
medullary canal to create an indentation for the 4T1 breast carci-
noma cells. Seven sham-treated animals were injected with HBSS
into the intramedullary space of the left femur, and 28 animals
were injected with 4T1 breast carcinoma cells at a density of 104
cells. A small amount of dental amalgam was used to seal the
injection site, thereby confining the tumor cells or HBSS within
the medullary space.
Treatment of C3H/Scid mice with 4T1 intrafemoral tumors
Treatments were administered 7 days after intrafemoral injec-
tion of 4T1 cells into C3H/Scid mice. Animals with documented
radiogram lytic lesions were randomized into 1 of the 4 groups
consisting of 7 mice per group: (i) single vein tail injection of bev-
acizumab (AvastinTM; Genentech) at 15 mg/kg; this dose was cho-
sen based on its activity in a human clinical trial20; (ii) single dose
of local megavoltage irradiation (2-cm-diameter field) of the
femur at 30 Gy in 1 fraction; (iii) single vein tail injection of beva-
cizumab at 15 mg/kg and single dose of local megavoltage irradia-
tion at 30 Gy in 1 fraction; and (iv) a control group receiving a sin-
gle vein tail injection with 0.9% NaCl solution in equivalent vol-
Megavoltage radiation (X-rays)
A Varian 2100c linear accelerator (Varian Medical Systems,
Palo Alto, CA) equipped with a floor stand radiosurgery system
(Varian Medical Systems) was used. The irradiated area was iden-
tified using a tertiary collimator of 20 mm diameter. Alignment of
the system was verified using the methodology described by Lutz
et al.24Animals were irradiated one at a time using a 6-MV beam.
A dose of 30 Gy was delivered to the femur by placing it 100 cm
from the radiation source and covering the target area with 1 cm
of bolus to place the femur at an approximate depth of 1.5 cm, cor-
responding to the depth of maximum dose of the 6-MV beam. A
30 Gy dose was selected based on our data that this dose can
achieve a decrease in 4T1 bone tumor size of about 50%. In our
experience, the growth of 4T1 cancer cells could not be controlled
by lower doses of radiation delivered over several days (data not
shown); the dosing regimen that is similar to the low dosing regi-
mens used to treat patients with cancer.
Radiographic evaluation of bone destruction
Radiographs of the femur were taken on days 7, 13 and 16 after
4T1-cell implantation using an X-ray machine (Faxitron X-ray
Corp., Wheeling, IL). Images were captured on Kodak Min-R
2000 mammography film (Eastman Kodak, Rochester, NY; expo-
sure settings: 7 s, 21 kVp). Assessments of the extent of bone
destruction were performed on day 16 by a single investigator
blinded to group assignments. A grading system of bone lysis with
numeric values ranging from 0 to 4 was used as described previ-
ously.6A grade of 0 represents no lysis; 1, minimal but visible
bone lysis within the medullary canal; 2, moderate bone lysis in
the medullary canal with preservation of the cortex; 3, severe bone
lysis with cortical disruption; and 4, massive destruction with soft
tissue extension of the tumor. This grading system of bone
destruction has been previously proposed and used clinically.25–27
Quantification of radiolucent tumor lesions in bone
X-ray films taken on day 16 were scanned on an Epson 2450
flatbed scanner in transmission mode at a fixed exposure setting.
The resulting images were opened in Adobe Photoshop CS2
(Adobe Corp., Mountain View, CA), and individual femur sec-
tions were outlined, copied to a layer and then a threshold was set
to select the entire femur area (reference area). By binarizing an
image in this way, areas of the image representing bone became
white and areas representing background became black. Areas of
the bone that were not converted to white were selected using
Color Range, and the selection was expanded and contracted to
include nonwhite areas. The whole femur was then selected (refer-
ence area), and the total number of pixels comprising that area
was recorded from the Histogram readout. The background (orig-
inal) image was selected and copied to a second layer constrained
to only the bone area by the femur selection. This image was
thresholded in order to convert all of the darker areas to black, and
Color Range was used to make a selection around this area. The
total number of pixels comprising the bone density area was
recorded from the Histogram readout. The ratio of total pixels
from the reference area to the bone density area was determined
using Microsoft Excel. Osteolytic lesions on bone radiographs
were confirmed by hematoxylin and eosin (H&E) staining to have
been caused by tumor cells.
Necropsy procedures and histological studies of bone tumor
Half of the tumor was fixed in formalin, decalcified and embed-
ded in paraffin for immunohistochemistry studies, using H&E
staining and tartrate-resistant acid phosphatase (TRAP). The other
half was fixed using periodate–lysine–paraformaldehyde, decalci-
fied and embedded in optimal cutting temperature compound
(Miles, Elkhart, IN). The bones were decalcified by exposure for
10 days to phosphate buffer solution (PBS) containing 10% ethyl-
enediaminetetraacetic acid. Tissue sections (4–6 lm) were
mounted on positively charged Superfrost slides (Fisher Scientific,
Houston, TX) and dried overnight. Sections were deparaffinized in
xylene, treated with a graded series of alcohol [100, 95, 80% etha-
nol/double distilled H2O (v/v)] and then rehydrated in PBS (pH
7.5). Immunostaining for the endothelial cell marker, CD31, was
performed as described previously.6Frozen tissues used for the
identification of platelet endothelial cell adhesion molecule-1
ZWOLAK ET AL.
(CD31/PECAM-1) were sectioned (8–10 lm), mounted on posi-
tively charged Plus slides (Fisher Scientific, Houston, TX) and air-
dried for 30 min, after which immunohistochemical studies were
performed. TRAP (osteoclasts number) staining was done in ace-
tate buffer (pH 5.0) containing 50 mM sodium L-tartrate (Sigma)
and 0.1 mM levamisole (to inhibit alkaline phosphatase activity).
All sections were counterstained with 1% neutral red solution and
Quantification of osteoclasts
The number of osteoclasts was determined by quantifying the
osteoclasts of femoral sections with fast red stain activated by
TRAP (only TRAP-positive cells were counted) as previously
described.23Results are expressed as the number of osteoclasts per
square millimeter of tumor area. The number of osteoclasts dem-
onstrating clear zones was assessed as a measure of osteoclast
activation, maturation and bone resorption.28The osteoclasts with
clear zones were quantified using a 1003 objective on H&E-
stained sections. Results are expressed as a percent of mature
osteoclasts with >3 nuclei per square millimeter.
Quantification of microvessel density
Tumor microvessels were counted in sections after staining
with antibodies specific to CD31 according to methods described
previously.29Briefly, areas containing the highest number of
capillaries were identified by scanning the tumor sections at low
power (403). After the areas of high vascular density were identi-
fied, the number of microvessels was assessed by counting micro-
vessels in 5 random 0.0321-mm2fields taken of each section at
2003 magnification. Based on the criteria described by Weidner
et al.,30observation of a vessel lumen was not required.
Behavioral assessment of C3H/Scid mice injected with
4T1 tumor cells and HBSS
Baseline behaviors were recorded for all mice prior to 4T1 tu-
mor cell and sham injections. Behavioral assessments were next
recorded on day 7, prior to administration of control and experi-
mental treatments. Subsequent assessments were recorded on days
10, 13 and 16. Mice were allowed to habituate on the observation
platform for 30 min before testing commenced. Spontaneous and
movement-evoked measurements of pain were recorded for each
mouse. Behavioral testing included limb use in an open field,
guarding during forced ambulation, number of flinches and dura-
tion of spontaneous guarding. Limb use scores were based on a
scoring system23that ranged from 0 to 4, with 4 being normal
limb use; 3, pronounced limping; 2, limping and guarding; 1, par-
tial nonuse; and 0, complete loss of limb use. Guarding scores dur-
ing forced ambulation were determined using the Economex
Rotarod (Columbus Instruments, Columbus, OH) set to run at a
constant speed of 6 rpm. Guarding scores were defined as the time
for which the animal held the affected limb completely aloft. A
score of 5 indicated normal limb use; 4, presence of minimal
guarding; 3, pronounced guarding; 2, pronounced guarding
accompanied by limping; 1, partial nonuse; and 0, complete loss
of limb use. Times for spontaneous guarding were obtained using
a stopwatch to record the duration of guarding behavior during a
2-min observation period.
Immunohistochemical assessment of pain
Expression of substance P (SP) was examined in cells of the
spinal cord after the completion of treatments (i.e., 16 days).
Mouse spinal cords (axial sections), 60 lm thick, were immersed
in Zamboni’s fixative [0.03% (w/v) picric acid and 2% (w/v) para-
formaldehyde] for 48 hr at 4?C and transferred to a 20% sucrose
solution with 0.05% sodium azide in PBS for storage. Vertebrae
were carefully excised to expose the spinal cord, and 1-cm cord
sections of L4/L5 were removed. Processing and staining of the
cord sections were carried out according to a previously published
procedure.31Sections were incubated in polyclonal SP (1:500)
made in rabbit (Chemicon International, Temecula, CA) and anti-
rabbit IgG conjugated to Cy3 (1:1,000). Samples were mounted
on slide covers in agar, dehydrated in ethanol, cleared with methyl
salicylate and mounted in DEPEX (Electron Microscopy Science,
Image capture and integrated density quantification of
Low magnification (Olympus UPlanFl 43/0.13) brightfield
images of the entire spinal cord were captured for subsequent
quantification analysis of cord total area. Dorsal horn epifluores-
cent images of the L4-L5 spinal cord slices were acquired at 203
(Olympus LCPlanFl 203/0.40) using a DVC 1412M-FW-TE200
12-bit digital camera (DVC, Austin, TX) mounted on an Olympus
IX-70 Inverted microscope (Olympus America, Lake Success,
NY). Images were recorded using Image Pro Plus 2.2.8 (Media
Cybernetics, Silver Spring, MD) and their integrated densities
were subsequently analyzed using MetaMorph (Molecular Devi-
ces Corp., Sunnyvale, CA).
The integrated density was deemed the most biologically appro-
priate measurement because it takes into account the area of inter-
est, containing dorsal horn innervation and the intensity of epifluo-
rescence. This technique, with slight modifications, was used for
the quantification of nerves in several previous studies.32–35This
volumetric approach conveys information about the depth of
innervation, whereas simple average gray level measurements do
not. Biological variation in relation to any differences in stain
absorption of individual spinal cords was accounted for by setting
the gray level lower threshold at ?50 above the cord’s observed
background gray level.
Statistical data analysis
Statistical significance was determined using 1-way ANOVA or
Student’s t test, as appropriate. A p-value of <0.05 was considered
to be statistically significant. For analysis of images of spinal cord
sections, initial counts of integrated density were divided by the
cord’s total area to account for spinal cord size variation. Inte-
grated density counts in each group were analyzed using a 1-way
ANOVA test with Fisher’s protected least significant difference
post hoc test using p < 0.05.
Irradiated 4T1 cells secrete VEGF protein
4T1 breast cancer cells were treated with a single dose of local
megavoltage irradiation at 0 and 25 Gy, and then incubated up to
96 hr. At 96 hr VEGF concentration was 1.6 times higher in me-
dium collected from irradiated cancer cells than from nonirradi-
ated cells (p < 0.001; Fig. 1).
Bevacizumab binds to murine VEGF
Murine VEGF levels were significantly reduced in supernatant
from 4T1 tumor cells treated with bevacizumab. An invasion
assay of murine lung endothelial cells revealed a 7% reduction
over 48 hr in the relative invasiveness of endothelial cells stimu-
lated by 4T1 tumor cells in the presence of bevacizumab at 15 lg/
ml in 5% FBS compared with 5% FBS alone (p 5 0.01). Simi-
larly, an invasion assay of murine lung endothelial cells showed a
16% reduction over 48 hr in the migration of endothelial cells
stimulated by murine VEGF (164) in the presence of bevacizumab
at 15 lg/ml in 1 lg/ml of murine VEGF (164) as compared to
stimulation of endothelial cell migration by murine VEGF (164)
alone (p < 0.05).
Treatment effects on bone destruction
Intrafemoral injections with 0.9% NaCl solution had no signifi-
cant influence on bone destruction except that related to needle
insertion (Bone destruction score: 0.28 6 0.286; Figs. 2aI and 2b).
Intrafemoral injection with 4T1 tumor cells had a significant effect
BEVACIZUMAB PLUS IRRADIATION FOR METASTATIC BONE CANCER
on bone destruction with involvement of surrounding soft tissue
(Bone destruction score: 3.0 6 0.30; Figs. 2aII and 2b). Bone
destruction in mice treated with either localized radiation or beva-
cizumab was not significantly different from bone destruction in
nontreated mice injected with 4T1 tumor cells (Figs. 2aII, 2aIII
and 2b); however, it was significantly reduced in tumor-bearing
mice treated with a combination of bevacizumab and radiation
(Bone destruction score: 2.0 6 0.26; p < 0.05) as compared to
tumor-bearing mice that received no treatment, radiation-only
treatment or bevacizumab-only treatment (Figs. 2aIV and 2b).
Treatment effect on tumor area
To examine the tumor area in femora, the percentage of radiolu-
cent lesions caused by tumor on day 16 was determined using
image analysis and presence of tumor was confirmed by H&E
staining. Mean percentage density of radiolucent bone in control
animals was increased (29.4 6 10.0)%. Mean percentage density
of radiolucent bone for bevacizumab-only, radiation-only and
combination treatments were all significantly decreased to (10.1 6
9.5)%, (6.0 6 14.0)% and (4.3 6 11.9)%, as compared to tumor-
bearing control femora (p < 0.034; p < 0.024; p < 0.028), respec-
tively (Fig. 3). Differences were not statistically significant
between each treatment arm.
Treatment effects on tumor microvessel density
Mean microvessel density was 203.1 microvessels per field in
tumors of control mice treated with 0.9% NaCl solution. The num-
ber of microvessels declined to 186.9 in bevacizumab-treated
mice and declined further to 142 in mice treated with a combina-
tion of radiation and bevacizumab (p < 0.013 and p < 0.001,
respectively; Figs. 4a and 4b). Mice treated with radiation had no
FIGURE 1 – Irradiation of 4T1 cells enhanced secretion of VEGF by
tumor cells. 4T1 breast cancer cells were treated with a single dose of
local megavoltage irradiation at 0 and 25 Gy, and then incubated for
up to 96 hr. At 96 hr VEGF concentration was 1.6 times higher in me-
dium collected from irradiated cancer cells than from nonirradiated
cells (p < 0.001).
FIGURE 2 – Treatment effect on
bone destruction after intrafemoral
injection of 4T1 tumor cells. Nega-
tive controls were injected with sa-
line solution only. Panel (aI) repre-
sents radiographs taken on day 7 af-
ter saline injection, but not 4T1
cells. Animals injected with tumor
were randomly assigned to the 4
study groups. Positive controls were
injected with 4T1 tumor cells and
received no treatment. Radiographs
were taken on days 13 and 16 in all
4 treatment groups. Note the mainte-
nance of bony cortex on day 16 in
mice treated with a combination of
bevacizumab and radiation (Panel
aV). Large radiolucent lesions were
noted in control, radiation-treated
(head arrows) with extension into
surrounding soft tissues (Panels aII,
aIII and aIV). (b) Bone destruction
scores for treatment and control
groups were assessed on days 13
ZWOLAK ET AL.
significant reduction in microvessel count compared to nontreated,
Treatment effect on osteoclast activity and number
We assessed the number of osteoclasts to investigate the mecha-
nism by which bevacizumab reduced tumor-caused bone destruc-
tion. On day 16, there was an increased number of osteoclasts in
tumor-bearing animals (72.4 6 17.6 osteoclasts/mm2of tumor) as
compared to tumor-free controls (27.3 6 14.2 osteoclasts/mm2of
tumor). Treatment with bevacizumab, radiation and their combi-
nation did not decrease tumor-induced osteoclast numbers (62.3 6
24.2, 78.2 6 28.4 and 68.8 6 23.8 osteoclasts/mm2of tumor,
respectively), despite a significant reduction in bone destruction in
combination-treated mice (Table I).
To examine the effects of bevacizumab, radiation and combina-
tion treatment on osteoclast activity, we quantified the percentage
of multinucleated osteoclasts exhibiting a clear zone at the osteo-
clast-mineralized bone interface. Randomized counts of osteo-
clasts attached to cortical bone indicated a greater number of
actively resorbing osteoclasts with a clear zone in tumor-bearing
animals (75.0 6 3.0% of mature osteoclasts with >3 nuclei) com-
pared to tumor-free controls (15.10 6 3.4% of mature osteoclasts
with >3 nuclei; p < 0.05). Bevacizumab alone and combination
treatment also decreased the number of active osteoclasts with
clear zones (52.4 6 6.0% and 65.2 6 3.0% of mature osteoclasts
with >3 nuclei, respectively) compared to tumor-free mice (p <
0.05). There was no difference in osteoclast activity between
radiation-only and nontreated, tumor-bearing mice (Table I).
Treatment effects on bone cancer pain
All treatment groups showed a significant decrease in spontane-
ous guarding and flinching, but combination treatment showed the
greatest decrease in ongoing pain behavior in comparison to all
other groups (Figs. 5a and 5b). All treatment groups receiving
radiation, bevacizumab and combination treatment had signifi-
cantly improved limb use on day 16 (Scores: 2.429 6 0.36, 2.0 6
0.218 and 2.85 6 0.21, respectively), while ambulatory pain in the
combination-treated group displayed the greatest decrease as com-
pared to nontreated, tumor-bearing controls (Fig. 5c). Similarly,
all groups receiving radiation, bevacizumab or combination treat-
ment showed significantly less guarding during forced movement
on the rotarod wheel on day 16 (Rotarod scores: 2.57 6 0.36,
2.14 6 0.26 and 3.0 6 0.309, respectively), while movement-
evoked guarding in combination-treated animals displayed the
greatest reduction in guarding as compared to nontreated, tumor-
bearing controls (Rotarod score: 1.14 6 0.34; Fig. 5d).
Immunohistochemical analysis of spinal cords
The expression of SP in the spinal cord was increased due to tu-
mor formation in the femora of tumor-bearing mice. Mean density
count of immunoreactive neurons in the spinal cord of tumor-bear-
ing mice was 59.1 on the ipsilateral side of the spinal cord and
33.2 on the contralateral side (p < 0.006) (Figs. 6a–6d). Mean
density counts for radiation-only, bevacizumab-only and combina-
tion treatments were all significantly decreased to 26.8, 22.0, 32.2
as compared to tumor-bearing controls on the ipsilateral side (p <
0.002, p < 0.001 and p < 0.02), respectively (Fig. 6e).
We have demonstrated that VEGF blockade with in vivo beva-
cizumab enhances the effect of clinical megavoltage radiation by
further reducing tumor burden in comparison to single modality
treatments. Combined modality treatment reduced the frequency
of pain-associated behaviors, decreased levels of nociceptive pro-
tein expression in the spinal cord, maintained cortical integrity
and decreased the density of microvessels as compared to single
FIGURE 3 – Mean percentage density of radiolucent bone lesions.
Tumor area in the femora of animals in bevacizumab-only, radiation-
only and combination treatments were all significantly decreased to
(10.1 6 9.5)%, (6.0 6 14.0)% and (4.3 6 11.9)%, as compared to
tumor-bearing control femora (p < 0.034; p < 0.024; p < 0.028),
femoral tumor tissue stained using
immunohistochemical analysis with
antibodies to CD31 (a). Quantifica-
tion of immunohistochemical data
from tissue stained for CD31 (b).
4 – Cross-sections of
BEVACIZUMAB PLUS IRRADIATION FOR METASTATIC BONE CANCER
modality treatments. For this study we did not have access to
A4.6.1, which is the murine equivalent to the humanized anti-
VEGF monoclonal antibody bevacizumab. The murine monoclo-
nal antibody A4.6.1 binds more avidly to murine VEGF36; there-
fore, we have also demonstrated that bevacizumab binds murine
VEGF, albeit less avidly than human VEGF.
Our findings are consistent with other reports that have found
that radiation stimulates tumor cells to produce and secrete VEGF,
which in turn increases endothelial cell radioresistance.37,38Elimi-
nation of the protective action of radiation-stimulated VEGF with
a specific antibody enhances endothelial cell radiosensitivity. Fur-
thermore, radiation stimulates the expression of other factors
involved in tumor angiogenesis, such as bFGF and matrix-degrad-
ing proteases.39Our findings support the hypothesis that tumor
neovascularization and radioresistance mediated by radiation-
increased VEGF levels can be mitigated by use of antiangiogenic
Similarly, mice receiving combination treatment had even
greater reduction of intratumoral vascular density than mice
treated only with bevacizumab. Moreover, tumor vessels in ani-
mals receiving combination treatment appeared larger as com-
pared to other treatment and control groups.
Radiographic analysis revealed a significant reduction in bone
destruction scores of mice treated with combination therapy. In
this same group, behavioral and neuroimmunochemical assess-
ment showed decreased ongoing pain. Other reports have also sug-
gested that tumor-induced osteolysis correlates with increased
pain assessed by behavioral tools and neurochemical changes in
the spinal cord and dorsal root ganglia.40–42The neurochemical
effects of osteolysis begin with the Ca21-related signaling cascade
in the presynaptic primary afferent nerve fiber that stimulates the
release of such excitatory peptides as SP. In postsynaptic dorsal
horn neurons, it stimulates the release of prostaglandin and NO, as
well as N-Methyl-D-Aspartate receptor upregulation. In our study,
radiation and bevacizumab treatment reduced SP signals 9 days
after treatment. SP is a primary afferent peptide with the strongest
evidence for a role in pain perception. It is well known that inflam-
mation of peripheral tissues upregulates the production of SP in
TABLE I – EFFECT OF IRRADIATION, BEVACIZUMAB AND COMBINATION TREATMENT ON BONE REMODELING
(OSTEOCLAST NUMBER AND ACTIVITY) IN ORTHOTOPIC MODEL OF BREAST CANCER IN BONE
1 NaCl treatment
1 bevacizumab treatment
1 radiation treatment
1 combination treatment
Number of osteoclasts/mm2
% of activated multinucleated osteoclasts
27.3 6 14.2
15.10 6 3.4%
72.4 6 17.6
75.0 6 3.0%1
62.3 6 24.2
52.4 6 6.0%1,2
78.2 6 28.4
71.8 6 2.2%1
68.8 6 23.8
65.2 6 3.0%1,2
Values represent mean 6 SEM.
1p value < 0.05 vs. HBSS-injected controls.–2p value < 0.05 vs. 4T1 injected 1 NaCl-treated animals.
FIGURE 5 – Pain-associated behaviors of animals injected with 4T1 tumor cells. Duration of guarding and number of spontaneous flinches
over a 2-min observation period were used as a measure of ongoing pain on days 7, 11, 13 and 16 (Panels a and b). Parameters of movement-
evoked pain were assessed using spontaneous limb use and forced ambulation (guarding) on a Rotarod (Panels c and d). Note that combined
treatment significantly reduced guarding and flinching behaviors on days 13 and 16 as compared to positive controls (Panels a and b).
Movement-evoked pain behaviors on days 13 and 16 were also significantly alleviated as compared to positive controls (Panels c and d). Asterisks (*)
indicate significantly different treatment groups as compared to positive controls (p < 0.05). Values represent means divided by SEM.
ZWOLAK ET AL.
the sensory ganglia, concurrent with the development of hyperal-
gesia. It has been suggested that radiation-mediated pain reduction
may be associated with decreased glial activity due to lower neu-
rotransmitter levels.43Our results are in agreement with these
Recent reports have indicated that bevacizumab may decrease
bone destruction by inhibiting osteoclast activity and recruit-
ment.17,44Our findings revealed no reduction in osteoclast num-
ber, although osteoclast activity at the tumor-bone interface was
significantly decreased in mice receiving bevacizumab or combi-
nation treatment. Notably, we observed a reduction in expression
of SP in mice treated with bevacizumab, which correlated with be-
havioral pain studies. This suggests that radiation and antiangio-
genic treatments might relieve pain by decreasing bone destruc-
tion, although another possible mechanism is an anti-inflammatory
reaction. This latter possibility was not investigated in this study.
Although radiation therapy alone is associated with high rates
of complete response of osseous metastases, as well as palliation
of pain,45,46there is still a large number of patients whose bone
tumors are relatively radioresistant or quickly progress after initial
radiation therapy. These patients may benefit from enhancement
of radiation therapy, such as that used in our murine model. Irradi-
ation of tumor tissue decreases the number of tumor and endothe-
lial cells47,48and subsequent tumor regrowth depends on activa-
tion of angiogenesis. For this reason, a strategy that uses antian-
giogenic agents in combination with local radiation warrants
further investigation. To our knowledge, our study is the first that
reports on the additive effect of megavoltage radiation and bevaci-
zumab in a bone tumor model.
The authors thank Mrs. Audrey and Mr. Dennis Anderson for
their support through the Experimental Therapeutics Fund and
also thank Mr. Michael Franklin for editorial support.
Guise TA, Mundy GR. Cancer and bone. Endocr Rev 1998;19:18–54.
Body JJ. Metastatic bone disease: clinical and therapeutic aspects.
Bone 1992;13 (Suppl 1):S57–S62.
Hartsell WF, Scott CB, Bruner DW, Scarantino CW, Ivker RA, Roach
M, III, Suh JH, Demas WF, Movsas B, Petersen IA, Konski AA,
Cleeland CS, et al. Randomized trial of short- versus long-course
radiotherapy for palliation of painful bone metastases. J Natl Cancer
Kerbel R, Folkman J. Clinical translation of angiogenesis inhibitors.
Nat Rev Cancer 2002;2:727–39.
Ramaswamy B, Shapiro CL. Bisphosphonates in the prevention and
treatment of bone metastases. Oncology (Williston Park) 2003;17:
1261–70; discussion 1270–2,1277–8,1280.
Weber KL, Doucet M, Price JE, Baker C, Kim SJ, Fidler IJ. Blockade of
Guidi AJ, Berry DA, Broadwater G, Perloff M, Norton L, Barcos MP,
Hayes DF. Association of angiogenesis in lymph node metastases
with outcome of breast cancer. J Natl Cancer Inst 2000;92:486–92.
Spiro IJ, Rice GC, Durand RE, Stickler R, Ling CC. Cell killing,
radiosensitization and cell cycle redistribution induced by chronic hy-
poxia. Int J Radiat Oncol Biol Phys 1984;10:1275–80.
Gorski DH, Beckett MA, Jaskowiak NT, Calvin DP, Mauceri HJ,
Salloum RM, Seetharam S, Koons A, Hari DM, Kufe DW,
Weichselbaum RR. Blockage of the vascular endothelial growth fac-
tor stress response increases the antitumor effects of ionizing radia-
tion. Cancer Res 1999;59:3374–8.
10. Tsai JH, Makonnen S, Feldman M, Sehgal CM, Maity A, Lee WM.
Ionizing radiation inhibits tumor neovascularization by inducing inef-
fective angiogenesis. Cancer Biol Ther 2005;4:1395–400.
11. Jain RK. Normalization of tumor vasculature: an emerging concept in
antiangiogenic therapy. Science 2005;307:58–62.
12. Dvorak HF. Vascular permeability factor/vascular endothelial growth
factor: a critical cytokine in tumor angiogenesis and a potential target
for diagnosis and therapy. J Clin Oncol 2002;20:4368–80.
13. Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases.
14. Teicher BA, Sotomayor EA, Huang ZD. Antiangiogenic agents poten-
tiate cytotoxic cancer therapies against primary and metastatic dis-
ease. Cancer Res 1992;52:6702–4.
15. Grunstein J, Roberts WG, Mathieu-Costello O, Hanahan D, Johnson
RS. Tumor-derived expression of vascular endothelial growth factor
is a critical factor in tumor expansion and vascular function. Cancer
16. Engsig MT, Chen QJ, Vu TH, Pedersen AC, Therkidsen B, Lund LR,
Henriksen K, Lenhard T, Foged NT, Werb Z, Delaisse JM. Matrix
metalloproteinase 9 and vascular endothelial growth factor are essen-
tial for osteoclast recruitment into developing long bones. J Cell Biol
17. Aldridge SE, Lennard TW, Williams JR, Birch MA. Vascular endo-
thelial growth factor receptors in osteoclast differentiation and func-
tion. Biochem Biophys Res Commun 2005;335:793–8.
18. Nakagawa M, Kaneda T, Arakawa T, Morita S, Sato T, Yomada T,
Hanada K, Kumegawa M, Hakeda Y. Vascular endothelial growth
factor (VEGF) directly enhances osteoclastic bone resorption and sur-
vival of mature osteoclasts. FEBS Lett 2000;473:161–4.
19. Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, Hainsworth
J, Heim W, Berlin J, Baron A, Griffing S, Holmgren E, Ferrara N,
Fyfe G, et al. Bevacizumab plus irinotecan, fluorouracil, and leucovo-
rin for metastatic colorectal cancer. N Engl J Med 2004;350:2335–
20. Sandler A, Gray R, Perry MC, Brahmer J, Schiller JH, Dowlati A,
Lilenbaum R, Johnson DH. Paclitaxel-carboplatin alone or with beva-
cizumab for non-small-cell lung cancer. N Engl J Med 2006;355:
21. Goblirsch MJ, Zwolak P, Clohisy DR. Advances in understanding
bone cancer pain. J Cell Biochem 2005;96:682–8.
22. Sevcik MA, Ghilardi JR, Peters CM, Lindsay TH, Halvorson KG,
Jonas BM, Kubota K, Kuskowski MA, Boustany L, Shelton DL,
Mantyh PW. Anti-NGF therapy profoundly reduces bone cancer pain
and the accompanying increase in markers of peripheral and central
sensitization. Pain 2005;115:128–41.
23. Honore P, Luger NM, Sabino MA, Schwei MJ, Rogers SD, Mach DB,
O’Keefe PF, Ramnaraine ML, Clohisy DR, Mantyh PW. Osteoprote-
gerin blocks bone cancer-induced skeletal destruction, skeletal pain
and pain-related neurochemical reorganization of the spinal cord. Nat
24. Lutz W, Winston KR, Maleki N. A system for stereotactic radiosur-
gery with a linear accelerator. Int J Radiat Oncol Biol Phys 1988;14:
FIGURE 6 – Immunoreactivity of SP in spinal cord sections at lum-
bar levels 4 and 5. The expression of SP was increased by tumor for-
mation in tumor-bearing mice (a) and was decreased 9 days after bev-
acizumab (b), radiation (c) and combination treatment (d). Scale bar:
100 lm. (e) Quantification of immunohistochemical data from spinal
cords stained for SP. Amount of SP in spinal cord ipsilateral to tumor
in nontreated animals was significantly larger than in treated animals
(p < 0.05). Values represent means divided by SEM.
BEVACIZUMAB PLUS IRRADIATION FOR METASTATIC BONE CANCER
25. Lodwick GS. Radiographic diagnosis and grading of bone tumors, Download full-text
with comments on computer evaluation. Proc Natl Cancer Conf 1964;
26. Taconis WK, Mulder JD. Fibrosarcoma and malignant fibrous histio-
cytoma of long bones: radiographic features and grading. Skeletal
27. Link TM, Hillmann A, Erlemann R, Gronefeld A, Haussler M, Heppe
AE, Vestring T, Peters PE. Imaging of bone tumors: evaluation of
direct magnification radiography. Skeletal Radiol 1996;25:441–7.
28. Lees RL, Heersche JN. Macrophage colony stimulating factor
increases bone resorption in dispersed osteoclast cultures by increas-
ing osteoclast size. J Bone Miner Res 1999;14:937–45.
29. Yoneda J, Kuniyasu H, Crispens MA, Price JE, Bucana CD, Fidler IJ.
Expression of angiogenesis-related genes and progression of human
ovarian carcinomas in nude mice. J Natl Cancer Inst 1998;90:447–54.
30. Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis
and metastasis—correlation in invasive breast carcinoma. N Engl J
31. Wacnik PW, Baker CM, Herron MJ, Kren BT, Blazar BR, Wilcox
GL, Hordinsky MK, Beitz AJ, Ericson ME. Tumor-induced mechani-
cal hyperalgesia involves CGRP receptors and altered innervation and
vascularization of DsRed2 fluorescent hindpaw tumors. Pain 2005;
32. Christmas TJ, Rode J, Chapple CR, Milroy EJ, Turner-Warwick RT.
Nerve fibre proliferation in interstitial cystitis. Virchows Arch A
Pathol Anat Histopathol 1990;416:447–51.
33. Engeland WC, Levay-Young BK, Rogers LM, Fitzgerald D. Differen-
tial gene expression of cytochrome P450 11b-hydroxylase in rat adre-
nal cortex after in vivo activation. Endocrinology 1997;138:2338–46.
34. Mihm MJ, Schanbacher BL, Wallace BL, Wallace LJ, Uretsky NJ,
Bauer JA. Free 3-nitrotyrosine causes striatal neurodegeneration in
vivo. J Neurosci 2001;21:RC149.
35. Pan J, Yeger H, Cutz E. Innervation of pulmonary neuroendocrine
cells and neuroepithelial bodies in developing rabbit lung. J Histo-
chem Cytochem 2004;52:379–89.
36. Gerber HP, Ferrara N. Pharmacology and pharmacodynamics of beva-
cizumab as monotherapy or in combination with cytotoxic therapy in
preclinical studies. Cancer Res 2005;65:671–80.
37. Park JS, Qiao L, Su ZZ, Hinman D, Willoughby K, McKinstry R,
Yacoub A, Duigou GJ, Young CS, Grant S, Hagan MP, Ellis E, et al.
Ionizing radiation modulates vascular endothelial growth factor
(VEGF) expression through multiple mitogen activated protein kinase
dependent pathways. Oncogene 2001;20:3266–80.
38. Gupta VK, Jaskowiak NT, Beckett MA, Mauceri HJ, Grunstein J,
Weichselbaum RR. Vascular endothelial growth factor enhances endo-
39. Hartford AC, Gohongi T, Fukumura D, Jain RK. Irradiation of a pri-
mary tumor, unlike surgical removal, enhances angiogenesis suppres-
sion at a distal site: potential role of host-tumor interaction. Cancer
40. Schwei MJ, Honore P, Rogers SD, Salak-Johnson JL, Finke MP,
Ramnaraine ML, Clohisy DR, Mantyh PW. Neurochemical and cellu-
lar reorganization of the spinal cord in a murine model of bone cancer
pain. J Neurosci 1999;19:10886–97.
41. Honore P, Rogers SD, Schwei MJ, Salak-Johnson JL, Luger NM,
Sabino MC, Clohisy DR, Mantyh PW. Murine models of inflammatory,
neuropathic and cancer pain each generates a unique set of neuro-
chemical changes in the spinal cord and sensory neurons. Neuro-
42. Sabino MA, Ghilardi JR, Jongen JL, Keyser CP, Luger NM, Mach
DB, Peters CM, Rogers SD, Schwei MJ, de Felipe C, Mantyh PW. Si-
multaneous reduction in cancer pain, bone destruction, and tumor
growth by selective inhibition of cyclooxygenase-2. Cancer Res 2002;
43. Vit JP, Ohara PT, Tien DA, Fike JR, Eikmeier L, Beitz A, Wilcox
GL, Jasmin L. The analgesic effect of low dose focal irradiation in a
mouse model of bone cancer is associated with spinal changes in
neuro-mediators of nociception. Pain 2006;120:188–201.
44. Aldridge SE, Lennard TW, Williams JR, Birch MA. Vascular endo-
thelial growth factor acts as an osteolytic factor in breast cancer me-
tastases to bone. Br J Cancer 2005;92:1531–7.
45. Tong D, Gillick L, Hendrickson FR. The palliation of symptomatic
osseous metastases: final results of the Study by the Radiation Ther-
apy Oncology Group. Cancer 1982;50:893–9.
46. Wu JS, Wong R, Johnston M, Bezjak A, Whelan T. Meta-analysis of
dose-fractionation radiotherapy trials for the palliation of painful bone
metastases. Int J Radiat Oncol Biol Phys 2003;55:594–605.
47. Solesvik OV, Rofstad EK, Brustad T. Vascular changes in a human
malignant melanoma xenograft following single-dose irradiation.
Radiat Res 1984;98:115–28.
48. Hast J, Schiffer IB, Neugebauer B, Teichman E, Schreiber W, Brieger
J, Kim DW, Gebhard S, Born CJ, Strugala M, Sagemuller J, Brenner
W, et al. Angiogenesis and fibroblast proliferation precede formation
of recurrent tumors after radiation therapy in nude mice. Anticancer
ZWOLAK ET AL.