Bladder Cancer

Molecular and Immunohistochemical Analysis of thePrognostic Value of Cell-Cycle Regulators in UrothelialNeoplasms of the Bladder

Andriy O. Yurakh a,*, David Ramos b, Silvia Calabuig-Farinas b,Jose Antonio Lopez-Guerrero c, Jose Rubio d, Eduardo Solsona d,Alina M. Romanenko a, Alexander F. Vozianov a, Antonio Pellin b,Antonio Llombart-Bosch b

a Institute of Urology, Academy of Medical Science of Ukraine, Kyiv, UkrainebDepartment of Pathology, School of Medicine, University of Valencia, Valencia, SpaincUnit of Molecular Biology, Instituto Valenciano de Oncologıa, Valencia, SpaindDepartment of Urology, Instituto Valenciano de Oncologıa, Valencia, Spain

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5

avai lable at www.sc iencedi rect .com

journal homepage: www.europeanurology.com

Article info

Article history:Accepted March 13, 2006Published online ahead ofprint on March 31, 2006

Keywords:Bladder cancerCyclin D1Disease progressionp14ARFp15INK4Bp16INK4Ap21Waf1p27kip1p53MDM2RecurrenceRetinoblastoma Protein

Abstract

Objective: To evaluate the prognostic and predictive value of molecular andimmunohistochemical markers related to cell-cycle control in terms of recur-rence, progression, and survival in urothelial neoplasms of the bladder (UNB).Patients and Methods: Clinical and pathological findings of 84 patients with UNBwere assessed. Homozygous deletion (HD) and promoter methylation of p14ARF,p15INK4B, p16INK4A, loss of heterozygosity of the locus 9p21, p53 mutations, andimmunohistochemical expression of p53, p16, p14, p21, p27, pRb, Ki67, MDM2,and cyclin D1 proteins were evaluated in relation to overall survival (OS),recurrence-free survival (RFS), and progression-free survival (PFS).Results: In the univariate analysis, RFS was shorter in cases with p14ARF

( p = 0.006), p15INK4B (p = 0.003), p16INK4A (p = 0.03) HD, low p14 immunoreactivityindex (IRI) ( p = 0.01) and high Ki67 IRI ( p = 0.04); HD of the 9p21 locus genes andp14 IRI remained as independent prognostic factors for early UNB recurrence(p = 0.006) whereas tumour stage (p = 0.00001) and cyclin D1 IRI ( p = 0.049) wererelated to worse PFS in the multivariate analysis. In the univariate analysis, IRI forKi67 (p = 0.002), cyclin D1 ( p = 0.06), p53 (p = 0.00008), p16 (p = 0.02), p27(p = 0.0005) MDM2 (p = 0.01) and p53 mutations (p = 0.03) were related to poorOS, and only the Ki67 IRI retained their independent value in the multivariateanalysis.Conclusion: : 9p21 HD and p14 IRI constitute independent predictive factors forUNB recurrence and cyclin D1 IRI and tumour stage for progression. In addition,Ki67 IRI and tumour stage are independent prognostic factors for overallsurvival in UNB.# 2006 European Association of Urology. Published by Elsevier B.V. All rights reserved.

* Corresponding author. Institute of Urology 9a, Yu. Kotsubinsky Str. 04053, Kiev, Ukraine.Tel. +380 44 4866731; Fax: +380 44 2551565.E-mail address: yurakh@list.ru (A.O. Yurakh).

0302-2838/$ – see back matter # 2006 European Association of Urology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.eururo.2006.03.027

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5 507

Table 1 – Antibodies used in the study

Antigen Clone Source Retrieval Dilution

p53 DO7 DAKO yes 1/50

Ki67 MIB1 DAKO yes 1/50

p14ARF AB-4 Neomarkers yes 1/100

p16INK4A F-12 Santa Cruz yes 1/500

p21WAF1 F-5 Santa Cruz yes 1/50

p27KIP1 SX53G8 Novocastra yes 1/50

pRb 1F8 Neomarkers yes 1/25

MDM2 1B-10 Novocastra yes 1/50

Cyclin D1 P2D11F11 Novocastra yes 1/20

1. Introduction

Molecular alterations of genes implicated in cellcycles are important molecular events in carcino-genesis of urothelial neoplasm of the bladder (UNB).The prognostic significance of alterations of the 9p21chromosomal region, where p15INK4B, p16INK4A, andp14ARF genes are located, as well as alteration ofp21Cip, p27Kip1, p53, MDM2, and CCND1 genes andproteins have been intensively studied during thelast decade. Numerous studies show them as aprognostic markers of UNB clinical outcomes, butvery often results of studies are dissimilar and thecomparative prognostic value of these proteins andgenes are not clear.

Alterations of 9p21 locus, especially p16INK4A/ARF

gene, are very promising prognosticators of clinicaloutcomes because they affect both the p53 and pRbcell-cycle regulatory pathways [1]. The significanceof this locus for tumourigenesis was shown withanimal models of cancer [2].

The aim of the present study was to evaluate thesignificance of alterations in 10 cell-cycle regulatorsto predict recurrence, progression, and overallsurvival (OS) in a series of patients with UNB.

2. Materials and methods

We analyzed material from 84 patients with UNB who had

been treated in the Fundacion Instituto Valenciano de

Oncologıa (I.V.O.), Valencia, Spain. The classification of

tumour stage and grade was defined according to Interna-

tional Society of Urological Pathology and World Health

Organization criteria [3]. No special patient selection was

performed, but many Ta tumours were too small to obtain

adequate samples and were not included in the study. For the

molecular study, fresh-frozen material from the tumour and

from bladder urothelium of normal appearance were used.

The frozen sections were studied with hematoxylin and eosin.

The immunohistochemical (IHC) study was carried out on

formalin-fixed, paraffin-embedded tumoural tissues from the

same samples.

Tumour progression was defined as: (a) progression from

superficial UNB when it becomes a muscle-infiltrating tumour

during follow-up; (b) progression from muscle-infiltration

UNB when after treatment (cystectomy) the disease has

spread to lymph nodes, metastasis, or has had local

recurrence.

2.1. Immunohistochemical study

An IHC study for p53, MDM2, pRb, Cyclin D1, Ki67, p14ARF,

p16INK4A, p21WAF1/Cip1, and p27Kip1 proteins was performed

with the standard avidin-biotin peroxidase complex (ABC)

method with the LSAB kit (DAKO, Denmark). The antibodies

used are listed in Table 1. As controls, known positive tissue

sections were used, and for negative controls, exposure to the

primary antibody was omitted. Colour was developed with

30,30-aminobenzidine and counter-stained with Mayer’s

hematoxylin.

To evaluate the extent of immunostaining we used the

immunoreactivity index (IRI), which was calculated as the

percentage of stained cells by counting 1000 cells per slide.

2.2. DNA extraction

DNA was extracted from 0.10–0.15 g of fresh frozen tissue.

Lysis solution that contained proteinase K (Life Technologies)

in 100 mg/ml concentration was used with subsequent

incubation at 37 8C for 16 h. Standard phenol/chloroform

protocol, followed by precipitation with ethanol, was used for

DNA purification [4]. DNA was quantified by spectrophoto-

metry at 260 nm and quality was checked by measuring the

ratio between the 260 nm and 280 nm values.

2.3. Homozygous deletions of p15INK4B, p16INK4A andp14ARF genes

Differential polymerase chain reaction (dPCR) was used to

detect homozygous deletions (HD) of 9p21 locus genes. PCR

parameters and primer sequences were designed as pre-

viously described by our group [5,6]. The PCR products were

visualized and photographed in ultraviolet light after electro-

phoresis in 1.5% agarose (Fig. 1). The gel image was then

analyzed by 1D Image Analysis Software (Eastman Kodak). The

intensities of the target band and the control band (b-globin)

were compared. The presence of HD was determined by

decreasing the ratio of target band intensity to b-globin band

intensity compared with the normal control samples analysed

in the same gel.

2.4. Methylation-specific PCR

The methylation status of the 50CpG islands in genes promoter

regions was determined by following the procedure described

by Herman et al. [7]. The primer used and the PCR parameters

have been described by our group in previous studies [5,6]. PCR

products were analyzed in 2% ethidium-bromide stained gel.

We confirmed the presence of hypermethylation by identify-

ing the PCR bands in the samples amplified with the

methylated primers. The presence of bands in cases that

were amplified with unmethylated primer was used to

confirm the quality of the modified DNA.

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5508

Fig. 1 – dPCR analysis of HDs of genes p15INK4B (exon 1),

p16INK4A (exon 1a and 2), and p14ARF (exon 1b). Cases #1, #2,

and #6 have all investigated exons deleted, which was

common for most cases with deletions. NC – negative

control.

2.5. p53 Mutation analysis

We used the non-radioisotopic method of single-strand

conformation polymorphism (SSCP) described by Ainsworth

et al. [8] to screen mutations in exons 5–8 of p53 with primers

previously described [6].

All SSCP positive cases underwent sequencing analysis to

confirm the presence of a mutation. PCR products were

purified with Centricon columns (Amicon) and sequenced

with the Big Dye terminator kit v1.1 (Applied Biosystem)

according to standard protocol on an automatic sequencer ABI

Prism 310 (Applied Biosystem). Sequences obtained were

compared with a consensus p53 sequence (GeneBank acces-

sion number: AF307851).

2.6. Loss of heterozygosity

For the lossofheterozygosity (LOH) analysis of the9p21region, a

small tandem repeat testing of DNA from tumoural and

corresponding normal tissue was used by employing fluores-

cent-labelled primers for the following microsatellite markers:

D9S162 (forward: 50AGCAAGGCA AGCCACATTTC 30; reverse: 50

TGGGGGATGCCCAGATAACTATATC 30), D9S942 (forward: 50

GCAAGATTCCAAACAGTA 30; reverse: 50 CTCATCCTGCG-

GAAACCA TT 30), D9S157 (forward: 50 AGCAAGGCAAGCCA-

CATTTC 30; reverse: 50 TGGGGA TGCCCAGATAACTATATC 30).

After DNA amplification, the PCR products were electrophor-

esed in an ABI Prism 310 sequencer with a subsequent

quantitative evaluation of the allele ratio with Gene Scan and

Geno Typer software (Applied Biosystem). Loss of one allele in

the tumour sample was estimated as LOH, and loss of both

alleles, when present in the normal sample, as HD. Cases that

presented an additional peak (allele) in a tumour sample were

considered as having microsatellite instability.

2.7. Statistical analysis

The STATISTICA version 6.0 (Stat Soft Inc.) statistical software

package was used for statistical analysis. We applied the

Pearson Chi-square test for statistical analysis between

qualitative variables. The nonparametric ANOVA Mann-

Whitney U Test or, depending on the number of groups, the

Kruskal-Wallis test, was used between qualitative and

quantitative variables.

All pathological, molecular, and IHC data were analyzed in

relation to tumour recurrence and progression as end point

variables, and to findings of the three-month control cysto-

scopy, as we had previously found this to be an important

prognostic factor for recurrence and progression [9]. Survival

analysis, including OS, progression-free survival (PFS), and

recurrence-free survival (RFS) was performed in relation to

clinical, pathological, molecular, and IHC data. The Kaplan-

Meier proportional risk model with the log-rank test for

comparing survival between groups was used. Optimal cut-

points of immunostaining extent, which predict tumour

recurrence or progression with the smallest quantity of

false-positive and false-negative results (Table 2), were

calculated with ROC-analysis. Multivariate Cox proportional

regression analysis was performed to reveal independent

survival predictors. All markers that showed a relationship

with survival in the univariate analysis were included in the

regression model. Models were then adjusted by using forward

and backward stepwise methods.

3. Results

Clinical and pathological data of the patients arepresented in Table 3. Positive results at the three-month cystoscopy in 21.8% (12 of 55 patients) andtumour recurrence in 29.8% (25 of 63 patients) wereobserved. Disease progression occurred only in19.1% (16 of 84 patients) patients because of highfrequency of superficial tumours in the series.Fourteen of 84 patients (16.7%) died of tumourprogression. Median follow-up was 36.4 months(range: 1.3–49.2).

Some pathological parameters show relationshipwith RFS, PFS, and OS (Table 4). Patterns of IHCstaining of UNB are illustrated in Fig. 2. Afterunivariate analysis, we found no relationshipbetween tumour recurrence and IRI of investigatedproteins, excluding Ki67 and p14 expression. Asignificant relationship between all IRI of investi-gated proteins and both PFS and OS was founded,except p14ARF and pRb for PFS and p14ARF for OS(Table 2 and Fig. 3).

Deletion of p15INK4B or p16INK4A/ARF (exons 1a, 1b,and 2) genes was present in 23.8% (20 of 84) tumours.Furthermore, 85% (17 of 20) cases with HD had allinvestigated exons deleted (Fig. 1). Recurrence wasdiagnosed in 64.7% (11 of 17) tumours with HD of atleast one of the investigated genes. Patients with HD

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5 509

Table 3 – Baseline clinical and pathological data of thestudy group

Gender Male 72 (85.7%)

Female 12 (14.3%)

Tumour type Single 46 (54.8%)

Multicentric 38 (44.2%)

Papillary 63 (75.0%)

Sessile 21 (25.0%)

Primary 57 (67.9%)

Recurrent 27 (32.1%)

Stage pTa 8 (9.5%)

pT1 55 (65.5%)

pT2-4 21 (25.0%)

Grade G1 15 (17.6%)

G2 43 (51.2%)

G3 26 (31.0%)

Tumour size <1 cm 5 (6.0%)

1–3 cm 41 (48.8%)

>3 cm 38 (45.2%)

CIS simultaneous yes 8 (9.5%)

no 76 (90.5%)

Ta

ble

2–

Un

iva

ria

tea

na

lysi

so

fth

ece

ll-c

ycl

ep

rote

ins

imm

un

ore

act

ivit

yin

dex

es

wit

hR

FS

,P

FS

,a

nd

OS

Pro

tein

RFS

PFS

OS

Cu

t-p

oin

t3-y

ea

rsu

rviv

ala

log-

ran

kte

stp

Cu

t-p

oin

t3-y

ea

rsu

rviv

ala

log-

ran

kte

stp

Cu

t-p

oin

t3-y

ea

rsu

rviv

ala

log-

ran

kte

stp

p53

25%

47.1

/67.1

1.4

70.1

460%

82.9

/40.1

�2.8

50

.00

460%

89.8

/46.0

�3.9

60

.00

00

8

Ki6

710%

75.2

/51.1

�2.0

10

.04

435%

94.3

/61.1

�2.9

40

.00

335%

97.2

/71.2

�3.1

50

.00

16

p14

AR

F30%

20.0

/65.7

2.5

70

.01

50%

81.5

/70.5

�1.1

00.2

750%

84.5

/92.3

0.1

50.8

7

p16

INK

4A

10%

75.0

/53.6

1.0

10.3

110%

56.8

/85.1

2.5

10

.01

210%

68.8

/86.5

2.3

00

.02

2

p21

Wa

f150%

51.1

/73.4

0.7

80.4

310%

64.4

/82.3

1.9

30

.05

10%

59.6

/93.9

2.9

50

.00

32

p27

Kip

150%

49.2

/59.7

0.7

40.4

620%

24.6

/84.6

3.7

40

.00

02

20%

35.8

/90.3

3.4

50

.00

05

5

pR

b5%

75.0

/53.6

�1.0

90.2

720%

67.9

/80.0

1.4

10.1

510%

68.5

/88.5

2.0

10

.04

4

MD

M2

25%

58.7

/55.5

�0.4

40.6

515%

63.5

/82.5

1.9

60

.05

15%

62.7

/91.8

2.5

40

.01

1

Cy

clin

D1

8%

76.2

/52.9

�1.5

40.1

210%

62.4

/88.3

2.7

90

.00

53

10%

71.5

/92.4

2.7

80

.00

55

Nega

tiv

elo

g-r

an

kv

alu

es

resu

ltfr

om

inv

ers

ere

lati

on

ship

betw

een

IRI

an

dsu

rviv

al.

aP

rop

ort

ion

surv

ivin

gin

gro

up

wit

hIR

Ilo

wer/

hig

her

tha

ncu

t-p

oin

t.

in the 9p21 locus showed a significantly higher rateof positive results at the three-month controlcystoscopy: 46.7% (7 of 15) vs. 12.5% (5 of 40)( p = 0.006). Risk of disease recurrence was signifi-cantly higher in patients with 9p21 HD compared topatients without HD of these genes (p = 0.01) (Fig. 3).Univariate analysis showed no significant relation-ship between HD and OS or PFS (Table 4).

No relationship between LOH in locus 9p21 andOS, PFS and RFS was found.

Promoter hypermethylation of the 50CpG islandswas observed in six (7.1%) cases for p15INK4B, 22(26.1%) for p14ARF, and 5 (6%) for p16INK4A. Simulta-neous hypermethylation of p15INK4B and p14ARF wasobserved in one (1.2%) case, whereas p16INK4A andp14ARF were simultaneously hypermethylated inthree cases (3.6%). No differences in OS, PFS, andRFS between groups and p14ARF promoter methyla-tion were found. The same was observed for p16INK4A

and p15INK4B promoter methylation (Table 4).Progression was found in 44.4% (4 of 9 patients)

with p53 mutation, higher than in the group with awild-type p53—16.0% (12 of 75 patients) ( p = 0.04).Patients with mutated p53 showed a worse three-year OS and PFS when compared to patients withoutp53 mutation ( p = 0.03 and p = 0.05 respectively); norelationship was found regarding RFS ( p = 0.81).

Cox proportional regression analysis showedloss of p14 expression and HD in 9p21 locus asindependent prognostic factors for a shorterrecurrence-free interval (Table 5). Both tumourstage and cyclin D1 expression were independentmarkers for PFS (Table 6) and tumour stage and

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5510

Ta

ble

4–

Un

iva

ria

tea

na

lysi

so

fth

ep

ath

olo

gic

al

da

taa

nd

mo

lecu

lar

alt

era

tio

nw

ith

RFS

,P

FS

,a

nd

OS

Pa

tien

tsgro

up

RFS

PFS

OS

3y

ea

rsu

rviv

al

log-r

an

kte

stp

3y

ear

surv

iva

llo

g-r

an

kte

stp

3y

ea

rsu

rviv

al

log-r

an

kte

stp

Mo

no

foca

lv

s.m

ult

ifo

cal

tum

ou

r69.8

/50.9

�1.5

60.1

268.2

/90.7

2.3

80

.02

76.2

/94.3

1.8

90.0

57

Tu

mo

ur

size

<3

cmv

s.>

3cm

65.2

/43.6

�1.7

00.0

997.8

/55.6

�4.3

00

.00

00

5100/6

6.0

�4.4

80

.00

00

5

Pri

ma

ryv

s.re

curr

en

ttu

mo

ur

65.8

/37.3

�2.0

00.0

478.3

/79.3

0.0

90.9

286.4

/80.4

�1.0

30.3

0

Gra

de

G1–2

vs.

G3

55.5

/83.3

1.2

50.2

194.3

/43.0

�4.9

40

.00

00

596.1

/58.3

�4.8

30

.00

00

5

Sta

ge

Ta

-1v

s.T

2–4

––

–92.7

/28.9

�5.4

20

.00

00

596.5

/44.3

�5.2

40

.00

00

5

Sta

ge

Ta

vs.

T1

62.5

/57.7

�0.1

40.8

8100/9

1.6

�1.7

90.4

3100/9

6.0

�0.6

40.5

2

Wit

ho

ut

HD

vs.

HD

p16IN

K4A

ex

on

s1

a,

267.5

/34.3

�2.4

60

.01

476.1

/88.7

1.0

20.3

183.1

/89.3

0.7

30.4

6

Wit

ho

ut

HD

vs.

HD

p15IN

K4B

ex

on

166.7

/32.7

�2.3

60

.01

876.8

/87.2

0.8

00.4

383.6

/88.1

0.5

10.6

1

Wit

ho

ut

HD

vs.

HD

p14ARF

ex

on

1b

67.5

/34.3

�2.4

60

.01

477.3

/84.1

0.4

20.6

784.4

/84.9

0.0

90.9

2

Wit

ho

ut

LO

Hv

s.LO

H9p

21

60.6

/54.4

�0.7

40.4

674.8

/87.0

1.3

60.1

780.5

/92.5

1.0

90.2

7

p16

pro

mo

ter

un

meth

yla

ted

vs.

meth

yla

ted

57.1

/75.0

0.5

30.6

078.6

/80.0

0.0

40.9

784.7

/80.0

�0.1

90.8

4

p15

pro

mo

ter

un

meth

yla

ted

vs.

meth

yla

ted

59.6

/33.3

�1.0

10.3

179.8

/66.0

�1.1

50.2

586.1

/62.5

�1.3

10.1

9

p14

pro

mo

ter

un

meth

yla

ted

vs.

meth

yla

ted

54.4

/66.7

0.5

90.5

580.2

/78.4

�0.3

90.6

990.4

/82.4

�1.2

70.2

0

P53

wil

d-t

yp

ev

s.m

uta

ted

58.2

/60.0

0.2

50.8

182.3

/50.7

�1.9

50

.05

87.3

/62.5

�2.1

80

.02

9

Nega

tiv

elo

g-r

an

kv

alu

es

resu

ltfr

om

inv

ers

ere

lati

on

ship

betw

een

ma

rker

an

dsu

rviv

al.

Ki67 were independent predictors for OS (Table 7)(Fig. 3).

4. Discussion

p53 immunostaining is the most studied predictor ofpoor prognosis in bladder cancer. Esrig et al. [10]asserted that p53 accumulation in the tumour-cellnuclei detected by IHC methods is an independentpredictor of increased risk of recurrence and deathin patients with UNB. Some studies that usedunivariate analysis support this opinion [11–13],but others, including the present study, reject thepredictive value of p53 immunostaining for recur-rence [14,15].

p53 has been also described as a prognosticmarker for UNB progression [13,16]. Nevertheless,some studies have shown a lack of associationbetween p53 expression and progression [17,18]. Inthe present study, IRI higher than 60% for p53 wasrelated to a shorter PFS in the univariate analysis,but was not an independent prognostic factor in themultivariate analysis. The cut-point for p53 IRI ishigher in our study than those used in otherpublications, but was optimal for predicting tumourprogression according to the ROC analysis.

Our results agree with studies [10,19–21] that havedescribed the relationship of OS to p53 expression inthe univariate analysis. However, contrary to otherstudies that demonstrated the independent prog-nostic value of p53 for OS [10,21,22], we have notbeen able to show this in our series of cases.

The high prognostic value of Ki67 labelling indexin terms of recurrence has been described inunivariate [13] and multivariate [14,15,23] studies.We found a significantly shorter RFS in cases with aKi67 IRI higher than 10%, but multivariate analysisdid not confirm the independent prognostic value ofKi67 for RFS. Ki67 expression was related withtumour stage (r = 0.45 p = 0.00005) and grade (r = 0.55p < 0.00005). This relationship and the low varia-bility of stage and grade in the studied series canexplain the borderline p-value of Ki67 in univariateanalysis. Furthermore, previous studies did notcompare Ki67 expression with HD of p16INK4A andp15INK4B genes in multivariate analysis.

Ki67 expression as a marker of tumour progres-sion has not been extensively studied and contra-dictory data have been described [13,22,24,25]. In ourstudy, higher Ki67 expression was related toprogression and shorter PFS. As other authors[22,26,27], we found a high prognostic value forKi67 in multivariate analysis, which was alsodemonstrated in terms of OS.

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5 511

Fig. 2 – IHC detection of pRb (A), p27Kip1 (B), p53 (C), p14ARF (D), MDM2 (E), Ki67 (F) proteins in UNB. (magnification: A, B, �90; C,

F, �120; D, E, �230). In cases C, E, and F progression was diagnosed during follow-up time.

Fig. 3 – RFS, OS, and PFS in relation to HD in 9p21 locus, level of Ki67, p21, and cyclin D1 immunoreactivity.

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5512

Table 7 – Cox multivariate analysis of OS

Wald OR 95%CI of OR p

Ki67 IRI 3.71 1.03 1.00–1.07 0.054

CD1 IRI 2.10 0.96 0.92–1.01 0.15

Stage 7.09 6.83 1.66–28.07 0.008

Table 5 – Cox multivariate analysis of RFS

Wald OR 95%CI of OR p

9p21 HD 10.85 4.38 1.82–10.56 0.001

p14ARF IRI 7.33 0.97 0.95–0.99 0.007

Table 6 – Cox multivariate analysis of PFR

Wald OR 95%CI of OR p

p21Waf1 IRI 2.89 1.02 1.00–1.05 0.089

CD1 IRI 3.88 0.97 0.93–1.00 0.049

Stage 19.57 17.59 4.94–62.65 0.00001

The remainder of the IHC markers showed norelationship to tumour recurrence, even thoughp21WAF1/Cip1 – [28,29], p27Kip1 – [20,30], MDM2 – [11],pRb [29], and cyclin D1-expression [27] have allbeen described as prognostic markers for recu-rrence in UNB. Our opinion supports data fromIoachim et al. [31], which rejected p21WAF1/Cip1,p27Kip1, and cyclin D1 expressions as predictors ofrecurrence.

Numerous studies have addressed the role ofp21WAF1/Cip1 and p27Kip1 expression as prognosticmarkers for UNB progression and OS, but mostprovided no additional prognostic information com-pared with already established prognostic factors[27,31,32]. In our study, univariate analysis showedshorter PFS and OS in groups with p21WAF1/Cip1

IRI lower than 10% and p27Kip1 IRI lower than20%, but multivariate analysis failed to confirmp21WAF1/Cip1 and p27Kip1 as independent predictorsfor PFS and OS, in contrast to findings from otherstudies [28,29,33].

A relationship between cyclin D1 and UNBprogression has been described in some studies[20,27], but only Lopez-Beltran et al. [20] haveevaluated cyclin D1 as an independent progressionmarker for superficial UNB. In our series, cyclin D1was significantly related to progression and OS inthe univariate analysis. However, multivariate ana-lysis showed that cyclin D1 plays an important roleonly as an independent predictor for tumourprogression. Our data support those from Liukkonenet al. [27] and Takagi et al. [34], but are contrary tothose described by Tut et al. [35].

Results of our study agree with recent investiga-tions [36,37] that described expression of p16INK4A

protein as an independent predictor for UNBprogression, but not for RFS. We observed signifi-cantly lower RFS in the group with p14ARF IRI than30% and multivariate analysis confirmed the pre-dictive value of this protein expression. The prog-nostic value of p14ARF has not been studied inbladder cancer previously, but has been described asa marker of clinical outcomes for other humancancers [38].

MDM2 expression showed a relationship with PFSand OS, but was not an independent predictor, inagreement with other authors [24,39].

Our study did not support the findings observedby others [29,40] about the relationship between pRbexpression and UNB progression and OS.

Orlow et al. [41] were the first to describe thatp16INK4A/ARF exon 1a and exon 1b homozygousdeletion constitute a predictive factor for superficialbladder cancer recurrence. They found a correlationbetween HD and shorter RFS, but not with PFS andOS. In our study, we found a significant relationshipbetween p16INK4A/ARF exons 1a, 1b, 2 or p15INK4B exon 1HD and tumour recurrence as well as with a positiveresult at the three-month control cystoscopy. How-ever, no relationship of 9p21 HD with tumourprogression and OS was demonstrated. Interest-ingly, multivariate analysis showed that HD in 9p21locus could be used as a prognostic marker for earlytumour recurrence. The same result was describedby Kruger et al. [42].

Tsukamoto et al. [43] reported that monosomy ofchromosome 9 might be a prognostic marker forearly tumour recurrence in patients with negative orequivocal cytology specimens. Some publicationsshowed no relationship of tumour recurrence with9p21 LOH [36,44], but a significant relationship with9q34 LOH [44]. In our study, we confirmed theabsence of a relationship between 9p21 LOH andtumour recurrence, and found no relationship withOS or tumour progression.

We found no relationship between tumourrecurrence, progression, or OS and 50CpG islandshypermethylation in p14ARF, p15INK4B, or p16INK4A ofthe promoter regions of these genes. Our analysisconfirms earlier studies [41,45]. However, Domin-guez et al. [46] showed that plasma p14ARF promoterhypermethylation was associated with relapse ofthe disease.

Our work agreed with Goudopoulou et al. [47] inthat we did not find a relationship between p53mutation and p53 protein overexpression, althoughopposite results have been extensively published[12,48]. No relationship between p53 mutation and

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5 513

tumour recurrence was found that confirmed pre-vious studies [49]. However, it was directly relatedwith tumour progression, shorter PFS, and OS in theunivariate analysis, but were not independentprognostic factors in the multivariate analysis.According to published data, there is no consensusabout the prognostic importance of p53 mutation forUNB progression and OS [49,50].

5. Conclusions

After our extensive molecular and IHC study of somecritical cell cycle regulators in relationship to clinicaldata we found that:

� L

oss of p14ARF expression and HD in 9p21 locus areindependent prognostic factors for early UNBrecurrence.

� T

umour stage and CD1 IRI are independentpredictors for UNB progression. Ki67 expressionis strongly related to progression, but does notreveal additional information in comparison withtumour stage.

� K

i67 IRI is a very good prognosticator for OS, butdoes not reveal additional information comparedwith tumour stage.

Acknowledgements

This work was supported by the European School ofUrology Program Grant from the European Associa-tion of Urology (2003) and by Fundacion para laInvestigacion en Urologia from the Spanish Urologi-cal Association. The technical assistance of EstelaPons and Laura Martinez is gratefully acknowledged.

References

[1] Chin L, Pomerantz J, DePinho RA. The INK4a/ARF tumor

suppressor: one gene—two products—two pathways.

Trends Biochem Sci 1998;23:291–6.

[2] Serrano M. The INK4a/ARF locus in murine tumorigen-

esis. Carcinogenesis 2000;21:865–9.

[3] Epstein JI, Amin MB, Reuter VR, Mostofi FK. The World

Health Organization/International Society of Urological

Pathology consensus classification of urothelial (transi-

tional cell) neoplasms of the urinary bladder. Bladder

Consensus Conference Committee. Am J Surg Pathol

1998;22:1435–48.

[4] Sambrook J, Fritch EF, Maniatis T. Molecular cloning: a

laboratory manual. Cool Spring Harbor, NY: Cool Spring

Harbor Laboratory Press; 1989, pp. 1659.

[5] Romanenko A, Morell-Quadreny L, Lopez-Guerrero JA,

et al. The INK4a/ARF locus: role in cell cycle control for

renal cell epithelial tumor growth after the Chernobyl

accident. Virchows Arch 2004;445:298–304.

[6] Lopez-Guerrero JA, Pellin A, Noguera R, Carda C, Llom-

bart-Bosch A. Molecular analysis of the 9p21 locus and

p53 genes in Ewing family tumors. Lab Invest 2001;81:803–

14.

[7] Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB.

Methylation-specific PCR: a novel PCR assay for methyla-

tion status of CpG islands. Proc Natl Acad Sci USA

1996;93:9821–6.

[8] Ainsworth PJ, Surh LC, Coulter-Mackie MB. Diagnostic

single strand conformational polymorphism, (SSCP): a

simplified non-radioisotopic method as applied to a

Tay-Sachs B1 variant. Nucleic Acids Res 1991;19:405–6.

[9] Solsona E, Iborra I, Dumont R, Rubio-Briones J, Casanova J,

Almenar S. The 3-month clinical response to intravesical

therapy as a predictive factor for progression in patients

with high risk superficial bladder cancer. J Urol 2000;

164:685–9.

[10] Esrig D, Elmajian D, Groshen S, et al. Accumulation of

nuclear p53 and tumor progression in bladder cancer. N

Engl J Med 1994;331:1259–64.

[11] Tuna B, Yorukoglu K, Tuzel E, Guray M, Mungan U, Kirkali

Z. Expression of p53 and mdm2 and their significance in

recurrence of superficial bladder cancer. Pathol Res Pract

2003;199:323–8.

[12] Friedrich MG, Riethdorf S, Erbersdobler A, et al. Relevance

of p53 gene alterations for tumor recurrence in patients

with superficial transitional cell carcinoma of the bladder.

Eur Urol 2001;39:159–66.

[13] Kilicli-Camur N, Kilicaslan I, Gulluoglu MG, Esen T, Uysal

V. Impact of p53 and Ki-67 in predicting recurrence

and progression of superficial (pTa and pT1) urothelial

cell carcinomas of urinary bladder. Pathol Int 2002;52:

463–9.

[14] Gontero P, Casetta G, Zitella A, et al. Evaluation of P53

protein overexpression, Ki67 proliferative activity and

mitotic index as markers of tumour recurrence in super-

ficial transitional cell carcinoma of the bladder. Eur Urol

2000;38:287–96.

[15] Rodriguez Alonso A, Pita Fernandez S, Gonzalez-Carrero J,

Nogueira March JL. Multivariate analysis of recurrence

and progression in stage T1 transitional-cell carcinoma of

the bladder. Prognostic value of p53 and Ki67. Actas Urol

Esp 2003;27:132–41.

[16] Sarkis AS, Dalbagni G, Cordon-Cardo C, et al. Nuclear

overexpression of p53 protein in transitional cell bladder

carcinoma: a marker for disease progression. J Natl Can-

cer Inst 1993;85:53–9.

[17] Vatne V, Maartmann-Moe H, Hoestmark J. The prognostic

value of p53 in superficially infiltrating transitional cell

carcinoma. Scand J Urol Nephrol 1995;29:491–5.

[18] Vatne V, Maartmann-Moe H, Hoestmark J. Flow cyto-

metric DNA and p53 analysis in superficially infiltrating

bladder carcinoma. Anticancer Res 1994; 14:2735–8.

[19] Irie A, Uchida T, Ishida H, Matsumoto K, Iwamura M, Baba

S. p53 Mutation in bladder cancer patients in Japan and

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5514

inhibition of growth by in vitro adenovirus-mediated

wild-type p53 transduction in bladder cancer cells. Mol

Urol 2001;5:53–8.

[20] Lopez-Beltran A, Luque RJ, Alvarez-Kindelan J, et al. Prog-

nostic factors in stage T1 grade 3 bladder cancer survival:

the role of G1-S modulators (p53, p21Waf1, p27kip1,

Cyclin D1, and Cyclin D3) and proliferation index (ki67-

MIB1). Eur Urol 2004;45:606–12.

[21] Rodriguez-Alonso A, Pita-Fernandez S, Gonzalez-Carrero

J, Nogueira-March JL. p53 and ki67 expression as prog-

nostic factors for cancer-related survival in stage T1 tran-

sitional cell bladder carcinoma. Eur Urol 2002;41:182–8,

discussion 8–9.

[22] Popov Z, Hoznek A, Colombel M, et al. The prognostic

value of p53 nuclear overexpression and MIB-1 as a pro-

liferative marker in transitional cell carcinoma of the

bladder. Cancer 1997;80:1472–81.

[23] Rodriguez-Alonso A, Pita-Fernandez S, Gonzalez-Carrero

J, Nogueira-March JL. Multivariate analysis of survival,

recurrence, progression and development of mestastasis

in T1 and T2a transitional cell bladder carcinoma. Cancer

2002;94:1677–84.

[24] Shiina H, Igawa M, Shigeno K, et al. Clinical significance of

mdm2 and p53 expression in bladder cancer. A compar-

ison with cell proliferation and apoptosis. Oncology

1999;56:239–47.

[25] Tsuji M, Kojima K, Murakami Y, Kanayama H, Kagawa S.

Prognostic value of Ki-67 antigen and p53 protein

in urinary bladder cancer: immunohistochemical analy-

sis of radical cystectomy specimens. Br J Urol 1997;79:

367–72.

[26] Lopez-Beltran A, Luque RJ, Alvarez-Kindelan J, et al.

Prognostic factors in survival of patients with stage Ta

and T1 bladder urothelial tumors: the role of G1-S mod-

ulators (p53, p21Waf1, p27Kip1, cyclin D1, and cyclin D3),

proliferation index, and clinicopathologic parameters.

Am J Clin Pathol 2004;122:444–52.

[27] Liukkonen T, Lipponen P, Raitanen M, et al. Evaluation of

p21WAF1/CIP1 and cyclin D1 expression in the progres-

sion of superficial bladder cancer. Finbladder Group. Urol

Res 2000;28:285–92.

[28] Stein JP, Ginsberg DA, Grossfeld GD, et al. Effect of

p21WAF1/CIP1 expression on tumor progression in blad-

der cancer. J Natl Cancer Inst 1998;90:1072–9.

[29] Chatterjee SJ, Datar R, Youssefzadeh D, et al. Combined

effects of p53, p21, and pRb expression in the progression

of bladder transitional cell carcinoma. J Clin Oncol

2004;22:1007–13.

[30] Del Pizzo JJ, Borkowski A, Jacobs SC, Kyprianou N. Loss

of cell cycle regulators p27(Kip1) and cyclin E in transi-

tional cell carcinoma of the bladder correlates with

tumor grade and patient survival. Am J Pathol 1999;155:

1129–36.

[31] Ioachim E, Michael M, Stavropoulos NE, et al. Expression

patterns of cyclins D1, E and cyclin-dependent kinase

inhibitors p21(Waf1/Cip1) and p27(Kip1) in urothelial car-

cinoma: correlation with other cell-cycle-related proteins

(Rb, p53, Ki-67 and PCNA) and clinicopathological fea-

tures. Urol Int 2004;73:65–73.

[32] Santos LL, Amaro T, Pereira SA, et al. Expression of cell-

cycle regulatory proteins and their prognostic value in

superficial low-grade urothelial cell carcinoma of the

bladder. Eur J Surg Oncol 2003;29:74–80.

[33] Shariat SF, Tokunaga H, Zhou J, et al. p53, p21, pRB, and

p16 expression predict clinical outcome in cystectomy

with bladder cancer. J Clin Oncol 2004;22:1014–24.

[34] Takagi Y, Takashi M, Koshikawa T, Sakata T, Ohshima S.

Immunohistochemical demonstration of cyclin D1 in

bladder cancers as an inverse indicator of invasiveness

but not an independent prognostic factor. Int J Urol

2000;7:366–72.

[35] Tut VM, Braithwaite KL, Angus B, Neal DE, Lunec J, Mellon

JK. Cyclin D1 expression in transitional cell carcinoma of

the bladder: correlation with p53, waf1, pRb and Ki67. Br J

Cancer 2001;84:270–5.

[36] Friedrich MG, Blind C, Milde-Langosch K, et al. Frequent

p16/MTS1 inactivation in early stages of urothelial carci-

noma of the bladder is not associated with tumor recur-

rence. Eur Urol 2001;40:518–24.

[37] Kruger S, Mahnken A, Kausch I, Feller AC. P16 immuno-

reactivity is an independent predictor of tumor progres-

sion in minimally invasive urothelial bladder carcinoma.

Eur Urol 2005;47:463–7.

[38] Kwong RA, Kalish LH, Nguyen TV, et al. p14ARF protein

expression is a predictor of both relapse and survival in

squamous cell carcinoma of the anterior tongue. Clin

Cancer Res 2005;11:4107–16.

[39] Schmitz-Drager BJ, Kushima M, Goebell P, et al. p53 and

MDM2 in the development and progression of bladder

cancer. Eur Urol 1997;32:487–93.

[40] Grossman HB, Liebert M, Antelo M, et al. p53 and RB

expression predict progression in T1 bladder cancer. Clin

Cancer Res 1998;4:829–34.

[41] Orlow I, LaRue H, Osman I, et al. Deletions of the INK4A

gene in superficial bladder tumors. Association with

recurrence. Am J Pathol 1999;155:105–13.

[42] Kruger S, Mess F, Bohle A, Feller AC. Numerical aberra-

tions of chromosome 17 and the 9p21 locus are indepen-

dent predictors of tumor recurrence in non-invasive

transitional cell carcinoma of the urinary bladder. Int J

Oncol 2003;23:41–8.

[43] Tsukamoto M, Matsuyama H, Oba K, Yoshihiro S, Taka-

hashi M, Naito K. Numerical aberrations of chromosome 9

in bladder cancer. A possible prognostic marker for

early tumor recurrence. Cancer Genet Cytogenet 2002;

134:41–5.

[44] Edwards J, Duncan P, Going JJ, Watters AD, Grigor KM,

Bartlett JM. Identification of loci associated with putative

recurrence genes in transitional cell carcinoma of the

urinary bladder. J Pathol 2002;196:380–5.

[45] Maruyama R, Toyooka S, Toyooka KO, et al. Aberrant

promoter methylation profile of bladder cancer and its

relationship to clinicopathological features. Cancer Res

2001;61:8659–63.

[46] Dominguez G, Carballido J, Silva J, et al. p14ARF promoter

hypermethylation in plasma DNA as an indicator of dis-

ease recurrence in bladder cancer patients. Clin Cancer

Res 2002;8:980–5.

e u r o p e a n u r o l o g y 5 0 ( 2 0 0 6 ) 5 0 6 – 5 1 5 515

[47] Goudopoulou A, Saetta A, Korkolopoulou P, et al. p53

mutations detection in urinary bladder cancer in the

Greek population: application of the NIRCA assay. J Exp

Clin Cancer Res 2003;22:99–105.

[48] Cordon-Cardo C, Dalbagni G, Saez GT, et al. p53 mutations

in human bladder cancer: genotypic versus phenotypic

patterns. Int J Cancer 1994;56:347–53.

Editorial Comment

Antonio Lopez-Beltran, Cordoba, Spainem1lobea@uco.es

This interesting paper deals with the molecularevaluation of cell cycle regulators in bladderurothelial carcinoma. It emphasises alterations atthe INK4 gene, cyclin D1, and tumour proliferation.The topic of cell cycle alterations in bladder canceris expanding rapidly, since some of these proteinsmight represent therapeutic targets in the nearfuture. Yurakh et al. found 9p21 homozygousdeletion and p14 immunoreactivity index-independent predictors of bladder cancer recur-rence. Although this is not an unexpected finding,the methodological approach that combines mole-cular genetic and phenotypic findings with knownclinical predictors give more consistency to theresults.

The role of tumour suppressor genes in cell cycleregulation is complex. The 9p21 gene locus lies atthe centre of the two major tumour suppressorpathways identified in bladder cancer, p53 andpRb.

The 9p21 locus is crucial in the regulation of thecell cycle because of the unusual situation in whichtwo functionally different genes, p16INK4A andp14ARF, are transcribed from the same locus. Thep16INK4A is a cyclin-dependent kinase inhibitorthat functions upstream of pRb to block cyclin-Ddirected phosphorylation of pRb, which induces G1arrest. Inactivation of p16INK4A by mutation,promoter methylation, or, as found in the currentstudy, by homozygous deletion will lead to phos-phorylation of pRb and subsequent cell prolifera-tion. The other gene product at 9p21 is p14ARF,which acts upstream of p53 to stimulate p21WAF1expression, but also plays a role in the regulation of

[49] Lorenzo Romero JG, Salinas Sanchez AS, Gimenez Bachs

JM, et al. p53 Gene mutations in superficial bladder can-

cer. Urol Int 2004;73:212–8.

[50] Oyasu R, Nan L, Szumel RC, Kawamata H, Hirohashi S. p53

gene mutations in human urothelial carcinomas: analysis

by immunohistochemistry and single-strand conforma-

tion polymorphism. Mod Pathol 1995;8:170–6.

the level of p53 by interacting with the proto-oncogene product MDM2; p14ARF can cause cellcycle arrest at any point in the cell cycle through itseffect on p21WAF1. Homozygous deletions of theINK4A gene at 9p21 resulted in lower recurrence-free survival, similar to the findings by Yurakhet al.

In line with recent studies, Yurakh et al. foundthe cyclin D1 immunoreactivity index to be anindependent predictor of progression. Cyclin D1 isan upstream regulator of the cell cycle progressionthrough the G1-S checkpoint. The cyclin D familyforms complexes with CDK 4 and 6 and promotesphosphorylation and inactivation of the pRb,which releases the promoter factor E2F-1 andmediates the progression of cells from the G1 tothe S phase. These authors also found that theKi-67 immunoreactivity index as an independentpredictor of overall survival. Ki-67 is a promisingprognostic biomarker in bladder cancer and awaitsvalidation in clinical trials. Noteworthy in thisstudy, stage remains the most significant singlepredictor of progression and overall survival, andsupports the value of conventional prognosticmarkers in bladder cancer. Despite limitations inthe study (sample size, stage selection, and shortfollow-up) this paper provides superior clinicallyuseful information by combining biomarker infor-mation with conventional clinical parameters.

Understanding the molecular mechanisms thatunderlie cell cycle alterations in bladder cancershould provide important insights into how nor-mal cells become tumourigenic, and how newtherapeutic strategies can be devised. The knowl-edge about cell cycle related markers as predictorsof local and systemic therapy will continue toexpand in the near future, but remains unsettled atpresent.