University of Southern California Dissertations and Theses

BMI-1 expression inversely correlates with differentiation but not clinical outcome or CDKN2A expression in Ewing sarcoma family of tumors

(USC Thesis Other)

BMI-1 expression inversely correlates with differentiation but not clinical outcome or CDKN2A expression in Ewing sarcoma family of tumors

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

BMI-1 EXPRESSION INVERSELY CORRELATES WITH DIFFERENTIATION BUT
NOT CLINICAL OUTCOME OR CDKN2A EXPRESSION IN EWING SARCOMA
FAMILY OF TUMORS

by

John Anthony van Doorninck

___________________________________________________________________

A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(CLINICAL AND BIOMEDICAL INVESTIGATIONS)

August 2009

Copyright 2009 John Anthony van Doorninck

ii
DEDICATION

This work is dedicated with greatest respect and thanks to the memory of Dr.
Stephen Qualman.

iii
ACKNOWLEDGEMENTS
The author would like to thank the staff at the Children’s Cooperative Tissue
Network at Columbus, OH, for their exceptional skill and service; Minerva Mongeotti,
Sue Ann Phung, Betty Schaub, Long Hung and Jasmine Jenabi for technical
assistance; the Columbus Children' s Research Institute for Childhood Cancer (Dr.
Qualman' s Bioinformatics group) for TMA scanning and imaging; Dr. Michele Wing,
Dr. Tim Triche and members of the Lawlor lab for helpful discussion; Dr. Chizuko
Okamatsu and Dr. Hiro Shimada for assistance with pathologic review; Dr. Marc
Ladanyi for graciously providing samples of Ewing' s tumors; Dr. Richard Womer, Dr.
Mark Krailo, and Children' s Oncology Group for providing clinical outcomes data;
Dr. Richard Sposto and Lingyun Ji for invaluable support with biostatistical analysis;
and Dr. Elizabeth Lawlor for her mentorship, support, friendship, passion for
science, and exemplification of excellence.

iv
TABLE OF CONTENTS
Dedication ii
Acknowledgements iii
List of Tables v
List of Figures vi
Abbreviations vii
Abstract x
Introduction 1
Materials and Methods 3
Sample Accrual 3
Immunohistochemical Analysis 4
Molecular Analysis of Gene Expression and p16 Deletion 5
Clinical Correlates and Outcomes Analysis 6

Results 7
Discussion 18
References 23

v
LIST OF TABLES
Table 1: Clinical and molecular features of ESFT samples 8
evaluated for BMI-1 expression

Table 2: Association of BMI-1 expression with molecular and 12
clinical features associated with differential presentation
and/or outcome in ESFT

Table 3: Clinical Outcomes Data 13

Table 4: Relationship between BMI-1 expression and p16/p53 status 15

Table 5: Analysis of correlation between BMI-1 expression and 18
markers of neural crest and mesenchymal differentiation

vi
LIST OF FIGURES
Figure 1: Primary ESFT display differential expression of BMI-1 9

Figure 2: Kaplan-Meier survival analysis of ESFT patients stratified 14
according to BMI-1 immunohistochemistry score

Figure 3: No correlation exists between BMI-1 and either CDKN2A or 16
CDKN1A in ESFT

vii
ABBREVIATIONS
ACHE Acetylcholinesterase
ADRA2B Adrenergic receptor, alpha 2b
ALPL Alkaline phosphatase, liver/bone/kidney
ARF ADP ribosylation factor
AZ Arizona
BMI-1 B-cell-specific Moloney murine leukemia virus integration site 1
CA California
CCI Committee for Clinical Investigations
CDKN1A Cyclin-dependent kinase inhibitor 1A
CDKN2A Cyclin-dependent kinase inhibitor 2A
CHLA Childrens Hospital Los Angeles
CHRNA2 Cholinergic receptor, nicotinic, alpha 2
CHRNA4 Cholinergic receptor, nicotinic, alpha 4
CHRND Cholinergic receptor, nicotinic, delta
CHRNG Cholinergic receptor, nicotinic, gamma polypeptide
CHTN Cooperative Human Tissue Network
COG Childrens Oncology Group
COL10A1 Collagen, type X, alpha 1
DNA Deoxyribonucleic acid
DP Digital Photography
DRD4 Dopamine receptor 4
EDTA Ethylenediaminetetraacetic acid
EFS Event free survival
viii
ERG ETS related gene
ESFT Ewing sarcoma family of tumors
ETS Erythroblastosis transforming sequence
ETV-1 ETS variant gene 1
EZH2 Enhancer of zeste homolog 2
EWS Ewing sarcoma
E1A Early region 1A
E1AF E1A enhancer binding protein
FABP4 Fatty acid binding protein 4
FDR False discovery rate
FEV Fifth Ewing variant
FFPE Formalin-fixed paraffin-embedded
FLI-1 Friend leukemia virus integration 1
GAPDH Glyceraldehyde-3-phosphate dehydrogenase
GFAP Glial fibrillary acidic protein
H & E Hematoxylin and eosin
HIPPA Health Insurance Portability and Accountability Act
Inc Incorporated
ISH In situ hybridization
LDH Lactate dehydrogenase
LPL Lipoprotein lipase
MA Massachusetts
MO Missouri
MSKCC Memorial Sloan Kettering Cancer Center
ix
NIH-3T3 National Institutes of Health - 3-day transfer, inoculum 3 x 10
5
cells
OLIG1 Oligodendrocyte transcription factor 1
OLIG2 Oligodendrocyte transcription factor 2
OS Overall survival
PCR Polymerase chain reaction
POU4F1 POU domain, class 4, transcription factor 1
PP Percentage of positive cells
PPARG Peroxisome proliferator activated receptor gamma
PRPH Peripherin
RB Retinoblastoma
RNA Ribonucleic acid
RT-PCR Reverse transcriptase polymerase chain reaction
RUNX2 Runt-related transcription factor 2
SI Staining intensity
SOX9 SRY (sex determining region Y)-box 9
SOX10 SRY (sex determining region Y)-box 10
SPP1 Secreted phosphoprotein 1
ST sense target
TH Tyrosine hydroxylase
TMA Tissue microarray
USA United States of America

x
ABSTRACT
Ewing sarcoma family tumors (ESFT) are highly undifferentiated bone and soft
tissue tumors that primarily affect children. Metastatic disease at diagnosis portends
a worse prognosis and remains the only reliable predictor of outcome. The
polycomb protein BMI-1 is associated with worse clinical outcome in some human
cancers. This study evaluates the clinical and biologic significance of BMI-1
expression in ESFT.
Using immunohistochemical analysis, 80% of ESFT were found to robustly
express BMI-1 in all tumor cells, while 15% lacked BMI-1 expression. Analysis of 79
tumors revealed no association between BMI-1 and clinical presentation, outcome,
loss of p16 or mutation of p53. Significantly, however, BMI-1 levels were highly
inversely correlated with markers of neural crest differentiation. Together these
findings demonstrate that ESFT cells nearly universally express BMI-1 and suggest
that high-level expression of BMI-1 underlies the highly undifferentiated phenotype
that characterizes this tumor family.
1
INTRODUCTION
The Ewing sarcoma family of tumors (ESFT) are highly malignant neoplasms
of bone and soft tissue that affect approximately one per 2.9 million people under 20
years of age (Bernstein et al., 2006). ESFT are identified by hallmark reciprocal
translocations which fuse part of the EWS gene on chromosome 22 with a fragment
of an ETS gene family (Burchill, 2003). In nearly 90% of cases, the EWS fusion
partner is the FLI1 gene on chromosome 11, leading to creation of an EWS-FLI1
fusion (de Alava et al., 1998). Other ETS family members that pair with EWS
include ERG and more rarely, ETV1, E1AF, and FEV (Burchill, 2003). The
resulting chimeric proteins function as oncogenic transcription factors and are
believed to be the initiating genetic events in ESFT tumorigenesis. Although their
histogenesis remains elusive, ESFT display varying degrees of neural differentiation
and recent studies suggest that tumors may arise from malignant transformation of
mesenchymal or neural crest stem or progenitor cells (Castillero-Trejo et al., 2005;
Riggi et al., 2005; Tirode et al., 2007; Meltzer, 2007; Cole et al., 2008).
Despite intensification of therapy, overall survival rates for ESFT have not
seen dramatic improvements in the past 20 years and current survival estimates are
70% for patients with localized disease and 15-20% for patients who present with
metastases (Rodriguez-Galindo et al., 2007). With the exception of metastasis, no
clinical features are reliable predictors of relapse or survival in newly diagnosed
patients although increased age (³14 or 18 years) (Rodriguez et al., 2007; Jenkin et
al., 2002), large tumor volume (Jenkin et al., 2002; Ahrens et al., 1999; Paulussen et
al., 2001), poor response to induction therapy (Jenkin et al., 2002), axial tumor
location (Rodriguez et al., 2007; Jenkin et al., 2002), and elevated peripheral blood
2
LDH (Bacci et al., 1999) have all been reported to be associated with worse
outcomes. In terms of tumor genetics, a worse clinical outcome has been reported
in association with presence of a variant EWS-FLI1 fusion type (de Alava et al.,
1998; Zoubek et al., 1996), gain of chromosome 1q or loss of 16q (Hattinger et al.,
2002), genomic loss of p16 (Wei et al., 2000; Huang et al., 2005), mutation of p53
(Huang et al., 2005; Abudu et al., 1999; de Alava et al., 2000). Currently, presence
or absence of metastasis is the only criterion that is used to stratify patients into
different treatment groups at the time of diagnosis. Accessible and reliable
pathologic tools for identification of high-risk, localized tumors are needed if
improvements in outcome are to be realized for the 30% of patients who eventually
succumb to their disease.
Stem cells are multipotent cells with unlimited capacity for self-renewal
(Marshak et al., 2001). The genes and pathways that control the self-renewal of
normal stem cells are often deregulated in cancer and cancer-associated genetic
lesions frequently induce malignant transformation by hijacking or inappropriately
reactivating normal stem cell programs (Krivtsov et al., 2006; Pardal et al., 2003).
BMI-1 (B-cell murine leukemia integration site 1) is a member of the polycomb group
gene family that promotes self-renewal of normal hematopoietic, neural and neural
crest stem cells through epigenetic repression of developmental, cell cycle and
senescence pathways (reviewed in Gil, Bernard & Peters, 2005). In addition, BMI-1
functions as an oncogene in many human cancers and has been implicated in the
self-renewal of tumor initiating cancer stem cells in leukemia as well as some solid
tumors (Lessard & Sauvageau, 2003; Liu et al., 2006; Prince et al., 2007).
Mechanistically, BMI-1 maintains stemness and promotes tumorigenesis primarily
3
through transcriptional repression of the p16/ARF-encoding CDKN2A locus which
results in functional suppression of the RB and p53 tumor-suppressor pathways
(Lessard & Sauvageau, 2003; Jacobs et al., 1999; Smith et al., 2003). Consistent
with this, analysis of some primary tumors has demonstrated an inverse correlation
between BMI-1 and expression of p16/ARF (Vonlanthen et al., 2001; Pietersen et
al., 2008). Importantly, recent studies have shown that over-expression of BMI-1 is
associated with a worse clinical outcome in several hematologic, epithelial, and
neural malignancies (Chowdhury et al., 2007; Feng et al., 2007; Kim et al., 2004; Liu
et al., 2008; van Galen et al., 2007; Wang et al., 2008).
We have previously reported that BMI-1 is highly expressed by ESFT and
functions to promote the tumorigenicity of ESFT cell lines independently of CDKN2A
repression (Douglas et al., 2008). In the current study we have evaluated the
clinical and pathologic implications of BMI-1 expression in primary tumors.

MATERIALS AND METHODS

Sample Accrual

Formalin-fixed paraffin-embedded (FFPE) and fresh-frozen ESFT sections
were acquired from tumor banks at the Children’s Oncology Group (COG)
Biorepository in Columbus, Ohio (Cooperative Human Tissue Network -- CHTN),
and Childrens Hospital Los Angeles (CHLA). All specimens were obtained from
chemotherapy-naïve tumors unless otherwise specified. In addition, tissue
microarray (TMA) sections of genetically characterized ESFT specimens (Huang et
al., 2005) were obtained from Memorial Sloan Kettering Cancer Center (MSKCC).
For CHTN specimens, samples from patients who had been registered on a single
4
clinical trial were preferentially obtained. This COG trial, AEWS0031, was a phase
III study that ran from 2001-2005 and was restricted to patients with localized ESFT.
The study was designed to determine if interval compression of chemotherapy to a
2-weekly from a 3-weekly regimen would improve outcome (Womer et al., 2008).
Diagnosis of ESFT was confirmed for all specimens by pathologic review at the
three contributing sites. Unstained FFPE sections were obtained as single tumor
slides (one sample per slide), or in the form of multiple-sample TMAs (30-35 unique
ESFT tumors/TMA). Available clinical outcomes data were obtained from chart
review for CHLA cases and through the Biostatistical Office of the COG for samples
acquired from the CHTN Biorepository. All specimens and correlative data were
obtained in compliance with HIPAA regulations and following protocol review and
approval by the Committee for Clinical Investigations (CCI) at CHLA.

Immunohistochemical Analysis
Four-micron sections were cut from banked single tumor FFPE or TMA blocks
and mounted on slides for staining. Sections were deparaffinized, pretreated with
CCI
TM
(Tris/Borate/EDTA buffer pH 8, Ventana Medical Systems, Inc., Tucson, AZ,
USA), and incubated with anti-BMI-1 antibody (1:50, Millipore, Billerica, MA, USA)
according to the BenchMark UHC/ISH Staining Module
R
BMI-1-T 1/50 protocol (32
minutes at 42
O
C). All sections were then treated with iView detection kit (Ventana
Medical Systems, Inc.) according to the manufacturer’s instruction. Adjacent
sections were stained with H & E to verify the presence of viable tumor tissue.
Individually stained tumor slides were directly visualized and scored by direct light
microscopy and photomicrographs acquired via digital camera (DP-11, Olympus,
5
Tokyo, Japan). Stained TMA sections were sent to digitally imaged at Columbus
Children’s Research Institute Biopathology Center using Aperio ScanScopeT2
software (Aperio, Vista, CA).
Stained sections were assigned a BMI-1 score using a previously published
criteria in which both percentage of positive cells (PP) and staining intensity (SI) are
considered. (Wang et al., 2008). In brief, the PP value (ranging from 0 to 3 for < 5%
to > 50% positive cells, respectively) is multiplied by the SI value (which ranges from
0 for absent to 3 for strongly positive) to generate a composite PPxSI value that
ranges from 0-9. Tumor sections with composite values of 0-1 are assigned a BMI-1
score of negative; 2-3 is designated 1+; 4-6 is 2+ and values >6 receive a score of
3+.

Molecular Analysis of Gene Expression and p16 Deletion
Total RNA was isolated from banked diagnostic ESFT biopsies containing at
least 50% viable tumor cells using Qiagen miRNA kit (Qiagen, Valencia, CA). BMI-1,
CDKN2A and CDKN1A expression levels were measured by quantitative RT-PCR
using validated Taqman Assays (Applied Biosystems, Foster City, CA) as described
(Douglas et al., 2008), and expression of each gene was normalized to the average
expression of 2-housekeeping genes (GAPDH and ACTIN) in the same sample
using the formula % expression = 2
-Ct
x 100. Statistical correlations between BMI-1
and cell-cycle genes were determined using Prism (GraphPad Software, Inc., La
Jolla, CA). Biopsies with 70% tumor content and RNA-integrity (RIN) values of
greater than 4.0 (determined by Agilent LabChip bioanalyzer, Agilent, Santa Clara,
CA) were subjected to whole genome expression profiling using Affymetrix
6
GeneChip Human Exon 1.0 ST oligonucleotide microarrays. Preparation and
labeling of samples for array hybridization was performed in the CHLA genome core
facility according to Affymetrix protocols. Signal intensities from core probesets were
quantile normalized by robust multichip averaging and transcript expression levels
median summarized using Partek Genomics Suite software (Partek, St. Louis, MO).
Statistical correlations between BMI-1 and mesenchymal and neural crest
differentiation genes were assessed with the Spearman’s rank correlation tests
(Hollander & Wolfe, 1973), with the expected false discovery rate (FDR) controlled
using the Benjamini and Hochberg procedure (Benjamini & Hochberg, 1995).
Genes with a Spearman correlation coefficient of > ±0.6 and a FDR adjusted p-
value of 0.05 were considered to be highly significantly correlated with BMI-1.
Statistical calculations were performed in R version 2.8.1 (http://www.R-project.org)
(Team, 2008).
To assess for p16

deletions, genomic DNA was isolated from tumors with
>70% viable tumor cell content using the EZ1 DNA tissue kit (Qiagen, Valencia,
CA). Real-time PCR analysis of the p16 gene locus was performed using previously
published primers and conditions (Labuhn et al., 2001). Calculation of exon 1a
(p16-specific) copy number was determined relative to the 2N GAPDH copy number
in each sample as described (Berggren et al., 2003). Samples with copy number
ratios of <0.4 were determined to be homozygously deleted for the p16 gene.

Clinical Correlates and Outcomes Analysis
Correlations between BMI-1 expression and patient or disease characteristics
were analyzed via Pearson’s chi-square test, or by Fisher’s exact test when the
7
number of cases was limited. Event free survival (EFS) was defined as the minimum
interval from the date of diagnosis to the date of tumor recurrence, progression,
occurrence of a second malignancy, death, or the last follow-up. Overall survival
(OS) was defined as the interval from the date of diagnosis to the date of death or
the last follow-up. Estimates of EFS and OS percent were based on the product-limit
(Kaplan-Meier) estimate with Greenwood standard errors (Cox & Oakes, 1984).
The association of EFS and OS with BMI-1 protein expression was tested using the
logrank test, either univariately or with stratification based on stage at presentation
(localized versus metastatic), patient age, tumor location (pelvic versus non-pelvic),
treatment era, and EWS-FLI-1 transcript type (Cox & Oakes, 1984). Survival
analyses were performed using STATA software version 9.2. All reported p-values
were two-sided and a p-value less than 0.05 was considered to be statistically
significant.

RESULTS

BMI-1 is diffusely expressed by the majority of ESFT. To assess the level
and pattern of BMI-1 protein expression in ESFT we analyzed tumor sections from
130 archived, formalin-fixed, paraffin-embedded tumor samples. The patient and
tumor characteristics of this large sample are, in the main, reflective of other
population-based ESFT studies in terms of both clinical and molecular features
(Table 1) (de Alava et al., 1998; Rodriguez-Galindo et al., 2007).
The over-representation of localized tumors (85% vs. 70% expected) is a
result of our intended bias to acquire samples from patients treated on a single
clinical trial (AEWS0031), which was restricted to patients with localized disease
8
Table 1: Clinical and molecular features of ESFT samples evaluated for BMI-1
expression.

Characteristic N=130
Median age at diagnosis, years (range) 13 (0-47)

Gender
Male (%) 84 (64.6)
Female (%) 46 (35.4)

Stage
Localized (%) 99 (84.6)
Metastatic (%) 18 (15.4)
Unknown 13

Region of tumor
Pelvic (%) 20 (15.4)
Non-pelvic (%) 110 (84.6)

Translocation status
EWS-FLI-1 Type 1 (%) 42 (56)
EWS-FLI-1 Type 2 (%) 15 (20)
EWS-FLI-1 other (%) 9 (12)
EWS-ERG (%) 4 (5.3)
No fusion detected (%) 5 (6.7)
Unknown 55

Chemotherapy Status
Chemotherapy-naïve 99 (87)
Post-chemotherapy 15 (13)
Unknown 16

(Womer et al., 2008). In terms of primary site of disease, our tumor cohort is also
relatively enriched for non-pelvic primaries (85% vs. 75% expected). This may also
be a reflection of the increased number of localized tumors since pelvic tumors more
commonly present with metastatic disease (Hense et al., 1999). In addition, the
growing practice of closed (needle) rather than open diagnostic biopsy may yield
less tissue for banking studies when tumors are in the pelvis rather than other, more
surgically accessible sites.
9
Immunohistochemical staining of these tumors revealed that although BMI-1
expression in ESFT varies widely (Figure 1), it is diffusely and robustly expressed in
over 80% of tumors (Figures 1C and D). Of the remaining cases, 5% show weak
and sporadic expression of BMI-1 (Figure 1B) while 14% do not express the protein

Figure 1: Primary ESFT display differential expression of BMI-1. 130 archived
ESFT biopsies were stained with hematoxylin and eosin (H&E) to identify regions of
viable tumor cells and adjacent sections stained with a BMI-1-specific antibody and
scored as described in the text. Over 85% of tumors were robustly and diffusely
positive for BMI-1 (2+ or 3+; C & D) while a minority of tumors displayed absent or
minimal staining (Negative or 1+; A & B). A representative example and the
frequency of each category is shown. BM-1 negative cells in (C) are non-tumor
stromal cells (see corresponding H&E section). Sections were imaged using
AperioScope Image software at low (left and middle panels; 44X) and high (right
panel; 150X) power.

1+
6/130
(4.6%)
2+
33/130
(25.4%)
3+
73/130
(56.2%)
H&E BMI-1
Neg
18/130
(13.8%)
BMI-1
1+
6/130
(4.6%)
2+
33/130
(25.4%)
3+
73/130
(56.2%)
H&E BMI-1
Neg
18/130
(13.8%)
BMI-1
10
to detectable levels (Figure 1A). Comparison with normal non-malignant tissues
confirmed that both the intensity and the distribution of BMI-1+ cells were unique to
the ESFT samples (not shown). Nine of 16 control tissues (56%) received scores of
0 or 1+ and only hematopoietic samples (eg thymus, lymph node) showed discrete
regions of robustly positive cells. No control tissues received a score of 3+.
Tumor-initiating cancer stem cells have been identified in some human
cancers (Clarke et al., 2006). These cells are generally believed to be more
chemoresistant than their more differentiated progeny and BMI-1 has been
implicated as a marker of this cell population in some tumor types (Prince et al.,
2007; Hayry et al., 2008). We reasoned that if BMI-1 is a marker of chemoresistant
cancer stem cells in ESFT, then BMI-1 expression would be higher in tumor
specimens acquired post-chemotherapy. Comparison of BMI-1 expression in 99
diagnostic biopsies and 15 samples obtained at the time of surgical resection (post-
neoadjuvant chemotherapy) revealed no difference in BMI-1 expression between
these two groups (p=0.52). Moreover, in 2 cases where a post-treatment specimen
was also available, BMI-1 expression was unchanged following chemotherapy.
Finally, and most importantly, our data show that in most ESFT BMI-1 expression is
diffuse and not limited to a minority subpopulation (Figure 1). Thus, if
chemoresistant, tumor-initiating cancer stem cells exist as a minority subpopulation
in ESFT, BMI-1 is unlikely to be useful as a distinguishing marker.
BMI-1 protein expression does not predict a worse clinical outcome. Recent
studies have indicated that high expression of BMI-1 is a marker of more clinically
aggressive disease in some human cancers and is associated with a worse clinical
outcome (Chowdhury et al., 2007; Feng et al., 2007; Kim et al., 2004; Liu et al.,
11
2008; van Galen et al., 2007; Wang et al., 2008). To evaluate whether BMI-1
expression correlates with outcome in ESFT we first assessed whether the 80% of
ESFT that diffusely and robustly express BMI-1 (2+/3+) display clinical or pathologic
features that have been previously identified as markers of poor prognosis disease
(de Alava et al., 1998; Rodriguez-Galindo et al., 2007; Jenkin et al., 2002; Zoubek et
al., 1996). Chi-square analysis was conducted between BMI-1 score and disease
stage (localized vs. metastatic), age at presentation (less than or greater than 14
years), tumor location (pelvis vs. non-pelvis), and EWS-FLI1 transcript type (Type 1
vs. non-Type 1). Not all information was available for all tumors and due to limited
numbers BMI-1 negative and 1+ groups were combined into a single BMI-1
Low
group
for this analysis. No relationship was found to exist between any of these variables
and BMI-1 expression (Table 2). Insufficient data were available for analysis of other
putative prognostic factors such as tumor size, LDH at diagnosis, and coincident
cytogenetic abnormalities (Jenkin et al., 2002; Ahrens et al., 1999; Paulussen et al.,
2001; Bacci et al., 1999; Hattinger et al., 2002).
Next, we evaluated whether high levels of BMI-1 expression were associated
with a worse EFS or OS in ESFT patients. Reliable clinical correlates and outcomes
data were available for 79 of the130 patients whose tumors were scored for BMI-1
protein expression (Table 3). Consistent with the complete cohort of 130 tumors,
BMI-1 was robustly positive (2+ or 3+) in 71 tumors in this subset and undetectable
in eight (Table 3). As shown (Figure 2), survival analysis showed no significant
difference in either EFS or OS between BMI-1 negative (EFS 57 +/- 25%; OS 86 +/-
13%); 2+ (EFS 73 +/- 11%, OS 77 +/- 12%) or 3+ (EFS 61+/- 8%, OS 81 +/- 7%)
tumors (p-values= 0.87 and 0.84 for EFS and OS, respectively). Similarly, when
12
Table 2. Association of BMI-1 expression with molecular and clinical
Features associated with differential presentation and/or outcome in ESFT.

BMI-1 Expression Level
Low (Neg/1+) 2+ 3+ p-value
Age
<14 14 14 43 0.27
³14 10 19 30
Stage
Localized 15 22 62 0.37
Metastatic 5 4 9
Location
Non-Pelvic 19 30 61 0.45
Pelvic 5 3 12
Tissue of Origin
Bone 15 19 41 0.59
Soft tissue 7 10 32
EWS-FLI-1 Translocation
Type 1 3 11 28 0.12
Non-type 1 6 6 12

analysis was restricted to patients with localized disease (N=72) no association was
seen between BMI-1 expression level and outcome (p-values= 0.55 and 0.80 for
EFS and OS, respectively) (Figures 2C and D). However, we should note that, since
the numbers of patients in the BMI-1 negative and 2+ groups are relatively small,
there is a low power to detect other than large associations between BMI-1 and
outcome. To determine if BMI-1 may be associated with outcome among discrete
subsets of this clinically annotated cohort we performed a stratified analysis for age
(< or 14 years), pelvic vs. non-pelvic primary, and treatment era (1884-1995, 1996-
2000, and 2001-2005). BMI-1 expression level did not correlate with outcome in any
of these groups (data not shown).

13
Table 3: Clinical Outcomes Data. Clinical correlative and outcome data were
available for 71 BMI-1 positive and 8 BMI-1 negative tumors. No significant
differences in clinical covariates exist between the two groups.

BMI-1 Expression Level

Negative

Positive
All patients

2+

3+

Total
(%)
79
(100)
8
(10.1)
16
(20.3)
55
(69.6)

Characteristic

Median age at diagnosis, years
(Range)
12
(2-21)
11
(9-18)
14
(2-21)
12
(3-18)

54 (68.4) 6 14 34
Gender
Male (%)
Female (%)
25 (31.6) 2 2 21

Stage
Localized (%) 72 (92.3) 8 13 51
Metastatic (%) 6 (7.7) 0 2 4
Unknown 1 0 1 0

Region of tumor
Pelvic (%) 10 (12.7) 0 2 8
Non-pelvic (%) 69 (87.3) 8 14 47

Translocation status
EWS-FLI1 Type 1 (%) 23 (56.1) 0 2 21
EWS-FLI1 Type 2 (%) 9 (22) 1 3 5
EWS-FLI1 Other (%) 3 (7.3) 1 1 1
EWS-ERG (%) 3 (7.3) 0 1 2
No fusion detected (%) 3 (7.3) 0 0 3
Unknown 38 6 9 23

Systemic therapy
CCG-7942 (%) 15 (19) 2 6 7
CCG-7881 (%) 1 (1.3) 0 0 1
CCG-7951 (%) 3 (3.8) 0 1 2
AEWS0031 (%) 57 (72.2) 5 8 44
Alkylator-based, other (%) 1 (1.3) 0 0 1
Unknown 2 (2.5) 1 1 0

Treatment era
1984-1995 (%) 4 (5.1) 1 1 2
1996-2000 (%) 10 (12.7) 1 3 6
2001-2006 (%) 65 (82.3) 6 12 47

Median follow up (months) 40 21 46 41
(range) (3-210) (3-106) (10-129) (11-210)

14
Figure 2: Kaplan-Meier survival analysis of ESFT patients stratified according
to BMI-1 immunohistochemistry score. A. Event free survival (EFS) and B.
Overall survival (OS) for all patients with clinical outcome data available. C. EFS
and D. OS for patients with localized disease at presentation.

BMI-1 expression is not associated with p16 deletion or p53 mutation or with
CDKN2A or CDKN1A expression levels in ESFT. We have found that ESFT can be
divided into 2 distinct categories on the basis of BMI-1 protein expression: tumors
that diffusely and robustly express BMI-1 (80-90% of cases) and tumors that
express little to no BMI-1 (10-20%) (Figure 1). Given the critical role of BMI-1 in
normal stem cells to functionally repress the p16-RB and p53 tumor suppressor
pathways, we hypothesized that BMI-1

negative tumors may have sustained
EFS (all patients) OS (all patients)
OS (localized patients)
B
A
C D
EFS (all patients) OS (all patients)
OS (localized patients)
B
A
C D
EFS (all patients) OS (all patients)
OS (localized patients)
B
A
C D
15
secondary mutations in these pathways, thereby obviating a need for BMI-1
expression. Although secondary mutations in ESFT are relatively uncommon,
genomic deletion of p16 occurs in approximately 20% of cases and mutations in p53
in 10-15% (Wei et al., 2000; Huang et al., 2005; Abudu et al., 1999; de Alava et al.,
2000). To address whether absence of BMI-1 expression was associated with
either of these genetic events we first evaluated BMI-1 expression in 36 primary
tumors in which the p16 genomic status had been determined by fluorescence in
situ hybridization (N=24) (Huang et al., 2005) or quantitative genomic PCR (N=12;
see Materials and Methods). Homozygous deletion of p16 was identified in 3 of 36
cases and although one of these cases was found to be BMI-1-negative, two were
diffusely positive for BMI-1 (Table 4). Next, we similarly evaluated 24 ESFT with
documented p53 status (Huang et al., 2005). Again, although two p53-mutant
samples did not express BMI-1, the protein was robustly and diffusely expressed in
the remaining 6 mutant tumors (Table 4). Taken together, these data show that
absence of BMI-1 expression in ESFT is not significantly associated with either loss
of p16 or mutation in p53.

Table 4: Relationship between BMI-1 expression and p16/p53 status. BMI-1
expression was assessed by immunohistochemistry in ESFT with known genotypes.
No correlation exists between absence of BMI-1 expression and either loss of p16
or mutation of p53.

BMI-1 IHC score
Genotype Total no. neg/1+ 2+ 3+ p value
p16
Positive 33 5 9 19 0.73
Negative 3 1 1 1
p53
Wild type 16 4 5 7 0.99
Mutant 8 2 2 4
16
In both lung and breast tumors, high levels of BMI-1 are correlated with
reduced expression of the p16 and p14ARF tumor suppressors (Vonlanthen et al.,
2001; Pietersen et al., 2008; Chowdhury et al., 2007). To evaluate whether
expression of BMI-1 inversely correlates with expression of thep16/p14ARF-
encoding CDKN2A locus in ESFT we used quantitative RT-PCR to evaluate gene
expression levels in 36 localized, chemotherapy-naïve ESFT. As shown (Figure 3A),
no significant correlation exists between BMI-1 and CDKN2A expression in these
tumors. In addition, no significant correlation was identified between BMI-1 and
CDKN1A (p21), a p53-induced target gene that is repressed by BMI-1 in developing
neural stem cells (Fasano et al., 2007) (Figure 3B). Importantly, these in vivo data
corroborate our in vitro studies which showed that BMI-1 does not repress CDKN2A
or CDKN1A in ESFT cell lines (Douglas et al., 2008).

Figure 3: No correlation exists between BMI-1 and either CDKN2A or CDKN1A
in ESFT. Quantitative RT-PCR was used to measure gene expression in primary
tumor samples. As shown, although expression of BMI-1, CDKN2A and CDKN1A is
highly variable, no correlation exists between BMI-1 expression and either A.
CDKN2A or B. CDKN1A (r=Spearman correlation coefficient).

17
BMI-1 expression negatively correlates with neural crest differentiation in
ESFT. BMI-1 prevents differentiation of stem cells through epigenetic repression of
developmental pathways (Lee et al., 2006; Tolhuis et al., 2006) and recent studies
suggest that ESFT may arise from malignant transformation of mesenchymal and/or
neural crest stem or progenitor cells (Castillero-Trejo et al., 2005; Riggi et al., 2005;
Tirode et al., 2007; Melter, 2007; Coles, Lawlor & Bronner-Fraser, 2008). Thus, we
hypothesized that high levels of BMI-1 in ESFT may contribute to maintaining their
highly undifferentiated phenotype. To assess the relationship between BMI-1
expression level and global differentiation state we performed whole genome
expression profiling on RNA that had been isolated from 20 chemotherapy-naïve,
non-metastatic ESFT. In keeping with quantitative RT-PCR studies, no correlation
was found to exist between BMI-1 and either CDKN2A or CDKN1A expression
levels (not shown). In contrast, a striking inverse correlation was identified between
BMI-1 and expression of cellular differentiation markers (Table 5). Intriguingly,
however, with the exception of chondrocytic markers, low levels of BMI-1 were most
significantly and consistently associated with increased expression of markers of
neural crest (Jiang et al., 2008) rather than mesenchymal (Tirode et al., 2007)
differentiation (Table 5). These findings are consistent with our in vitro studies which
showed that knockdown of BMI-1 in ESFT cells leads to significant induction of
markers of neural differentiation (Douglas et al., 2008) Together these data suggest
that high BMI-1 levels mediate the undifferentiated neural phenotype that
characterizes ESFT.

18
Table 5. Analysis of correlation between BMI-1 expression and markers of
neural crest and mesenchymal differentiation. Twenty chemotherapy-naïve
primary ESFT with >70% tumor cell content were analyzed by whole genome
expression profiling as described in the text. A significant inverse correlation exists
between BMI-1 expression and genes associated with neural crest differentiation.

Lineage Differentiation
Gene Symbol Spearman r p- value FDR
Significant?
Neural ACHE -0.7263 0.0004 0.0040 *
ADRA2B -0.8511 <0.0001 <0.0001 *
CHRNA2 -0.6586 0.0021 0.0098 *
CHRNA4 -0.8045 <0.0001 0.0009 *
CHRND -0.7323 0.0004 0.0037 *
CHRNG -0.6571 0.0021 0.0100 *
DRD4 -0.6180 0.0045 0.0159 *
POU4F1 -0.7925 <0.0001 0.0013 *
PRPH -0.5549 0.0124 0.0325 *
TH -0.5173 0.0210 0.0482 *
Glial GFAP -0.5624 0.0111 0.0301 *
OLIG1 -0.6256 0.0039 0.0145 *
OLIG2 -0.5865 0.0076 0.0230 *
Neural Crest
SOX10 -0.6226 0.0041 0.0151 *
Adipocytic FABP4 0.1083 0.6465 0.7184 Ns
LPL -0.2842 0.2238 0.3099 Ns
PPARG -0.5744 0.0092 0.0264 *
Osteocytic ALPL -0.1218 0.6079 0.6852 Ns
RUNX2 -0.0767 0.7479 0.8044 Ns
SPP1 -0.0015 0.9975 0.9980 Ns
Chondrocytic COL10A1 -0.7489 0.0002 0.0030 *
Mesenchymal
SOX9 -0.6346 0.0033 0.0131 *

*FDR 0.05; ns=not significant; shaded genes are highly significant with a correlation
coefficient > -0.6 and FDR of 0.05

DISCUSSION
In this clinico-pathologic correlates study of primary tumors we have confirmed
that high-level expression of the polycomb group protein BMI-1 is a nearly universal
feature of ESFT. However, in contrast to other tumor types, BMI-1 expression is not
confined to discrete cell populations, and higher levels of expression do not
19
correlate with clinical outcome or with reduced expression of the CDKN2A gene.
Intriguingly, however, BMI-1 transcript expression is inversely correlated with
markers of neural crest differentiation. As discussed below, these findings have
important implications for the cellular origin and molecular pathogenesis of ESFT.
In recent years the hypothesis that tumor-initiating cancer stem cells exist as
minority populations in established tumors has been rigorously tested and
compelling evidence now supports their existence in myeloid leukemia as well as
brain tumors and several epithelial cancers (reviewed in Clarke et al., 2006). In
some tumors these cells have been found to display increased expression of BMI-1
(Lessard & Sauvageau, 2003; Liu et al., 2006; Prince et al., 2007), and since normal
somatic stem cells also express high levels of BMI-1 it has been proposed that BMI-
1 may be useful as a marker of self-renewing cancer stem cells. To our knowledge,
tumor-initiating cancer stem cells have yet to be definitively isolated from bone and
soft tissue sarcomas. In this study we have shown that, in the 80% of ESFT that
robustly express BMI-1, expression of the protein is uniform among all tumor cells
and is not limited to discrete subpopulations. Thus, if a more tumorigenic cancer
stem cell population exists within established ESFT, BMI-1 will not be useful as
marker of these cells. Alternately, if BMI-1 truly is a definitive marker of tumor-
initiating cancer stem cells, this would suggest that all cells within primary ESFT
possess tumor-initiating potential. Limiting dilution studies and in vivo tumorigenicity
assays of freshly isolated primary tumor cells are needed to address this intriguing
possibility.
Stratification of patients into prognostic groups at the time of diagnosis allows
for the development and prescription of therapeutic regimens that are best tailored
20
to patients’ needs. Unfortunately, with the exception of metastatic disease, no
reliable clinical features exist that allow such stratification in patients who are newly
diagnosed with ESFT. High expression of BMI-1 has been reported to be associated
with an unfavorable prognosis in various malignancies including diffuse large B cell
lymphoma (van Galen et al., 2007), acute myeloid leukemia (Chowdhury et al.,
2007), hepatocellular carcinoma (Wang et al., 2008), gastric carcinoma (Liu et al.,
2008), and oligodendroglial tumors (Hayry et al., 2008). In contrast, in malignant
melanoma loss of BMI-1 expression is associated with disease progression
(Bachmann et al., 2008), and in breast cancer conflicting reports have demonstrated
associations of BMI-1 with either aggressive disease (Feng, 2007; Kim et al., 2004),
or more favorable outcome (Pietersen et al., 2008; Choi et al., 2008), identified no
association between BMI-1 expression and either EFS or OS in ESFT. As noted
above, however, given the small sample size of BMI-1 negative tumors in our
clinical cohort, there was low power to detect other than large associations between
BMI1 and outcome. Importantly, over 80% of the cases evaluated in this study were
treated using very recent clinical protocols (between 2001 and 2006) and the vast
majority of these were treated on a single COG trial, AEWS0031. Thus, although it
is conceivable that differential expression of BMI-1 may be associated with more
clinically aggressive disease in ESFT, our data provide no indication that it will be
useful as a predictor of outcome in patients with localized disease who will be
treated using current clinical protocols.
In contrast to other tumor types we have found that BMI-1 does not repress
expression of the CDKN2A locus or its protein product p16 in ESFT cells in vitro
(van Galen et al., 2007). In corroboration with these findings we have now
21
demonstrated that there is no correlation between BMI-1 and CDKN2A expression
levels in primary tumors. Thus, for as yet unexplained reasons, BMI-1 appears to be
dissociated from p16 regulation in ESFT. One possible explanation for this is that
BMI-1-mediated repression of p16 has been shown to depend on the presence of a
functional retinoblastoma protein (RB) (Kotake et al., 2007). Although RB is only
rarely mutated in ESFT (Kovar et al., 1997; Maitra et al., 2001), recent work
suggests that it may be functionally inactivated by the EWS-FLI1 fusion (Hu et al.,
2008). Further studies are now required to determine if RB inactivation is
responsible for dissociation of BMI-1 from CDKN2A repression in ESFT.
The cellular origin of ESFT remains an enigma. Histologically and genetically
they are largely undifferentiated tumors that demonstrate variable degrees of
parasympathetic neural differentiation. For this reason a neural crest progenitor was
long favored as a putative cell of origin (Kovar, 2005). Over the past few years,
however, evidence for a mesenchymal stem cell origin has accumulated (Tirode et
al., 2007; Riggi et al., 2008). We have shown that knockdown of BMI-1 in ESFT
cells in vitro results in induction of neural differentiation (Douglas, 2008). Similarly,
studies of mouse models of gliomagenesis have demonstrated that bmi-1
expression is a major determinant of the differentiation-state and differentiation-
capacity of these brain tumors (Bruggerman et al., 2007). In corroboration with
these studies, we now demonstrate that BMI-1 expression inversely correlates with
differentiation in primary ESFT. In particular, tumors with lower levels of BMI-1
display significantly higher levels of expression of genes associated with neural
crest differentiation, specifically neural and glial markers (Jiang et al., 2008). These
data lead us to hypothesize that BMI-1 while EWS-FLI1 may repress mesenchymal
22
differentiation of ESFT cells (Tirode et al., 2008), high levels of BMI-1 may prevent
neuroglial differentiation.
Whether the ESFT cell of origin expresses high levels of BMI-1, or BMI-1 is
induced during the process of malignant transformation, remains an unanswered
question. Interestingly, BMI-1 is highly expressed by neural crest stem cells
(Molofsky et al., 2003) but not by mesenchymal stem cells (Douglas et al., 2008).
We have found no evidence for genomic amplification of the BMI-1 locus in primary
tumors (unpublished data) and although the related polycomb gene EZH2 was
found to be induced in human mesenchymal stem cells following EWS-FLI1
transduction (Riggi et al., 2008), a similar upregulation of BMI-1 was not observed in
these cells. In contrast, BMI-1 was upregulated by EWS-FLI1 in immortalized
murine NIH-3T3 cells (Zwerner et al., 2008). Thus, in some cell types and contexts
BMI-1 may be induced downstream of EWS-FLI1; however, the nature of this
situational specificity remains to be elucidated.
Finally, although we have found that BMI-1 is likely to be integral to the
pathogenesis of most ESFT, approximately 15% of tumors do not express the
protein. This raises the intriguing possibility that alternate genetic events
compensate for the absence of BMI-1 expression in these tumors. Given the role of
BMI-1 in p16-RB and p53 pathway suppression in stem cells we speculated that
mutations in these pathways may be present in BMI-1-negative tumors. However,
our analysis did not reveal any significant correlation between absence of BMI-1 and
either loss of p16 or mutation of p53, the two most common secondary mutations in
ESFT (Burchill, 2003). Although alternate mechanisms for tumor suppressor
pathway inactivation, including p16 promoter hypermethylation and MDM2
23
amplification, have been reported in ESFT, these are rare events (Burchill, 2003).
Furthermore, we have identified no correlation between BMI-1 and CDKN2A
expression levels in ESFT suggesting that other mechanisms of BMI-1 action drive
its function as a cellular oncogene in this tumor family.
In summary, we have shown that BMI-1 is diffusely expressed by most ESFT
but that presence or absence of BMI-1 does not correlate with outcome. In addition,
we find no association between BMI-1 expression and either loss of p16 or mutation
of p53. Moreover, BMI-1 transcript expression does not correlate with either
CDKN2A or CDKN1A levels. Importantly, however, BMI-1 expression is highly
inversely correlated with genes associated with cellular differentiation, in particular
markers of neural crest differentiation. Together these findings demonstrate that
BMI-1 is a nearly universal characteristic of ESFT cells and suggest that high-level
expression of this polycomb protein contributes to ESFT pathogenesis in ways that
are independent of p16/ARF-repression.

REFERENCES
Abudu A, Mangham DC, Reynolds GM, Pynsent PB, Tillman RM, Carter SR, Grimer
RJ: Overexpression of p53 protein in primary Ewing' s sarcoma of bone:
relationship to tumour stage, response and prognosis, Br J Cancer 1999,
79:1185-1189.

Ahrens S, Hoffmann C, Jabar S, Braun-Munzinger G, Paulussen M, Dunst J, Rube
C, Winkelmann W, Heinecke A, Gobel U, Winkler K, Harms D, Treuner J,
Jurgens H: Evaluation of prognostic factors in a tumor volume-adapted
treatment strategy for localized Ewing sarcoma of bone: the CESS 86
experience. Cooperative Ewing Sarcoma Study, Med Pediatr Oncol 1999,
32:186-195.

Bacci G, Ferrari S, Longhi A, Rimondini S, Versari M, Zanone A, Forni C: Prognostic
significance of serum LDH in Ewing' s sarcoma of bone, Oncol Rep 1999,
6:807-811.

24
Bachmann IM, Puntervoll HE, Otte AP, Akslen LA: Loss of BMI-1 expression is
associated with clinical progress of malignant melanoma, Mod Pathol 2008,
21:583-590.

Benjamini Y, Hochberg Y: Controlling the false discovery rate - a practical and
powerful approach to multiple testing, journal of the royal statistical society
series B-methodological 1995, 57:289-300.

Berggren P, Kumar R, Sakano S, Hemminki L, Wada T, Steineck G, Adolfsson J,
Larsson P, Norming U, Wijkstrom H, Hemminki K: Detecting homozygous
deletions in the CDKN2A(p16(INK4a))/ARF(p14(ARF)) gene in urinary bladder
cancer using real-time quantitative PCR, Clin Cancer Res 2003, 9:235-242.

Bernstein M KH, and Paulussen M et al.: Ewing sarcoma family of tumors: ewing
sarcoma of bone and soft tissue and the peripheral primitive neuroectodermal
tumors. Edited by Pizzo PA PD, eds. Philadelphia, Lippincott Williams &
Wilkins, 2006, pp. 1002-1032.

Bruggeman SW, Hulsman D, Tanger E, Buckle T, Blom M, Zevenhoven J, van
Tellingen O, van Lohuizen M: Bmi1 controls tumor development in an
ink4a/arf-independent manner in a mouse model for glioma, Cancer Cell 2007,
12:328-341.

Burchill SA: Ewing' s sarcoma: diagnostic, prognostic, and therapeutic implications of
molecular abnormalities, J Clin Pathol 2003, 56:96-102.

Castillero-Trejo Y, Eliazer S, Xiang L, Richardson JA, Ilaria RL, Jr.: Expression of
the EWS/FLI-1 oncogene in murine primary bone-derived cells Results in
EWS/FLI-1-dependent, ewing sarcoma-like tumors, Cancer Res 2005,
65:8698-8705.

Choi YJ, Choi YL, Cho EY, Shin YK, Sung KW, Hwang YK, Lee SJ, Kong G, Lee
JE, Kim JS, Kim JH, Yang JH, Nam SJ: Expression of Bmi-1 protein in tumor
tissues is associated with favorable prognosis in breast cancer patients, Breast
Cancer Res Treat 2008.

Chowdhury M, Mihara K, Yasunaga S, Ohtaki M, Takihara Y, Kimura A: Expression
of Polycomb-group (PcG) protein BMI-1 predicts prognosis in patients with
acute myeloid leukemia, Leukemia 2007, 21:1116-1122.

Clarke MF, Dick JE, Dirks PB, Eaves CJ, Jamieson CH, Jones DL, Visvader J,
Weissman IL, Wahl GM: Cancer Stem Cells--Perspectives on Current Status
and Future Directions: AACR Workshop on Cancer Stem Cells, Cancer Res
2006, 66:9339-9344.

Coles EG, Lawlor ER, Bronner-Fraser M: EWS-FLI1 Causes Neuroepithelial
Defects and Abrogates Emigration of Neural Crest Stem Cells, Stem Cells
2008.
25

Cox DR, Oakes D: Analysis of survival data. Edited by CRC Press, 1984, p. 201.

de Alava E, Antonescu CR, Panizo A, Leung D, Meyers PA, Huvos AG, Pardo-
Mindan FJ, Healey JH, Ladanyi M: Prognostic impact of P53 status in Ewing
sarcoma, Cancer 2000, 89:783-792.

de Alava E, Kawai A, Healey JH, Fligman I, Meyers PA, Huvos AG, Gerald WL,
Jhanwar SC, Argani P, Antonescu CR, Pardo-Mindan FJ, Ginsberg J, Womer
R, Lawlor ER, Wunder J, Andrulis I, Sorensen PH, Barr FG, Ladanyi M: EWS-
FLI1 fusion transcript structure is an independent determinant of prognosis in
Ewing' s sarcoma, J Clin Oncol 1998, 16:1248-1255.

Douglas D, Hsu JH, Hung L, Cooper A, Abdueva D, van Doorninck J, Peng G,
Shimada H, Triche TJ, Lawlor ER: BMI-1 promotes ewing sarcoma
tumorigenicity independent of CDKN2A repression, Cancer Res 2008,
68:6507-6515.

Fasano CA, Dimos JT, Ivanova NB, Lowry N, Lemischka IR, Temple S: shRNA
knockdown of Bmi-1 reveals a critical role for p21-Rb pathway in NSC self-
renewal during development, Cell Stem Cell 2007, 1:87-99.

Feng Y, Song L, Guo B, Liao W, Li M, Liu W, Zeng M, Zhang L: Expression and
significance of Bmi-1 in breast cancer, Chinese journal of cancer 2007,
26:154-157.

Gil J, Bernard D, Peters G: Role of polycomb group proteins in stem cell self-
renewal and cancer, DNA Cell Biol 2005, 24:117-125.

Hattinger CM, Potschger U, Tarkkanen M, Squire J, Zielenska M, Kiuru-Kuhlefelt S,
Kager L, Thorner P, Knuutila S, Niggli FK, Ambros PF, Gadner H, Betts DR:
Prognostic impact of chromosomal aberrations in Ewing tumours, Br J Cancer
2002, 86:1763-1769.

Hayry V, Tynninen O, Haapasalo HK, Wolfer J, Paulus W, Hasselblatt M, Sariola H,
Paetau A, Sarna S, Niemela M, Wartiovaara K, Nupponen NN: Stem cell
protein BMI-1 is an independent marker for poor prognosis in oligodendroglial
tumours, Neuropathol Appl Neurobiol 2008.

Hense HW, Ahrens S, Paulussen M, Lehnert M, Jurgens H: Factors associated with
tumor volume and primary metastases in Ewing tumors: results from the
(EI)CESS studies, Ann Oncol 1999, 10:1073-1077.

Hollander M, Wolfe DA: Nonparametric statistical methods. Edited by New York,
John Wiley, 1973, p. 503.

26
Hu HM, Zielinska-Kwiatkowska A, Munro K, Wilcox J, Wu DY, Yang L, Chansky HA:
EWS/FLI1 suppresses retinoblastoma protein function and senescence in
Ewing' s sarcoma cells, J Orthop Res 2008, 26:886-893.

Huang HY, Illei PB, Zhao Z, Mazumdar M, Huvos AG, Healey JH, Wexler LH,
Gorlick R, Meyers P, Ladanyi M: Ewing sarcomas with p53 mutation or
p16/p14ARF homozygous deletion: a highly lethal subset associated with poor
chemoresponse, J Clin Oncol 2005, 23:548-558.

Jacobs JJ, Kieboom K, Marino S, DePinho RA, van Lohuizen M: The oncogene and
Polycomb-group gene bmi-1 regulates cell proliferation and senescence
through the ink4a locus, Nature 1999, 397:164-168.

Jacobs JJ, Scheijen B, Voncken JW, Kieboom K, Berns A, van Lohuizen M: Bmi-1
collaborates with c-Myc in tumorigenesis by inhibiting c-Myc-induced apoptosis
via INK4a/ARF, Genes Dev 1999, 13:2678-2690.

Jenkin RD, Al-Fawaz I, Al-Shabanah M, Allam A, Ayas M, Khafaga Y, Memon M,
Rifai S, Schultz H, Younge D: Localised Ewing sarcoma/PNET of bone--
prognostic factors and international data comparison, Med Pediatr Oncol 2002,
39:586-593.

Jiang X, Gwye Y, McKeown SJ, Bronner-Fraser M, Lutzko C, Lawlor ER: Isolation
and characterization of neural crest stem cells derived from in vitro
differentiated human embryonic stem cells, Stem Cells Dev 2008.

Kim JH, Yoon SY, Jeong SH, Kim SY, Moon SK, Joo JH, Lee Y, Choe IS, Kim JW:
Overexpression of Bmi-1 oncoprotein correlates with axillary lymph node
metastases in invasive ductal breast cancer, Breast 2004, 13:383-388.

Kotake Y, Cao R, Viatour P, Sage J, Zhang Y, Xiong Y: pRB family proteins are
required for H3K27 trimethylation and Polycomb repression complexes binding
to and silencing p16INK4alpha tumor suppressor gene, Genes Dev 2007,
21:49-54.

Kovar H: Context matters: the hen or egg problem in Ewing' s sarcoma, Semin
Cancer Biol 2005, 15:189-196.

Kovar H, Jug G, Aryee DN, Zoubek A, Ambros P, Gruber B, Windhager R, Gadner
H: Among genes involved in the RB dependent cell cycle regulatory cascade,
the p16 tumor suppressor gene is frequently lost in the Ewing family of tumors,
Oncogene 1997, 15:2225-2232.

Krivtsov AV, Twomey D, Feng Z, Stubbs MC, Wang Y, Faber J, Levine JE, Wang J,
Hahn WC, Gilliland DG, Golub TR, Armstrong SA: Transformation from
committed progenitor to leukaemia stem cell initiated by MLL-AF9, Nature
2006, 442:818-822.

27
Labuhn M, Jones G, Speel EJ, Maier D, Zweifel C, Gratzl O, Van Meir EG, Hegi ME,
Merlo A: Quantitative real-time PCR does not show selective targeting of
p14(ARF) but concomitant inactivation of both p16(INK4A) and p14(ARF) in
105 human primary gliomas, Oncogene 2001, 20:1103-1109.

Lee TI, Jenner RG, Boyer LA, Guenther MG, Levine SS, Kumar RM, Chevalier B,
Johnstone SE, Cole MF, Isono K, Koseki H, Fuchikami T, Abe K, Murray HL,
Zucker JP, Yuan B, Bell GW, Herbolsheimer E, Hannett NM, Sun K, Odom
DT, Otte AP, Volkert TL, Bartel DP, Melton DA, Gifford DK, Jaenisch R, Young
RA: Control of developmental regulators by Polycomb in human embryonic
stem cells, Cell 2006, 125:301-313.

Lessard J, Sauvageau G: Bmi-1 determines the proliferative capacity of normal and
leukaemic stem cells, Nature 2003, 423:255-260.

Liu JH, Song LB, Zhang X, Guo BH, Feng Y, Li XX, Liao WT, Zeng MS, Huang KH:
Bmi-1 expression predicts prognosis for patients with gastric carcinoma, J
Surg Oncol 2008, 97:267-272.

Liu S, Dontu G, Mantle ID, Patel S, Ahn NS, Jackson KW, Suri P, Wicha MS:
Hedgehog signaling and Bmi-1 regulate self-renewal of normal and malignant
human mammary stem cells, Cancer Res 2006, 66:6063-6071.

Maitra A, Roberts H, Weinberg AG, Geradts J: Aberrant expression of tumor
suppressor proteins in the Ewing family of tumors, Arch Pathol Lab Med 2001,
125:1207-1212.

Marshak DR, Gardner RL, Gottlieb D: Stem Cell Biology. Edited by Cold Spring
Harbor, NY, Cold Spring Harbor Laboratory Press, 2001, p. 550.

Meltzer PS: Is Ewing' s sarcoma a stem cell tumor? Cell Stem Cell 2007, 1:13-15.

Molofsky AV, Pardal R, Iwashita T, Park IK, Clarke MF, Morrison SJ: Bmi-1
dependence distinguishes neural stem cell self-renewal from progenitor
proliferation, Nature 2003, 425:962-967.

Pardal R, Clarke MF, Morrison SJ: Applying the principles of stem-cell biology to
cancer, Nat Rev Cancer 2003, 3:895-902.

Paulussen M, Ahrens S, Dunst J, Winkelmann W, Exner GU, Kotz R, Amann G,
Dockhorn-Dworniczak B, Harms D, Muller-Weihrich S, Welte K, Kornhuber B,
Janka-Schaub G, Gobel U, Treuner J, Voute PA, Zoubek A, Gadner H,
Jurgens H: Localized Ewing tumor of bone: final results of the cooperative
Ewing' s Sarcoma Study CESS 86, J Clin Oncol 2001, 19:1818-1829.

28
Pietersen AM, Horlings HM, Hauptmann M, Langerod A, Ajouaou A, Cornelissen-
Steijger P, Wessels LF, Jonkers J, van de Vijver MJ, van Lohuizen M: EZH2
and BMI1 inversely correlate with prognosis and TP53 mutation in breast
cancer, Breast Cancer Res 2008, 10:R109.

Prince ME, Sivanandan R, Kaczorowski A, Wolf GT, Kaplan MJ, Dalerba P,
Weissman IL, Clarke MF, Ailles LE: Identification of a subpopulation of cells
with cancer stem cell properties in head and neck squamous cell carcinoma,
Proc Natl Acad Sci U S A 2007, 104:973-978.

Riggi N, Cironi L, Provero P, Suva ML, Kaloulis K, Garcia-Echeverria C, Hoffmann
F, Trumpp A, Stamenkovic I: Development of Ewing' s sarcoma from primary
bone marrow-derived mesenchymal progenitor cells, Cancer Res 2005,
65:11459-11468.

Riggi N, Suva ML, Suva D, Cironi L, Provero P, Tercier S, Joseph JM, Stehle JC,
Baumer K, Kindler V, Stamenkovic I: EWS-FLI-1 expression triggers a Ewing' s
sarcoma initiation program in primary human mesenchymal stem cells, Cancer
Res 2008, 68:2176-2185.

Rodriguez-Galindo C, Liu T, Krasin MJ, Wu J, Billups CA, Daw NC, Spunt SL, Rao
BN, Santana VM, Navid F: Analysis of prognostic factors in ewing sarcoma
family of tumors: review of St. Jude Children' s Research Hospital studies,
Cancer 2007, 110:375-384.

Smith KS, Chanda SK, Lingbeek M, Ross DT, Botstein D, van Lohuizen M, Cleary
ML: Bmi-1 regulation of INK4A-ARF is a downstream requirement for
transformation of hematopoietic progenitors by E2a-Pbx1, Mol Cell 2003,
12:393-400.

Team RDC: R: A language and environment for statistical computing. Edited by
Computing RFfS. Vienna, 2008.

Tirode F, Laud-Duval K, Prieur A, Delorme B, Charbord P, Delattre O: Mesenchymal
stem cell features of Ewing tumors, Cancer Cell 2007, 11:421-429.

Tolhuis B, Muijrers I, de Wit E, Teunissen H, Talhout W, van Steensel B, van
Lohuizen M: Genome-wide profiling of PRC1 and PRC2 Polycomb chromatin
binding in Drosophila melanogaster, Nat Genet 2006, 38:694-699.

van Galen JC, Muris JJ, Oudejans JJ, Vos W, Giroth CP, Ossenkoppele GJ, Otte
AP, Raaphorst FM, Meijer CJ: Expression of the polycomb-group gene BMI1 is
related to an unfavourable prognosis in primary nodal DLBCL, Journal of
Clinical Pathology 2007, 60:167-172.

29
Vonlanthen S, Heighway J, Altermatt HJ, Gugger M, Kappeler A, Borner MM, van
Lohuizen M, Betticher DC: The bmi-1 oncoprotein is differentially expressed in
non-small cell lung cancer and correlates with INK4A-ARF locus expression,
Br J Cancer 2001, 84:1372-1376.

Wang H, Pan K, Zhang HK, Weng DS, Zhou J, Li JJ, Huang W, Song HF, Chen MS,
Xia JC: Increased polycomb-group oncogene Bmi-1 expression correlates with
poor prognosis in hepatocellular carcinoma, J Cancer Res Clin Oncol 2008,
134:535-541.

Wei G, Antonescu CR, de Alava E, Leung D, Huvos AG, Meyers PA, Healey JH,
Ladanyi M: Prognostic impact of INK4A deletion in Ewing sarcoma, Cancer
2000, 89:793-799.

Womer R, Krailo M, Dickman P, Pawel B: Randomized comparison of every-two-
week vs. every-three-week chemotherapy in Ewing family tumors (ESFT), J
Clin Oncol 2008, 26(15S):abs:10504.

Zoubek A, Dockhorn-Dworniczak B, Delattre O, Christiansen H, Niggli F, Gatterer-
Menz I, Smith TL, Jurgens H, Gadner H, Kovar H: Does expression of different
EWS chimeric transcripts define clinically distinct risk groups of Ewing tumor
patients? J Clin Oncol 1996, 14:1245-1251.

Zwerner JP, Joo J, Warner KL, Christensen L, Hu-Lieskovan S, Triche TJ, May WA:
The EWS/FLI1 oncogenic transcription factor deregulates GLI1, Oncogene
2008, 27:3282-3291.

Abstract (if available)

Abstract Ewing sarcoma family tumors (ESFT) are highly undifferentiated bone and soft tissue tumors that primarily affect children. Metastatic disease at diagnosis portends a worse prognosis and remains the only reliable predictor of outcome. The polycomb protein BMI-1 is associated with worse clinical outcome in some human cancers. This study evaluates the clinical and biologic significance of BMI-1 expression in ESFT.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Harnessing the power of stem cell self-renewal pathways in cancer: dissecting the role of BMI-1 in Ewing’s sarcoma initiation and maintenance

PDF

A pilot study of chemotherapy intensification by adding vincristine, topotecan and cyclophosphamide to standard chemotherapy agents with an interval compression schedule in newly diagnosed patien...

PDF

DNA methylation groups determined by GATA5 gene methylation level are correlated with tumor subtype, sex, smoking status, and body mass index in esophageal and gastric adenocarcinoma

PDF

Effects of demographic and tumor characteristics on outcomes in children with malignant peripheral nerve sheath tumors

$A phase I study of vincristine, escalating doses of irinotecan, temozolomide and bevacizumab (VIT-B) in pediatric and adolescent patients with recurrent or refractory solid tumors of non-hematopo...$

PDF

A phase I study of vincristine, escalating doses of irinotecan, temozolomide and bevacizumab (VIT-B) in pediatric and adolescent patients with recurrent or refractory solid tumors of non-hematopo...

Asset Metadata

Creator van Doorninck, John Anthony (author)

Core Title BMI-1 expression inversely correlates with differentiation but not clinical outcome or CDKN2A expression in Ewing sarcoma family of tumors

School Keck School of Medicine

Degree Master of Science

Degree Program Preventive Medicine (Health Behavior)

Publication Date 07/11/2009

Defense Date 03/18/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag BMI-1,CDKN2A,Ewing sarcoma,OAI-PMH Harvest,prognosis,stem cell

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Seeger, Robert (committee chair), Aldrovandi, Grace (committee member), Sposto, Richard (committee member)

Creator Email jvandoorninck@chla.usc.edu,jvandoorninck@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2350

Unique identifier UC190187

Identifier etd-Doorninck-2672 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-251016 (legacy record id),usctheses-m2350 (legacy record id)

Legacy Identifier etd-Doorninck-2672.pdf

Dmrecord 251016

Document Type Thesis

Rights van Doorninck, John Anthony

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu