Magnetic Resonance Imaging Determination of Tumor Grade

and Early Response to Temozolomide in a Genetically

Engineered Mouse Model of Glioma
Patrick McConville, Dolores Hambardzumyan, Jonathan B. Moody, et al.
Clin Cancer Res 2007;13:2897-2904.

Updated version

Access the most recent version of this article at:

http://clincancerres.aacrjournals.org/content/13/10/2897

Cited Articles

This article cites by 48 articles, 13 of which you can access for free at:
http://clincancerres.aacrjournals.org/content/13/10/2897.full.html#ref-list-1

Citing articles

This article has been cited by 8 HighWire-hosted articles. Access the articles at:
http://clincancerres.aacrjournals.org/content/13/10/2897.full.html#related-urls

E-mail alerts
Reprints and
Subscriptions
Permissions

Sign up to receive free email-alerts related to this article or journal.

To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at pubs@aacr.org.
To request permission to re-use all or part of this article, contact the AACR Publications
Department at permissions@aacr.org.

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

Imaging, Diagnosis, Prognosis

Magnetic Resonance Imaging Determination of Tumor Grade

and Early Response to Temozolomide in a Genetically
Engineered Mouse Model of Glioma
Patrick McConville,1 Dolores Hambardzumyan,3 Jonathan B. Moody,1 Wilbur R. Leopold,1 Alicia R. Kreger,1
Michael J. Woolliscroft,1 Alnawaz Rehemtulla,2 Brian D. Ross,2 and Eric C. Holland3

Abstract

Purpose: The median survival for patients diagnosed with glioblastoma multiforme, the most
common type of brain tumor, is less than 1 year. Animal glioma models that are more predictive
of therapeutic response in human patients than traditional models and that are genetically and
histologically accurate are an unmet need. The nestin tv-a (Ntv-a) genetically engineered mouse
spontaneously develops glioma when infected with ALV-A expressing platelet-derived growth
factor, resulting in autocrine platelet-derived growth factor signaling.
Experimental Design: In the Ntv-a genetically engineered mouse model, T2-weighted and
T1-weighted, contrast-enhanced magnetic resonance images were correlated with histology,
glioma grade (high or low), and survival. Magnetic resonance imaging (MRI) was therefore used
to enroll mice with high-grade gliomas into a second study that tested efficacy of the current
standard of care for glioma, temozolomide (100 mg/kg qdx5 i.p., n = 13).
Results: The Ntv-a model generated a heterogeneous group of gliomas, some with high-grade
growth rate and histologic characteristics and others with characteristics of lower-grade gliomas.
We showed that MRI could be used to predict tumor grade and survival. Temozolomide treatment
of high-grade tv-a gliomas provided a 14-day growth delay compared with vehicle controls.
Diffusion MRI measurement of the apparent diffusion coefficient showed an early decrease in
cellularity with temozolomide, similar to that observed in humans.
Conclusions: The use of MRI in the Ntv-a model allows determination of glioma grade
and survival prediction, distribution of mice with specific tumor types into preclinical trials,
and efficacy determination both by tumor growth and early apparent diffusion coefficient
response.

Gliomas are the most common form of brain tumor, with

more than half belonging to the most aggressive subtype
(i.e., glioblastoma multiforme) at initial diagnosis. Although
aggressive combination therapies, including surgery, radiation,
and chemotherapy, are commonly used, median survival for
patients diagnosed with glioblastoma multiforme is less than
1 year (1). Part of the reason for the deficit in significant
progress has been the lack of predictive preclinical models.
Development and validation of models that accurately recapitulate the complex-activated signaling pathways involved in glial

Authors Affiliations: 1MIR Preclinical Services, Inc.; 2Center for Molecular

Imaging, University of Michigan School of Medicine, Ann Arbor, Michigan; and
3
Departments of Surgery (Neurosurgery), Neurology, and Cancer Biology
Genetics, Memorial Sloan Kettering Cancer Center, NewYork, NewYork
Received 12/27/06; revised 2/21/07; accepted 2/28/07.
Grant support: National Cancer Institute grant 5R01CA099489 (E. Holland).
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
with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: Eric C. Holland, Department of Cancer Biology and
Genetics, Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York,
NY10021. Phone: 212-639-3017; Fax: 646-422-2062; E-mail: hollande@ mskcc.org.
F 2007 American Association for Cancer Research.
doi:10.1158/1078-0432.CCR-06-3058

www.aacrjournals.org

transformation and tumor progression has been hindered due

to lack of available methods for selecting and placing tumored
mice on study and for efficient end point determination.
However, recent developments in preclinical imaging technologies are facilitating the development of genetically engineered
mouse (GEM) models of glioma (2, 3).
RCAS/tv-a technology relies on somatic gene transfer through
infection by RCAS viral vectors derived from the avian
retrovirus A (ALV-A) in mice expressing the gene for the RCAS
receptor (tv-a). The nestin tv-a (Ntv-a) mouse expresses tv-a
under the control of the nestin promoter in glial progenitors.
The inherent flexibility of this model with respect to choice
of the oncogene(s) that induce(s) tumor formation is a key
advantage, providing opportunities not only for study of
cytotoxic therapies but also new classes of anticancer drugs
that target specific oncogenic pathways. When infected with
ALV virus encoding platelet-derived growth factor (PDGF), the
Ntv-a mice spontaneously develop gliomas by 3 weeks of age in
almost 100% of appropriately infected mice, with 30% of
these tumors displaying high-grade histologic features (4).
Recent results showing the similarity between Ntv-a gliomas
and human gliomas (5) suggest that the PDGF-driven tv-a
model may be a more predictive model for preclinical
evaluation of therapies than the currently available human
xenograft or syngeneic rodent models.

2897

Clin Cancer Res 2007;13(10) May 15, 2007

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

Imaging, Diagnosis, Prognosis

This model has been used in several preclinical trials looking

at the effects of blockade of PDGF receptor, mammalian target
of rapamycin, and Akt (3, 4, 6). Finally, This model has also
been used to look at Akt inhibition alone or in combination
with temozolomide (6). However, all of these trials were short
(7 days), and the effect of these inhibitors and temozolomide
over this period was primarily loss of proliferation and not cell
kill or apoptosis.
Preclinical evaluation of agents using these models would
greatly benefit from the establishment of additional quantitative end points for growth and treatment response that could
be correlated with survival and translated into clinical trials.
Magnetic resonance imaging (MRI) and spectroscopy offer
excellent opportunities for noninvasive determination of
transgenic tumor induction, and growth and response to
treatment dynamically over time in individual animals and
cohorts of animals. There are a multitude of quantifiable
MR-detectable variables that may provide valuable insights into
the pathophysiology and treatment response of the tumor.
Among these, proton MR spectroscopy (7 9), perfusion/
vascular permeability (10 13), and water-apparent diffusion
(14) measurements have shown promise. Perfusion/permeability MRI has been used for brain tumor characterization and
for monitoring vascular response to a variety of interventions
(15 18). Diffusion MRI allows the movement (Brownian
motion) of water within the tumor to be quantified (14).
Therapy-induced cell kill is reflected as a net decrease in the
restricted movement of water within a tumor, which can be
detected as an increase in apparent diffusion coefficient (ADC)
values (19). This has been recently reported in a variety of
tumor models and more recently in brain tumor clinical studies
(19 46). Thus, MRI biomarkers such as these provide
opportunities to investigate their sensitivity and specificity in
preclinical studies and to subsequently translate them along
with promising agents into clinical trials.
In this study, MRI was firstly used to characterize growth and
contrast enhancement (vascular permeability) during tv-a
glioma tumorigenesis. MRI identified high- and low-grade
brain tumors wherein high-grade tumors were distinguished by
faster growth rates and were contrast enhancing compared with
low-grade tumors that grew more slowly and did not enhance.
MRI was then used to enroll mice with high-grade tumors into
a preclinical trial using temozolomide. High-grade gliomas
were found to be responsive to temozolomide treatment as
quantified by a decrease in MRI-detected growth rate and
enhanced overall survival. Diffusion MRI revealed early
increases in tumor ADC values in treated tumors before
significant changes in tumor growth rates between treated
and control animals, revealing its potential as a clinically
translatable imaging biomarker for predicting therapeutic
outcome. These data further show the utility of the Ntv-a
model in preclinical trials and dMRI methodology to identify
patients responding to temozolomide.

Materials and Methods

Generation of mouse brain tumors. Ntv-a mice were injected i.c. with
104 DF-1 cells transfected with RCAS-PDGF retroviral vectors within
24 h of birth, as previously described (3). Two different groups of
animals were injected in the two separate studies described below: the
tumor grade study (group 1, n = 35) and the temozolomide efficacy

Clin Cancer Res 2007;13(10) May 15, 2007

study (group 2, n = 25). Mice were monitored daily for symptoms of

tumor development. From 4 to 5 weeks of age, all animals underwent
MRI screening for tumor development.
MRI examination of mice. Group 1 was used to characterize glioma
development using MRI. In this study, 35 mice confirmed by MRI to
have gliomas at 3 weeks of age underwent weekly T2-weighted and
contrast-enhanced T1-weighted (CE) MRI, using multi-slice fast spinecho and spin-echo sequences, respectively. MRI was done using a
Varian Unity Inova MRI system equipped with a Magnex 7T horizontal
bore magnet. dMRI was done using a motion-compensated, motioncorrected, isotropic diffusion sequence using a previously described
method (37). Briefly, a trace diffusion weighted multi-slice spin-echo
sequence (with motion compensation and a navigator echo) was used
to acquire 13 slices with two different diffusion weightings [b 1 = 100
s/mm2 (single acquisition) and b 2 = 1,248 s/mm2 (two averages)].
Images were acquired using contiguous transaxial slices with a
thickness of 0.75 mm, image matrix 128
128 (zero filled to 256),
field of view 20
20 mm, repetition time 2 s, and echo time 60 ms
for an acquisition time of 12 min. During all MRI procedures, animals
were anaesthetized using isofluorane, and body temperature was
maintained at 37jC using a heated re-circulating water pad. T2weighted images were acquired using a fast spin-echo sequence with
a TR/TE = 4,000/60, 8 echoes, echo spacing of 12.5 ms, 17 contiguous
transaxial slices, 0.5-mm slice thickness, and an acquisition time of
2 min. T1-weighted images were acquired using a spin echo sequence
with TR/TE = 600/20, 11 contiguous slices of 0.5-mm thickness, and
an acquisition time of 2.5 min. The average total exam time for each
animal was 20 to 25 min.
All tumors were analyzed using a customized volumetric regionof-interest drawing tool in Matlab to segment tumor from normal
tissue, with separate regions of interest drawn for the diffusion and
T2- and T1-weighted images. Tumor boundaries were manually
determined based on hypo-intense regions in T2-weighted images
and low b value diffusion images, with large cystic regions excluded.
In this model, T2-weighted hypo-intensities have been previously
confirmed by histology to correlate with tumor tissue. The enhancing
tumor volume was delineated based on pre- and post-contrast
T1-weighted images, and a consideration of the total tumor volume,
based on the inherently co-registered T2-weighted images. Tumor
and enhancing tissue volumes were derived from these multi-slice
MR data sets over time for each individual animal as previously
described (47). At 7.5 weeks of age, half of the mice were sacrificed for
histology, with the remaining mice continuing on study for survival
analysis.
The second group of 25 mice that were selected as having highgrade gliomas at 4.5 weeks of age were divided into two groups
(DMSO vehicle control: n = 12 and temozolomide treated: 100 mg/kg
qdx5 i.p. beginning on day 1, n = 13) with equal distribution of tumor
size based upon MRI measurements. Temozolomide was dissolved
in DMSO (33mg/mL) over brief heating and sonication. Baseline
dMRI measurements were taken just before treatment (day 0) and
again on days 3, 7, 10, and 14 and weekly thereafter. This protocol
allowed characterization of tumor growth and ADC response. Survival
data were collected under approval from MIRACUC guidelines
for oversight, housing, monitoring for condition, and euthanasia
requirements.
Histopathology. Animals used for histologic analysis were euthanized, and brains were removed and fixed in 10% neutral buffered
formalin for 72 h. Fixed tissues were then transferred to 70% ethanol
for 48 h and embedded in paraffin. Formalin-fixed, paraffin-embedded
specimens were serially sectioned and slide mounted. The sections were
deparaffinized in histo-clear (Richard-Allan Scientific) and were passed
through graded alcohols before staining with H&E reagent. For
proliferating cell nuclear antigen staining antigen retrieval was done
by submersing slides in 0.01 mol/L citric acid (pH 6.0) in a staining jar
in a boiling water bath for 30 min. After a 0.3% H2O2 (Sigma)
treatment for 15 min, sections were incubated with blocking reagent

2898

www.aacrjournals.org

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

MRI of a GEM Model for Glioma

at room temperature, followed by 60-min incubation with biotinylated
anti-rabbit IgG at 1:200 dilution (Vectastain ABC kit; rabbit IgG) then
reacted with DAB Detection kit containing Blocker D, Copper D,
Inhibitor D, streptavidin-HRP D, DAB D, and DAB-H2O2 D (Ventana
Medical Systems) according to the manufacturer instructions.

Results

Fig. 1. ExampleT2-weighted MRI for high-grade glioma (A) and low-grade

glioma (B) with corresponding whole-mount H&E-stained sections. There was a
correlation between hypo-intense regions inT2-weighted MRI (white arrow) and
pseudopalisading necrosis. C, H&E, Hif1a, and proliferating cell nuclear antigen
(PCNA) staining showing hypoxia and lack of proliferation in and around
pseudopalisading necrosis. Bar, 100 Am. D, mean tumor burden for low-grade
(4) and high-grade (.) Ntv-a glioma; error bars represent SE. High-grade tumors
were greater in size at 4 wks after RCAS injection and grew at a greater rate than
low-grade tumors, as evidenced by the greater slope on the log-linear scale
compared with low-grade tumors.

Imaging characteristics divide gliomas into two groups and

predict tumor growth rate and aggressiveness. Intracerebral
tumors could be detected using MRI in about 80% of animals at
4 to 5 weeks of age. Tumors were delineated from normal brain
tissue as a hyperintense region in T2-weighted images. Tumor
volumes were defined for each image slice by drawing a region
of interest encompassing the hyperintense region that was
found to correspond well with tumor tissue as identified on
histologic sections as shown in Fig. 1. Tumor volumes at 4 to
5 weeks of age were found to range from 5 to 30 AL. Weekly
MRI exams revealed two distinct types of tumor profiles as
identified by differences in growth rate as well as differences in
contrast enhancement and spontaneous large cystic regions.
The first tumor type was classified as high grade (f40% of
tumors imaged), and the other tumor type was identified as
low grade, which occurred in the remaining 60% of animals
(Fig. 1). Table 1 summarizes the characteristics of these two
tumor types. High-grade tumors had twice the growth rate
compared with low-grade tumors and were typically thrice
larger at the initial MRI screening time point (4.5 weeks after
RCAS injection). High-grade tumors also consistently showed a
more invasive growth pattern and exhibited gadolinium
contrast enhancement in T1-weighted images (Fig. 2), whereas
low-grade tumors generally did not enhance, except for a small
leakage into the ventricular space. The observed contrast
enhancement for high-grade tumors was heterogeneous
throughout the tumor mass and occurred in only about
20% of the tumor volume in early-stage tumors, increasing to
about 50% of the tumor volume for late-stage tumors (Fig. 2).
This localized enhancement commonly manifested itself in a
ring pattern (Fig. 2). These contrast-enhancing regions often
corresponded with localized hypo-intense regions in T2weighted images (Fig. 3). ADC values for both tumor types
were similar (Table 1) at the initial 4- to 5-week imaging time
point and slightly higher than the ADC for normal brain, as
measured on the contralateral side of the brain (81
10-5
mm2/s). However, whereas ADC values remained constant over

from MOM kit (Vector Laboratories). Subsequently, sections were

incubated with primary antibody anti proliferating cell nuclear
antigen (Oncogene, 0.8 Ag/mL) overnight at 4jC. After incubation
with secondary biotinylated anti-mouse IgG for 1 h at room
temperature, sections were incubated for 30 min with ABC (Vectastain)
reagent at room temperature and reacted with 4jC 3,3-diaminobenzidine (DAB; Vector Laboratories). Slides were finally counterstained with
hematoxylin then dehydrated in graded ethanol and mounted for
microscopic analysis.
For Hif1a and CD31 detection, an automated staining processor was
used (Discovery, Ventana Medical Systems, Inc.). The protocols were
established at the Molecular Cytology Core Facility at Memorial Sloan
Kettering Cancer Center. The protocol involves 30-min blocking using
10% normal goat serum (Vector Laboratories) and 2% bovine serum
albumin followed by incubation with Avidin and Biotin blocking
reagent (according to manufacturer instruction, this is required only
for Hif1a). Incubation with anti-Hif1 primary antibody (Chemicon;
6 Ag/mL) and anti-CD31 (Santa Cruz; 2 Ag/mL) was carried out for 3 h

www.aacrjournals.org

2899

Table 1. Characteristics of low- and high-grade

PDF-induced gliomas in the tv-a mouse
Characteristic
Incidence
Mean tumor size at 4-5 wk (AL)
Mean tumor volume doubling
time (d)
Mean contrast-enhancing
volume at 4-5 wk
Mean ADC at 4-5 wk (10-5 mm2/s)
Median survival (d)

Low-grade
glioma

High-grade
glioma

22/37
11 F 28
20 F 2

13/37
31 F 7
10 F 1

8F2

84 F 2
158

84 F 2
54

NOTE: Values are expressed as mean F SE.

Clin Cancer Res 2007;13(10) May 15, 2007

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

Imaging, Diagnosis, Prognosis

Fig. 2. A, example of T1-weighted post-contrast MRI time courses for high-grade glioma. Low-grade tumors did not enhance and are not shown. B, mean enhancing
volume for high-grade Ntv-a glioma. The enhancing volume of high-grade tumors increased at a rate exceeding tumor growth (see Fig. 1D). The enhancing fraction of
the tumor, therefore, increased with tumor growth, indicating aggressive vascular development. C, example of late-stageT1-weighted post-contrast MRI, in which the
contrast-enhancing volume exceeded 50% of the total tumor volume. For late-stage, high-grade tumors, the mean proportion of the contrast-enhancing volume increased
from 20% to 50%. D, example of a ring-enhancement pattern on aT1-weighted post-contrast MRI, typical of contrast-enhancing behavior of high-grade human glioblastoma.
T1w, T1weighted; T2w,T2 weighted.

time for the high-grade tumor, low-grade tumors were found to

have increasing ADC values over time (0.62 F 0.08
10-5
mm2/s/d; Fig. 4). A significant difference was found in survival
between mice bearing high- and low-grade tumors (Fig. 4),
with median survival of 54 and 158 days, respectively.
Contrast enhancement of gliomas in mice correlates with
histologic features defining glioma grade in humans. The highand low-grade tumor classifications were confirmed by
histology from animals sacrificed at 7 to 8 weeks. The tumors
classified as low grade by imaging criteria uniformly showed
diffuse invasion of the normal brain structures by cells with
oligodendroglioma-like morphologic characteristics. The
tumors classified as high grade by imaging criteria had
histologic evidence of pseudopalisading microvascular proliferation, high cell density, and hemorrhage. The T2-weighted
MRI hypo-intense regions in these tumors were confirmed by
histopathology to correspond with regions of high vascular
proliferation and associated high cellular density in six of six
cases (Fig. 3). In all six cases, these regions were also associated
with contrast enhancement observed in the post-contrast,
T1-weighted images. These features were not found in the
lower-grade gliomas in this study.
The heterogeneous growth rate seen within the population of
GEM gliomas could cause substantial noise in clinical trials
with patients harboring these tumors. However, the above data
indicate that the presence of contrast enhancement on initial
MRI scan predicts gliomas with a relatively uniform aggressive
behavior. Therefore, we used the presence of regions with high
vascular proliferation in images to stratify tumors and enroll
these mice with high-grade gliomas into a preclinical trial using
temozolomide treatment.

Clin Cancer Res 2007;13(10) May 15, 2007

Characterization of response of GEM high-grade gliomas to

temozolomide. We mimicked the treatment of human highgrade gliomas by enrolling on trial a cohort of 25 mice
identified as harboring high-grade gliomas with abnormal,
vascular proliferate regions. The mice were treated with vehicle
or a single course of temozolomide (100 mg/kg i.p.) daily for
5 days and then followed by MRI (T2 weighted and diffusion)
to noninvasively measure tumor volumes over time and in
response to therapy (Fig. 5). The mice tolerated this treatment
well, showing a maximum of 12.6% mean body weight loss,
3 days after the end of treatment that was fully recovered 9 days
later. Previous data have indicated that this treatment protocol
did not result in increased apoptosis within 5 days (6).
Consistent with this observation, temozolomide therapy did
not result in regression of any of the tumors; however, this
treatment resulted in a growth delay of tumor mass of f14
days followed by a return to pretreatment growth levels at the
time of progression. The net cell kill was 0.4 log unit,
corresponding to an estimated surviving tumor fraction of
44%.
Diffusion MRI was also evaluated in this study as a potential
imaging biomarker for detection of early treatment response
because the water-apparent diffusion coefficient is sensitive to
changes in tissue celullarity induced by cell kill. The treated
group was found to have a significant and early increase in ADC
values, which preceded significant deviation in volumetric
tumor growth rates between treated mice and controls (Fig. 5).
Changes in diffusion values were detectable as early as 4 to 5
days following initiation of treatment. In the treated group,
ADC values continued to increase during and for 2 weeks
following conclusion of the fractionated dosage schedule.

2900

www.aacrjournals.org

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

MRI of a GEM Model for Glioma

The subsequent decrease in tumor ADC values at 4 weeks

posttreatment correlated with the onset of tumor progression. A
significant difference in animal survival was observed between
the two groups, with a median survival in the control and
treated groups of 52 and 70 days, respectively (Fig. 5).

Discussion
RCAS/tv-a based GEM glioma models represent an opportunity for studying correlates between histology and imaging
characteristics and for developing and testing of new therapeutic
strategies for this disease. We characterized tumor development
characteristics of PDGF-induced Ntv-a gliomas using a multimodal MRI approach and quantified the response of GEM
gliomas to temozolomide, the current standard of care
chemotherapy for gliomas in humans.
Infection of Ntv-a mice with RCAS-PDGF viral vectors
encoding PDGF resulted in glioma formation in 80% of
animals that could be classified as high or low grade using
MRI at 4 to 5 weeks of age. High-grade tumors, defined by

vascular proliferative regions and contrast enhancement, were

characterized by more rapid growth compared with low-grade
tumors (Fig. 1; Table 1) and a substantial reduction in the
median survival relative to the non-enhancing low-grade
tumors (Fig. 4). Tumor contrast enhancement is generally
indicative of malformed, leaky microvasculature and high
vascular permeability, and has been shown to correlate with
histologic grade in human brain tumors (47). In the tv-a
tumors, the enhancing volume was localized within the tumor
volume as defined by T2 hyper-intensity (Fig. 3), in some cases
showing a ring enhancement pattern (Fig. 2). These properties
are typical of the contrast-enhancing properties of high-grade
human glioma (48, 49), which commonly shows non-uniform,
localized appearance that is frequently observed in a ring pattern,
presumably due to the presence of necrotic, non-enhancing
tissue inside the contrast ring and a periphery of abnormally
vascularized, leaky tissue, with associated high vascular permeability. Similar to the human pathology, contrast-enhancing
regions were shown by histology to be associated with vascular
proliferative, cellular dense tissue (Fig. 3).

Fig. 3. Comparison of T2-weighted MRI

(A),T1-weighted MRI, pre-contrast (B) and
post-contrast (C), and histologic sections
(D and E) of a high-grade glioma, showing
vascular proliferation. Four contiguous
slices are shown for the MRI. The histologic
sections (magnification,
100) correspond
with the region of contrast enhancement
in (C) (white arrow). The vascular
abnormalities as verified by H&E-stained
(D) and CD31-stained (E) histologic
sections, appeared as hypo-intense regions
on theT2-weighted scans (A) and contrastenhancing regions in theT1-weighted,
post-contrast scans (C). These features
were found in all high-grade tumors.
Bar, 100 mm (D and E).

www.aacrjournals.org

2901

Clin Cancer Res 2007;13(10) May 15, 2007

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

Imaging, Diagnosis, Prognosis

Fig. 4. Example of ADC maps, shown as

color overlays onT2-weighted anatomical
scans, for (A) low-grade and (B) highgrade gliomas at 52 d after RCAS
injection. The ADC was lower and showed
greater heterogeneity in high-grade tumors.
C, mean ADC time courses for low- and
high-grade gliomas. ADC values were
relatively constant in high-grade gliomas
but showed a linear increase in time for
low-grade gliomas, indicating lower
cellular density than in high-grade tumors.
D, survival curves for low- and high-grade
gliomas. Median survival was 8 wks
for mice with high-grade gliomas and
23 wks for mice with low-grade gliomas.

In the high-grade tumors, not only did the contrast-enhancing

volume increase with tumor growth, but the contrast-enhancing
fraction of the tumor also increased from 20% by 4.5 weeks after
RCAS injection to >50% by 7.5 weeks after RCAS injection and,
in some cases, as high as 60% to 70%. This progression indicates
an increase in the fraction of tissue with high vascular
permeability or leakiness and implies regional conversion from
low to high grade within a given tumor. By comparison, tumors
initially identified as low grade (lacking contrast enhancement
consistent with low-grade human gliomas; ref. 48) remained so
and did not acquire enhancing high-grade characteristics over
the period of this study.

Diffusion MRI characterization showed that the ADC

increased with tumor growth in low-grade tumors but
remained comparatively constant in high-grade tumors. The
tumor ADC is sensitive to the average translational water
mobility, which depends on the tissue cellularity. Both lowand high-grade tumors generally showed greater ADC compared with normal brain, consistent with a greater extracellular
fraction in the tumor tissue, compared with normal tissue.
Histologic assessment of low-grade tumors revealed diffuse
tumor cells scattered throughout the normal brain tissue. An
increasing ADC suggests that low-grade tumor cells were
continuing to displace normal brain cells and were becoming

Fig. 5. A, tumor growth time course shown for vehicle control (n = 12) and temozolomide (qdx5 100 mg/kg i.p.) treated high-grade gliomas. B, corresponding ADC time
course for the vehicle control and treated animals. A significant early increase in ADC was observed in the treated group, which corresponded to significant growth inhibition,
compared with the control group. The treated group ADC continued to increase until it reached a maximum at 52 d, at which time there was sharp decrease in ADC, and
tumor progression was observed. C, survival plot for temozolomide and control groups. Median survival was 52 d for the control group versus 70 d for the treatment group.

Clin Cancer Res 2007;13(10) May 15, 2007

2902

www.aacrjournals.org

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

MRI of a GEM Model for Glioma

a larger percentage of the tumor tissue mass over time due to

the proliferation of the tumor cells. In contrast, normal brain
tissue was completely displaced by the presence of a high-grade
tumor, and as such, the mean diffusion values for high-grade
tumors showed a small, early increase (Fig. 4C), then remained
much more constant. The early ADC increase was presumably
due to early displacement of normal cells by tumor cells. The
greater ADC of the advanced low-grade tumors compared with
the high-grade tumors is consistent with inherently greater
cellular density/lower extracellular fraction in the high-grade
tumors, compared with the most advanced low-grade tumors.
The ability of MRI to not only determine tumor burden but
also to distinguish the vascular and cellular characteristics of
high- and low-grade tv-a gliomas suggests that these methods
may also be used to determine response to therapy. An initial
validation of this hypothesis was done by temozolomide
treatment of mice identified as having high-grade tv-a gliomas
by MRI characteristics. Treatment with temozolomide resulted in
a tumor growth delay of 14 days, compared with vehicle-treated
controls. In addition, an early increase in ADC was observed,
consistent with decreasing cellularity in response to treatment.

A later decrease in ADC correlated with tumor regrowth and was

likely due to increasing cellularity with tumor repopulation.
The Ntv-a model of glioma provides for a recapitulation of
specific biological processes involved in oncogene-specific
neoplastic transformation, along with subsequent tumor development and progression. These features constitute a model that
is potentially more predictive of the clinical effectiveness of an
experimental agent and that may significantly facilitate preclinical testing of new therapeutic strategies, including combination
treatments. However, despite the use of a single oncogenic
pathway (in this case, the PDGF autocrine loop), GEM gliomas
consistently show heterogeneities both between different
tumors and within a single tumor, similar to that observed in
the human condition. MRI is therefore an important, highresolution imaging tool for determining tumor grade and type in
this model, enrolling subjects into preclinical trials, and for
characterizing tumor heterogeneities. The additional use of
diffusion MRI allows sensitive detection of early therapeutic
effects in tv-a glioma, predicting later gross effects on tumor
growth. This highlights the potential for tumor ADC as a
surrogate marker for efficacy in human glioma patients.

References
1. Behin A, Hoang-Xuan K, Carpentier AF, Delattre JY.
Primary brain tumours in adults. Lancet 2003;361:
323 ^ 31.
2. McConville P, Moody JB, Moffat BA. High-throughput magnetic resonance imaging in mice for phenotyping and therapeutic evaluation. Curr Opin Chem
Biol 2005;9:413 ^ 20.
3. Uhrbom L, Nerio E, Holland EC. Dissecting tumor
maintenance requirements using bioluminescence
imaging of cell proliferation in a mouse glioma model.
Nat Med 2004;10:1257 ^ 60.
4. Shih AH, Dai C, Hu X, Rosenblum MK, Koutcher JA,
Holland EC. Dose-dependent effects of plateletderived growth factor-B on glial tumorigenesis.
Cancer Res 2004;64:4783 ^ 9.
5. Koutcher JA, Hu X, Xu S, et al. MRI of mouse models
for gliomas shows similarities to humans and can be
used to identify mice for preclinical trials. Neoplasia
2002;4:480 ^ 5.
6. Momota H, Nerio E, Holland EC. Perifosine inhibits
multiple signaling pathways in glial progenitors and
cooperates with temozolomide to arrest cell proliferation in gliomas in vivo. Cancer Res 2005;65:
7429 ^ 35.
7. Murphy PS, Viviers L, Abson C, et al. Monitoring
temozolomide treatment of low-grade glioma with
proton magnetic resonance spectroscopy. BrJ Cancer
2004;90:781 ^ 6.
8. AlexanderA, Murtha A, Abdulkarim B, et al. Prognostic significance of serial magnetic resonance spectroscopies over the course of radiation therapy for
patients with malignant glioma. Clin Invest Med
2006;29:301 ^ 11.
9. Grand S,Tropres I, Hoffmann D, Ziegler A, Le Bas JF.
[Proton magnetic resonance spectroscopy (1H-MRS)
for the diagnosis of brain tumors and the evaluation of
treatment]. Neurochirurgie 2005;51:299 ^ 308.
10. HarrerJU, Parker GJ, Haroon HA, et al. Comparative
study of methods for determining vascular permeability and blood volume in human gliomas. J Magn Reson
Imaging 2004;20:748 ^ 57.
11. Haroon HA, Buckley DL, Patankar TA, et al. A
comparison of Ktrans measurements obtained with
conventional and first pass pharmacokinetic models
in human gliomas. J Magn Reson Imaging 2004;19:
527 ^ 36.
12. Aronen HJ, Perkio J. Dynamic susceptibility con-

www.aacrjournals.org

trast MRI of gliomas. Neuroimaging Clin N Am 2002;

12:501 ^ 23.
13. Provenzale JM, Wang GR, Brenner T, Petrella JR,
Sorensen AG. Comparison of permeability in highgrade and low-grade brain tumors using dynamic
susceptibility contrast MR imaging. AJR Am J Roentgenol 2002;178:711 ^ 6.
14. Le Bihan D, Breton E, Lallemand D, Grenier P,
Cabanis E, Laval-Jeantet M. MR imaging of intravoxel
incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 1986;161:
401 ^ 7.
15. Fayed N, Modrego PJ. The contribution of magnetic
resonance spectroscopy and echoplanar perfusionweighted MRI in the initial assessment of brain
tumours. J Neurooncol 2005;72:261 ^ 5.
16. Cha S. Update on brain tumor imaging. Curr Neurol
Neurosci Rep 2005;5:169 ^ 77.
17. Fan G, Zang P, Jing F, Wu Z, Guo Q. Usefulness of
diffusion/perfusion-weighted MRI in rat gliomas:
correlation with histopathology. Acad Radiol 2005;
12:640 ^ 51.
18. Lupo JM, Cha S, Chang SM, Nelson SJ. Dynamic
susceptibility-weighted perfusion imaging of highgrade gliomas: characterization of spatial heterogeneity. AJNR Am J Neuroradiol 2005;26:1446 ^ 54.
19. Ross BD, ChenevertTL, Kim B, Ben-Yoseph O. Magnetic resonance imaging and spectroscopy: application to experimental neuro-oncology. Q Magn Reson
Biol Med 1994;1:89 ^ 106.
20. Zhao M, Pipe JG, Bonnett J, Evelhoch JL. Early detection of treatment response by diffusion-weighted
1H-NMR spectroscopy in a murine tumour in vivo.
Br J Cancer 1996;73:61 ^ 4.
21. Chenevert TL, McKeever PE, Ross BD. Monitoring
early response of experimental brain tumors to therapy
using diffusion magnetic resonance imaging. Clin
Cancer Res 1997;3:1457 ^ 66.
22. Ross BD, Stegman LD, Chenevert TL, Rehemtulla
A. The role of magnetic resonance in the evaluation
of cancer therapeutics. Clin Cancer Res 1999;5:
3870 ^ 1s.
23. Galons JP, Altbach MI, Paine-Murrieta GD, Taylor
CW, Gillies RJ. Early increases in breast tumor xenograft water mobility in response to paclitaxel therapy
detected by non-invasive diffusion magnetic resonance imaging. Neoplasia 1999;1:113 ^ 7.

2903

24. Lyng H, Haraldseth O, Rofstad EK. Measurement of

cell density and necrotic fraction in human melanoma
xenografts by diffusion weighted magnetic resonance
imaging. Magn Reson Med 2000;43:828 ^ 36.
25. Chinnaiyan AM, Prasad U, Shankar S, et al. Combined effect of tumor necrosis factor-related apoptosis-inducing ligand and ionizing radiation in breast
cancer therapy. Proc Natl Acad Sci U S A 2000;97:
1754 ^ 9.
26. Evelhoch JL, Gillies RJ, Karczmar GS, et al. Applications of magnetic resonance in model systems: cancer
therapeutics. Neoplasia 2000;2:152 ^ 65.
27. Stegman LD, Rehemtulla A, Hamstra DA, et al.
Diffusion MRI detects early events in the response of
a glioma model to the yeast cytosine deaminase gene
therapy strategy. GeneTher 2000;7:1005 ^ 10.
28. Ross BD, Chenevert TL, Rehemtulla A. Magnetic
resonance imaging in cancer research. Eur J Cancer
2002;38:2147 ^ 56.
29. Chenevert TL, Stegman LD, Taylor JM, et al. Diffusion magnetic resonance imaging: an early surrogate
marker of therapeutic efficacy in brain tumors. J Natl
Cancer Inst 2000;92:2029 ^ 36.
30. Rehemtulla A, Hall DE, Stegman LD, et al. Molecular
imaging of gene expression and efficacy following
adenoviral-mediated brain tumor gene therapy. Mol
Imaging 2002;1:43 ^ 55.
31. ChenevertTL, Meyer CR, Moffat BA, et al. Diffusion
MRI: a new strategy for assessment of cancer therapeutic efficacy. Mol Imaging 2002;1:336 ^ 43.
32. Jennings D, Hatton BN, Guo J, et al. Early response
of prostate carcinoma xenografts to docetaxel chemotherapy monitored with diffusion MRI. Neoplasia
2002;4:255 ^ 62.
33. Ross BD, Moffat BA, LawrenceTS, et al. Evaluation
of cancer therapy using diffusion magnetic resonance
imaging. Mol CancerTher 2003;2:581 ^ 7.
34. Moffat BA, Chenevert TL, Ross BD. Diffusion MR
imaging in adult neoplasia. In: Gillard J, Waldman A,
Barker P, editors. Physiological MR in clinical neuroscience. Cambridge: CUP; 2004.
35. Hall DE, Moffat BA, Stojanovska J, et al. Therapeutic efficacy of DTI-015 using diffusion magnetic
resonance imaging as an early surrogate marker. Clin
Cancer Res 2004;10:7852 ^ 9.
36. Hamstra DA, Lee KC,Tychewicz JM, et al.The use of
19F spectroscopy and diffusion-weighted MRI to

Clin Cancer Res 2007;13(10) May 15, 2007

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.

Imaging, Diagnosis, Prognosis

evaluate differences in gene-dependent enzyme prodrug therapies. Mol Ther 2004;10:916 ^ 28.
37. Moffat BA, Hall DE, Stojanovska J, et al. Diffusion
imaging for evaluation of tumor therapies in preclinical
animal models. Magma 2004;17:249 ^ 59.
38. Rodrigues LM, Stubbs M, Robinson SP, Newell B,
Mansi J, Griffiths JR. The C-neu mammary carcinoma
in Oncomice; characterization and monitoring response
to treatment with Herceptin by magnetic resonance
methods. Magma 2004;17:260 ^ 70.
39. Theilmann RJ, Borders R,Trouard TP, et al. Changes
in water mobility measured by diffusion MRI predict response of metastatic breast cancer to chemotherapy.
Neoplasia 2004;6:831 ^ 7.
40. Schepkin VD, Ross BD, Chenevert TL, et al. Sodium
magnetic resonance imaging of chemotherapeutic responseinaratglioma.MagnResonMed2005;53:85^ 92.
41. Lee KC, Hamstra DA, Bullarayasamudram S, et al.
Fusion of the HSV-1 tegument protein vp22 to cytosine deaminase confers enhanced bystander effect

and increased therapeutic benefit. Gene Ther 2006;

13:127 ^ 37.
42. SchepkinVD, ChenevertTL, Kuszpit K, et al. Sodium
and proton diffusion MRI as biomarkers for early
therapeutic response in subcutaneous tumors. Magn
Reson Imaging 2006;24:273 ^ 8.
43. Jordan BF, Runquist M, Raghunand N, et al.
Dynamic contrast-enhanced and diffusion MRI show
rapid and dramatic changes in tumor microenvironment in response to inhibition of HIF-1alpha using
PX-478. Neoplasia 2005;7:475 ^ 85.
44. Moffat BA, Chenevert TL, LawrenceTS, et al. Functional diffusion map: a noninvasive MRI biomarker for
early stratification of clinical brain tumor response.
Proc Natl Acad Sci U S A 2005;102:5524 ^ 9.
45. Hamstra DA, Chenevert TL, Moffat BA, et al. Evaluation of the functional diffusion map as an early
biomarker of time-to-progression and overall survival
in high-grade glioma. Proc Natl Acad Sci U S A 2005;
102:16759 ^ 64.

Clin Cancer Res 2007;13(10) May 15, 2007

2904

46. Lee KC, Hall DE, Hoff BA, et al. Dynamic imaging of
emerging resistance during cancer therapy. Cancer
Res 2006;66:4687 ^ 92.
47. Ross BD, Zhao YJ, Neal ER, et al. Contributions of
cell kill and posttreatment tumor growth rates to the
repopulation of intracerebral 9L tumors after chemotherapy: an MRI study. Proc Natl Acad Sci U S A1998;
95:7012 ^ 7.
48. Roberts HC, Roberts TP, Brasch RC, Dillon WP.
Quantitative measurement of microvascular permeability in human brain tumors achieved using dynamic contrast-enhanced MR imaging: correlation
with histologic grade. AJNR Am J Neuroradiol
2000;21:891 ^ 9.
49. Roberts HC, Roberts TP, Ley S, Dillon WP, Brasch
RC. Quantitative estimation of microvascular permeability in human brain tumors: correlation of dynamic Gd-DTPA-enhanced MR imaging with
histopathologic grading. Acad Radiol 2002;9 Suppl
1:S151 ^ 5.

www.aacrjournals.org

Downloaded from clincancerres.aacrjournals.org on September 15, 2014. 2007 American Association for
Cancer Research.