Biomaterials 33 (2012) 5414e5422

Contents lists available at SciVerse ScienceDirect

Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials

Noninvasive monitoring of orthotopic glioblastoma therapy response using

RGD-conjugated iron oxide nanoparticles
Fan Zhang a,1, Xinglu Huang b,1, Lei Zhu b, c, Ning Guo c, Gang Niu b, Magdalena Swierczewska b,
Seulki Lee b, Hong Xu d, Andrew Y. Wang d, Khalid A. Mohamedali e, Michael G. Rosenblum e,
Guangming Lu a, **, Xiaoyuan Chen b, c, *
a
Department of Radiology, Nanjing Jinling Hospital, Clinical School of Medical College of Nanjing University, Nanjing 210002, China
b
Laboratory of Molecular Imaging and Nanomedicine (LOMIN), National Institute of Biomedical Imaging and Bioengineering (NIBIB), National Institutes of Health (NIH),
Bethesda, MD 20892, USA
c
Center for Molecular Imaging and Translational Medicine, School of Public Health, Xiamen University, Xiamen 361005, China
d
Ocean NanoTech, LLC 2143 Worth Lane Springdale, AR 72764, USA
e
Department of Experimental Therapeutics, The University of Texas M.D. Anderson Cancer Center, Houston, TX 77230, USA

a r t i c l e i n f o a b s t r a c t

Article history: Noninvasive imaging techniques have been considered important strategies in the clinic to monitor
Received 27 March 2012 tumor early response to therapy. In the present study, we applied RGD peptides conjugated to iron oxide
Accepted 11 April 2012 nanoparticles (IONP-RGD) as contrast agents in magnetic resonance imaging (MRI) to noninvasively
Available online 3 May 2012
monitor the response of a vascular disrupting agent VEGF121/rGel in an orthotopic glioblastoma model.
RGD peptides were firstly coupled to IONPs coated with a crosslinked PEGylated amphiphilic triblock
Keywords:
copolymer. In vitro binding assays confirmed that cellular uptake of particles was mainly dependent on
Magnetic resonance imaging (MRI)
the interaction between RGD and integrin avb3 of human umbilical vein endothelial cells (HUVEC). The
Iron oxide nanoparticles (IONPs)
RGD peptides
tumor targeting of IONP-RGD was observed in an orthotopic U87 glioblastoma model. Finally, nonin-
Tumor targeting vasive monitoring of the tumor response to VEGF121/rGel therapy at early stages of treatment was
Therapy response successfully accomplished using IONP-RGD as a contrast agent for MRI, a superior method over common
anatomical approaches which are based on tumor size measurements. This preclinical study can accel-
erate anticancer drug development and promote clinical translation of nanoprobes.
Published by Elsevier Ltd.

1. Introduction monitor response to therapies with novel mechanisms of action,

often predicting the success of therapy before conventional
The detailed information of tumor progression in response to measurements have changed [3,4]. The development of advanced
therapy is important to improve patient selection for specific imaging strategies, although still challenging, not only allow the
treatment strategies and guides adaptation of treatment at an early detection and monitoring of tumor development, but also facilitate
stage [1]. Anatomical approaches based on measurements of tumor a broad understanding of the cellular and molecular events that
size are extensively applied for assessing therapy response so far, propagate tumor angiogenesis, as well as those occurring in
but significant limitations exist, such as the presence of tumors that response to therapy.
cannot be measured, poor measurement reproducibility, and mass A highly versatile device in monitoring tumor progression and
lesions that persist following therapy [2]. Noninvasive imaging therapy response is magnetic resonance imaging (MRI), since this
techniques are emerging more and more as important tools to technique provides a high spatial resolution and excellent contrast
of opaque soft tissue [5]. However, the low sensitivity of MRI often
reduces the success of imaging approaches. Recently, iron oxide
nanoparticles (IONPs) have shown their suitability for use as
* Corresponding author. Laboratory of Molecular Imaging and Nanomedicine sufficiently high tissue contrast agents for MRI in terms of their
(LOMIN), National Institute of Biomedical Imaging and Bioengineering (NIBIB), intrinsic properties and versatile surface functionality [6e9].
National Institutes of Health (NIH), Bethesda, MD 20892, USA. Tel.: þ1 301 451 4246.
Among these applications, the detection of initial and further
** Corresponding author.
E-mail addresses: cjr.luguangming@vip.163.com (G. Lu), shawn.chen@nih.gov
development of tumors using IONPs as contrast agents is of
(X. Chen). particular interest. Glioblastoma (GBM) is the most common form
1
These authors contributed equally to this work. of primary brain tumor, and its aggressive nature and evasiveness

0142-9612/$ e see front matter Published by Elsevier Ltd.

doi:10.1016/j.biomaterials.2012.04.032
 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422 5415

to treatments make it one of the most lethal cancers [10,11]. To respectively. The slides were washed twice with PBS and then observed with an
better develop new theranostic strategies in experimental research, Olympus microscope.

orthotopic GBM models are considered more realistic than

2.5. In vivo MRI
common subcutaneous xenografts. However, current strategies are
mainly limited by the bloodebrain barrier (BBB) of models. Several MRI studies were conducted in a 7 T horizontal bore small animal MRI scanner
recent reports have demonstrated the feasibility of IONPs to target (Bruker Biospin). All mice were anesthetized with 1e2% isoflurane mixed with pure
orthotopic GBM across the BBB [12,13], but it still remains a chal- oxygen via a nose cone and were placed in a stretched supine position with
a respiratory sensor. Axial and coronal two-dimensional (2D) fast spin-echo
lenge to noninvasively assess tumor response to therapeutics using sequence images were first acquired to ensure the imaging position of the
IONPs. implanted tumor. The following parameters were adopted in data acquisition: ① T1-
VEGF121/rGel is a fusion protein containing VEGF121 linked by weighted multislice gradient-echo images: TR/TE ¼ 250/4.5 m s,
a flexible G4S tether to the toxin gelonin (rGel) [14]. Such novel matrix ¼ 256  256, FA ¼ 30 , 9 contiguous slices; ② T2-weighted multislice spin-
echo images: TR/TE ¼ 2000/48 m s, matrix ¼ 256  256, 9 contiguous slices; ③ T2*-
fusion construct has high specificity and cytotoxicity to endothelial
weighted images: TR/TE ¼ 1500/4 m s, matrix ¼ 256  256, FA ¼ 30 , 9 contiguous
cells, and has been successfully used for antiangiogenic therapies in slices; ④ T2-map: TR ¼ 2000/48 m s, matrix ¼ 256  256, 1 slice, T2 relaxation
various tumor models [14e16]. In this work, we applied an measurements with TE of 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150,
amphiphilic PEG lipid coated, IONP-RGD conjugate for noninvasive and 160 ms. For each data set, one slice with comparable locations within the tumor
monitoring of tumor progression during VEGF121/rGel therapy of was selected to determine signal intensities. Signal intensities were measured in
defined regions of interest (ROIs) with Image J (National Institutes of Health).
orthotopic GBM. The IONPs were firstly functionalized by conju- For IONP-RGD binding evaluation, the mice with U87MG tumors were injected
gation with RGD peptides. Subsequently, the specific binding of with IONP-RGD at a dose of 5 mg Fe/kg. Two control experiments were also per-
IONP-RGD to integrin avb3 was studied in vitro and in vivo. The formed by injection with IONPs and another with IONP-RGD after blocking the
damaging effect of VEGF121/rGel on angiogenic tumor blood vessels target by adding free RGD (IONP-RGD þ block) at the same dose. For treatment
evaluation, two doses of 12 mg/kg VEGF121/rGel were intraperitoneally adminis-
was also confirmed in this study. Finally, monitoring the thera-
tered, and then the animals were imaged at 4 days.
peutic response was explored using IONP-RGD as MRI contrast Dynamic T2*-weighted images were performed using the T2*-weighted method
agents in an orthotopic U87MG model. with a temporal resolution of 3.4 s over 70 min. The imaging parameters were: TR/
TE ¼ 10/3 m s, flip angle ¼ 45 , FOV ¼ 35  35 mm, matrix size ¼ 128  128, slice
2. Materials and methods thickness ¼ 1 mm, and NEX ¼ 1. Thirty seconds after the start of the TR-MRI
sequence, the particles were administered manually over 30 s. The serial images
2.1. Materials were normalized by dividing signal intensity (SI) of each image by an average value
from 4 consecutive pre-injection images collected within the first 40 s of the T2*-MRI
Iron oxide nanoparticles were obtained from Ocean NanoTech. BS3 crosslinker sequence. For quantitative analysis, pre-injection images within the first 40 s were
was purchased from Thermo Scientific. Arginine-glycine-aspartic acid peptide averaged to calculate the initial SI (0). SI (t) represents the intensity yield from post-
c(RGDyK) was bought from Anaspec. Disposable PD-10 desalting column was injection images. The IONP-RGD concentrations (CIONP-RGD) were estimated using
acquired from GE healthcare. The centrifugal filter (30 k cutoff) was bought from the following equation: CINOP-RGD ¼ -ln[SI (t)/SI(0)]/TE [18].
Millipore. Mounting medium with DAPI was purchased from Vector Laboratories.
Human umbilical vein endothelial cell line (HUVEC) and U87MG human 2.6. Histological study
glioblastoma-astrocytoma cell line were purchased from ATCC. Rat anti-mouse CD31,
hamster anti-mouse b3 antibody (CD61), biotinylated anti-rat and anti-hamster IgG 2.6.1. Double staining of Prussian blue and CD31/integrin b3
secondary antibodies were purchased from BD Bioscience. VEGF121/rGel was Mouse brains were collected in optimal-cutting-temperature (O.C.T.) compound
prepared according to a reported procedure [17]. and stored in the freezer at 80  C. Tissue samples were later cut into 5 mm thick
slices. Slides were warmed for 20 min at room temperature after removal from 80  C
2.2. Synthesis and characterization of IONP-RGD freezer and were then fixed with ice-cold acetone for 5 min. After fixation, slides were
incubated with 0.3% H2O2 solution in PBS for 10 min to block endogenous peroxidase
Ten nm PEGylated amphiphilic copolymer coated IONPs with amino groups activity. Then slides were rinsed 3 times with PBS (2 min each). Rat anti-mouse CD31
were obtained from Ocean NanoTech. For surface modification of RGD peptide, 2 mg primary antibody diluent (1:50) was subsequently applied to the tissue sections, and
of IONPs was dispersed in 4 ml of borate buffer (pH 9.0). Then, 10 ml of BS3 crosslinker the incubation was left at room temperature for 1 h in a humid chamber. After rinsing
(10 mg/ml) was added to the solution. After stirring for 30 min at room temperature, with PBS (3  2 min), a biotinylated anti-rat IgG secondary antibody solution (1:50)
the mixture was centrifugated and collected via centrifugal filter (30 k cutoff). The was applied, and the mixture was incubated for 30 min at room temperature. The
particles were redispersed in PBS buffer solution and collected after centrifuging slides were rinsed again with PBS and incubated with streptavidineHRP solution for
three times. Subsequently, 0.1 ml of RGD peptide (2 mg/ml) was added in 1 ml of 30 min at room temperature. After another washing cycle, the slides were developed
particle solution. The mixture was stirred for 2 h at room temperature. Finally, the with DAB substrate solution until the desired color intensity was reached. Then the
particles were purified by PD-10 desalting column and kept in PBS solution. The resulting slides were subjected to Prussian blue staining. The slides were immersed in
IONPs were observed by transmission electron microscopy (TEM). The hydrody- staining solution (20% hydrochloric acid and 10% potassium ferrocyanide solution
namic diameter of IONPs and IONP-RGD were also analyzed by dynamic light mixture, 1:1 volume ratio) for 20 min, and counterstained with nuclear fast red for
scattering (DLS). 5 min. Afterwards, the slides were dehydrated consecutively with 90%, 95% and 100%
EtOH (3 min each), cleared with xylene, and mounted with Permount medium.
2.3. Orthotopic brain tumor model Double staining with Prussian blue and murine integrin b3 (CD61) was conducted
using the same protocol with the exception of the primary antibodies used. For CD61
The procedures for developing an orthotopic brain tumor model were per- staining, hamster anti-mouse CD61 mAb and biotinylated anti-hamster IgG
formed according to a protocol approved by the National Institutes of Health Clinical secondary antibody solution were used.
Center Animal Care and Use Committee (NIH CC/ACUC). Briefly, female athymic
nude mice (4e6 weeks) were intracranially injected with 1  105 U87MG cells in the 2.6.2. Fluorescence staining and image analysis
right frontal lobe at coordinates 1.5 mm lateral from the bregma, 0.5 mm anterior, Frozen tumor tissue slices (5 mm) were fixed with cold acetone for 20 min and
and 2.5 mm intraparenchymal. Tumor cells were allowed to engraft for the indicated dried in the air for 30 min at room temperature. After blocking with 2% BSA for
time points, and then in vivo MRI were performed to monitor tumor growth. 30 min, the sections were incubated with primary antibody for 2 h at room
temperature and then visualized with dye-conjugated secondary antibodies (1:400).
2.4. Prussian blue staining of IONP-RGD labeled HUVECs The following primary antibodies against different target antigens were used: rat
anti-mouse CD31 antibody (1:300) and hamster anti-mouse CD61 antibody (1:100).
Approximately 1  105 HUVECs were seeded in each cell culture chamber for After washing 3  5 min with PBS, the whole slides were mounted with DAPI-
growth overnight. After incubation with 50 mg/ml IONPs or IONP-RGD for the containing mounting medium. Fluorescence images were acquired with an epi-
indicated time points, the cells were washed three times with PBS buffer. Subse- fluorescence microscope (Olympus, X81). Quantitative analysis of positive areas of
quently, the cells were stained with Prussian blue solution containing 20% (v/v) CD31 and CD 61epositive vessels was done using Image J. In this context, at least
hydrochloric acid and 10% (v/v) potassium ferrocyanide solution. After incubation three randomly selected vision fields of each tumor were analyzed.
for 40 min at room temperature, the cells were washed twice with PBS buffer and To visualize and quantify integrin avb3 expressed by the murine tumor vessels
were subjected to incubation with a fast red nuclear staining solution for 10 min. and the tumor stroma, immnohistochemistry was performed following a previous
Then, the consecutive dehydrations were performed by 70%, 90% and 100% ethanol, report [19].
5416 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422

2.7. Statistical evaluation 3.2. IONPs and IONP-RGD uptake in HUVEC

Statistical analysis was performed using the Student’s two-tailed t test with p
values 0.05 considered statistically significant.
To assess the specific binding of particles, the uptake of IONPs,
IONP-RGD, and IONP-RGD plus free RGD to block the target (IONP-
RGD þ block) was performed in vitro using Prussian blue staining
3. Results
(Fig. 3). After incubation with HUVECs for 10 min, uptake of small
amounts of IONP-RGD was observed, whereas there was no
3.1. Fabrication and characterization of IONPs with RGD peptides
significant uptake during IONPs and IONP-RGD þ block treatments.
The uptake of particles was obviously increased in all three groups
The conjugation of IONPs and c(RGDyK) (RGD) was successfully
after incubation for 1 h, showing the internalization to be time-
performed based on the procedures of Fig. 1. Briefly, the IONPs were
dependent. However, IONP-RGD þ block effectively reduced the
firstly synthesized by modifying the previous methods [19e21].
amount of blue granules in the cytoplasm of HUVECs.
The amphiphilic copolymer with amino groups was subsequently
To further confirm the uptake of particles, an MRI phantom
added to interact with the original oleate coating and form
study was also performed after incubation with HUVECs for 10 min
a bilayered structure, rendering the particles water soluble. Such
and 1 h. Quantitative analysis of the relaxation time of samples
IONPs can be dispersed in various buffer solutions for months
showed that the results were consistent with the Prussian blue
without precipitation, indicating good stability. The as-synthesized
staining. These results indicate that the accumulation of the
IONPs were highly-ordered and spherical with a diameter of
particles was specifically mediated by the integrin avb3 binding.
around 10 nm (Fig. 2A). To evaluate the MRI contrast effect of these
particles, a phantom study was also performed with IONPs of
elevated concentrations. As shown in Fig. 2B, a clear T2 signal 3.3. Evaluation of integrin avb3 binding in an orthotopic
reduction effect was displayed in a concentration-dependent glioblastoma model
manner. The linear relation of the concentration of IONPs and T2
signal was further constructed and demonstrated that a r2 value of To better understand the capability of the particles to bind
around 187 mM1s1, which is much higher than that of Feridex integrin avb3 in vivo, an orthotopic glioblastoma model was firstly
(130 mM1s1) under the same condition (data not shown). To established by inoculating 1  105 U87MG cells in the right frontal
conjugate the IONPs with RGD peptides, a crosslinker with NHS lobe. The tumor model was confirmed by H&E histological exami-
esters on both ends was introduced to link the amino groups on the nation after tumor inoculation for 3e4 weeks (Supporting infor-
particles with RGD (Fig. 1). The successful conjugation of IONPs and mation Fig. S1). The mice were subsequently subjected to MRI
RGD (IONP-RGD) was confirmed by DLS analysis. The average studies on a 7.0 T small animal MRI system. IONP-RGD, IONPs and
hydrodynamic diameter was increased from 33.4  0.3 nm of the IONP-RGD þ block were intravenously administrated at a dose of
IONPs to 44.7  0.6 nm of the IONPs-RGD (Fig. 2C). 5 mg Fe/kg, and T2*-weighted fast spin-echo images pre- and 6 h

Fig. 1. Schematic illustration of the conjugation of IONPs with RGD peptides.

 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422 5417

Fig. 2. Characterization of IONPs and IONP-RGD (A) TEM of IONPs (B) T2*-weighted phantom images of IONPs at different Fe concentration. Top, T2*-weighted phantom images;
Bottom, 1/T2 vs. Fe concentration curve of IONPs.

Fig. 3. Cellular uptake of particles in HUVECs. The IONPs, IONP-RGD and IONP-RGD þ block were incubated with HUVECs for 10 min and 1 h, respectively. Cellular uptake was
evaluated by (A) Prussian blue staining and (B) T2-weighted phantom images (C) Quantitative analysis of relaxation time of T2*-weighted phantom images.
5418 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422

post-injection were acquired. As displayed in Fig. 4A, a dramatic surface. These results indicate that the particle homing is indeed
signal drop was observed at the tumor area (indicated by white specifically mediated by RGDeblood vessel integrin interaction.
circles) after the injection of IONP-RGD, while only a marginal
signal drop was witnessed in mice injected with IONPs. Moreover, 3.4. Effect of VEGF121/rGel on the integrin avb3 expression of tumor
there was no obvious signal drop observed in the IONP- angiogenic blood vessels
RGD þ block group. Quantitative analysis of T2 relaxation time also
indicated that the post-injection signal was definitely decreased To study the effect of VEGF121/rGel on orthotopic U87MG tumor
compared to the pre-injection signal of the IONP-RGD group growth, two doses of VEGF121/rGel were firstly administered
(p < 0.01), whereas the signal of both IONPs and IONP-RGD þ block intraperitoneally (i.p.) at day 0 and day 2 after tumor inoculation.
group exhibited no obvious changes. The results demonstrate that Then, the animals were imaged post-treatment for 4 days. Fig. 6A
the binding capacity of IONPs was dependent on the interaction shows a time-dependent increase of tumor volume in both
between RGD peptides and integrin avb3. VEGF121/rGel treated and untreated groups. However, U87MG
To further confirm the targeting specificity, the mice were tumor was inhibited by VEGF121/rGel compared with the control
sacrificed 6 h post-injection. After preparing frozen tissue slices, group. A significant difference in tumor volume was observed at 7
CD31/CD61 (integrin avb3) and Prussian blue double staining was days between the treatment group and the control group
performed. The perfect overlap of IONPs (blue) and CD61 (brown) (p < 0.01).Furthermore, there was no significant body weight
positive staining was observed in tumor blood vessels of the IONP- loss observed during the treatment process, indicating that
RGD group, while little superimposition of IONPs and CD61 were VEGF121/rGel had no observable side effects at the dosage used
found in the IONPs and IONP-RGD þ block groups (Fig. 5). Addi- during this study.
tionally, Prussian blue and CD31 double staining confirmed the To validate the inhibition effect of VEGF121/rGel histologically,
particles were mainly attached to the interior of the blood vessel the presence of tumor angiogenic blood vessels and integrin avb3

Fig. 4. T2*-weighted MR images of nude mice bearing orthotopic U87MG glioblastoma (A) T2*-weighted MR images were acquired before and after injection of IONPs, IONP-RGD and
IONP-RGD þ block, respectively (B) Quantitative analysis of T2*-weighted MR images in tumor areas.
 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422 5419

Fig. 5. Prussian blue and CD31/CD61 double staining of the tumor sections. At 6 h post-injection, the mice were sacrificed and frozen tissue slices were prepared.

expression were investigated. As demonstrated by immunohisto- and quantify murine integrin avb3 expression by immunofluores-
chemistry (Fig. 6B), the untreated group exhibited significantly cence in frozen tissue slices (Fig. 6C). Intact tumor blood vessels
higher CD31-positive area fractions, which are highly related to were observed in the untreated group, whereas an interrupted
tumor angiogenic blood vessels, than in the VEGF121/rGel treated distribution was shown in the treated group. It was also observed
group. The anti-mouse-CD61 antibody was also used to visualize that the integrin expression was co-localized with the distribution

Fig. 6. The effect of VEGF121/rGel on the integrin avb3 expression of tumor angiogenic blood vessels (A) Tumor growth curves of untreated and VEGF121/rGel treated groups were
analyzed by MRI (B) CD31 staining of tumor angiogenic blood vessels by immnohistochemistry (C) CD31/CD61 double staining of tumor angiogenic blood vessels and integrin avb3
expression on frozen tissue slices (D) Quantitative analysis of CD31- and CD 61-positive area using Image J software (5 mice each group).
5420 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422

of tumor blood vessels. For the untreated group, the positive area of demonstrate that IONP-RGD can be used for noninvasive moni-
both CD31 and CD61 expression was significantly higher than the toring of therapeutic response by orthotopic GBM using MRI.
treated group (p < 0.05) (Fig. 6D). These results confirm the
VEGF121/rGel damaged tumor angiogenic blood vessels and 4. Discussion
inhibited integrin expression.
In the present study, an amphiphilic copolymer was utilized to
3.5. Dynamic monitoring of orthotopic glioblastoma therapy coat oleate-coated IONPs and render the particles water soluble
response by MRI and functionalizable. Our previous report demonstrates that such
phase transfer and surface modification are straightforward, with
To noninvasively monitor the therapy response of orthotopic high throughput and high yield [19]. In view of amino groups
GBM, VEGF121/rGel was used to treat GBM by destroying tumor coating the IONPs surface, BS3 crosslinker was introduced to
angiogenic blood vessels. After treatment for 4 days, T2*-weighted conjugate the amino groups of the RGD peptide to the IONPs. To
images of pre- and 6 h post-injection were acquired in both reduce the chance of crosslinking among nanoparticles, the
untreated and VEGF121/rGel groups. As shown in Fig. 7A, compared concentration of IONPs was kept at 0.5 mg/ml during the conju-
with pre-injection signals of IONP-RGD, the post-injection signal in gation procedure. With such a controllable conjugation, we found
tumor areas was obviously decreased in the treated group. the diameter of about 90% of IONPs was less than 50 nm (Fig. 2C),
However, there was no significant signal change in the untreated which is suitable for in vivo tumor targeting.
group. Quantitative analysis found that the relaxation time (1/T2) of Cell adhesion molecules such as integrin avb3 are overexpressed
the untreated group was 1.3  0.27 fold of the treated group in activated and proliferating endothelial cells [22]. Various tumor
(p < 0.01) (Fig. 7B). cells, such as gliomas [23,24], ovarian carcinomas [25], and breast
To further understand the monitoring information by MRI, carcinomas [26], also highly express integrin avb3. Therefore, the
a dynamic signal reduction was investigated in the untreated and ability to noninvasively detect integrin avb3 expression in living
treated groups. Dynamic T2*-weighted data was collected at various subjects would allow a better characterization of tumors and help
time points from 0 to 70 min. The representative images over time to identify tumor regions with higher aggressiveness. RGD peptides
are displayed in Fig. 8A. A fast and obvious accumulation of IONP- are one of the most popular ligands that target integrin avb3 [27].
RGD was observed in tumor areas of the untreated group, Therefore, RGD conjugates can be designed for specific tumor tar-
showing a clear profile of GBM with time-dependent signal geting without the need of extravasation and diffusion in the tumor
intensity. Maximum T2 signal of the untreated group was observed interstitial space. When conjugated to radiotracers [28], fluorescent
at 2 min after injection of IONP-RGD, which subsequently markers [29,30], or magnetic NPs [23,31], RGD has been used for
decreased over time (Fig. 8B). Compared to baseline, only a small the detection of tumor angiogenesis and tumor metastasis. We and
signal decrease in the treated group was detected. These results others have shown the specific targeting of angiogenic blood

Fig. 7. Monitoring of therapeutic response of VEGF121/rGel using IONP-RGD in orthotopic U87MG glioblastoma model (A) T2*-weighted MR images of untreated and VEGF121/rGel
treated groups were acquired before and after injection of IONP-RGD (B) Quantification analysis of T2*-weighted MR images. The white circle indicates location of the implanted
tumor (4 mice each group).
 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422 5421

Fig. 8. Dynamic monitoring of therapeutic response of VEGF121/rGel using IONP-RGD in orthotopic U87MG glioblastoma model (A) T2*-weighted MR dynamic images of untreated
and VEGF121/rGel treated groups were acquired before and after injection of IONP-RGD (B) Quantitative analysis of T2*-weighted MR dynamic images (4 mice each group).

vessels [32,33] or tumor cells [19] with IONP-RGD, which was cause large reductions in tumor volume in a short time scale, even if
specifically dependent on RGD peptides binding to integrin avb3 they are effective, which reduces the usefulness of anatomic
(Fig. 4). Most importantly, IONP-RGD can specifically target ortho- measures in monitoring response to such agents. To better monitor
topic GBM across the blood brain barrier [12]. The BBB represents treatment efficacy, a critical challenge in future clinical application
one of the most exclusive biological barriers encountered in the of VEGF121/rGel and other antiangiogenic therapies is to develop
treatment of neurologic diseases, limiting the delivery of a vast effective noninvasive in vivo imaging techniques. Recently, we
majority of potential diagnostic agents and therapeutics. Recent applied PET probes to noninvasively monitor glucose metabolism,
studies have indicated that BBB passage by NPs was dictated by cellular proliferation, tumor hypoxia and angiogenesis during
hydrodynamic size. For example, a recent study indicated that VEGF121/rGel therapy of breast cancer [35]. However, it is still
dendrimer nanoparticles with sizes less than 12 nm were able to a challenge to monitor therapeutic outcomes of orthotopic brain
permeate the tumor BBB, while larger ones could not [34]. tumors. This study explored the feasibility of monitoring the ther-
However, Veiseh et al. demonstrated that IONPs with diameters of apeutic response of VEGF121/rGel therapy in an orthotopic GBM
33 nm have the ability to cross the BBB and specifically target brain model using IONPs as MRI contrast agents. To better demonstrate
tumors [12]. Herein, we showed the excellent tumor targeting the advantages of noninvasive monitoring of antiangiogenic
effect of IONP-RGD with a hydrodynamic diameter of 45 nm (Fig. 2). therapy response over common anatomical approaches that are
Therefore, the permeability of NPs to BBB is dependent not only on based on measurements of tumor size, a time point of 4 days post-
the hydrodynamic diameter, but also on other properties of the NP. treatment was chosen to study the therapeutic outcome of
The specific mechanisms for BBB penetration by NPs have yet to be VEGF121/rGel therapy by MRI. As shown in Fig. 7, the definite
elucidated. One possible explanation is that the integrity of the BBB therapeutic outcome of VEGF121/rGel therapy is observed, even
may be damaged in orthotopic GBM. though the tumor volume was not significantly changed compared
VEGF121/rGel fusion protein is a vascular disruptive agent with the control (p > 0.05). Such results indicate the potential
composed of the VEGF-A isoform VEGF121 and a recombinant plant advantages of IONPs to noninvasively monitor the early response of
toxin gelonin (rGel) [14]. This novel vasculature-targeting fusion tumors to therapy in the clinic. The monitoring ability of the
protein has been shown to inhibit various tumor growth with low therapeutic response was further confirmed by dynamic MRI
systemic toxicity, including melanoma, GBM, prostate, breast, and (Fig. 8). Fig. 8A showed a significant difference of 1/T2 signal
bladder tumor models [15]. Our previous study assessed the anti- between the untreated and VEGF121/rGel therapy treated group.
tumor effects of VEGF121/rGel in an orthotopic GBM mouse model The difference in the 1/T2 signal occurs because the signal is
by use of noninvasive in vivo bioluminescence imaging, MRI, and dependent on the density of tumor blood vessels. In other words,
PET [16]. In this current study, the growth of orthotopic GBM was the results also demonstrate an excellent antiangiogenic therapy of
also inhibited by VEGF121/rGel (Fig. 6A). It was shown the VEGF121/ VEGF121/rGel. Furthermore, the 1/T2 signal of the tumor area in the
rGel disrupted tumor angiogenic blood vessels and inhibited untreated group was immediately increased post-injection of
integrin avb3 expression (Fig. 6BeD). However, many new anti- IONP-RGD, and then decreased until reaching equilibrium over
angiogenic drugs are principally cytostatic and do not necessarily time. An explanation for this phenomenon is that the signal was
5422 F. Zhang et al. / Biomaterials 33 (2012) 5414e5422

mainly derived from the high concentration of IONP-RGD in blood [11] Tzeng SY, Guerrero-Cazares H, Martinez EE, Sunshine JC, Quinones-
Hinojosa A, Green JJ. Non-viral gene delivery nanoparticles based on poly(b-
at early time points post-injection, and not due to the particles
amino esters) for treatment of glioblastoma. Biomaterials 2011;32(23):
binding to integrin avb3 of the blood vessels. With the clearance of 5402e10.
particles, only IONP-RGD binding to blood vessels remained in [12] Veiseh O, Sun C, Fang C, Bhattarai N, Gunn J, Kievit F, et al. Specific targeting of
tumor areas and the signal did not change over time. brain tumors with an optical/magnetic resonance imaging nanoprobe across
the blood-brain barrier. Cancer Res 2009;69(15):6200e7.
[13] Dilnawaz F, Singh A, Mewar S, Sharma U, Jagannathan NR, Sahoo SK. The
5. Conclusion transport of non-surfactant based paclitaxel loaded magnetic nanoparticles
across the blood brain barrier in a rat model. Biomaterials 2012;33(10):
2936e51.
We have successfully accomplished in vivo noninvasive moni- [14] Veenendaal LM, Jin H, Ran S, Cheung L, Navone N, Marks JW, et al. In vitro and
toring of therapeutic response to VEGF121/rGel therapy using IONPs in vivo studies of a VEGF121/rGelonin chimeric fusion toxin targeting the
neovasculature of solid tumors. Proc Natl Acad Sci U S A 2002;99(12):
as MRI contrast agents in an orthotopic GBM model. The results 7866e71.
show that IONP-RGD not only target GBM by binding to integrin [15] Ran S, Mohamedali KA, Luster TA, Thorpe PE, Rosenblum MG. The vascular-
avb3 of angiogenic blood vessels, but also cross BBB to penetrate the ablative agent VEGF121/rGel inhibits pulmonary metastases of MDA-MB-231
breast tumors. Neoplasia 2005;7(5):486e96.
tumor. Most importantly, we also demonstrated the feasibility of
[16] Hsu AR, Cai W, Veeravagu A, Mohamedali KA, Chen K, Kim S, et al. Multi-
using IONP-RGD for noninvasive monitoring of early tumor modality molecular imaging of glioblastoma growth inhibition with
responses to antiangiogenic therapies, showing higher sensitivity vasculature-targeting fusion toxin VEGF121/rGel. J Nucl Med 2007;48(3):
than common approaches based on measurements of tumor size. 445e54.
[17] Kim S, Mohamedali KA, Cheung LH, Rosenblum MG. Overexpression of bio-
Undoubtedly, such results are important in clinical oncology to logically active VEGF121 fusion proteins in Escherichia coli. J Biotechnol 2007;
reduce side effects and save costs during treatment, especially with 128(3):638e47.
the growing number of alternative treatment regimens that are [18] Rosen BR, Belliveau JW, Vevea JM, Brady TJ. Perfusion imaging with NMR
contrast agents. Magn Reson Med 1990;14(2):249e65.
only effective in select subgroups of patients. [19] Chen K, Xie J, Xu H, Behera D, Michalski MH, Biswal S, et al. Triblock copol-
ymer coated iron oxide nanoparticle conjugate for tumor integrin targeting.
Acknowledgment Biomaterials 2009;30(36):6912e9.
[20] Park J, An K, Hwang Y, Park JG, Noh HJ, Kim JY, et al. Ultra-large-scale
syntheses of monodisperse nanocrystals. Nat Mater 2004;3(12):891e5.
This work was supported in part, by the Intramural Research [21] Yu WW, Falkner JC, Yavuz CT, Colvin VL. Synthesis of monodisperse iron oxide
Program of the NIBIB, NIH, NCI R41CA137960-01, the Henry nanocrystals by thermal decomposition of iron carboxylate salts. Chem
Commun (Camb) 2004;(20):2306e7.
M. Jackson Foundation, the International Cooperative Program of [22] Weis SM, Cheresh DA. Tumor angiogenesis: molecular pathways and thera-
the National Science Foundation of China (NSFC) (81028009), and peutic targets. Nat Med 2011;17(11):1359e70.
the NSFC grant No. 30930028. [23] Kiessling F, Huppert J, Zhang C, Jayapaul J, Zwick S, Woenne EC, et al. RGD-
labeled USPIO inhibits adhesion and endocytotic activity of avb3-integrin-
expressing glioma cells and only accumulates in the vascular tumor
Appendix A. Supplementary material compartment. Radiology 2009;253(2):462e9.
[24] Schnell O, Krebs B, Carlsen J, Miederer I, Goetz C, Goldbrunner RH, et al.
Imaging of integrin avb3 expression in patients with malignant glioma by
Supplementary data related to this article can be found online at [18F] Galacto-RGD positron emission tomography. Neuro Oncol 2009;11(6):
doi:10.1016/j.biomaterials.2012.04.032. 861e70.
[25] Cruet-Hennequart S, Maubant S, Luis J, Gauduchon P, Staedel C, Dedhar S. av
integrins regulate cell proliferation through integrin-linked kinase (ILK) in
References ovarian cancer cells. Oncogene 2003;22(11):1688e702.
[26] Yang C, Hayashida T, Forster N, Li C, Shen D, Maheswaran S, et al. The integrin
[1] Weber WA. Assessing tumor response to therapy. J Nucl Med 2009;1(50 avb3/5 ligand MFG-E8 is a p63/p73 target gene in triple-negative breast
Suppl):1Se10S. cancers but exhibits suppressive functions in ER(þ) and erbB2(þ) breast
[2] Husband JE, Schwartz LH, Spencer J, Ollivier L, King DM, Johnson R, et al. cancers. Cancer Res 2011;71(3):937e45.
Evaluation of the response to treatment of solid tumours - a consensus [27] Ye Y, Chen X. Integrin targeting for tumor optical imaging. Theranostics 2011;
statement of the International Cancer Imaging Society. Br J Cancer 2004; 1:102e26.
90(12):2256e60. [28] Zhou Y, Chakraborty S, Liu S. Radiolabeled Cyclic RGD peptides as radiotracers
[3] Han Z, Fu A, Wang H, Diaz R, Geng L, Onishko H, et al. Noninvasive assessment for imaging tumors and Thrombosis by SPECT. Theranostics 2011;1:58e82.
of cancer response to therapy. Nat Med 2008;14(3):343e9. [29] Hsu AR, Hou LC, Veeravagu A, Greve JM, Vogel H, Tse V, et al. In vivo near-
[4] Day SE, Kettunen MI, Gallagher FA, Hu DE, Lerche M, Wolber J, et al. Detecting infrared fluorescence imaging of integrin avb3 in an orthotopic glioblas-
tumor response to treatment using hyperpolarized 13C magnetic resonance toma model. Mol Imaging Biol 2006;8(6):315e23.
imaging and spectroscopy. Nat Med 2007;13(11):1382e7. [30] Chen X, Conti PS, Moats RA. In vivo near-infrared fluorescence imaging of
[5] de Langen AJ, van den Boogaart V, Lubberink M, Backes WH, Marcus JT, van integrin avb3 in brain tumor xenografts. Cancer Res 2004;64(21):8009e14.
Tinteren H, et al. Monitoring response to antiangiogenic therapy in non-small [31] Gianella A, Jarzyna PA, Mani V, Ramachandran S, Calcagno C, Tang J, et al.
cell lung cancer using imaging markers derived from PET and dynamic Multifunctional nanoemulsion platform for imaging guided therapy evaluated
contrast-enhanced MRI. J Nucl Med 2011;52(1):48e55. in experimental cancer. ACS Nano 2011;5(6):4422e33.
[6] Lee N, Hyeon T. Designed synthesis of uniformly sized iron oxide nano- [32] Zhang C, Jugold M, Woenne EC, Lammers T, Morgenstern B, Mueller MM, et al.
particles for efficient magnetic resonance imaging contrast agents. Chem Soc Specific targeting of tumor angiogenesis by RGD-conjugated ultrasmall
Rev 2012;41(7):2575e89. superparamagnetic iron oxide particles using a clinical 1.5-T magnetic reso-
[7] Huang J, Zhong X, Wang L, Yang L, Mao H. Improving the magnetic resonance nance scanner. Cancer Res 2007;67(4):1555e62.
imaging contrast and detection methods with engineered magnetic nano- [33] Li J, Zhang C, Yang K, Liu P, Xu LX. SPIO-RGD nanoparticles as a molecular
particles. Theranostics 2012;2(1):86e102. targeting probe for imaging tumor angiogenesis using synchrotron radiation.
[8] Zhen Z, Xie J. Development of manganese-based nanoparticles as contrast J Synchrotron Radiat 2011;18(Pt 4):612e6.
probes for magnetic resonance imaging. Theranostics 2012;2(1):45e54. [34] Sarin H, Kanevsky AS, Wu H, Brimacombe KR, Fung SH, Sousa AA, et al.
[9] Huang X, Zhuang J, Chen D, Liu H, Tang F, Yan X, et al. General strategy for Effective transvascular delivery of nanoparticles across the blood-brain tumor
designing functionalized magnetic microspheres for different bioapplications. barrier into malignant glioma cells. J Transl Med 2008;6:80.
Langmuir 2009;25(19):11657e63. [35] Yang M, Gao H, Sun X, Yan Y, Quan Q, Zhang W, et al. Multiplexed PET probes
[10] Gajbhiye V, Jain NK. The treatment of Glioblastoma Xenografts by surfactant for imaging breast cancer early response to VEGF/rGel treatment. Mol Pharm
conjugated dendritic nanoconjugates. Biomaterials 2011;32(26):6213e25. 2011;8(2):621e8.