c-Myc Overexpression Causes Anaplasia in Medulloblastoma · manufacturer’s technical report. Cells were plated in six-well plates at a density of 50,000 per well on day 0 and collected

c-Myc Overexpression Causes Anaplasia in Medulloblastoma

Duncan Stearns,1,3Aneeka Chaudhry,

1Ty W. Abel,

1Peter C. Burger,

1Chi V. Dang,

2

and Charles G. Eberhart1

Departments of 1Neuropathology and 2Hematology, Johns Hopkins University School of Medicine, Baltimore, Maryland and3Department of Pediatrics, Drexel University School of Medicine, Philadelphia, Pennsylvania

Abstract

Both anaplasia and increased c-myc gene expression havebeen shown to be negative prognostic indicators for survivalin medulloblastoma patients.myc gene amplification has beenidentified in many large cell/anaplastic medulloblastoma, butno causative link between c-myc and anaplastic changes hasbeen established. To address this, we stably overexpressedc-myc in two medulloblastoma cell lines, DAOY and UW228,and examined the changes in growth characteristics. Whenanalyzed in vitro, cell lines with increased levels of c-myc hadhigher rates of growth and apoptosis as well as significantlyimproved ability to form colonies in soft agar compared withcontrol. When injected s.c. into nu/nu mice, flank xenografttumors with high levels of c-myc in DAOY cell line backgroundwere 75% larger than those derived from control. Over-expression of c-myc was required for tumor formation byUW228 cells. Most remarkably, the histopathology of the Myctumors was severely anaplastic, with large areas of necrosis/apoptosis, increased nuclear size, and macronucleoli. Indicesof proliferation and apoptosis were also significantly higher inMyc xenografts. Thus, c-myc seems to play a causal role ininducing anaplasia in medulloblastoma. Because anaplasticchanges are often observed in recurrent medulloblastoma, wepropose that c-myc dysregulation is involved in the progressionof these malignant embryonal neoplasms. (Cancer Res 2006;66(2): 673-81)

Introduction

Medulloblastoma is the most common malignant tumor of thecentral nervous system in childhood. Treatment has improvedgreatly over the past few decades due to innovations inneurosurgical technique, better delivery of radiation therapy, andthe advent of chemotherapy [reviewed by Rood et al. (1)]. For asubset of patients, overall survival is relatively high, approaching70% to 80% (2–5). However, for patients with metastatic disease,with unresectable tumors, or who are very young, prognosis isworse (5, 6). Even for survivors, there is a high incidence of post-therapy morbidities (7). Current clinical trials are designed, in part,to identify biological features that will allow future stratification oftherapy. This may improve outcomes for high-risk patients andreduce side effects in patients with lower-risk disease.Although the presence of metastatic disease continues to be a

strong predictor of poor outcome (3, 6), several additional prog-nostic markers have been identified. One histopathologic feature,

large cell/anaplastic (LC/A) change, has been associated with poorprognosis [refs. 8–11; reviewed by Eberhart and Burger (12)]. TheLC/A phenotype is defined by the presence of large pleomorphicnuclei, numerous mitoses, and an increased apoptotic rate; it isfound in 20% to 25% of medulloblastoma (10, 13). Moderate orsevere anaplasia was associated with a statistically significantdifference in clinical outcome in a comprehensive retrospectivereview of Pediatric Oncology Group tumor specimens (10). RecentEuropean studies have noted a similar incidence of medulloblas-tomas with LC/A features and confirmed the association with poorsurvival (11, 13).The molecular mechanisms underlying the anaplastic phenotype

are poorly understood. Several groups have found amplification ofeither the c-myc or the N-myc gene locus to be associated with theLC/A phenotype (9, 14–16). However, myc gene amplification isnot universally detected in tumors with LC/A features, calling intoquestion the linkage between increased myc activity and anaplasticchange (13, 16–18). Additionally, when analyzed at a single-cell levelby in situ methods, c-myc or N-myc amplification is sometimespresent in only a subset of cells in a given tumor, including cellsthat do not appear anaplastic (15, 16). Anaplastic tumors oftenhave complex karyotypes, suggesting they have a higher degree ofgenomic instability than their nonanaplastic counterparts (14, 16).It is therefore possible that genomic instability itself drivesanaplastic change, and increased myc gene dosage is a secondaryeffect.To further explore the hypothesis that c-Myc overexpression

causes anaplastic changes in medulloblastoma, we developed amodel system using medulloblastoma cell lines engineered to stablyoverexpress c-Myc. These lines had increased rates of growth andapoptosis in vitro and in vivo and exhibited a markedly anaplastichistopathologic appearance in flank xenograft tumors. Analysis of alarge set of pathologically diverse primary medulloblastomas alsoshowed a significant association between elevated myc RNA levelscomparable with those in our model system and anaplasia.

Materials and Methods

Constructs and cell culture. The plasmid pMYCEGFP was constructedby cloning the human c-myc cDNA (19) downstream of the cytomegalovirus

(CMV) promoter in the EcoRI and SalI sites of pIRESEGFP (Clontech,

Palo Alto, CA). pLDHluc (20) was used to assess myc activity; pTK-ren was

purchased from Promega (Madison, WI). Qiagen (Valencia, CA) plasmidpurification kits were used according to the manufacturer’s instructions in

preparing DNA.

DAOY, D283, and D425 medulloblastoma (American Type Culture

Collection, Manassas, VA) cell lines were grown in MEM Zn2+ option(Richter’s modification) medium supplemented with 10% fetal bovine serum

(FBS). UW228 medulloblastoma cells were grown in DMEM/F-12, 10% FBS

and all lines were incubated at 37jC, 5% CO2. Medium and serum werepurchased from Invitrogen (Carlsbad, CA).

Transfections were carried out using the FuGene reagent (Roche

Diagnostics, Indianapolis, IN) at a ratio of 3:2 (FuGene/DNA, volume/mass)

Requests for reprints: Duncan Stearns, Department of Neuropathology, JohnsHopkins University School of Medicine, 558 Ross Building, 720 Rutland Av., Baltimore,MD 21205. Phone: 410-502-5157; Fax: 410-955-9777; E-mail: steardu@ jhmi.edu.

I2006 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-05-1580

www.aacrjournals.org 673 Cancer Res 2006; 66: (2). January 15, 2006

Research Article

Research. on March 17, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

according to the manufacturer’s instructions. Cells were plated at a densityof 20,000 to 40,000 per well in 24-well plates and collected 24 to 48 hours

following transfection. The relative luciferase activities were measured using

the Promega Dual Luciferase Reporter Assay kit. The standard luciferase

activity was normalized using the Renilla luciferase activity.Stable clones were derived from DAOY or UW228 cells transfected with

either pMYCEGFP or pIRESEGFP. Following 48 hours of incubation, G418

was added to a concentration of 500 Ag/mL. Cells were cultured in the

selection medium for an additional 2 weeks. Clones were screened for green

fluorescence by microscopy following single-cell plating in 96-well plates.

To avoid unique insertion events, multiple clones from two different

transfection experiments were derived and characterized. All myc+ lines

have been continuously passaged for >18 months without any obvious

change in phenotype.

In vitro growth analysis. Cells were plated at a density of 5,000 to15,000 viable cells per well in 24-well plates on day 0. Cell number and

viability were assessed using the Guava PCA and Viacount reagent accor-

ding to instructions (Guava Technologies, Hayward, CA). The cells werecollected by trypsin-EDTA treatment, pelleted by gentle centrifugation,

and resuspended in Guava Viacount reagent for measurement on the

Guava PCA. The Guava Viacount reagent contains two fluorescent dyes,

one cell permeable and one impermeable, except in cells that have lostmembrane integrity [propidium iodide (PI)]. This allows assessment of the

fraction of cells that are viable (membrane intact, no PI staining), dead

(high PI staining), and intermediate or midapoptotic. Multiple replicate

wells were assayed per time point in each experiment, and each subclonewas examined in two to three independent experiments. Each time point

represents the mean fold increase in the day 0 cell number from all scored

wells; error bars display SE.Cell cycle was also assessed on the Guava PCA following the

manufacturer’s technical report. Cells were plated in six-well plates at a

density of 50,000 per well on day 0 and collected on day 3. After overnight

fixation in 70% ethanol, the cells were pelleted, washed, incubated in PIstaining solution (25 Ag/mL PI in PBS, 20 Ag/mL RNase A, 0.1% Triton

X-100) for 30 minutes, and analyzed on the Guava PCA. Each experiment

represents a minimum of 5,000 gated events from each of three wells. Cell

cycle fractions were assigned using Guava software.Soft agar assays. Viable cells (n = 3,000) were plated in the appropriate

medium in 0.5% agar over a 1% agar base and covered in medium with 10%

serum. Medium was changed every 4 days for 10 to 21 days. For colony

assessment, the plates were washed with PBS, stained with 0.5% Wright’sstain in PBS, and counted with Chemidoc XRS (Becton Dickinson, San

Diego, CA) at 10-sensitivity.

Flank xenograft tumors. Viable cells (4 � 106 DAOY-derived or 2 � 106

UW228-derived) were mixed 1:1 with Matrigel (BD Biosciences, San Jose,

CA) and a total volume of 0.25 mL was injected s.c. into each flank of ether-

anaesthetized 5-week-old female nu/nu mice. Measurements of tumor size

in long and short axes were made weekly and estimates of volume weremade by the formula: width2 � length � 0.52 (21). At 8.5 to 9.5 weeks, the

mice were sacrificed and tumors were excised. Following weighing, the

tumors were split, half was snap frozen for subsequent protein and

RNA analysis and half was fixed in formalin. To assess the apoptotic andproliferative indices, primary antibodies were used on formalin-fixed,

paraffin-embedded xenograft tumor tissue. 3,3V-Diaminobenzidine immu-

noperoxidase staining was done on formalin-fixed, paraffin-embeddedxenograft tumor tissue using the Vectastain Elite system according to the

manufacturer’s instructions (Vector Laboratories, Burlingame, CA). Cleaved

caspase-3 (rabbit monoclonal, Cell Signaling Technologies, Beverly, MA) was

diluted 1:100 and MIB-1 (DAKO, Carpinteria, CA) was diluted 1:1,000;primary antibody incubations were for 45 minutes at room temperature.

Positive cell counts were determined from randomly selected high-power

fields in all tumors by a neuropathologist (T.W.A.) blinded to tumor

genotype.RNA and protein preparation and analysis. Total RNA from cultured

cells or tumor xenografts was prepared using Qiagen RNeasy kits according

to instructions. A RNase-free DNase (Qiagen) step was added to reducegenomic DNA contamination. Primary medulloblastomas were obtained

from the Department of Pathology, Johns Hopkins University School ofMedicine, with institutional review board approval. Classification of

histopathologic subtype was determined by one of two board-certified

neuropathologists (C.G.E. or P.C.B.). RNA from medulloblastoma tumor

samples snap-frozen at the time of surgery was prepared either with aQiagen kit or using Trizol reagent (22). Whole-cell protein extracts were

prepared as described previously (23).

Quantitation of c-myc mRNA was done using a two-step real-time PCR

method. Total RNA (500-2,000 ng) was reverse transcribed to cDNA usingrandom hexamers and Moloney murine leukemia virus reverse transcrip-

tase (Applied Biosystems, Foster City, CA). h-Actin and c-myc levels were

determined using purchased Applied Biosystems PDARs and Taqman

Master Mix on a Bio-Rad (Hercules, CA) Icycler (50jC � 2 minutes, 95jC �10 minutes, 45 cycles of 95jC � 15 seconds, 60jC � 60 seconds) in 20 ALreaction volumes. A standard curve was generated using dilutions of DAOY

cDNA and the unknown values were determined using the Icycler software.Each unknown (1-10 ng starting material) was run in triplicate or greater.

A SYBR Green–based real-time PCR assay was used to quantify levels of

N-myc mRNA. Primers ( forward 5V-TGAAGAGGAAGATGAAGAGGAAGA-3Vand reverse 5V-GTGACAGCCTTGGTGTTGGA-3V) were designed usingPrimer3 express software to cross an intron/exon boundary and give a

product of 80 bp. Template cDNA was added to 20 AL reactions containing

200 primer concentrations and 1� SYBR Green Master Mix (Applied

Biosystems) and run on a Bio-Rad Icycler (95jC � 10 minutes, 45 cycles of95jC � 15 seconds, 60jC � 60 seconds). Standard curves were generated

using D283 cDNA serial dilutions. Relative expression values were

normalized to h-actin levels, determined separately.Whole-cell protein extracts (30-50 Ag) were run on SDS-PAGE in Tris-

glycine buffer, electrophoretically transferred to Nytran membranes, and

blocked overnight at 4jC overnight in PBST with 5% NFDM. c-Myc protein

levels were determined using 9E10 or N262 primary antibody hybridization(Santa Cruz Biotechnology, Santa Cruz, CA; 1:500 and 1:250 dilution,

respectively) followed by anti-mouse or anti-rabbit secondary antibody

(KPL, Inc., Gaithersburg, MD; 1:5,000 dilution) and visualization by

chemiluminescence (Western Lightning, Perkin-Elmer, Foster City, CA).Densitometric quantitation was carried out using Chemidoc XRS software.

Statistical analysis. GraphPad Prism 4 statistical software (GraphPad

Software, San Diego, CA) was used for statistical analysis. Error barsrepresent SE unless indicated otherwise.

Results

Characterization of medulloblastoma cells stably over-expressing c-myc. To understand the relationship betweenc-Myc activity and medulloblastoma growth characteristics, weselected two medulloblastoma cell lines, DAOY and UW228, whichhave not been described to have amplifications in the myc locus(24, 25). As outlined, the c-myc cDNA was cloned downstream ofthe CMV immediate-early promoter in pMYCEGFP. A series ofstable transfectants carrying either this plasmid or the parentvector were derived via G418 selection and characterized. Duringthe selection process, it was noted that pMYCEGFP transfectionresulted in more EGFP-positive cells than pEGFP transfections,implying that there may be some selective advantage in c-Mycoverexpression (data not shown). As shown in Fig. 1A , c-mycmRNA levels as measured by a Taqman real-time PCR assay wereinduced 10- to 30-fold in the various DAOY-myc and UW228-myclines (designated DAOYM1-2, 13-15, UWM13-15) compared witheither the parent or the vector-transfected lines (DAOYV11,UWV1). These increases in mRNA level are similar to those seenin the upper quartile of primary human tumors (Fig. 2) andapproach levels of other cell lines (D425 and D283) known tohave high c-Myc expression (data not shown), suggesting thatour model falls into a biologically relevant range. Protein levelsrose in parallel with mRNA expression (Fig. 1B). This protein was

Cancer Research

Cancer Res 2006; 66: (2). January 15, 2006 674 www.aacrjournals.org



functional as evidenced by 1.5- to 2-fold increases in transcrip-tional activity measured by transient transfection of a c-Myc–sensitive promoter-reporter construct (Fig. 1C). The morphologyof the transfected cells was not significantly altered in culture,although the UW-myc lines tended to have a more spindledappearance than their parent line. Both DAOYM and UWM lineshad increases in the fraction of nonadherent cells and lowerplating efficiencies when compared with the non-myc lines (datanot shown) but have been passaged over 18 months withoutdifficulty.

Overexpression of c-myc alters cell growth characteristicsin vitro . As shown in Fig. 3A and B , the cell lines overexpressingc-myc had increased growth rates when cultured in standardconditions. The DAOY parent line basal growth rate was higherthan that of UW228, with confluence achieved in 6 days versus 7to 8 days, respectively, in these experiments. Myc overexpressionresulted in higher numbers of cells throughout the logarithmicgrowth phase. Statistical analysis of grouped control DAOY versusDAOYM lines on day 5, just before growth plateau, revealed a 40%increase in fold change of cell number, with a mean of 9.1 for DAOYcontrols and 12.8 for DAOYM lines (P = 0.002, unpaired, two-tailedt test). Similar analysis of control versus UWM lines on day 6showed a 2-fold increase (P < 0.001), with a mean of 12.7 for UWcontrols and 25.2 for UWM lines. Although the data shown inFig. 3A and B represent measurements of viable cells, the increasein growth rate was even more marked when comparing total cell

Figure 1. Levels and activity of c-myc are increased in a series of stablytransfected DAOY and UW228 medulloblastoma cell lines. A, c-myc mRNAlevels, determined by quantitative reverse transcription-PCR (RT-PCR), areincreased 10- to 30-fold in c-myc–transfected cell lines (DAOYM2, DAOYM14,and UWM13-15) compared with parent (DAOY and UW228) or vector-transfected (DAOYV11 and UWV1) cells. Columns, averages of more than threeindependent determinations, normalized to h-actin levels; bars, SE. B, Westernblots of protein extracts, with glyceraldehyde-3-phosphate dehydrogenase(GAPDH ) included as a loading control, document similar increases in c-Mycprotein levels. C, c-Myc transcriptional activity was measured using a transienttransfection assay with the pLDHluc plasmid positively regulated by c-Myc.Results of two experiments run in triplicate. Bars, SE.

Figure 2. Increased myc mRNA levels in anaplastic medulloblastoma.Quantitative real-time RT-PCR was used to measure c-myc (A) or N-myc (B)transcript levels in primary human medulloblastoma, other primitiveneuroectodermal tumors (PNET ), and normal human cerebelli. The highestc-myc levels were observed in three tumors with diffuse anaplasia. Interestingly,three additional anaplastic medulloblastoma with low levels of c-myc mRNA(A, arrows) had high levels of N-myc mRNA (B, arrows). The three cerebelli withhigher levels of N-myc were fetal samples. Bars, averages of values in eachhistopathologic group.

c-myc Causes Anaplasia in Medulloblastoma




number (including viable, midapoptotic, and dead cells; data notshown). As the cells reached confluence, lines overexpressing c-mychad a significantly larger fraction of nonviable cells as seen inFig. 3C . After 7 days in culture, well after the growth plateau, theDAOYM lines had a 2-fold larger percentage of PI staining deadcells than vector or parent lines, which was consistent with thesubjective microscopic finding of increased numbers of floatingcells. This phenomenon was similar in the UWM lines. In additionto its effects, increasing rates of apoptosis, c-Myc is a knownregulator of cell cycle entry. However, the fraction of cells in S orG2-M phases was only slightly larger in the DAOYM lines than incontrols (Fig. 3D). The finding that the parent line had a largeproportion of cells in active cell cycle likely represents a featureof standard culture conditions, where growth factors and nutrientsupply are not limiting. Increased c-myc activity may therefore onlyresult in a small decrease in the quiescent cell population.Overexpression of c-Myc greatly enhanced colony-forming ability

in nonadherent conditions (Fig. 4). Both parent DAOY and UW228cell lines inefficiently form colonies in soft agar. Experiments withDAOYM lines resulted in three to eight times more coloniescompared with parent DAOY or the vector-transfected lineDAOYV11. UW228M lines had a remarkably higher rate of colonyformation than the parent line, with a 10- to 40-fold increase innumber. The colonies formed by the myc-transfected DAOY lineswere larger and generally more robust, a phenomenon that wasalso more pronounced in the UW228-derived cell lines (Fig. 4,inset). In both DAOYM and UW228M lines, the relative level ofc-Myc in each subclone did not precisely correlate with theirgrowth characteristics or clonogenicity. This suggests the possibil-ity of a threshold level above which c-Myc does not have anadditional additive effect on growth.Increased c-Myc results in more rapid tumor growth

in vivo . Given the limitations of in vitro analysis, we nextexamined the effect of c-Myc overexpression on growth of s.c.flank xenograft tumors in nu/nu mice by comparing the growth ofsix DAOYM-2 tumors to eight DAOY tumors. As shown in Fig. 5,

Figure 3. Characteristics of cell lines in vitro. A and B, growth curves forDAOY- and UW228-derived cell lines expressed as fold increase from startingcell count. Points, mean of two to three independent experiments, whichincluded two to three wells per time point; bars, SE. All cell lines overexpressingc-Myc had increased rates of growth. C, high levels of c-myc were associatedwith an increased fraction of dead and apoptotic cells measured by flowcytometry after 7 days in culture. Columns, mean of three wells; bars, SE. D, cellcycle analysis done after 48 hours in culture comparing two DAOYM cell lineswith the parent DAOY. Overexpression of c-Myc marginally increases thepercentage of cells in S and G2-M phases compared with the parent cell line.

Figure 4. Myc increases colony-forming ability. Top, number of colonies formedafter 2 weeks in a series of soft agar experiments; bottom, a representativewell and a colony (inset ), highlighting the increased size of colonies formed bycells with elevated c-Myc. Columns, mean of a single experiment run in triplicate;bars, SE.

Cancer Research




the DAOYM-2 line resulted in significantly larger tumors asmeasured by either weekly volumetric assessment (Fig. 5A) orfinal tumor weight (P < 0.05, unpaired, two-tailed t test; Fig. 5Band C) compared with those generated by the parent DAOY cellline. Other DAOYM lines also resulted in larger tumors andsimilar histopathologic changes (data not shown). High levels ofc-myc RNA (Fig. 5D) and protein (Fig. 5E) were maintained inthese tumors, confirming the stability of pMYCEGFP activity. Asin the soft agar assays, overexpression of c-Myc had a much moreprofound effect on UW228 xenograft tumor formation. The parentline did not form tumors in any of six injections (tissue excisedafter 8.5 weeks showed only residual Matrigel and scar tissue).However, UWM14 or UWM15 cells resulted in tumors in 2 of 2and 4 of 6 injections, respectively, as documented by histopa-thology (see Fig. 6E). These tumors were on average much smallerthan those formed even by the parent DAOY line (Fig. 5C) butwere clearly neoplasms on slide review, indicating that c-Mycoverexpression in the UW228 cell line was required for efficientxenograft tumor formation.As shown in Fig. 6A , c-myc markedly altered the histopathologic

appearance of the DAOY xenograft tumors. Whereas the parentDAOY tumors consisted of a relatively homogeneous population ofcells with bland nuclei and small nucleoli, the DAOYM-2 tumors

had a more malignant appearance. Large areas of confluentapoptosis or necrosis very similar to the ‘‘apoptotic lakes’’ seen inprimary LC/A medulloblastoma tumors were present in all sectionsexamined. The nuclear morphology was notable for increasedatypia and macronucleoli, again similar to primary tumors with theLC/A phenotype. Interestingly, scattered ‘‘cellular wrapping’’ wasseen in the DAOYM-2 tumors (data not shown). Histologic analysisof H&E stains of the UWM tumors revealed remarkably similarfeatures (Fig. 6E).Unlike the in vitro analysis that only showed a mild increase in

the cell fraction outside of G1-G0, DAOYM2 tumors had an f2-foldincrease (mean = 83) in the number of mitotic figures counted per10 high-power fields compared with DAOY (mean = 48; Fig. 6B ,with additional comparison with a primary LC/A tumor). Similarly,Ki-67 positivity was nearly doubled (Fig. 6C). Myc also increasedapoptosis f4-fold as measured by cleaved caspase-3 staining(Fig. 6D). All of these measurements between the two groups oftumors were different with a statistical significance of P < 0.05(mitotic figures) or P < 0.01 (Ki-67 and cleaved caspase-3 staining)by unpaired, two-tailed t test. The lack of xenograft tumorformation by any UW228 control cell line precluded comparativeanalysis; however, the mitotic rate was robust in the UWM-derivedtumors, averaging 25 mitotic figures per 10 high-power fields.Primary medulloblastomas with diffuse anaplasia have high

levels of c-myc or N-myc mRNA. Amplification of either thec-myc or the N-myc locus has been strongly associated withmedulloblastoma anaplasia. We have reported previously thatpositive in situ staining for c-myc RNA correlated significantly withanaplasia (26). To further examine the association between mycand anaplasia in human tumors and to determine if c-myc levels inour model fell into the range observed in human tumors,we analyzed both c-myc and N-myc mRNA levels for 56 primarysamples by real-time PCR: 8 medulloblastoma with diffuseanaplasia, 24 medulloblastoma with nonanaplastic (classic)histology or only focal anaplasia, 8 nodular-desmoplastic medul-loblastoma, 6 primitive neuroectodermal tumors (PNETs; including2 medulloepithelioma), and 10 normal cerebella. As shown in Fig. 2,high levels of c-myc mRNA were commonly found in the tumorswith diffuse anaplasia (4 of 8 tumors, with values ranging 0.06-12.42) but infrequently in other nonanaplastic medulloblastomasand PNETs. Nodular medulloblastomas and normal cerebellum hadvery low levels of c-myc (with values ranging 0.17-1.13). This wasa statistically significant finding both by one-way ANOVA (P =0.0025) comparing all five groups and by unpaired t test (P =0.0012) comparing diffusely anaplastic tumors with all othermedulloblastoma samples. High levels of N-myc mRNA were foundin many tumor samples and fetal cerebelli. There was no significantcorrelation with elevated N-myc and any one tumor subtype;however, several patterns did emerge. Medulloblastomas with anodular phenotype, which is commonly associated with dysregu-lation of the Hedgehog signaling pathway, had high N-myc levels,consistent with previous reports of N-myc being a downstreamtarget of Hedgehog signaling (27). Additionally, of the fourmedulloblastomas with diffuse anaplasia and low c-myc levels,three had markedly elevated N-myc (Fig. 2, arrows). Thus, very highlevels of c-myc or N-myc are strongly associated with diffuseanaplastic features in medulloblastoma.

Discussion

There are several molecular markers under investigation fortheir prognostic significance in medulloblastoma. Molecular

Figure 5. Flank xenograft tumor growth in athymic nu/nu mice is increased byMyc. A, growth of eight DAOY and six DAOYM2 tumors assessed by weeklymeasurements. B, representative mice carrying DAOY or DAOYM2 tumors at8.5 weeks. C, average final excised tumor weights. Bars, SE. The differencebetween DAOY versus DAOYM2 average values was statistically significant(P < 0.05, unpaired, two-tailed t test). D, tumors maintain high levels of c-mycmRNA expression as determined in eight representative xenografts. E, c-Mycprotein levels in representative xenografts were also significantly elevated.





Figure 6. Overexpression of c-Myc promotes anaplastic changes in flank xenograft tumors. A, H&E-stained sections of DAOY- and DAOYM2-derived tumors(left and middle ) and a human LC/A medulloblastoma (right ). Myc overexpression in DAOY cells results in increased nuclear size, prominent nucleoli, and regions ofnecrosis/confluent apoptosis, all features also seen in human LC/A medulloblastoma. Original magnification, �40 (left to right ) and �600 (insets ). B, as in humanLC/A medulloblastoma (left), mitotic activity was brisk in xenografts with high c-Myc (middle ). Original magnification, �600. The number of mitotic figures per10 high-power fields in DAOYM2 xenografts was significantly higher than in DAOY xenografts (P < 0.01, unpaired, two-tailed t test). C, the Ki-67 proliferation indexwas also significantly higher in DAOYM2 tumors (P < 0.001, unpaired, two-tailed t test. Original magnification, �100. D, immunohistochemical stains for cleavedcaspase-3 were used to highlight apoptotic cells in DAOY and DAOYM2 xenografts. Original magnification, �100. Both single apoptotic cells and confluentregions of apoptosis were detected. The number of single apoptotic cells per 10 high-power fields was significantly higher in high c-Myc xenografts (P < 0.001,unpaired, two-tailed t test). E, photomicrographs of H&E-stained sections of UW228- and UWM14-derived tumors. Original magnification, �40 (left and middle) and�600 (right ). The UW228 ‘‘tumor’’ is actually normal tissue with residual Matrigel. The low-power view highlights an area of necrosis/confluent apoptosis similarto that seen in the DAOYM tumors. The high-power view shows the similar nuclear features seen in primary LC/A medulloblastoma and DAOYM-derived tumors.

Cancer Research




features, such as isolated chromosome 17p loss, HIC1 genehypermethylation, and elevated ERBB2 protein expression, havebeen associated with poor outcomes, whereas high TRKCexpression has been associated with improved survival [brieflyreviewed by Fisher et al. (28)]. Immunopositivity for p53 has alsobeen shown to be a negative prognostic factor (6, 29, 30). Inaddition, there is ample evidence that myc plays an importantrole in medulloblastoma biology. Increased c-myc activity inmedulloblastoma cell lines has been shown to occur throughboth gene amplification (31) and altered transcriptional regu-lation (32). In primary tumor specimens, two large retrospectiveanalyses showed that high levels of c-myc mRNA were found ina much larger proportion of cases than could be explainedsolely by gene amplification and were associated with pooroutcome (33, 34). These studies did not evaluate anaplasia, so adirect link between expression of c-myc and histologic subtypecould not be made. However, we have shown recently that elevationin c-myc mRNA level as identified by in situ hybridization isassociated not only with poor clinical outcome but also withanaplasia (26).Here, we show that there is a causative link between increased

c-Myc and anaplasia in medulloblastoma. We created a series ofstably transfected medulloblastoma cell lines expressing high levelsof c-Myc. The higher levels of c-myc mRNA level were mirrored byincreases in protein and activity, implying that mRNA measure-ment was a reasonable surrogate for relative activity. In vitro , onlysubtle changes in cellular morphology were noted, but increasedrates of growth and apoptosis occurred. The ability to formcolonies in soft agar was greatly increased in terms of bothquantity and colony size. Myc overexpression also resulted insignificantly larger flank xenograft tumors and, most remarkably, amicroscopic phenotype akin to the LC/A subtype of medulloblas-toma. Similarities to human LC/A medulloblastoma includedincreases in nuclear atypia, apoptosis, and mitotic index alongwith macronucleoli and occasional ‘‘cellular wrapping.’’ Mycoverexpression was required for efficient establishment of flankxenograft tumors in the UW228 cell line. Moreover, increased Mycactivity in UW228 cells resulted in highly similar LC/A changes,indicating that the effects of Myc on phenotype are not confinedto the DAOY cell line. Finally, analysis of a large set of primarymedulloblastoma tumors showed strong correlation betweenc-myc mRNA level and diffuse anaplasia. Taken together, thesedata strongly suggest that increased c-myc activity not only isassociated with anaplasia in medulloblastoma but also inducesanaplasia.The role of N-myc is potentially very interesting and one that

we have addressed before (26). It may be that it fulfills a similarrole as c-Myc in the development of anaplasia, as there does seemto be some functional redundancy between the two proteins.N-myc can restore viability in c-myc�/� mice (35) and a modifiedornithine decarboxylase promoter can detect both c-myc andN-myc activity (36). However, N-myc , unlike c-myc , has beenidentified as a crucial downstream target of the sonic hedgehog(shh)/PTCH pathway both in the normal developing cerebellumand in nodular medulloblastomas (27, 37–40). This is consistentwith our finding of higher levels of N-myc mRNA in fetal cerebelliand nodular tumors. c-myc mRNA, conversely, is not found inhigh levels in either of these tissues. This dichotomy, theprognostic value of c-myc expression in nonanaplastic tumors,and the role of N-myc in anaplastic progression will requirefurther elucidation.

The distinctive changes in the histopathologic appearance andin vivo aggressiveness induced by c-Myc are much more profoundthan the effects in vitro. Some of the differences are a matter ofdegree; there are increases in growth and apoptosis in culture thatare more pronounced when analyzed in vivo . However, the ability ofMyc to increase growth in nonadherent conditions implies thatthere are changes that cannot be quantified by standard cell culturealone. In our xenograft model, additional factors, such asangiogenesis and/or decreased dependence on limiting growthfactors, may also play a role. This seems highly probable, given thatthe cell intrinsic changes leading to an increased growth rate in

Figure 7. Proposed multistep model for progression to anaplasia.





UW228M cells is very unlikely to be the sole explanation for theirability to establish tumors in vivo . Given the wide range of functionsthat Myc regulates (extensively reviewed in ref. 41), this is notsurprising.The role of c-myc in oncogenesis has proven to be complex. In

some cases, such as lymphoma, c-Myc overexpression alone leadsto transformation (42). In other cases, such as breast cancer, c-mycdysregulation leads to more aggressive disease but is insufficient tocause tumorigenesis alone [reviewed by Jamerson et al. (43)]. Theability of Myc alone to induce medulloblastomas may be limited aswell. In a mouse model system, Rao et al. found that combiningc-Myc overexpression with ectopic shh production using theirRCAS-TVA system resulted in more tumors than shh alone (44).In their hands, however, c-Myc by itself did not cause tumors,resulting only in areas of subependymal cellular proliferation. Incontrast, the Li-Fraumeni, Gorlin, and Turcot tumor predispositionsyndromes as well as mouse medulloblastoma models indicatethat the Wnt, Hedgehog, and p53 pathways can drive medullo-blastoma initiation (45, 46).We believe that, although c-myc may play a role in the initiation

of some medulloblastomas, its increased expression and activity is

primarily observed as tumors progress and become more ag-gressive. This can occur either at recurrence or in anaplasticsubclones present in the specimen resected at initial presentation.As outlined in Fig. 7, we propose a multistep progression model,with myc oncogenes playing a critical role in the development ofthe LC/A phenotype in medulloblastoma (12). Genes or pathwaysassociated in other studies with anaplasia, including TP53/ARF,ERBB2, OTX2, and hTERT, may also play a role in medulloblastomaprogression (3, 47–49) How these specific molecular events relateto anaplastic progression has yet to be determined, but they maysubstitute in some way for myc dysregulation. Regardless, it is clearthat c-Myc overexpression is sufficient to induce anaplasia in atleast two medulloblastoma cell lines and is strongly associated withthe anaplastic phenotype in primary tumors.

Acknowledgments

Received 5/6/2005; revised 10/19/2005; accepted 11/3/2005.Grant support: Children’s Brain Tumor Foundation (P.C. Burger) and NIH training

grants T32CA-009574 (D. Stearns) and K08 NS43279 (C.G. Eberhart).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 accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Cancer Research


References1. Rood BR, Macdonald TJ, Packer RJ. Current treatmentof medulloblastoma: recent advances and future chal-lenges. Semin Oncol 2004;31:666–75.

2. Packer RJ, Goldwein J, Nicholson HS, et al. Treatmentof children with medulloblastomas with reduced-dosecraniospinal radiation therapy and adjuvant chemo-therapy: a Children’s Cancer Group study. J Clin Oncol1999;17:2127–36.

3. Gajjar A, Hernan R, Kocak M, et al. Clinical,histopathologic, and molecular markers of prognosis:toward a new disease risk stratification system formedulloblastoma. J Clin Oncol 2004;22:984–93.

4. Taylor RE, Bailey CC, Robinson KJ, et al. Impact ofradiotherapy parameters on outcome in the Interna-tional Society of Paediatric Oncology/United KingdomChildren’s Cancer Study Group PNET-3 study ofpreradiotherapy chemotherapy for M0-M1 medulloblas-toma. Int J Radiat Oncol Biol Phys 2004;58:1184–93.

5. Rutkowski S, Bode U, Deinlein F, et al. Treatment ofearly childhood medulloblastoma by postoperativechemotherapy alone. N Engl J Med 2005;352:978–86.

6. Ray A, Ho M, Ma J, et al. A clinicobiological modelpredicting survival in medulloblastoma. Clin Cancer Res2004;10:7613–20.

7. Ris MD, Packer R, Goldwein J, Jones-Wallace D, BoyettJM. Intellectual outcome after reduced-dose radiationtherapy plus adjuvant chemotherapy for medulloblas-toma: a Children’s Cancer Group study. J Clin Oncol2001;19:3470–6.

8. Giangaspero F, Rigobello L, Badiali M, et al. Large-cellmedulloblastomas. A distinct variant with highlyaggressive behavior. Am J Surg Pathol 1992;16:687–93.

9. Brown HG, Kepner JL, Perlman EJ, et al. ‘‘Large cell/anaplastic’’ medulloblastomas: a Pediatric OncologyGroup study. J Neuropathol Exp Neurol 2000;59:857–65.

10. Eberhart CG, Kepner JL, Goldthwaite PT, et al.Histopathologic grading of medulloblastomas: a Pediat-ric Oncology Group study. Cancer 2002;94:552–60.

11. McManamy CS, Lamont JM, Taylor RE, et al.Morphophenotypic variation predicts clinical behaviorin childhood non-desmoplastic medulloblastomas.J Neuropathol Exp Neurol 2003;62:627–32.

12. Eberhart CG, Burger PC. Anaplasia and grading inmedulloblastomas. Brain Pathol 2003;13:376–85.

13. Lamont JM, McManamy CS, Pearson AD, Clifford SC,Ellison DW. Combined histopathological and molecularcytogenetic stratification of medulloblastoma patients.Clin Cancer Res 2004;10:5482–93.

14. Eberhart CG, Kratz JE, Schuster A, et al. Comparativegenomic hybridization detects an increased number ofchromosomal alterations in large cell/anaplastic medul-loblastomas. Brain Pathol 2002;12:36–44.

15. Aldosari N, Bigner SH, Burger PC, et al. MYCC andMYCN oncogene amplification in medulloblastoma. Afluorescence in situ hybridization study on paraffinsections from the Children’s Oncology Group. ArchPathol Lab Med 2002;126:540–4.

16. Leonard JR, Cai DX, Rivet DJ, et al. Large cell/anaplastic medulloblastomas and medullomyoblasto-mas: clinicopathological and genetic features. J Neuro-surg 2001;95:82–8.

17. Badiali M, Pession A, Basso G, et al. N-myc and c-myconcogenes amplification in medulloblastomas. Evidenceof particularly aggressive behavior of a tumor withc-myc amplification. Tumori 1991;77:118–21.

18. Gilbertson R, Wickramasinghe C, Hernan R, et al.Clinical and molecular stratification of disease risk inmedulloblastoma. Br J Cancer 2001;85:705–12.

19. Stone J, de Lange T, Ramsay G, et al. Definition ofregions in human c-myc that are involved in transfor-mation and nuclear localization. Mol Cell Biol 1987;7:1697–709.

20. Shim H, Dolde C, Lewis BC, et al. c-Myc trans-activation of LDH-A: implications for tumor metabolismand growth. Proc Natl Acad Sci U S A 1997;94:6658–63.

21. Boehm T, Folkman J, Browder T, O’Reilly MS.Antiangiogenic therapy of experimental cancer doesnot induce acquired drug resistance. Nature 1997;390:404–7.

22. Sambrook J, Fritsch EF, Maniatis, T. Molecularcloning: a laboratory manual. New York: Cold SpringHarbor Laboratory Press; 1989.

23. Fan X, Mikolaenko I, Elhassan I, et al. Notch1 andnotch2 have opposite effects on embryonal brain tumorgrowth. Cancer Res 2004;64:7787–93.

24. Jacobsen PF, Jenkyn DJ, Papadimitriou JM. Establish-ment of a human medulloblastoma cell line and itsheterotransplantation into nude mice. J NeuropatholExp Neurol 1985;44:472–85.

25. Keles GE, Berger MS, Srinivasan J, Kolstoe DD,Bobola MS, Silber JR. Establishment and characteriza-tion of four human medulloblastoma-derived cell lines.Oncol Res 1995;7:493–503.

26. Eberhart CG, Kratz J, Wang Y, et al. Histopathologicaland molecular prognostic markers in medulloblastoma:c-myc, N-myc, TrkC, and anaplasia. J Neuropathol ExpNeurol 2004;63:441–9.

27. Kenney AM, Cole MD, Rowitch DH. Nmyc upregula-

tion by sonic hedgehog signaling promotes proliferationin developing cerebellar granule neuron precursors.Development 2003;130:15–28.

28. Fisher PG, Burger PC, Eberhart CG. Biologic riskstratification of medulloblastoma: the real time is now.J Clin Oncol 2004;22:971–4.

29. Woodburn RT, Azzarelli B, Montebello JF, Goss IE.Intense p53 staining is a valuable prognostic indicatorfor poor prognosis in medulloblastoma/central nervoussystem primitive neuroectodermal tumors. J Neurooncol2001;52:57–62.

30. Eberhart CG, Chaudhry A, Daniel RW, Khaki L, ShahKV, Gravitt PE. Increased p53 immunopositivity inanaplastic medulloblastoma and supratentorial PNETis not caused by JC virus. BMC Cancer 2005;5:19.

31. Bigner SH, Mark J, Friedman HS, Biegel JA, Bigner DD.Structural chromosomal abnormalities in human medul-loblastoma. Cancer Genet Cytogenet 1988;30:91–101.

32. Siu IM, Lal A, Blankenship JR, Aldosari N, Riggins GJ.c-Myc promoter activation in medulloblastoma. CancerRes 2003;63:4773–6.

33. Herms J, Neidt I, Luscher B, et al. C-MYC expressionin medulloblastoma and its prognostic value. Int JCancer 2000;89:395–402.

34. Grotzer MA, Hogarty MD, Janss AJ, et al. MYCmessenger RNA expression predicts survival outcome inchildhood primitive neuroectodermal tumor/medullo-blastoma. Clin Cancer Res 2001;7:2425–33.

35. Malynn BA, de Alboran IM, O’Hagan RC, et al. N-myccan functionally replace c-myc in murine development,cellular growth, and differentiation. Genes Dev 2000;14:1390–9.

36. Iyengar RV, Pawlik CA, Krull EJ, et al. Use of amodified ornithine decarboxylase promoter to achieveefficient c-MYC- or N-MYC-regulated protein expres-sion. Cancer Res 2001;61:3045–52.

37. Knoepfler PS, Cheng PF, Eisenman RN. N-myc isessential during neurogenesis for the rapid expansion ofprogenitor cell populations and the inhibition ofneuronal differentiation. Genes Dev 2002;16:2699–712.

38. Pomeroy SL, Tamayo P, Gaasenbeek M, et al.Prediction of central nervous system embryonal tumouroutcome based on gene expression. Nature 2002;415:436–42.

39. Oliver TG, Grasfeder LL, Carroll AL, et al. Transcrip-tional profiling of the sonic hedgehog response: acritical role for N-myc in proliferation of neuronalprecursors. Proc Natl Acad Sci U S A 2003;100:7331–6.

40. Kenney AM, Widlund HR, Rowitch DH. Hedgehogand PI-3 kinase signaling converge on Nmyc1 to





promote cell cycle progression in cerebellar neuronalprecursors. Development 2004;131:217–28.

41. Oster SK, Ho CS, Soucie EL, Penn LZ. The myconcogene: MarvelouslY Complex. Adv Cancer Res 2002;84:81–154.

42. Adams JM, Harris AW, Pinkert CA, et al. The c-myconcogene driven by immunoglobulin enhancers induceslymphoid malignancy in transgenic mice. Nature 1985;318:533–8.

43. Jamerson MH, Johnson MD, Dickson RB. Of mice andMyc: c-Myc and mammary tumorigenesis. J MammaryGland Biol Neoplasia 2004;9:27–37.

44. Rao G, Pedone CA, Coffin CM, Holland EC, Fults DW.c-Myc enhances sonic hedgehog-induced medulloblas-toma formation from nestin-expressing neural progen-itors in mice. Neoplasia 2003;5:198–204.

45. Taylor MD, Mainprize TG, Rutka JT. Molecularinsight into medulloblastoma and central nervoussystem primitive neuroectodermal tumor biology fromhereditary syndromes: a review. Neurosurgery 2000;47:888–901.

46. Hesselager G, Holland EC. Using mice to decipher themolecular genetics of brain tumors. Neurosurgery 2003;53:685–94; discussion 95.

47. Frank AJ, Hernan R, Hollander A, et al. The TP53-ARFtumor suppressor pathway is frequently disrupted inlarge/cell anaplastic medulloblastoma. Brain Res MolBrain Res 2004;121:137–40.

48. Di C, Liao S, Adamson DC, et al. Identification ofOTX2 as a medulloblastoma oncogene whose productcan be targeted by all-trans retinoic acid. Cancer Res2005;65:919–24.

49. Fan X, Wang Y, Kratz J, et al. hTERT geneamplification and increased mRNA expression in centralnervous system embryonal tumors. Am J Pathol 2003;162:1763–9.



2006;66:673-681. Cancer Res Duncan Stearns, Aneeka Chaudhry, Ty W. Abel, et al. c-Myc Overexpression Causes Anaplasia in Medulloblastoma

Updated version

http://cancerres.aacrjournals.org/content/66/2/673

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/66/2/673.full#ref-list-1

This article cites 48 articles, 18 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/66/2/673.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. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://cancerres.aacrjournals.org/content/66/2/673To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/66/2/673.full#ref-list-1

http://cancerres.aacrjournals.org/content/66/2/673.full#related-urls

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

mailto:[email protected]



Documents

c-Myc Overexpression Causes Anaplasia in Medulloblastoma · manufacturer’s technical report. Cells were plated in six-well plates at a density of 50,000 per well on day 0 and collected