Tumor and Stem Cell Biology Cancer Research Preclinical Evaluation

2548

Published OnlineFirst March 2, 2010; DOI: 10.1158/0008-5472.CAN-09-2503

Tumor and Stem Cell Biology
Cancer
Research

Preclinical Evaluation of Radiation and Perifosinein a Genetically and Histologically AccurateModel of Brainstem Glioma

Oren J. Becher1,3,5, Dolores Hambardzumyan2,5, Talia R. Walker1,5, Karim Helmy1,5, Javad Nazarian6,Steffen Albrecht7, Rebecca L. Hiner3,5, Sarah Gall8, Jason T. Huse1,4,5, Nada Jabado7,Tobey J. MacDonald6, and Eric C. Holland1,2,5

Abstract

Authors' A2NeurosurgMemorial6Children's7Montréal8Carolinas

Note: SupResearch O

CorresponCancer CePhone: 212

doi: 10.115

©2010 Am

Cancer R

Dow

Brainstem gliomas (BSG) are a rare group of central nervous system tumors that arise mostly in childrenand usually portend a particularly poor prognosis. We report the development of a genetically engineeredmouse model of BSG using the RCAS/tv-a system and its implementation in preclinical trials. Using immu-nohistochemistry, we found that platelet-derived growth factor (PDGF) receptor α is overexpressed in 67% ofpediatric BSGs. Based on this observation, we induced low-grade BSGs by overexpressing PDGF-B in the pos-terior fossa of neonatal nestin tv-a mice. To generate high-grade BSGs, we overexpressed PDGF-B in combi-nation with Ink4a-ARF loss, given that this locus is commonly lost in high-grade pediatric BSGs. We show thatthe likely cells of origin for these mouse BSGs exist on the floor of the fourth ventricle and cerebral aqueduct.Irradiation of these high-grade BSGs shows that although single doses of 2, 6, and 10 Gy significantly increasedthe percent of terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL)–positive nu-clei, only 6 and 10 Gy significantly induce cell cycle arrest. Perifosine, an inhibitor of AKT signaling, signifi-cantly induced TUNEL-positive nuclei in this high-grade BSG model, but in combination with 10 Gy, it did notsignificantly increase the percent of TUNEL-positive nuclei relative to 10 Gy alone at 6, 24, and 72 hours. Sur-vival analysis showed that a single dose of 10 Gy significantly prolonged survival by 27% (P = 0.0002) butperifosine did not (P = 0.92). Perifosine + 10 Gy did not result in a significantly increased survival relativeto 10 Gy alone (P = 0.23). This PDGF-induced BSG model can serve as a preclinical tool for the testing of novelagents. Cancer Res; 70(6); 2548–57. ©2010 AACR.

Introduction

Gliomas may arise anywhere in the central nervous system(CNS). An uncommon location for gliomagenesis in the CNS isthe brainstem. Brainstem gliomas (BSG) develop predomi-nantly in the pediatric population where they constitute upto 20% of childhood brain tumors and are the leading causeof death for children with brain tumors (1). According to his-topathologic characteristics, BSGs are divided into high grade(WHO grades 3 and 4), which comprise ∼85% of BSGs [alsocalled diffuse intrinsic pontine gliomas (DIPG) if the tumor

ffiliations: Departments of 1Cancer Biology and Genetics,ery, 3Pediatrics, and 4Pathology and 5Brain Tumor Center,Sloan-Kettering Cancer Center, New York, New York;National Medical Center, Washington, District of Columbia;Children's Hospital, Montréal, Quebec, Canada; andMedical Center, Charlotte, North Carolina

plementary data for this article are available at Cancernline (http://cancerres.aacrjournals.org/).

ding Author: Eric C. Holland, Memorial Sloan-Ketteringnter, Z1304, 408 East 69th Street, New York, NY 10021.-639-3005; Fax: 646-422-0231; E-mail: [email protected].

8/0008-5472.CAN-09-2503

erican Association for Cancer Research.

es; 70(6) March 15, 2010

Researcon Decembecancerres.aacrjournals.org nloaded from

is primarily localized to the pons], with the remaining beinglow grade (WHO grades 1 and 2). By contrast, BSGs in adultsare much less common, accounting for <2% of gliomas. Fur-thermore, BSGs in adults are less aggressive than those inchildren and are thought to be biologically different (2).Focal radiation consisting of daily 2 Gy doses for ∼30 doses

over 6 weeks is the current standard of care for these tumors—a treatment modality that unfortunately provides onlytemporary relief of symptoms. Despite numerous clinicalinvestigations to date, there have been no chemotherapeuticor biological agents that have proven clearly beneficial for thetreatment of high-grade BSGs in children. Themedian survivalfor these children is <1 year after diagnosis, although it hasbeen observed that children <3 years of age may fare better (3).There is limited understanding of the biology of BSGs in

children due to limited availability of clinical specimens. Asthese tumors are not routinely resected and autopsy ratesare quite low, such clinical specimens are extremely precious(4). The limited studies published on this entity show thatthese tumors have similar genetic alterations as pediatricgliomas arising in other parts of the brain but distinct fromthe molecular characteristics of adult gliomas (5–10). Erbb1is amplified in a small subset of BSGs and overexpressed ina larger subset (11). Kras and AKT activation have been

h. r 19, 2018. © 2010 American Association for Cancer

http://cancerres.aacrjournals.org/

Mouse Model of Brainstem Glioma


described in 60% of pediatric high-grade gliomas includingseveral BSGs (12). p53 mutations have been described rangingfrom 29% to 71% of pediatric BSGs (8, 10, 13–16). The mostcommon reported molecular alterations occur in platelet-de-rived growth factor receptor (PDGFR), which is overexpressedin about 30% to 80% of high-grade gliomas including BSGs(17–19), and loss of expression of Ink4a-ARF, which has beendescribed ranging from 17% to 100% of pediatric glioblasto-mas in the brainstem (8, 16, 20).There are two published rat models of BSG, both are allo-

graft models using cell lines derived from rat cortical gliomasthat are stereotactically implanted into the brainstem. Therats develop tumors in the brainstem, but it is not clear towhat extent these accurately model the histology and genet-ics of the human disease (21, 22). The cell lines are derivedfrom a carcinogen-induced cortical glioma and are main-tained in serum conditions before implantation, culture con-ditions that are inferior to culturing in basic fibroblastgrowth factor and epidermal growth factor in maintainingthe phenotype and genotype of primary gliomas (23). In ad-dition, a human xenograft BSG model in rats has also beenrecently reported, which uses adult glioma cell lines thatarose from adult cortical gliomas and are implanted into im-munodeficient rats (24). To our knowledge, there is no genet-ically engineered mouse model for this disease.Here, we initially observed that PDGFRα is overexpressed in

67% of high-grade BSGs overall, including 87% of biopsy speci-mens. We used the RCAS/tv-a system to overexpress PDGF-Bin nestin-expressing cells lining the fourth ventricle and aque-duct. This gene transfer resulted in the formation of low-gradeBSGs, whereas PDGF overexpression in conjunction withInk4a-ARF loss in the same periventricular cells resulted inthe formation of high-grade BSGs. These tumors were histo-logically similar to diffuse pediatric BSGs in many respects. Ashort-term analysis of radiation therapy (RT) in this BSGmod-el showed that higher doses of RT are more effective in induc-ing cell cycle arrest. As the AKT pathway is thought to mediateradioresistance (25), we tested perifosine, an inhibitor of AKTsignaling, alone and in combination with RT in this BSG mod-el. Perifosine treatment significantly increased terminal deox-ynucleotidyl transferase–mediated dUTP nick end labeling(TUNEL)–positive nuclei, but in combination with RT, perifo-sine did not significantly increase TUNEL-positive nucleirelative to RT alone at 6, 24, or 72 hours after RT. Survival anal-ysis showed that a single dose of 10 Gy increased survival ofhigh-grade BSG-bearing mice by 27% (57 days versus 45 days;P = 0.0002), whereas perifosine monotherapy did not. Pretreat-ment of 10 Gy RT with perifosine did not result in a signifi-cantly increased median survival relative to 10 Gy alone (64days versus 57 days; P = 0.23). In summary, our PDGF-inducedBSGmodel is the first genetically engineered BSGmousemod-el that may serve as useful preclinical model for testing of nov-el treatment regimens.

Materials and Methods

Generation of BSGs. To generate BSGs, we injected 1 μL(105 cells) of RCAS-PDGF-B–expressing DF1 cells into nestin

www.aacrjournals.org

Researcon Decembecancerres.aacrjournals.org Downloaded from

tv-a (Ntv-a) or Ntv-a;Ink4a-ARF−/− mice. The genetic back-ground of our Ntv-a mice is a mix of BALB/c, FVB, B6, 129,and SW strains. Ink4a-ARF−/−;Ntv-a strain is a mix of BALB/c,FVB, B6, and 129 strains. Greater than 20 generations ofbackcrossing were done to generate the lines. Injections wereintracranial (2 mm posterior to the bregma at the midlineposition within 72 h of birth using a Hamilton syringe). Micewere monitored carefully for signs of tumor development(hydrocephalus, lethargy, head tilt). On the appearance ofbrain tumor symptoms, mice were euthanized with CO2,and brain tissue was extracted, fixed in formalin, and paraffinembedded.Investigation of the cell of origin of BSGs. To determine

the cell of origin, we injected 1 μL (105 cells) of DF1 cells ex-pressing RCAS–green fluorescent protein (GFP) or RCAS–redfluorescent protein (RFP) into Ntv-a mice as described above.Mice were euthanized 24 h later (n = 2) or at 4 wk of age (n =6). Their brain tissue was then extracted, postfixed for 30 minin 4% paraformaldehyde, transferred into 30% sucrose at 4°Cfor cryoprotection, embedded in OCT, frozen on dry ice, andstored at −80°C. Brains were sectioned in 10-μm sections andinspected with fluorescent microscope for GFP- or RFP-labeled cells.Treatment of BSG-bearing mice. We used total body irra-

diation (TBI) for all the irradiation experiments, with the ex-ception of the survival analysis. TBI was delivered with aGamma Cell 40 Irradiator (MDS Nordion) that delivers 106cGy/min. For survival analysis, mice were sedated with keta-mine and xylazine, and irradiation of the head was done us-ing a X-RAD 320 from Precision X-Ray at 115 cGy/min (therest of the mouse was shielded with a lead jig). Perifosinetreatment was i.p. at a dose 30 mg/kg/d (in normal saline/5% dextrose) once daily and was for a minimum of three dailydoses. Combination of perifosine and RT was done as follows:perifosine was given daily for 3 d as monotherapy followed bysingle-dose RT. At the end of the treatment, mice were eutha-nized, and brains were dissected and placed in formalin.Immunohistochemistry. Immunohistochemistry on

mouse and human formalin-fixed, paraffin-embedded sec-tions was performed using Discovery XT (Ventana MedicalSystems). Human BSG tissue was obtained from MemorialSloan-Kettering Cancer Center, Children's National MedicalCenter, Montreal Children's Hospital, and Carolinas Health-care System. Analysis of human BSG tissue was InstitutionalReview Board approved at all institutions. The following anti-bodies were used: PDGFRα (1:100; Cell Signaling), Ki-67(1:100 human; vector for mouse; DAKO), phospho-H3-Ser10

(1:800; Upstate), Olig-2 (1:250; Millipore), and GFAP (1:1,000;DAKO). For immunofluorescence, we used anti-RFP antibodyfrom Rockland at 1:200 followed by Alexa Fluor 488 anti-rabbit secondary antibody as per the manufacturer's protocol.The TUNEL assay was performed per Roche's instructions.Quantitative analysis. All immunohistochemical slides

were scanned. One observer (O.J.B.) performed the quantifi-cation of TUNEL, phospho-H3, and Ki-67 using the programMetaMorph, where a threshold was set for TUNEL, phos-pho-H3, and Ki-67 and applied to all sections analyzed. Atleast five high-magnification 20× representative areas were

Cancer Res; 70(6) March 15, 2010 2549



Becher et al.

2550


analyzed per tumor. For quantification of phospho-H3,positive cells were at least 5 pixels of positive area. Forquantification of TUNEL-positive nuclei, we excluded pseu-dopalisades.Magnetic resonance imaging. Magnetic resonance imag-

ing (MRI) of murine gliomas has been previously de-scribed (26).

Results

PDGFRα is overexpressed in pediatric high-grade BSGs.To specifically investigate the PDGFRα expression in pediatrichigh-grade BSGs biopsy and autopsy specimens, we immunos-tained paraffin sections of high-grade BSGs for PDGFRα andnoted that 12 of 18 (67%) of high-grade BSGs contained tumorcells that express PDGFRα (Table 1). Interestingly, we notedthat 7 of 8 surgical biopsies of BSGs had PDGFRα overexpres-sion (87.5%), whereas only 5 of 10 autopsies expressedPDGFRα (50%). The difference in expression of PDGFRα be-tween surgical biopsies and autopsies was not statisticallysignificant (P = 0.15, two-tailed Fisher's exact test).Generation of PDGF-induced BSGs in mice. PDGFRα

overexpression is a common event in pediatric high-gradeBSGs, and so, we were interested to determine if overexpres-sion of PDGF in the posterior fossa of neonatal Ntv-a micecan induce BSGs. We infected RCAS-PDGF into the posteriorfossa of Ntv-a neonatal mice within 72 hours of birth (∼2mm posterior and midline to the bregma) and observedthe mice for signs and symptoms of brain tumor formation.PDGF overexpression in Ntv-a mice resulted in low-grade

Cancer Res; 70(6) March 15, 2010


BSGs (grade 2; Fig. 1B). The incidence of tumor formationwas 80% (12 of 15), and most of the mice were asymptomaticuntil euthanasia at 12 weeks (Fig. 1A). Of note, no embryonaltumors (such as medulloblastomas) were observed, and thecerebellar external granule layer was not altered.The tumor cells of these low-grade gliomas were prolifer-

ating at a low rate with a Mib1 index of 0.83 ± 0.16%. Therewas minimal baseline apoptosis as observed with cleavedcaspase-3 immunohistochemistry (data not shown). Thelow-grade gliomas were not purely astrocytic as most ofthe cells in these lesions immunostained for Olig-2. Theselesions were not simply reactive gliosis, as they were quitelarge at 12 weeks with several tumors occupying ∼50% ofthe pons. These lesions were easily visible on MRI (Fig. 1B).Mice injected with vector alone did not show any lesions.The single mouse that died developed obstructive hydro-cephalus due to tumor. Characteristic immunostains of theselesions are illustrated in Fig. 2C (Ki-67, Olig-2, and PDGFR)and can be contrasted with immunostains of normal brain-stem (Fig. 2D). In addition there is a GFAP immunostain of alow-grade BSG in Supplementary Fig. S1A.High-grade BSGs in humans show Ink4a-ARF loss, and

therefore, we added this genetic alteration to our model.Ntv-a mice with deletion of Ink4a-ARF infected with RCAS-PDGF developed symptoms of brain tumor formation suchas macrocephaly, lethargy, and hemiparesis between 4 and8 weeks. We sacrificed these symptomatic animals and ex-tracted their brains. All of the mice had grade 4 BSGs withpathologic elements of microvascular proliferation and pseu-dopalisading necrosis (Fig. 1C). The incidence and latency for

Table 1. Human high-grade BSGs overexpress PDGFRα

BSG grade
PDGFR IHC*
h. r 19,

Biopsy or autopsy

C

2018. © 2010 American Association

Age of patient

Pontine grade 4
1+ Autopsy 15 y Pontine grade 4 Negative Autopsy 4 y Pontine grade 4 1+ Autopsy 7 y Pontine grade 3 Negative Autopsy 9 y Pontine grade 4 Negative Autopsy 7 y Pontine grade 2 Negative Autopsy 8 mo Pontine grade 4 1+ focal Autopsy 7 y Pontine grade 4 Negative Autopsy 10 y Pontine grade 4 1+ focal Autopsy 5 y Pontine grade 4 2+ Autopsy 6 y Pontine grade 4 1+ Biopsy 10 y Pontine grade 4 2+ Biopsy 4 y Pontine grade 2 1+ Biopsy 4 y Pontine grade 3 1+ Biopsy 4 y Pontine grade 4 Negative Biopsy 8 y Pontine grade 4 2+ Biopsy 4 y Pontine grade 4 2+ Biopsy 10 y Midbrain grade 3 2+ Biopsy 2 y
Abbreviation: IHC, immunohistochemistry.*Negative = no staining; 1+ = weak staining of tumor cells; 2+ = strong immunostaining of tumor cells.

ancer Research

for Cancer




PDGF overexpression in Ink4a-ARF–null mice was 96% (68 of71), with a median survival of 31.5 days (Fig. 1A). The survivalof mice with high-grade BSGs induced in Ink4a-ARF−/− back-ground was significantly shorter than that of mice with low-grade BSGs (P < 0.0001, log-rank test). We obtained a MRI ofa subset of symptomatic mice before sacrificing the animalsand could easily visualize the brainstem tumors in all cases(Fig. 1C). We examined the brainstems of at least 50 unin-jected Ntv-a mice with deletion of Ink4a-ARF and did not ob-serve any preneoplastic lesions.The anatomic extent of the tumors at 4 weeks was that

80% of the tumors localized completely within the brainstem(based on T2-weighted MRI images). The natural course ofthese tumors based on serial MRIs was that the tumorswould go on to invade into the cerebellum posteriorly andinvade anteriorly and superiorly into the thalamus and topof the brainstem, respectively. Eventually, all the mice devel-oped obstructive hydrocephalus that likely resulted in ob-struction of the cerebrospinal fluid flow at the level of thefourth ventricle and aqueduct and its associated symptoms.Mice were euthanized at that point. The tumor volume esti-mate at 4 weeks (measurements were done on T2-weightedMRI images of at least 10 high-grade BSGs) was 11.9 ± 4.7 mm3.Histologic examination revealed that 70% of the tumorsoccupied at least 50% of the pons.



The histologic appearance of our murine BSGs was similarto human pediatric BSGs, particularly with regard to the dif-fusely infiltrating arrangement of the tumor cells (Fig. 2A andB). We immunostained our murine BSGs for Ki-67, Olig-2,GFAP, and PDGFRα and noted further similarities betweenthese tumors and pediatric BSGs. Pediatric BSGs stained pos-itive for Olig-2 but with large variability in the percent of pos-itive cells in different tumors ranging from 10% to ∼100%,whereas >90% of the tumor cells of our PDGF-induced BSGsin Ink4a-ARF−/− mice stained strongly for PDGFRα and Olig-2(Fig. 2A and B; Supplementary Fig. S2). GFAP immunostain-ing was positive in both low- and high-grade murine BSGs,with similarity to pediatric BSGs (Supplementary Figs. S1and S2). Of note, Mib1 labeling was 10.2 ± 5.6% for PDGF-induced BSGs in Ink4a-ARF−/− background.RCAS infections target cells on the ventricular surface of

the fourth ventricle. The RCAS/tv-a system allows for line-age tracing of tv-a–expressing cells infected with RCAS. Aswe were using the Ntv-a mouse, we were infecting only nes-tin-expressing cells. Nestin expression in the ependymal layerof the fourth ventricle has been previously described beforein humans (27). In addition, it has been hypothesized thatcells lining the floor of the fourth ventricle may serve as cellsof origin for BSGs (28). We therefore decided to investigatewhich cells we were targeting with our infections (29, 30).

Figure 1. Generation of PDGF-inducedBSGs. A, Kaplan-Meier of 15 Ntv-a miceand 71 Ntv-a;Ink4a-ARF−/− mice injectedwith RCAS-PDGF-B. B, low-magnification(left) and high-magnification (middle) H&E ofa PDGF-induced BSG in Ntv-a mouse(sacrificed at 12 wk of age). Right,T2-weighted image at 12 wk of age of aBSG induced by overexpression of PDGFin Ntv-a mouse. C, low-magnification(left) and high-magnification (middle)H&E of a PDGF-induced BSG in Ntv-a;Ink4a-ARF−/− mouse (sacrificed at 5 wkof age). Right, T2-weighted image at 4 wk ofage of a BSG induced by overexpressionof PDGF in Ntv-a;Ink4a-ARF–null mice.

Cancer Res; 70(6) March 15, 2010 2551



Becher et al.

2552


We injected DF1 cells producing RCAS-RFP into the pos-terior fossa of neonatal Ntv-a mice at postnatal day 2 (n = 2).Twenty-four hours later, we euthanized both neonatal mice,extracted their brains, sectioned, and looked for RFP-labeledcells. Both mice had RFP-labeled cells floating in the cerebro-spinal fluid and were localized around the fourth ventricleand aqueduct (Supplementary Fig. S3). In parallel, we in-jected DF1 cells producing RCAS-GFP in a similar fashionat postnatal day 2 (n = 6). We monitored the mice for 4 weeks(DF1 cells have been previously reported to survive up to 3weeks after injection), euthanized the mice, extracted theirbrains, and looked for GFP-labeled cells. The cells expressingGFP were located mostly in the outer layer of the fourth ven-tricle (Fig. 3A) or aqueduct. Of note, the distribution was notrandom (Fig. 3B). We found that >90% of the GFP-labeledcells were located symmetrically at specific locations onthe ventricular surface of the floor of the fourth ventricle,whereas only 5% were from the roof of the fourth ventricle(266 GFP-labeled cells counted). In addition, we immunos-tained sections of neonatal pups at postnatal days 1 to 3



and noted strong nestin expression in the cells lining thefourth ventricle during this early postnatal age (both roofand floor of fourth ventricle as seen in Fig. 3C).To confirm that the nestin-expressing cells of the fourth

ventricle and aqueduct are the cells of origin for our murineBSG model, we infected RCAS-PDGF as described above, ob-served the mice for 3 weeks, euthanized them, and looked forneoplastic lesions. We noted that all of the mice had smallprecursor lesions and 80% of the precursor lesions (12 of15) were arising from the floor of the fourth ventricle or aq-ueduct (Fig. 3D).In summary, we have developed a genetically engineered

BSG model that has histologic similarities to human BSGsand has comparable immunostaining characteristics. OurBSG model arises from the cells of the fourth ventricle or aq-ueduct. Next, we were interested to use this BSG model inpreclinical trials.Preclinical trials on the mouse BSG model. RT is the only

known therapeutic modality with antitumor activity for chil-dren with BSGs. Therefore, we investigated this treatment in

Figure 2. PDGF-induced murine BSGs are histopathologically similar to human BSGs. A, H&E, Ki-67, PDGFRα, and Olig-2 of a pediatric high-gradeBSG (human). B, H&E, Ki-67, PDGFRα, and Olig-2 of PDGF-induced BSG generated in an Ntv-a;Ink4a-ARF−/− mouse (high-grade BSG). C, H&E, Ki-67,PDGFRα, and Olig-2 of PDGF-induced BSG generated in an Ntv-a mouse (low-grade BSG). D, H&E, Ki-67, PDGFRα, and Olig-2 of normal brainstemof an Ntv-a;Ink4a-ARF−/− mouse.

Cancer Research





our mouse model. We irradiated PDGF-induced high-gradeBSG-bearing mice with single-dose RT at 2 Gy (approximate-ly the dose used in the clinic), 6 Gy, and 10 Gy, and sacrificedthe mice at three time points: 6, 24, and 72 hours (minimumof five mice per group). We analyzed the response of thesetumors to irradiation using TUNEL and evaluated the prolif-eration rate using phospho-H3. Doses of 10 and 6 Gy induceda near-complete cell cycle arrest 6 hours after treatment asevaluated by phospho-H3, but 2 Gy did not significantly af-fect phospho-H3 (Fig. 4A and B). We also found significantlyincreased TUNEL-positive nuclei relative to unirradiatedBSGs at 24 hours after RT for all doses of RT analyzed (Fig.4C and D). Time course (10 Gy) showed that the peak inhi-bition of proliferation was 6 hours after RT and that the high-est percent of TUNEL-positive nuclei was 24 hours after RT(Supplementary Fig. S4A and B). To determine if RT prolongssurvival in this BSG model, we treated BSG-bearing micewith a single dose of 10 Gy (by only irradiating the head).Treatment was begun after pretreatment MRIs at 4 weeksso that similarly sized tumors were assigned to each treat-



ment group (including untreated group). A single dose of10 Gy prolonged survival by 12 days (57 days versus 45 days;P = 0.0002; Fig. 5D). A single dose of 2 Gy did not prolongsurvival significantly (P = 0.14). In summary, high-dose RTis effective in prolonging survival in this BSG model.Combination of perifosine and RT in this high-grade

BSG model. We recently noted that perifosine, an inhibitorof AKT signaling, cooperates with radiation in medulloblas-tomas (31). In addition, the AKT pathway is active in radio-resistant glioma cells, and we have previously shown thatperifosine cooperates with temozolomide in gliomas (25,32). Therefore, we determined whether perifosine has a ther-apeutic effect in BSGs as a single agent and in combinationwith 10 Gy RT, the most effective single radiation dose. Wetreated a group of PDGF-induced high-grade BSGs with peri-fosine alone for 3 to 5 days, as well as a second group with 3days of perifosine followed by a single dose of 10 Gy. Perifo-sine treatment did not affect proliferation significantly as asingle agent, and there was no statistically significant differ-ence in proliferation with the addition of perifosine to 10 Gy

Figure 3. Cell of origin for murine BSGs. A, high magnification of a cross section from a 4-wk-old Ntv-a mouse infected with RCAS-GFP. White arrowspoint to clusters of GFP-labeled cells. B, quantification of the number of GFP-labeled cells at particular locations of a schematic of the cross section of thefourth ventricle. Y axis, number of GFP-positive cells; X axis, location of the GFP-positive cells. This cartoon illustrates the cross section of the fourthventricle as a triangle. Each side of the triangle (where the floor of the fourth comprises two sides of the triangle) was divided into four equal areas, andthe number of GFP-labeled cells was counted in each area and tabulated. C, high magnification of a cross section through the fourth ventricle at postnatalday 2 immunostained with nestin. Black arrows point to positive immunostaining. D, examples of precursor lesions arising from the fourth ventricle oraqueduct of Ntv-a mice infected with RCAS-PDGF. Bottom left, cross section of aqueduct (AQ).

Cancer Res; 70(6) March 15, 2010 2553



Becher et al.

2554


RT at 6, 24, and 72 hours as measured with phospho-H3(P = 0.7, P = 0.28, and P = 0.06, Mann-Whitney, respectively;Fig. 5A). The percent of TUNEL-positive nuclei increasedsignificantly with treatment with perifosine (P = 0.005,Mann-Whitney), but the combination of perifosine + 10 Gydid not significantly affect the percent of TUNEL-positivenuclei relative to 10 Gy alone at 6, 24, and 72 hours afterRT (P = 0.46, P = 0.15, and P = 0.09, Mann-Whitney, respec-tively; Fig. 5B). To investigate the long-term effect of thesetreatments, survival analysis was carried out to determineif perifosine will prolong survival as monotherapy or incombination with RT (using head irradiation) in this high-grade BSG model. Single-agent daily perifosine treatmentfor 1 week did not prolong survival (43 days versus 45 days;



P = 0.92), and pretreatment with perifosine before 10 Gy RTdid not result in a significantly increased survival relative to10 Gy RT alone (64 days versus 57 days; P = 0.23; Fig. 5C and D).Our results suggest that perifosine, when used a single agentor with a single dose of 10 Gy, does not have sufficient anti-tumor activity to prolong survival in this model.

Discussion

In this study, we applied RCAS/tv-a technology to generateBSGs that can be used for preclinical trials. There are numer-ous advantages to this preclinical model. First, it is consis-tent with the genetic alterations of a subset of the humandisease according to the published literature and our analysis

Figure 4. Dose-response curve for RT of high-grade PDGF-induced BSGs. A, graph of % phospho-H3 (Ph3) quantification at 6 h after RT at differentdoses: 2 Gy (n = 6), 6 Gy (n = 5), and 10 Gy (n = 6) and unirradiated controls (n = 7). Mann-Whitney analysis was significant at P = 0.0025 between controlgroup and 6 Gy group and was significant at P = 0.0012 between control and 10 Gy. We used the following data more than once: control BSGs and10 Gy 6 h in Figs. 4A and 5A and Supplementary Fig. S4A. B, high-magnification sections immunostained with phospho-H3. Top left, control; top right,2 Gy; bottom left, 6 Gy; bottom right, 10 Gy. C, graph of % TUNEL-positive nuclei of 2 Gy (n = 5), 6 Gy (n = 5), and 10 Gy (n = 7) and unirradiatedcontrols (n = 5) at 24 h after RT. Mann-Whitney analysis was significant at P = 0.008 between control and 2 Gy, P = 0.008 between control and 6 Gy, andP = 0.0243 between control and 10 Gy. We used the following data more than once: control BSGs and 10 Gy 24 h in Figs. 4C and 5B and SupplementaryFig. S4B. D, high-magnification TUNEL of 2, 6, and 10 Gy 24 h after RT and unirradiated control.

Cancer Research





of human specimens. Second, it is histologically similar to thehuman disease. Third, the tumors arise in immunocompetentmice. Fourth, themodel has high penetrance and short latencyand forms in the appropriate microenvironment—the brain-stem. Fifth, this model responds to, but is not cured with, RTlike the clinical experience of treating children with DIPGswith focal RT. Sixth, the pattern of invasion of the PDGF-induced high-grade BSGs is similar to the pattern of invasionof human high-grade BSGs, as a subset of the latter has beendescribed to contiguously involve the cerebellar peduncles,cerebellum, and/or thalamus (33). It thus may serve as a pre-clinical model for screening of novel agents. There are only afew hundred new cases of pediatric BSGs in the world eachyear, and as there are now an increasingly large number ofnovel agents and exponential combinations of these agents,a genetic mouse model may be useful to screen for the mostactive combinations and schedules.



The cell of origin for supratentorial NF-1–deleted gliomashas been proposed to be neural stem cells from the subven-tricular zone (34). However, other modeling systems indicatethat nestin-expressing cells in other areas of the brain arealso able to serve as the cell of origin (26). Here, we presentevidence that the cells of origin for this murine BSG modelare mostly localized to the surface of the fourth ventricle andaqueduct. Other cells in the brainstem may also serve as thecells of origin for the development of murine BSGs, but as weused the Ntv-a mouse, we restricted our investigation to nes-tin-expressing cells in the brainstem in the neonatal period,which are the cells lining the fourth ventricle and aqueduct.We evaluated external beam irradiation in this BSG model

and noted that, similar to the clinical experience of treatingchildren with DIPGs with focal RT, RT is active in this model.It is important to note that the results of our preclinical test-ing with single doses of irradiation do not suggest that higher

Figure 5. Treatment of PDGF-induced Ink4a-ARF–null BSGs with radiation and perifosine. A, graph of % phospho-H3 quantification at 6, 24, and 72 hafter RT with and without perifosine. Mice per group are denoted in parentheses: untreated (n = 7), perifosine (n = 7), RT 6 h (n = 6), RT 6 h + perifosine(n = 6), RT 24 h (n = 5), RT 24 h + perifosine (n = 8), RT 72 h (n = 6), and RT 72 h + perifosine (n = 6). B, graph of % TUNEL-positive nuclei at 6, 24,and 72 h after RT with and without perifosine. Mice per group are denoted in parentheses: untreated (n = 5), perifosine (n = 7), RT 6 h (n = 8), RT 6 h +perifosine (n = 7), RT 24 h (n = 7), RT 24 h + perifosine (n = 5), RT 72 h (n = 6), and RT 72 h + perifosine (n = 6). C, low-magnification H&E (left) andTUNEL (right) of high-grade BSG treated with perifosine + 10 Gy 6 h after RT. D, survival analysis of PDGF-induced Ink4a-ARF−/− BSG treated with a singledose of 2 Gy (10 mice), a single dose of 10 Gy (12 mice), perifosine (7 mice), perifosine + 10 Gy (9 mice), and untreated mice (9 mice).

Cancer Res; 70(6) March 15, 2010 2555



Becher et al.

2556


doses of irradiation will be more beneficial to treat childrenwith DIPGs. A single dose of 10 Gy is a potentially toxic doseto the normal cells in the brainstem, and higher total dosesof radiation have been tested in clinical trials for DIPGs usinghyperfractionated radiation and did not show increased effi-cacy (35). As BSGs are heterogeneous tumors, observationswith this mouse model may only apply to BSGs that overex-press PDGFR and have lost Ink4a-ARF. If this was the case,pretreatment tissue to determine genetic status would beneeded, and obtaining biopsies for DIPGs has recently beenreported to have minimal complications (36). However, it isimportant to note that to biopsy DIPGs is a controversialtopic, its success is highly dependent on who performs it,and it is hard to obtain approval for it in clinical trials (37).We extended the preclinical testing by evaluating perifo-

sine, an inhibitor of AKT signaling, alone and in combinationwith RT using this model. Our analysis showed that althoughperifosine monotherapy has modest activity in inducing celldeath in short-term analysis in this model, it does not haveenough antitumor activity to increase survival. In addition,perifosine did not provide additional survival benefit tohigh-dose RT. This is perhaps not surprising, as there is nodrug that has been shown to increase survival of childrenwith high-grade BSGs in numerous clinical trials even incombination with RT and may be due to the poor drug



delivery into the brainstem due to the blood-brain barrier.This observation reinforces our recommendation that novelcombinations should be tested in a genetically and histolog-ically accurate preclinical model first before their translationto the clinic.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Katia Manova, Sho Fujiwama, Jim Finney, Quancho Zhang, CarlLe, Dov Winkelman, Monica Garcia-Barros, Adriana Haimovitz-Friedman, andMihaela Lupu.

Grant Support

O.J. Becher is a St. Baldrick's scholar and is supported by Memorial Sloan-Kettering Cancer Center Brain Tumor Center, Matthew Larson Foundation,and Witmer Foundation K12 Award. D. Hambardzumyan and J.T. Huse areLeon Levy Foundation Young Investigators. E.C. Holland is supported by KirbyFoundation grants R01 CA100688, U01 CA141502, and U54 CA126518.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 07/06/2009; revised 01/03/2010; accepted 01/11/2010; publishedOnlineFirst 03/02/2010.

References
1. Hargrave D, Bartels U, Bouffet E. Diffuse brainstem glioma in chil-
dren: critical review of clinical trials. Lancet Oncol 2006;7:241–8.2. Guillamo JS, Monjour A, Taillandier L, et al. Brainstem gliomas in

adults: prognostic factors and classification. Brain 2001;124:2528–39.3. Broniscer A, Laningham FH, Sanders RP, Kun LE, Ellison DW, Gajjar

A. Young age may predict a better outcome for children with diffusepontine glioma. Cancer 2008;113:566–72.

4. Leach PA, Estlin EJ, Coope DJ, Thorne JA, Kamaly-Asl ID. Diffusebrainstem gliomas in children: should we or shouldn't we biopsy?Br J Neurosurg 2008;22:619–24.

5. Puputti M, Tynninen O, Sihto H, et al. Amplification of KIT, PDGFRA,VEGFR2, and EGFR in gliomas. Mol Cancer Res 2006;4:927–34.

6. Di Sapio A, Morra I, Pradotto L, Guido M, Schiffer D, Mauro A. Mo-lecular genetic changes in a series of neuroepithelial tumors of child-hood. J Neurooncol 2002;59:117–22.

7. Wong KK, Tsang YT, Chang YM, et al. Genome-wide allelic imbal-ance analysis of pediatric gliomas by single nucleotide polymorphicallele array. Cancer Res 2006;66:11172–8.

8. Sung T, Miller DC, Hayes RL, Alonso M, Yee H, Newcomb EW. Pref-erential inactivation of the p53 tumor suppressor pathway and lackof EGFR amplification distinguish de novo high grade pediatric astro-cytomas from de novo adult astrocytomas. Brain Pathol 2000;10:249–59.

9. Liang ML, Ma J, Ho M, et al. Tyrosine kinase expression in pediatrichigh grade astrocytoma. J Neurooncol 2008;87:247–53.

10. Cheng Y, Ng HK, Zhang SF, et al. Genetic alterations in pediatrichigh-grade astrocytomas. Hum Pathol 1999;30:1284–90.

11. Gilbertson RJ, Hill DA, Hernan R, et al. ERBB1 is amplified andoverexpressed in high-grade diffusely infiltrative pediatric brain stemglioma. Clin Cancer Res 2003;9:3620–4.

12. Faury D, Nantel A, Dunn SE, et al. Molecular profiling identifiesprognostic subgroups of pediatric glioblastoma and shows in-creased YB-1 expression in tumors. J Clin Oncol 2007;25:1196–208.

13. Badhe PB, Chauhan PP, Mehta NK. Brainstem gliomas—a clinico-pathological study of 45 cases with p53 immunohistochemistry.Indian J Cancer 2004;41:170–4.

14. Louis DN, Rubio MP, Correa KM, Gusella JF, von Deimling A. Molec-ular genetics of pediatric brain stem gliomas. Application of PCRtechniques to small and archival brain tumor specimens. J Neuro-pathol Exp Neurol 1993;52:507–15.

15. Zhang S, Feng X, Koga H, Ichikawa T, Abe S, Kumanishi T. p53 genemutations in pontine gliomas of juvenile onset. Biochem Biophys ResCommun 1993;196:851–7.

16. Sure U, Ruedi D, Tachibana O, et al. Determination of p53 mutations,EGFR overexpression, and loss of p16 expression in pediatric glio-blastomas. J Neuropathol Exp Neurol 1997;56:782–9.

17. Thorarinsdottir HK, Santi M, McCarter R, et al. Protein expression ofplatelet-derived growth factor receptor correlates with malignant his-tology and PTEN with survival in childhood gliomas. Clin Cancer Res2008;14:3386–94.

18. Nakamura M, Shimada K, Ishida E, et al. Molecular pathogenesis ofpediatric astrocytic tumors. Neuro-Oncol 2007;9:113–23.

19. Geoerger B, Morland B, Ndiaye A, et al. Target-driven exploratorystudy of imatinib mesylate in children with solid malignancies bythe Innovative Therapies for Children with Cancer (ITCC) EuropeanConsortium. Eur J Cancer 2009;45:2236–8.

20. Wakabayashi T, Natsume A, Hatano H, et al. p16 promoter methyl-ation in the serum as a basis for the molecular diagnosis of gliomas.Neurosurgery 2009;64:455–61.

21. Jallo GI, Volkov A, Wong C, Carson BS Sr., Penno MB. A novel brain-stem tumor model: functional and histopathological characterization.Childs Nerv Syst 2006;22:1519–25.

22. Liu Q, Liu R, Kashyap MV, et al. Brainstem glioma progression in ju-venile and adult rats. J Neurosurg 2008;109:849–55.

23. Lee J, Kotliarova S, Kotliarov Y, et al. Tumor stem cells derived fromglioblastomas cultured in bFGF and EGF more closely mirror thephenotype and genotype of primary tumors than do serum-culturedcell lines. Cancer Cell 2006;9:391–403.

24. Hashizume R, Ozawa T, Dinca EB, et al. A human brainstem gliomaxenograft model enabled for bioluminescence imaging. J Neurooncol2009;96:151–9.

Cancer Research





25. Eyler CE, Foo WC, LaFiura KM, McLendon RE, Hjelmeland AB, RichJN. Brain cancer stem cells display preferential sensitivity to Akt in-hibition. Stem Cells 2008;26:3027–36.

26. Hambardzumyan D, Amankulor NM, Helmy KY, Becher OJ, HollandEC. Modeling adult gliomas using RCAS/t-va technology. TransplOncol 2009;2:89–95.

27. Takano T, Rutka JT, Becker LE. Overexpression of nestin and vi-mentin in ependymal cells in hydrocephalus. Acta Neuropathol1996;92:90–7.

28. Freeman CR, Farmer JP. Pediatric brain stem gliomas: a review. Int JRadiat Oncol Biol Phys 1998;40:265–71.

29. Holland EC, Celestino J, Dai C, Schaefer L, Sawaya RE, Fuller GN.Combined activation of Ras and Akt in neural progenitors inducesglioblastoma formation in mice. Nat Genet 2000;25:55–7.

30. Dai C, Celestino JC, Okada Y, Louis DN, Fuller GN, Holland EC.PDGF autocrine stimulation dedifferentiates cultured astrocytesand induces oligodendrogliomas and oligoastrocytomas from neuralprogenitors and astrocytes in vivo. Genes Dev 2001;15:1913–25.

31. Hambardzumyan D, Becher OJ, Rosenblum MK, Pandolfi PP, Man-ova-Todorova K, Holland EC. PI3K pathway regulates survival ofcancer stem cells residing in the perivascular niche following radia-tion in medulloblastoma in vivo. Genes Dev 2008;22:436–48.



32. Momota H, Nerio E, Holland EC. Perifosine inhibits multiple signal-ing pathways in glial progenitors and cooperates with temozolo-mide to arrest cell proliferation in gliomas in vivo. Cancer Res2005;65:7429–35.

33. Yoshimura J, Onda K, Tanaka R, Takahashi H. Clinicopathologicalstudy of diffuse type brainstem gliomas: analysis of 40 autopsycases. Neurol Med Chir 2003;43:375–82.

34. Alcantara Llaguno S, Chen J, Kwon CH, et al. Malignant astrocyto-mas originate from neural stem/progenitor cells in a somatic tumorsuppressor mouse model. Cancer Cell 2009;15:45–56.

35. Mandell LR, Kadota R, Freeman C, et al. There is no role for hyper-fractionated radiotherapy in the management of children with newlydiagnosed diffuse intrinsic brainstem tumors: results of a PediatricOncology Group phase III trial comparing conventional vs. hyper-fractionated radiotherapy. Int J Radiat Oncol Biol Phys 1999;43:959–64.

36. Roujeau T, Machado G, Garnett MR, et al. Stereotactic biopsy of dif-fuse pontine lesions in children. J Neurosurg 2007;107:1–4.

37. Albright AL, Packer RJ, Zimmerman R, Rorke LB, Boyett J,Hammond GD. Magnetic resonance scans should replace biopsiesfor the diagnosis of diffuse brain stem gliomas: a report from theChildren's Cancer Group. Neurosurgery 1993;33:1026–9.

Cancer Res; 70(6) March 15, 2010 2557



2010;70:2548-2557. Published OnlineFirst March 2, 2010.Cancer Res Oren J. Becher, Dolores Hambardzumyan, Talia R. Walker, et al. GliomaGenetically and Histologically Accurate Model of Brainstem Preclinical Evaluation of Radiation and Perifosine in a

Updated version

10.1158/0008-5472.CAN-09-2503doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2010/03/01/0008-5472.CAN-09-2503.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/70/6/2548.full#ref-list-1

This article cites 37 articles, 8 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/70/6/2548.full#related-urls

This article has been cited by 12 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/70/6/2548To request permission to re-use all or part of this article, use this link

Research. on December 19, 2018. © 2010 American Association for Cancercancerres.aacrjournals.org Downloaded from


http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-09-2503

http://cancerres.aacrjournals.org/content/suppl/2010/03/01/0008-5472.CAN-09-2503.DC1

http://cancerres.aacrjournals.org/content/70/6/2548.full#ref-list-1

http://cancerres.aacrjournals.org/content/70/6/2548.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/70/6/2548


Documents

Tumor and Stem Cell Biology Cancer Research Preclinical Evaluation