Bladder Cancer

Protein Expression Patterns of Ezrin Are Predictors of

Progression in T1G3 Bladder Tumours Treated with

Nonmaintenance Bacillus Calmette-Guerin

Joan Palou a, Ferran Algaba a, Irene Vera b, Oscar Rodriguez a, Humberto Villavicencio a,Marta Sanchez-Carbayo b,*

a Fundacio Puigvert, Barcelona, Spainb Tumor Markers Group, Spanish National Cancer Research Center, Madrid, Spain

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6

ava i lable at www.sciencedirect .com

journal homepage: www.europeanurology.com

Article info

Article history:

Accepted September 30, 2008Published online ahead ofprint on October 9, 2008

Keywords:

Ezrin

BCG

Bladder cancer

Immunohistochemistry

Tissue arrays

Abstract

Background: Bacillus Calmette-Guerin (BCG) is a standard treatment for reducing tumour

recurrence and delaying progression of high-risk, non–muscle-invasive bladder tumours.

However, it is not clear yet which patients are more likely to be responders to BCG.

Objective: To evaluate the role of ezrin expression in bladder cancer (BCa) progression in

T1G3 bladder tumours treated with BCG.

Design, setting, and participants: Ezrin protein expression patterns were analysed on

tumour specimens belonging to 92 patients with T1G3 non–muscle-invasive BCa

undergoing nonmaintenance BCG treatment. Re-resection was not performed. The

median follow-up was 90.5 mo (range: 3.0–173.0). A specific tissue array was created

containing three representative cores of each of the tumour specimens belonging to

these patients.

Measurements: Ezrin protein expression patterns were assessed by immunohistochem-

istry on this tissue array. Proliferation rates were assessed by means of Ki67 staining.

Recurrence, progression into muscle-invasive tumours, and disease-specific overall sur-

vival (OS) rates were analysed using univariate and multivariate tests.

Results and limitations: Among the 92 patients analysed, 40 recurred (43.5%), 17

progressed (18.5%), and 14 died of the disease (15.2%). Log-rank survival analyses

revealed that an ezrin membrane expression <20% was significantly associated with

increased progression ( p = 0.009) and shorter disease-specific OS ( p = 0.006). Multi-

variate analyses showed that ezrin was an independent prognostic marker of progres-

sion ( p = 0.031) and disease-specific survival ( p = 0.035). Interestingly, the low ezrin

membrane expression correlated with high proliferation rates ( p = 0.033).

Conclusions: Immunohistochemistry analyses revealed that the membrane expression

of ezrin is associated with the clinical outcome of patients with T1G3 tumours under-

going BCG treatment. Protein expression patterns of ezrin were associated with tumour

progression in T1G3 disease. The differential expression of ezrin distinguished patients

responding to BCG from those who may require a more aggressive therapeutic approach.

# 2009 Published by Elsevier B.V. on behalf of European Association of Urology.

* Corresponding author. Tumor Markers Group, 308A, Centro Nacional de Investigaciones Oncologicas,Melchor Fernandez Almagro 3, E-28029 Madrid, Spain. Tel. +34 91 732 8053; Fax: +34 91 224 6972.E-mail address: mscarbayo@cnio.es (M. Sanchez-Carbayo).

0302-2838/$ – see back matter # 2009 Published by Elsevier B.V. on behalf of European Association of Urology. doi:10.1016/j.eururo.2008.09.062

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6830

1. Introduction

Bladder cancer (BCa) is the fourth most frequent neoplasia

in men, clinically characterised by high recurrent rates and

poor prognosis when tumours invade the muscularis propia

[1,2]. The intravesical bacillus Calmette-Guerin (BCG)

immunotherapy represents a highly successful therapy

for patients with non–muscle-invasive BCa [3–6]. Despite

the superior efficacy of BCG over transurethral resection

(TUR) alone or TUR plus intravesical chemotherapy,

eventually, >50% of non–muscle-invasive BCas either

persist or recur. This is a particularly acute problem for

patients with high-risk disease, such as those with

carcinoma in situ (CIS), submucosa invasion (stage T1),

and high-grade papillary disease, because the risk of

progression increases proportionally with a narrowing

window of opportunity for conservative bladder therapy

before cystectomy is mandatory. A significant portion of

these patients fails to respond to BCG therapy; their

tumours not only persist or recur but may also become

invasive or metastatic [6,7]. None of the current tumour

biomarkers evaluated to date has provided sufficient

sensitivity and specificity to predict response to BCG

immunotherapy in any of these three subtypes of high-

risk BCa patients in clinical routine practice [2,6,7].

Improved prognostic biomarkers are required to ultimately

distinguish indolent cancers from those that are potentially

lethal so that therapeutic procedures could be tailored to

each individual patient [2–11].

The ezrin, radixin, and moesin (ERM) and merlin proteins

are closely related members of the band 4.1 superfamily of

proteins that, when activated, interact with both membrane

proteins and the actin cytoskeleton [12–15]. By organising

membrane–cytoskeleton-associated complexes and creat-

ing specialized membrane domains, the ERM proteins

regulate cellular activities such as survival, adhesion,

migration, and invasion, all of which are important during

tumour progression [12–16]. The membrane-linking pro-

tein ezrin is expressed in several types of human cancers,

and correlations of its immunoreactivity and clinicohisto-

pathologic data as well as patient outcome have previously

been shown [17–25]. However, to the best of our knowl-

edge, ezrin expression has not been reported to date in

bladder tissues. In this study, immunohistochemical

analyses were performed in T1G3 bladder tumours treated

with nonmaintenance BCG, aiming to evaluate the role of

ezrin expression in BCa progression and as a therapeutic

predictive marker of BCG response in this type of patient.

2. Methods

2.1. Patient population

Between 1989 and 1996, 92 patients were treated for primary stage T1

grade 3 urothelial carcinoma with complete transurethral resection

(TUR) at the Fundacio Puigvert. Of those, 83 were male (90.2%) and 9

were female (9.8%). The median age was 67.5 yr, ranging from 25 to 81

yr. Among these cases, 55 had concomitant CIS (59.8%), and 45 had

multifocal disease (48.9%). T1 substaging was defined following

previously reported criteria [26]. Primary bladder tumours were

collected following the guidelines for the protection of human subjects

and recruited under Institutional Review Board (IRB)–approved proto-

cols at the Fundacio Puigvert. They all received a 6-wk course of BCG

without maintenance therapy. Criteria for inclusion were that patients

had no prior intravesical treatment, histologic diagnosis (T1G3) with a

wide and deep primary resection including muscle in the specimen, and

paraffin-embedded tissue material adequate for analysis. Patients

having presented previously a tumour of the upper urinary tract were

excluded from the study. Overall, 92 patients were identified and

selected for the present analysis. Second-look TUR was not performed.

Follow-up consisted of cystoscopy with cytology every 3 mo for the first

2 yr and every 6 mo thereafter. Progression was defined as muscular

invasion (stage T2 or higher) or metastatic disease. Patients with

recurrence were treated either with another course of BCG or with

cystectomy when the disease progressed.

2.2. Tissue microarrays

Paraffin-embedded tissues belonging to these 92 patients were used for

the construction of the tissue microarray, as previously described

[27,28]. One section was stained with haematoxylin and eosin to

evaluate the presence of the tumour by light microscopy. After carefully

choosing the morphologically representative region on the paraffin-

embedded blocks (donor blocks), a core tissue biopsy specimen of

0.6 mm was punched and transferred to the donor paraffin-wax–

embedded block (recipient block). The tissue array included three cores

representative of each of the whole tumour for a total of 92 primary

T1G3 non–muscle-invasive bladder tumours that underwent BCG

treatment. Clinicopathologic and annotated follow-up information

available for BCG clinical outcome predictive analyses of the tumours

spotted onto the tissue microarray allowed the evaluation of the

histopathologic and clinical properties as well as the outcome

assessment of the protein expression patterns of ezrin, Ki67, and p53.

For progression analyses, only patients with available follow-up (either

‘‘progressing into invasive disease’’ or ‘‘alive with no evidence of

disease’’) were considered. For overall survival (OS) analyses, only

patients with available follow-up (either ‘‘dead as a result of disease’’ or

‘‘alive with no evidence of disease’’) were included. Cases with unknown

follow-up were excluded from these analyses.

2.3. Immunohistochemistry

The protein expression patterns of ezrin were assessed at the

microanatomic level using both cytospins from cancer cell lines (data

not shown) and the tissue microarrays. Standard avidin-biotin

immunoperoxidase procedures were applied for immunohistochemistry

[28]. Antigen retrieval methods (0.01% citric acid for 15 min under

microwave treatment) were used prior to incubation with primary

antibodies overnight at 4 8C. Ezrin staining was assessed using a mouse

monoclonal antibody diluted at 1:2000 (Sigma; Saint Louis, MO, USA).

P53 was assessed using a mouse monoclonal antibody at 1:50 dilution

(clone D07; Novocastra, Newcastle, UK). Ki67 was assessed using a

mouse monoclonal antibody diluted at 1:100 (clone MIB-1; DAKO,

Glostrup, Denmark). Secondary biotinylated antimouse antibodies

(Vector Laboratories, Burlingame, CA, USA) were used at 1:500 dilution.

The absence of primary antibody was used as a negative control.

Diaminobenzidine was used as the final chromogen and haematoxylin as

the nuclear counterstain. Ezrin immunoreactivity was evaluated in the

membrane and in the cytoplasm. Membrane staining was scored based

on the number of cancer cells presenting this protein sublocalisation.

Cytoplasmic staining was scored 0–3 (absent, weak, moderate, or strong

intensity). Scores were recorded for membrane stainings of ezrin, using

normal brain as positive control. P53 and Ki67 immunoreactivity was

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6 831

evaluated in the nucleus, as previously reported [9,10,28]. The individual

scores were reviewed, and the agreement between two independent

observers was calculated. Whenever a discrepancy was noted between

the first and the second interpretations, the pathologist decided on the

final scoring.

2.4. Statistical analysis

All cases (n = 92) were used for the analysis of association among ezrin,

Ki67, and p53 with clinicopathologic variables. The consensus (mean)

value of the three representative cores from each tumour sample arrayed

was used for statistical analyses. The association of the expression of

ezrin, Ki67, and p53 with age, sex, and histopathologic variables such as

T1 substaging, tumour size, multifocality, or the presence of associated

CIS was evaluated using the nonparametric Mann-Whitney and

Kruskall-Wallis tests [27]. The distribution of the protein expression

patterns of ezrin, Ki67, and p53 depending on these variables was

described by means of their median and range values. There is no

consensus on the cut-off of the immunohistochemical expression of

ezrin. Thus, the cut-off value for weak- and strong-expressing cases was

specified at the median percentage score of positive membrane tumour

cells, resulting in a value of 20%. Ezrin was then analysed continuously,

taking the cut-off of 20% when considered as a categoric variable [29].

The associations of membrane protein expression patterns of ezrin with

OS were evaluated using this set of 92 cases for which follow-up was

available. OS time was defined as the years elapsed between TUR and

death as a result of disease (or the last follow-up date). Patients who

were alive at the last follow-up or lost to follow-up were censored. The

association of ezrin expression levels with OS was analysed using the

log-rank test to examine its relationship, taking the 20% cut-off

mentioned above [29]. The survival curves were plotted using the

standard Kaplan-Meier methodology. Multivariate Cox regression

analyses were performed to evaluate how ezrin behaves compared

with other clinical and pathologic factors such as sex, age, the presence

of CIS, Ki67, p53, multifocality, tumour size, and T1 substaging [29]. The

simultaneous analysis of the status of Ki67 and p53 in these specimens

allowed comparison of the clinical relevance of the associations of ezrin

Table 1 – Distribution of the protein expression of ezrin, Ki67, and p53on tissue arrays.

Characteristic Ezrin % median (range)

Age p = 0.459

Cut-off �65 (n = 36) 32.5 (0.0–95.0)

Cut-off >65 (n = 56) 0.0 (0.0–100.0)

Sex p = 0.778

Male (n = 83) 12.1 (0.0–100.0)

Female (n = 9) 16.2 (0.0–85.0)

Substaging p = 0.356

1a (n = 44) 0.0 (0.0–95.0)

1b (n = 20) 42.5 (0.0–85.0)

1c (n = 28) 0.0 (0.0–95.0)

Tumour size p = 0.374

<3 (n = 61) 37.5 (0.0–95.0)

>3 (n = 31) 0.0 (0.0–100.0)

Focality p = 0.674

1 (n = 48) 25.0 (0.0–95.0)

2 (n = 17) 22.5 (0.0–100.0)

3 (n = 3) 0.0 (0.0–85.0)

Multiple (n = 24) 1.65 (0.0–95.0)

CIS presence p = 0.847

Yes (n = 55) 17.5 (0.0–100.0)

No (n = 37) 11.7 (0.0–95.0)

with these proteins previously described altered along BCa progression

[28]. Associations among ezrin with Ki67 and p53 were analysed using

Kendall tß test [29]. Only p values <0.05 were considered statistically

significant. Statistical analyses were performed using the SPSS v.11.0

(SPSS Inc, Chicago, IL, USA).

3. Results

3.1. Clinical follow-up of the cases under analyses: overall

results

At the last follow-up, of the 92 patients analysed, 40

recurred (43.5%), 17 progressed (18.5%), and 38 died

(39.1%); 14 of these cases died of BCa (15.2%). Fifty-two

of the patients analysed were alive at the end of the study

(56.5%). Two patients who experienced recurrence over the

time of the study were lost at the last follow-up. Patients

who recurred had recurrence at a median follow-up time of

15.5 mo (range: 1.0–116.0). Patients who progressed had

the progression at a median follow-up time of 17.0 mo

(range: 2.0–126.0). The median time at which patients died

of BCa was 33.5 mo (range: 7.0–136.0). The median time at

which patients died of causes other than BCa was 45.5 mo

(range: 6.0–147.0). The median follow-up time of cases free

of disease was 106.5 mo (recurrence-free survival; range:

51.0–173.0). The median follow-up time considering all the

cases under analyses was 90.5 mo (range: 3.0–173.0).

3.2. Ezrin, p53, and Ki67 expression: clinicopathologic

associations

The cellular expression patterns of ezrin were found in the

cytoplasm and in the membrane in T1G3 bladder tumours.

The immunoreactivity of ezrin in the cytoplasm was

, depending on the clinical characteristics of T1G3 patients spotted

Ki67% median (range) p53 median (range)

p = 0.271 p = 0.884

20.0 (0.0–75.0) 9.2 (0.0–90.0)

25.0 (2.5–65.0) 15.0 (0.0–96.7)

p = 0.007 p = 0.014

20.0 (0.0–75.0) 9.0 (0.0–95.0)

39.8 (25.0–60.0) 55.0 (3.3–96.7)

p = 0.817 p = 0.235

26.3 (0.0–95.0) 7.5 (4.2–75.0)

27.5 (2.5–75.0) 60.0 (0.0–96.7)

25.4 (5.0–60.0) 17.5 (0.0–95.0)

p = 0.465 p = 0.469

24.4 (0.0–95.0) 7.5 (0.0–75.0)

25.8 (3.3–75.0) 15.8 (0.0–95.0)

p = 0.321 p = 0.926

29.2 (0.0–75.0) 15.8 (0.0–96.7)

10.0 (4.2–55.0) 5.0 (0.0–95.0)

20.0 (20.0–35.0) 26.3 (0.0–85.0)

15.8 (2.5–75.0) 8.7 (0.0–85.0)

p = 0.535 p = 0.297

21.2 (4.2–75.0) 8.3 (0.0–90.0)

29.6 (0.0–75.0) 17.5 (0.0–96.7)

Table 2 – Summary of the associations among the proteinexpression patterns of ezrin, Ki67, and p53 in the cases understudy.

Association Kendall tß p value

Ezrin–Ki67 �0.172 0.033

Ezrin–p53 �0.007 0.935

Ki67–p53 0.249 0.001

Fig. 1 – Differential protein expression patterns of ezrin and Ki67 in PT1G3 blamembrane protein expression patterns of ezrin by immunohistochemistry onrepresentative immunostainings of Ki67 by immunohistochemistry on tissue aabove; (E, F) representative immunostainings of p53 by immunohistochemistrtumours shown above.

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6832

detected in all the cases under study, with various

intensities that did not differ in a statistically significant

manner regarding any of the clinicopathologic variables

under study. Ezrin expression in the membrane differed

among cases, with a wide range from only a few cells

positive from ezrin to nearly 100% of positive tumours cells.

The cut-off value for weak and strong expressing cases was

specified at the median percentage score of positive

dder tumours; (A, B) representative immunostainings of the differentialtissue arrays in non–muscle-invasive (T1G3) bladder tumours; (C, D)rrays in the paired non–muscle-invasive (T1G3) bladder tumours showny on tissue arrays in the paired non–muscle-invasive (T1G3) bladder

Fig. 2 – Ezrin membrane expression was associated with bladder cancerprogression in PT1G3 primary bladder tumours, with Kaplan-Meiercurve survival analysis indicating that a membrane protein expression ofezrin <20% measured by immunohistochemistry on tissue arrays wasassociated with high progression rate into muscle-invasive disease (logrank, p = 0.009).

Fig. 3 – Ezrin membrane expression was associated with poor survival inPT1G3 primary bladder tumours, with Kaplan-Meier curve survivalanalysis indicating that a membrane protein expression of ezrin <20%measured by immunohistochemistry on tissue arrays was associatedwith shorter overall survival (log rank, p = 0.006).

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6 833

membrane tumour cells, resulting in a value of 20%. Tumour

stroma was always negative for ezrin staining. A lack of

significant correlations was observed among p53, Ki67, and

ezrin expression, with any of the clinicopathologic variables

mentioned above, such as histopathologic T1 substaging,

tumour size, multifocality, or the presence of associated CIS

(Table 1). The statistical association found between the

expression of Ki67 and p53 regarding sex should be

mentioned, although its clinical implications is limited

because of the low number of females included in this

study. Interestingly, the low membrane protein expression

patterns of ezrin correlated with high proliferation rates, as

measured by Ki67 staining in a significant manner

(tß = �0.172, p = 0.033). The correlations among the protein

expression patterns of these proteins are summarised in

Table 2. Thus, although tumour cells showed strong

immunoreactivity in their cytoplasmic patterns, they

showed loss of membrane ezrin reactivity in highly

proliferative cases. Representing immunostainings of ezrin,

Ki67, and p53 are shown in Fig. 1.

3.3. Protein expression patterns of ezrin are associated with

progression into muscle-invasive disease and OS

The progression rate was increased for patients with low

ezrin membrane-expressing tumours compared with those

with strong membrane expression. Log-rank survival

analyses revealed that an ezrin membrane expression

<20% was significantly associated with an increased

progression into muscle-invasive disease (log-rank,

p = 0.009; Fig. 2). Moreover, a shorter disease-specific OS

was observed in patients with low membrane expression

patterns of ezrin (log-rank, p = 0.006). The 5-yr survival rate

was decreased for patients with low ezrin membrane-

expressing tumours compared with those with strong

membrane expression (Fig. 3). Cox regression analyses

revealed that ezrin ( p = 0.031), p53 ( p = 0.018), T1 substa-

ging ( p = 0.018), the presence of concomitant CIS

( p = 0.017), tumour size ( p = 0.015), and sex ( p = 0.014)

were independent prognostic factors for progression into

invasive disease. Multivariate analysis revealed that ezrin

expression was the only variable with an independent

prognostic effect on disease-specific survival ( p = 0.035)

compared with other clinical and pathologic factors,

including sex, age, concomitant CIS, Ki67, p53, multifocality,

tumour size, and T1 substaging.

3.4. Protein expression patterns of ezrin are predictive

biomarkers of BCG immunotherapy

The membrane protein expression of ezrin and Ki67

staining was found to be differentially expressed between

early recurrent patients from those who were free of disease

after BCG immunotherapy ( p = 0.041 and p = 0.015, respec-

tively). Both Ki67 and the ezrin membrane expression

served to discriminate patients who experienced recur-

rence at 3 mo after BCG intravesical immunotherapy from

those who were free of disease as confirmed by cystoscopy.

Interestingly, a nearly significant association was found

between membrane ezrin expression and disease recur-

rence along the overall follow-up of the cases in this study

( p = 0.06).

4. Discussion

Ezrin is a member of the ERM cytoskeleton-associated

protein family, which was first described as links between

membrane proteins and actin filaments [12,13]. Following

the initial description, these proteins have become impli-

cated in several important cellular complexes and processes.

Ezrin has been reported to be involved in the regulation of

critical events associated with tumour progression, such as

cell survival, proliferation, and migration [12–15]. Altera-

tions of ezrin expression can mediate many changes in the

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6834

cell surface signals and intracellular signalling cascade that

confer the invasive capability in tumour cells. Therefore, it is

conceivable that ezrin deregulation could contribute to the

invasive behaviour of tumours. Evidence from both animal

models and prospective human studies showing correlations

between ezrin expression levels and tumour progression are

consistent with a crucial role for ezrin in tumour invasion and

dissemination [16–23]. The results presented in this report

are consistent with these observations in the context of the

progression of non–muscle-invasive tumours into invasive

disease.

Ezrin is physiologically expressed in a variety of

epithelial tissue, including intestine, lung, and kidney,

among others. Within these tissues, ezrin has been

suggested to play an essential role in controlling cell

motility [14–16]. The membrane-linking protein ezrin is

highly expressed in several types of human cancers, and

correlations of its immunoreactivity and histopathologic

data as well as patient outcome have previously been

shown [17–25]. The presence of diverged protein expres-

sion patterns along tumour progression and prognosis has

been described for ERM proteins. Although high ezrin

expression has been reported to be associated with poor

prognosis, low expression has also been related to adverse

outcome [17–25]. Even within the same tumour type, such

as in ovarian cancer, low [19] and high [20] ezrin expression

has been associated with poor prognosis. Our observations

with ezrin expression are in line with our previous findings

reporting that the low expression of moesin—another of the

ERM proteins—was associated with poor outcome in BCa

[28]. Supporting our results is the association of weak or

negative ezrin immunoreactivity with poor outcome in

serous ovarian carcinoma [19]. To the best of our knowl-

edge, neither the potential differential expression in

bladder tissues nor the biological role of ezrin in BCa has

been reported to date. The loss of membrane ezrin protein

expression was associated with increased progression of

T1G3 tumours into invasive disease and poorer survival.

Although we focused on T1G3 disease, not covering other

non–muscle-invasive or advanced PT2+ tumours, our

results are consistent with the described biologic role of

ezrin in cancer cells. Positioned at the submembranal-

cytoskeletal interface, ezrin is believed to play a central role

in the regulation of invasion in several cancer cell types [17–

23]. The data presented above support the role of ezrin in

the progression of non–muscle-invasive T1G3 lesions into

invasive disease. Our study also showed a poorer response

to BCG in patients with low ezrin expression. Only adequate

in vitro and in vivo experiments exposing BCa cells and

derived tumours with different controlled ezrin levels to

BCG would be able to assess the biological mechanisms by

which ezrin expression and subsequent cytoskeleton

rearrangements affect BCG immunotherapeutic response.

The cytoplasmic and membrane protein expression

identified in bladder tumours is consistent with the dual

localisation reported for this protein. The N-terminal

binding domain of ezrin mediates membrane attachment

by binding the cytoplasmic tail of CD44, CD43, or

intercellular adhesion molecules. On the other end, ezrin

through its C-terminal domain associates with F-actin and

contributes to its microfilament cytoplasmic organisation.

It is well known that ezrin is exchanged between cytoplasm

and membrane. Cytoplasmic ezrin exists in a ‘‘closed’’

conformation based on intra- or intermolecular interactions

between the N- and C-terminus. Threonine and tyrosine

phosphorylation induces an ‘‘open’’ conformation. In this

state, ezrin is localised towards the membrane, where it

modulates F-actin dynamics and tethers the microfilament

system to the cytoplasmic face of cell adhesion sites [23].

Disruption of actin filaments and a decrease in focal

adhesion are common features of epithelial–mesenchymal

transition, which is associated with the onset of invasion.

Besides a structural role, ezrin may act as a signalling or

scaffold molecule. It is believed that ezrin is involved in cell

adhesion functions through interactions with the Rho-

associated signal transduction [16] and the Akt-mediated

apoptotic pathway [30]. These observations suggest that in

the presence of low ezrin expression, the subsequent

altered organisation of the cytoskeleton, the impaired focal

adhesion, and signalling deregulation would lead to an

increased invasion of BCa cells. These cellular changes

would affect disease progression, a poorer therapeutic

response to BCG, and clinical outcome. Further studies are

warranted to dissect the specific signalling events regulated

by ezrin using in vitro and in vivo models in BCa.

The results presented in this report are clinically relevant

for the patients affected with bladder tumours. It is known

that urothelial carcinomas presenting T1G3 non–muscle-

invasive lesions are high-risk tumours because of the

possibility of progression to muscle-invasive or metastatic

disease [7]. The difficulty remains in deciding which

patients should be treated under a conservative approach

using intravesical treatment with BCG and which should

undergo radical cystectomy [5]. Critical clinical variables,

such as the presence of an associated CIS or an early BCG

failure [4–8], have been defined as factors related to

progression and must be considered before making such a

decision [4–8]. Lacking reliable prognostic and predictive

tumour markers, laboratory research is warranted to

identify candidate biomarkers of clinical behaviour. In this

regard, p53 has been widely studied, with results not

conclusive enough to make clinical decisions in T1G3

disease [9,10,31], similar to our findings. In our series, ezrin

and Ki67 stainings showed a prognostic ability to dis-

criminate patients more likely to progress into muscle-

invasive disease. In this report, several issues related to the

wide heterogeneity among BCG failures have been

addressed. The loss of ezrin membrane expression was

associated with early recurrence, progression into muscle-

invasive disease, and OS. Despite the promising clinical

relevance of these findings, which suggest the adjunct value

as a prognostic and predictive biomarker, our results cannot

conclude whether the urologist could decide to undertake a

radical treatment based on the membrane protein expres-

sion of ezrin. In a multivariate survival analysis, ezrin

immunoreactivity had an independent effect on tumour

progression and disease-specific survival when controlling

for sex, age, concomitant CIS, Ki67, p53, multifocality,

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6 835

tumour size, and T1 substaging. This observation is relevant

because it was found in a limited number of cases analysed

(n = 92) and controlling its prognostic effect for a high

number of clinical and pathologic variables (n = 9). Further,

larger and prospective studies are warranted to define the

adjunct value of ezrin measurement in stratifying the

clinical outcome of patients with T1G3 disease undergoing

BCG immunotherapy and to decide upon tailored, more

aggressive interventions.

5. Conclusions

In summary, immunohistochemistry analyses revealed that

the membrane expression of ezrin is associated with the

clinical outcome of patients with T1G3 tumours undergoing

BCG treatment. Protein expression patterns of ezrin were

associated with tumour progression in T1G3 disease. Thus,

the differential expression of ezrin discriminated patients

responding to BCG from those who may require a more

aggressive therapeutic approach. These findings warrant

further elucidation in prospective trials to evaluate whether

ezrin immunohistochemical evaluation could be an adjunct

tool for therapeutic decisions.

Author contributions: Marta Sanchez-Carbayo had full access to all the data

in the study and takes responsibility for the integrity of the data and the

accuracy of the data analysis.

Study concept and design: Sanchez-Carbayo.

Acquisition of data: Sanchez-Carbayo, Palou, Algaba, Vera.

Analysis and interpretation of data: Sanchez-Carbayo, Palou, Algaba.

Drafting of the manuscript: Sanchez-Carbayo, Palou, Algaba.

Critical revision of the manuscript for important intellectual content: Sanchez-

Carbayo, Palou, Algaba.

Statistical analysis: Sanchez-Carbayo, Palou.

Obtaining funding: Sanchez-Carbayo.

Administrative, technical, or material support: Rodriguez, Villavicencio.

Supervision: Sanchez-Carbayo.

Other (specify): None.

Financial disclosures: I certify that all conflicts of interest, including specific

financial interests and relationships and affiliations relevant to the subject

matter or materials discussed in the manuscript (eg, employment/affilia-

tion, grants or funding, consultancies, honoraria, stock ownership or

options, expert testimony, royalties, or patents filed, received, or pending),

are the following: None.

Funding/Support and role of the sponsor: None.

Acknowledgement statement: The authors acknowledge all members of the

laboratory of Dr Sanchez-Carbayo at the Tumor Markers Group at the CNIO

for their technical support and constructive suggestions in the preparation

of this manuscript. We would like to thank all the members of our clinical

collaborators at the Fundacio Puigvert involved in this study for their

support in facilitating the tumour specimens as well as the clinical fol-

low-up of the BCa cases analysed in this study.

References

[1] Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J

Clin 2008;58:71–96.

[2] Sanchez-Carbayo M, Cordon-Cardo C. Molecular alterations asso-

ciated with bladder cancer progression. Semin Oncol 2007;34:75–

84.

[3] Morales A, Eideger D, Bruce AW. Intracavitary bacillus Calmette-

Guerin in the superficial bladder tumor. J Urol 1976;116:180–3.

[4] Lamm DL. Preventing progression and improving survival with BCG

maintenance. Eur Urol 2000;37(Suppl 1):9–15.

[5] Patard J-J, Rodriguez A, Leray E, et al. Intravesical bacillus Calmette-

Guerin treatment improves patient survival in T1G3 bladder

tumors. Eur Urol 2002;41:635–42.

[6] Saint F, Salomon L, Quintela R, et al. Do prognostic parameters of

remission versus relapse after bacillus Calmette–Guerin (BCG)

immunotherapy exist? Analysis of a quarter century of literature.

Eur Urol 2003;43:351–61.

[7] Sylvester RJ, van der Meijden APM, Oosterlinck W, et al. Predicting

recurrence and progression in individual patients with stage Ta T1

bladder cancer using EORTC risk tables: a combined analysis of

2596 patients from seven EORTC trials. Eur Urol 2006;49:466–77.

[8] Fernandez-Gomez J, Solsona E, Unda M, et al. Prognostic factors in

patients with non–muscle-invasive bladder cancer treated with

bacillus Calmette-Guerin: multivariate analysis of data from four

randomized CUETO trials. Eur Urol 2008;53:992–1002.

[9] Lacombe L, Dalbagni G, Zhang ZF, et al. Overexpression of p53

protein in a high-risk population of patients with superficial blad-

der cancer before and after bacillus Calmette-Guerin therapy:

correlation to clinical outcome. J Clin Oncol 1996;14:2646–52.

[10] Esuvaranathan K, Chiong E, Thamboo TP, et al. Predictive value of

p53 and pRb expression in superficial bladder cancer patients

treated with BCG and interferon-alpha. Cancer 2007;109:1097–

105.

[11] Palou J, Rodriguez O, Segarra J, Rosales A. Restaging transurethral

resection of high-risk superficial bladder cancer improves the

initial response to bacillus Calmette-Guerin therapy. J Urol 2006;

176:407.

[12] Trofatter JA, MacCollin MM, Rutter JL, et al. A novel moesin-, ezrin-,

radixin-like gene is a candidate for the neurofibromatosis 2 tumor

suppressor. Cell 1993;72:791–800.

[13] Mackay DJ, Esch F, Furthmayr H, et al. Rho- and rac-dependent

assembly of focal adhesion complexes and actin filaments in per-

meabilized fibroblasts: an essential role for ezrin/radixin/moesin

proteins. J Cell Biol 1997;138:927–38.

[14] Mangeat P, Roy C, Martin M. ERM proteins in cell adhesion and

membrane dynamics. Trends Cell Biol 1999;9:187–92.

[15] Bretscher A, Edwards K, Fehon RG. ERM proteins and merlin. Nat

Rev Mol Cell Biol 2002;3:586–99.

[16] Curto M, McClatchey AI. Ezrin . . . a metastatic determinant? Cancer

Cell 2004;5:113–4.

[17] Ohtani K, Sakamoto H, Rutherford T, et al. Ezrin, a membrane-

cytoskeletal linking protein, is involved in the process of invasion of

endometrial cancer cells. Cancer Lett 1999;147:31–8.

[18] Geiger KD, Stoldt P, Schlote W, Derouiche A. Ezrin immunoreactiv-

ity is associated with increasing malignancy of astrocytic tumors

but is absent in oligodendrogliomas. Am J Pathol 2000;157:1785–

93.

[19] Moilanen J, Lassus H, Leminen A, et al. Ezrin immunoreactivity in

relation to survival in serous ovarian carcinoma patients. Gynecol

Oncol 2003;90:273–81.

[20] Kobel M, Langhammer T, Huttelmaier S, et al. Ezrin expression is

related to poor prognosis in FIGO stage I endometrioid carcinomas.

Mod Pathol 2006;19:581–7.

[21] Pang ST, Fang X, Valdman A, et al. Expression of ezrin in prostatic

intraepithelial neoplasia. Urology 2004;63:609–12.

[22] Valdman A, Fang X, Pang S-T, et al. Ezrin expression in prostate

cancer and benign prostatic tissue. Eur Urol 2005;48:852–7.

E U R O P E A N U R O L O G Y 5 6 ( 2 0 0 9 ) 8 2 9 – 8 3 6836

[23] Elliott BE, Meens JA, Sengupta SK, et al. The membrane cytoskeletal

crosslinker ezrin is required for metastasis of breast carcinoma

cells. Breast Cancer Res 2005;7:365–73.

[24] Weng W-H, Ahlen J, Astrom K, et al. Prognostic impact of immu-

nohistochemical expression of ezrin in highly malignant soft tissue

sarcomas. Clin Cancer Res 2005;11:6198–204.

[25] Ilmonen S, Vaheri A, Asko-Seljavaara S, et al. Ezrin in primary

cutaneous melanoma. Mod Pathol 2005;18:503–10.

[26] Younes M, Sussman J, True LD. The usefulness of the level of the

muscularis mucosae in the staging of invasive transitional cell

carcinoma of the urinary bladder. Cancer 1990;66:543–8.

[27] Kononen J, Bubendorf L, Kallioniemi A, et al. Tissue microarrays for

high-throughput molecular profiling of tumor specimens. Nat Med

1998;4:844–7.

Editorial Comment on: Protein Expression Patterns of

Ezrin Are Predictors of Progression in T1G3 Bladder

Tumours Treated with Nonmaintenance Bacillus

Calmette-Guerin

Ofer N. Gofrit

Department of Urology, Hadassah Hebrew

University Hospital, Jerusalem, Israel

ogofrit@gmail.com

Palou and colleagues’ [1] deal with one of the most

difficult issues in urologic oncology: the management of

patients with T1G3 urothelial carcinoma. Despite high risk

of recurrence, muscle invasion, metastases, and disease-

related death, many of these patients do well with

endoscopic resection and bacillus Calmette-Guerin

(BCG) immunotherapy alone. Can molecular markers

predict the biologic potential of this unpredictable tumor?

Can they improve our understanding of the biology of

bladder cancer?

The authors studied the expression of ezrin, a cytoplas-

mic membrane protein that acts as a link between the

membrane and the actin cytoskeleton, in tissue blocks from

92 patients given BCG immunotherapy for T1G3 tumors [1].

As expected, after a median follow-up of 90.5 mo, 17

patients (18.5%) progressed to muscle-invasive disease and

14 (15.2%) died of the disease. Multivariate analysis showed

that ezrin membrane expression in<20% of the tumor cells

is significantly associated with progression and survival.

So, can ezrin expression be used as a prognostic

marker? Can it teach us anything about the pathogenesis

of T1G3 bladder cancer? Unfortunately, the answer to both

questions is no. A valuable tumor marker must prove its

predictive capacity beyond the standard clinical and

pathologic parameters in a prospective way, not only on

archival material. Perhaps the most disappointing mole-

cule in this respect is p53. Initial results were promising:

Disease progression of 20.5% per year was found in

patients with T1 disease expressing nuclear p53 in>20% of

the tumor cells, compared with 2.5% in patients with lower

[28] Sanchez-Carbayo M, Socci ND, Charytonowicz E, et al. Molecular

profiling of bladder cancer using cDNA microarrays: defining

histogenesis and biological phenotypes. Cancer Res 2002;62:

6973–80.

[29] Dawson-Saunders B, Trapp RG. Basic and clinical biostatistics. 2nd

ed. Norwalk, Connecticut: Appleton & Lange; 1994.

[30] Gautreau A, Poullet P, Louvard D, et al. Ezrin, a plasma membrane-

microfilament linker, signals cell survival through the phosphati-

dylinositol 3-kinase/Akt pathway. Proc Natl Acad Sci USA 1999;

96:7300–5.

[31] Schmidt-Drager BJ, Goebell PJ, Ebert T, Fradet Y. p53

immunohistochemistry as a prognostic marker in bladder

cancer: playground for urology scientists? Eur Urol 2000;38:

691–700.

levels of p53 [2]. Further studies, however, including a

meta-analysis, could not confirm this finding [3–5]. After

15 yr of intense research, neither p53 nor any other

molecular marker has joined our diagnostic armamentar-

ium [6] and none is recommended in the European

Association of Urology (EAU) guidelines. Immunohisto-

chemistry studies are also very limited in their ability to

provide valuable information regarding the biologic role of

the molecule studied. Much more complicated studies at

the molecular level are needed for this purpose.

Nevertheless, ezrin is an interesting molecule that is

certainly involved in the pathogenesis of bladder cancer

and is definitely worth more investigation.

References

[1] Palou J, Algaba F, Vera I, Rodriguez O, Villavicencio H, Sanchez-

Carbayo M. Protein expression patterns of ezrin are predictors of

progression in T1G3 bladder tumours treated with nonmainte-

nance bacillus Calmette-Guerin. Eur Urol 2009;56:829–36.

[2] Sarkis AS, Dalbagni G, Cordon-Cardo C, et al. Marker for disease

progression. J Natl Cancer Inst 1993;85:53–9.

[3] Zlotta AR, Noel JC, Fayt I, et al. Correlation and prognostic sig-

nificance of p53, p21WAF1/CIP1 and Ki-67 expression in patients

with superficial bladder tumors treated with bacillus Calmette-

Guerin intravesical therapy. J Urol 1999;161:792–8.

[4] Shariat SF, Weizer AZ, Green A, et al. Prognostic value of P53 nuclear

accumulation and histopathologic features in T1 transitional cell

carcinoma of the urinary bladder. Urology 2000;56:735–40.

[5] Schmitz-Drager BJ, Goebell PJ, Ebert T, Fradet Y. p53 immunohis-

tochemistry as a prognostic marker in bladder cancer: playground

for urology scientists? Eur Urol 2000;38:691–700.

[6] Habuchi T, Marberger M, Droller MJ, et al. Prognostic markers for

bladder cancer: International Consensus Panel on bladder tumor

markers. Urology 2005;66(Suppl 1):64–74.

DOI: 10.1016/j.eururo.2008.09.063

DOI of original article: 10.1016/j.eururo.2008.09.062