Intensive expression of Bmi-1 is a new independent predictor of poor outcome in patients with ovarian carcinoma.

Abstract

It has been suggested that the B-cell specific moloney leukemia virus insertion site 1 (Bmi-1) gene plays an oncogenic role in several types of human cancer, but the status of Bmi-1 amplification and expression in ovarian cancer and its clinical/prognostic significance are unclear.
The methods of immunohistochemistry and fluorescence in situ hybridization were utilized to examine protein expression and amplification of Bmi-1 in 30 normal ovaries, 30 ovarian cystadenomas, 40 borderline ovarian tumors and 179 ovarian carcinomas.
Intensive expression of Bmi-1 was detected in none of the normal ovaries, 3% cystadenomas, 10% borderline tumors, and 37% ovarian carcinomas, respectively. Amplification of Bmi-1 was detected in 8% of ovarian carcinomas. In ovarian carcinomas, significant positive associations were found between intensive expression of Bmi-1 and the tumors ascending histological grade, later pT/pN/pM and FIGO stages (P < 0.05). In univariate survival analysis of the ovarian carcinoma cohorts, a significant association of intensive expression of Bmi-1 with shortened patient survival (mean 49.3 months versus 100.3 months, p < 0.001) was demonstrated. Importantly, Bmi-1 expression provided significant independent prognostic parameters in multivariate analysis (p = 0.005).
These findings provide evidence that intensive expression of Bmi-1 might be important in the acquisition of an invasive and/or aggressive phenotype of ovarian carcinoma, and serve as a independent biomarker for shortened survival time of patients.

Full text

Page 1

RESEARCH ARTICLEOpen Access
Intensive expression of Bmi-1 is a new
independent predictor of poor outcome
in patients with ovarian carcinoma
Guo-Fen Yang1,2, Wei-Peng He2, Mu-Yan Cai1, Li-Ru He1, Jun-Hang Luo2, Hai-Xia Deng1, Xin-Yuan Guan1,
Mu-Sheng Zeng1, Yi-Xin Zeng1, Dan Xie1*
Abstract
Background: It has been suggested that the B-cell specific moloney leukemia virus insertion site 1 (Bmi-1) gene
plays an oncogenic role in several types of human cancer, but the status of Bmi-1 amplification and expression in
ovarian cancer and its clinical/prognostic significance are unclear.
Methods: The methods of immunohistochemistry and fluorescence in situ hybridization were utilized to examine
protein expression and amplification of Bmi-1 in 30 normal ovaries, 30 ovarian cystadenomas, 40 borderline ovarian
tumors and 179 ovarian carcinomas.
Results: Intensive expression of Bmi-1 was detected in none of the normal ovaries, 3% cystadenomas, 10%
borderline tumors, and 37% ovarian carcinomas, respectively. Amplification of Bmi-1 was detected in 8% of ovarian
carcinomas. In ovarian carcinomas, significant positive associations were found between intensive expression of
Bmi-1 and the tumors ascending histological grade, later pT/pN/pM and FIGO stages (P < 0.05). In univariate
survival analysis of the ovarian carcinoma cohorts, a significant association of intensive expression of Bmi-1 with
shortened patient survival (mean 49.3 months versus 100.3 months, p < 0.001) was demonstrated. Importantly, Bmi-
1 expression provided significant independent prognostic parameters in multivariate analysis (p = 0.005).
Conclusions: These findings provide evidence that intensive expression of Bmi-1 might be important in the
acquisition of an invasive and/or aggressive phenotype of ovarian carcinoma, and serve as a independent
biomarker for shortened survival time of patients.
Background
Ovarian cancer is a major lethal gynecological malig-
nancy worldwide [1]. Its peak incidence is at the age 45
or above. Because of its insidious onset, approximately
70% of ovarian cancer patients were diagnosed at
advanced stage(FIGO III/IV stage) with a very poor
prognosis, whose 5-year survival rate is of <30% [2].
Ovarian carcinoma is the most common histopathologi-
cal type of ovarian cancer. The development and pro-
gression of ovarian carcinoma are presumed to be a
multi-step process involving multiple genetic changes
[3]. Thus, a substantial amount of research on ovarian
carcinoma has focused on the discovery of specific
molecular markers that are present in ovarian carcinoma
cells which could serve as reliable prognostic factors.
The B-cell specific moloney leukemia virus insertion
site 1 (Bmi-1) gene belongs to mammalian Polycomb-
group (PcG) family forming multimeric gene-repressing
complexes involved in axial patterning, hematopoiesis,
regulation of proliferation, and senescence. Bmi-1 was
first identified as a proto-oncogene that cooperated with
c-Myc in generating pre-B-cell lymphomas in a murine
model [4-8]. It has been discovered that Bmi-1 partici-
pates in cell cycle regulation by acting as a stable tran-
scriptional repressor of the Ink4a locus, which encodes
the tumor suppressor proteins p16Ink4a and p19Arf
(mouse homologue of human p14ARF). Inactivation of
the p16Ink4a-pRb pathway and p14ARF-MDM2-p53
pathway by Bmi-1 deregulation has been clearly
* Correspondence: xied@mail.sysu.edu.cn
1State Key Laboratory of Oncology in South China, Cancer Center, Sun Yat-
Sen University, No. 651, Dongfeng Road East, 510060 Guangzhou, China
Yang et al. BMC Cancer 2010, 10:133
http://www.biomedcentral.com/1471-2407/10/133
© 2010 Yang et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.

Page 2

implicated in lymphomagenesis [9,10] and oncogenesis
in nonsmall-cell lung cancer of human [11]. This sug-
gested that the Bmi-1 gene plays an important role in
cell proliferation and tumor progression. It has been
confirmed that Bmi-1 gene is widely expressed in
diverse human tumors, including non-small cell lung
cancer, hepatocellular carcinoma, B-cell non-Hodgkin’s
lymphoma, breast cancer, ovarian cancer, colorectal can-
cer, skin cancer and neuroblastoma [10-20], and has
been shown to be a useful prognostic marker in myelo-
dysplastic syndrome and many cancers, including naso-
pharyngeal carcinoma, bladder cancer and gastric cancer
[17-20].
To date, however, the status of Bmi-1 expression and
its clinical/prognostic relevance in ovarian cancer have
not been fully elucidated. In this study, the protein
expression and amplification status of Bmi-1 in a series
of human epithelial ovarian tissue, normal and patholo-
gical, non-neoplastic and neoplastic, were examined.
The clinico-pathological and prognostic significance of
expression of Bmi-1 in our ovarian carcinoma cohorts
was also assessed.
Methods
Patients and tissue specimens
In this study, a total of 249 epithelial ovarian tumors
(benign, borderline and carcinomatous) were obtained
from archives of paraffin-embedded tissues between
1996 and 2008 at the Department of Pathology, Cancer
Center and the First Affiliated Hospital, Sun Yat-Sen
University, Guangzhou, China. The cancer cases selected
were based on availability of resection tissue and follow-
up data. Patients whose cause of death remained
unknown were excluded from our study. The ovarian
tumor cases encompassed 179 histologically confirmed
invasive carcinomas, 40 borderline tumors and 30 cysta-
denomas. Data of survival time and clinico-pathological
parameters were collected. Ages of the 179 patients with
ovarian carcinoma ranged from 18 to 86 years (mean
age, 50.7 years) and their clinico-pathological character-
istics are summarized in Table 1. None of the cancer
patients in this study had received preoperative radiation
or chemotherapy. In addition, 30 specimens of normal
ovaries from exairesis for non-ovary diseases in the
Department of Gynaecology and Obstetrics of the First
Affiliated Hospital, Sun Yat-Sen University from 2005 to
2008 were used as control. For the use of these clinical
materials for research purposes, prior patient’s consent
and approval from the Institute Research Medical Ethics
Committee of Sun Yat-Sen University was obtained.
Construction of tissue microarrays (TMA)
The TMA was constructed according to a method
described previously [21]. Briefly, the individual donor
tissue block and the corresponding histological H&E
stained slides were overlaid for tissue TMA sampling.
The tissues (179 ovarian carcinoma, 40 borderline
tumor, and 30 cystadenoma tissues) were sampled using
a tissue arraying instrument (Beecher Instruments, Silver
Spring, MD); a 0.6-mm-diameter cylinder of tissue was
removed. Subsequently, the tissue cylinder was re-
embedded into a predetermined position in a recipient
paraffin block. In our constructed ovarian tumor tissue-
TMA, 3 cores of sample were selected from each tumor
tissue. Multiple sections (5 μm thick) were cut from the
TMA block and mounted on microscope slides.
Immunohistochemistry (IHC)
IHC studies were performed using a standard streptavi-
din-biotin-peroxidase complex method [22]. For antigen
retrieval, tissue slides were microwave-treated and
boiled in a 10 mM citrate buffer (pH 6.0) for 10 min.
The slides were incubated with rabbit monoclonal anti-
body against human Bmi-1 (Shangying Biotechnology
Inc. Wuhan, China, 1:2000 dilution), mouse monoclonal
antibodies against human p16Ink4a (1:200 dilution) or
p14ARF (1:300 dilution, Labvision Co., Neomarkers,
USA) overnight at 4°C in a moist chamber. A negative
control was obtained by replacing the primary antibody
with normal rabbit or mouse IgG. Known immunostain-
ing positive slides were used as positive controls.
Expression levels of Bmi-1 protein were visualized by
observing the stained tissues under a light microscope.
Positive expression of Bmi-1 was defined as the presence
of brown or yellowish brown granules in the nuclei,
though occasionally yellowish brown granules could also
be seen in the cytoplasm. For evaluation of the Bmi-1
IHC staining in different ovarian tissues, a semi-quanti-
tative scoring criterion for IHC of Bmi-1 described pre-
viously [16] was used, in which both staining intensity
and positive areas were recorded. A staining index
(values 1 to 16), obtained as the intensity of Bmi-1 posi-
tive staining (negative = 1, weak = 2, moderate = 3, or
strong = 4 scores) and the proportion of immunoposi-
tive cells of interest (≤ 10% = 1, >10% to ≤ 50% = 2,
>50% to ≤ 75% = 3, >75% = 4 scores) were calculated.
The two different scores of corresponding sample were
then multiplied. Points equal or less than 4 was marked
as (-); 4 points to 8 points was marked as (+); 8 points
to 12 points, (++); and 12 points to 16 points, (+++).
For statistical analysis, (-) and (+) were counted as low
expression of Bmi-1 (Fig. 1A, B and 1C), while (++) and
(+++) were counted as intensive expression of Bmi-1
(Fig. 1D and 1E). The evaluation of p16Ink4a and
p14ARF IHC staining was scored on a semi-quantitative
scale [14], as follows: negative expression (<10% of the
cells were positive), down-regulated expression (small
cell clusters, but 10-50% of the cells were positive) and
Yang et al. BMC Cancer 2010, 10:133
http://www.biomedcentral.com/1471-2407/10/133
Page 2 of 9

Page 3

diffuse expression (>50% of the cells were positive).
Results were observed and assessed by two independent
pathologic doctors, without knowing the identity of the
samples.
Fluorescence in situ hybridization (FISH)
Two-color FISH was performed using a Spectrum
Orange-labeled BAC clone (RP11-573G6) at 10p12 con-
taining the Bmi-1 gene and a Spectrum Green-labeled
reference centromeric probe on chromosome 10 (Vysis,
Downers Grove, IL). The FISH reaction was performed
as described previously [23] with slight modification.
Briefly, deparaffinized ovarian carcinoma tissue sections
were treated with proteinase K (400 μg/ml) at 37°C for
30 min, followed by denaturing in 70% formamide, 2×
SSC at 75°C for 6 min. Fifty nanograms of each probe
were mixed in a 20 μl hybridization mixture (containing
55% formamide, 2× SSC, and 2 ∝g human Cot1 DNA),
denatured at 75°C for 6 min and then hybridized to the
denatured tissue sections at 37°C for 24 hours. The
slides were counterstained with 1 μg/ml DAPI in an
anti-fade solution and were examined with a Zeiss Axio-
phot microscope equipped with a triple-band pass filter.
A minimum of 300 tumor cells was evaluated per speci-
men. Amplification of Bmi-1 was defined as presence of
either 6 (or more) Bmi-1 gene signals or at least 3 times
as many gene signals than centromere signals of chro-
mosome 10 in tumor cells (Fig. 1F).
Statistical analyses
Statistical analysis was carried out with the SPSS statisti-
cal software package (SPSS Standard version 13.0, SPSS
Inc.). The association of Bmi-1 protein expression with
ovarian carcinoma patient’s clinico-pathological features
was assessed by the Chi-square test. Multivariate survival
analysis was performed on all parameters that were
Table 1 Association of Bmi-1 expression with patient’s clinico-pathological features in ovarian carcinomas
Bmi-1 protein
All casesLow expression Intensive expression
P valuea
Age at surgery (years)0.452
≤ 50.7b
> 50.7
92
87
60 (65%)
52 (60%)
32 (35%)
35 (40%)
Histological type0.038
Serous (grade 1)
Serous (grade2/3)
Mucinous
Endometrioid
Clear cell
Undifferentiated
11
107
23
8
7
23
9 (82%)
68 (64%)
15 (65%)
7 (88%)
5 (71%)
8 (35%)
2 (18%)
39 (36%)
8 (35%)
1 (12%)
2 (29%)
15 (65%)
Histological grade (Silveberg)0.011
G1
G2
G3
36
101
42
29 (81%)
63 (62%)
20 (48%)
7 (19%)
38 (38%)
22 (52%)
pT status0.037
pT1
pT2
pT3
51
35
93
39 (76%)
22 (63%)
51 (55%)
12 (24%)
13 (37%)
42 (45%)
pN status0.001
pN0
pN1
88
91
66 (75%)
46 (51%)
22 (25%)
45 (49%)
pM status0.006
pMX
pM1
153
26
102 (67%)
10 (38%)
51 (33%)
16 (62%)
FIGO stage0.002
I33
21
99
26
27 (82%)
17 (81%)
58 (59%)
10 38%)
6 (18%)
4 (19%)
41 (41%)
16 (62%)
II
III
IV
aChi-square test
bMean age
Yang et al. BMC Cancer 2010, 10:133
http://www.biomedcentral.com/1471-2407/10/133
Page 3 of 9

Page 4

Figure 1 Immunohistochemical staining of Bmi-1 protein and Fluorescence in situ hybridization of Bmi-1 gene in human ovarian
tissues. (A) Negative expression of Bmi-1 was observed in a normal surface epithelium of ovary (200×). (B) An ovarian cystadenoma showed low
expression of Bmi-1, in which about 10% of tumor cells was detected moderate positive staining of Bmi-1 (200×). (C) Low expression of Bmi-1
was observed in an borderline ovarian tumors, in which 5% of tumor cells showed weak positive staining of Bmi-1 (200×). (D) Intensive
expression of Bmi-1 was detected in an ovarian carcinoma (case 33), in which more than 90% of carcinoma cells showed strong positive staining
of Bmi-1 (200×). (E) Another ovarian carcinoma (case 161) showed intensive expression of Bmi-1, in which all of tumor cells had strong positive
staining of Bmi-1 (200×). (F) Amplification of Bmi-1 gene was observed by FISH in the same ovarian carcinoma case (161), in which Bmi-1 gene
signals (red) was detected at least 3 times more than centromere signals of chromosome 10 (green) (1000×).
Yang et al. BMC Cancer 2010, 10:133
http://www.biomedcentral.com/1471-2407/10/133
Page 4 of 9

Page 5

found to be significant on univariate analysis using the
Cox regression model. For univariate survival analysis, we
analyzed all ovarian carcinoma patients by Kaplan-Meier
analysis. Log rank test was used to compare different sur-
vival curves. P < 0.05 was considered significant.
Results
Bmi-1 expression in ovarian tissues
Bmi-1 expression could be evaluated informatively in
TMA tissues of 163/179 of ovarian carcinomas, 37/40 of
borderline tumors and 26/30 of cystadenomas. The non-
informative TMA samples included unrepresentative
samples, samples with too few tumor cells (<300 cells
per case) and lost samples. IHC staining of such non-
informative samples were replaced and performed by
using whole tissue slides. In our study, we defined that
(-) and (+) were counted as low expression of Bmi-1,
while (++) and (+++) were counted as intensive expres-
sion of Bmi-1. According to this definition, the intensive
expression of Bmi-1 was detected in 67/179 (37%) ovar-
ian carcinomas. The increasing frequencies of Bmi-1
intensive expression in normal ovarian epithelium (0), to
benign cystadenomas (3%), to borderline tumors (10%),
and to invasive carcinomas were significant (37%, P <
0.05, Table 2).
Association of Bmi-1 expression with ovarian carcinoma
patient clinico-pathologic features
The association between Bmi-1 expression in ovarian
carcinomas and several known clinico-pathological fea-
tures was further studied. Bmi-1 expression was posi-
tively correlated with tumors histological type, grade,
pT/pN/pM status, and FIGO stage (P < 0.05, Table 2).
No significant correlation was obtained between Bmi-1
expression and patient age (≤ 50.7 years vs > 50.7 years)
(P > 0.05, Table 1).
Relationship between clinicopathologic variables, Bmi-1
expression and ovarian carcinoma patient survival:
Univariate survival analysis
In univariate survival analyses, cumulative survival
curves were calculated according to the Kaplan-Meier
method. Differences in survival times were assessed with
the log-rank test. First, to confirm the representativeness
of the ovarian carcinomas in our study, we analyzed
established prognostic predictors of patient survival.
Kaplan-Meier analysis demonstrated a significant impact
of well-known clinical pathological prognostic para-
meters, such as tumor histological grade (p = 0.012),
pT/pN/pM status (p < 0.01) and FIGO stage (p < 0.001)
on patient survival (Table 3). The mean survival time
for patients with tumors having intensive expression of
Bmi-1 was 49.3 months compared to 100.3 months for
pateints with tumors having low expression of Bmi-1
Table 2 The expression of Bmi-1 in normal ovaries and in
benign and malignant epithelial ovarian tumorsa
Bmi-1 protein
All cases Low expression Intensive expression
Normal ovaries
Cystadenomas
Borderline tumors
Invasive carcinomas
30
30
40
179
30 (100%)
29 (97%)
36 (90%)
112 (63%)
0 (0)
1 (3%)
4 (10%)
67 (37%)
aValues are n (%). A significant increasing frequency of intensive expression of
Bmi-1 was observed in cystadenomas, in borderline tumors and in invasive
carcinomas (P < 0.01, Chi-Square Test for Trend)
Table 3 Clinical pathological parameters and expression
of Bmi-1 for prognosis of 179 patients with ovarian
carcinoma by univariate survival analysis (log-rank test)
VariableAll
cases
Mean
survival
(months)
Median
survival
(months)
P
value
Age at surgery (years)
≤ 50.7a
> 50.7
Histological type
Serous(grade 1)
Serous(grade2/3)
Mucinous
Endometrioid
Clear cell
Undifferentiated
Histological grade
(Silveberg)
G1
G2
G3
pT status
pT1
pT2
pT3
pN status
pN0
pN1
pM status
pMX
pM1
FIGO stage
0.388
92
87
83.4
81.7
136.0
55.0
0.491
11
107
23
8
7
23
110.4
63.4
75.8
116.7
102.8
32.7
136.0
52.0
NRb
NR
NR
NR
0.012
36
101
42
104.9
85.9
48.0
136.0
64.0
29.0
0.003
51
35
93
109.5
82.2
66.5
NR
NR
35.0
<0.001
88
91
100.9
52.8
136.0
28.0
<0.001
153
26
94.8
21.5
136.0
9.0
<0.001
I33
21
99
26
134.2
115.0
71.5
21.5
NR
NR
37.0
9.0
II
III
IV
Bmi-1 expression <0.001
Low112
67
100.3
49.3
136.0
22.0Intensive
aMean age
bNot reached
Yang et al. BMC Cancer 2010, 10:133
http://www.biomedcentral.com/1471-2407/10/133
Page 5 of 9