Ot(MpcRfswcw0pGCm

Tmtsoc

T(

TcCCKK

CC

1

ADULT UROLOGY

7

GENOTYPES OF TNF-�, VEGF, hOGG1, GSTM1, AND GSTT1:USEFUL DETERMINANTS FOR CLINICAL OUTCOME OF

BLADDER CANCER

EUN-JUNG KIM, PILDU JEONG, CHANGYI QUAN, JIYEON KIM, SUK-CHUL BAE,SEOK JUNG YOON, JONG-WON KANG, SANG-CHEOL LEE, JAE JUN WEE, AND WUN-JAE KIM

ABSTRACTbjectives. To determine whether polymorphisms of tumor necrosis factor-alpha (TNF-�), vascular endo-

helial growth factor (VEGF), human 8-oxoguanine DNA glycosylase 1 (hOGG1), glutathione S-transferase-�GSTM1), and glutathione S-transferase-� (GSTT1) are risk factors for bladder cancer among Koreans.

ethods. We performed polymerase chain reaction-restriction fragment length polymorphism and multi-lex polymerase chain reaction in blood genomic DNA of 153 patients with primary bladder cancer and 153ontrol subjects.esults. GSTM1-negative, GSTT1-positive, and hOGG1 Ser326Ser and Ser326Cys genotypes are risk

actors for bladder cancer (P � 0.020, P � 0.044, and P � 0.012, respectively). The cancer stage wasignificantly associated with the TNF-� genotype (GG versus GA and AA; P � 0.036). A notable correlationas observed between the VEGF genotype and grade (P � 0.015). In patients with superficial bladderancer, the hOGG1 genotype was related to recurrence. The hOGG1 Ser326Ser and Ser326Cys genotypesere risk factors for superficial bladder cancer recurrence compared with the Cys326Cys genotype (P �.033, adjusted odds ratio 5.580, 95% confidence interval 1.145 to 27.183). Patients with the GSTM1-ositive genotype were at a 3.3-fold increased risk of cancer progression compared with those with theSTM1-negative genotype (P � 0.009, adjusted odds ratio 0.303, 95% confidence interval 0.123 to 0.745).onclusions. Our data collectively suggest that these genetic polymorphisms may be useful as prognosticarkers for bladder cancer in the clinical setting. UROLOGY 65: 70–75, 2005. © 2005 Elsevier Inc.

dm

tttdt

mmtsgVmms

f

he precise reason why specific individuals getbladder cancer remains unknown. Host factors

ay be involved in tumorigenesis, including de-oxification-related enzymes,1,2 the gene repairystem,3 and cytokines.4–7 Important determinantsf a population’s risk of bladder cancer includeommon genetic traits (such as those influencing

his study was supported by Korean Research Foundation grantsKRF-2000-F00078 and KRF-2002-042-E00065).

From the Departments of Urology, Biochemistry, Institute forumor Research, Chungbuk National University College of Medi-ine, Cheongju, Chungbuk, South Korea, and Preventive Medicine,hungbuk National University College of Medicine, Cheongju,hungbuk, South Korea; and Division of Ginseng Pharmacology,orea Ginseng and Tobacco Research Institute, Daejon, SouthoreaReprint requests: Wun-Jae Kim, M.D., Department of Urology,

hungbuk National University College of Medicine, Cheongju,hungbuk 361-763, South Korea. E-mail: wjkim@chungbuk.ac.krSubmitted: May 11, 2004, accepted (with revisions): August

h1, 2004

© 2005 ELSEVIER INC.0 ALL RIGHTS RESERVED

etoxification or DNA repair) and promoter poly-orphisms of several cytokines.We have previously reported that glutathione S-

ransferase-� (GSTM1)-negative and glutathione S-ransferase-� (GSTT1)-positive genotypes, wherebyhe genes are involved in cellular metabolism andetoxification of carcinogenic products, are risk fac-ors for bladder cancer.1,2

Genes encoding a number of cytokines are poly-orphic. Tumor necrosis factor-alpha (TNF-�) is aultifunctional cytokine. Genotype changes at posi-

ion �308 of the TNF-� promoter are frequently ob-erved in several cancers.8–10 Vascular endothelialrowth factor (VEGF) is a major angiogenic factor.EGF mRNA is markedly upregulated in human tu-ors, such as kidney and bladder cancer.11 A poly-orphism at position �2578 of the promoter is as-

ociated with VEGF production.12

The mutagenic base, 8-oxoguanine, is removedrom damaged DNA by base excision repair. The

uman 8-oxoguanine DNA glycosylase 1 (hOGG1)

0090-4295/05/$30.00doi:10.1016/j.urology.2004.08.005

gefpl

pzbgfmhbg

tcsjdawvctSttp

tferv

G

pftr

bTfiSmb1

G

(taCttb

Ah

ICif2m(4t

PmG

MG

G

S

waye

T

V

h

G

G

K� ylase 1

U

ene encodes DNA glycosylase that catalyzes thexcision of the mutagenic lesion 8-oxoguaninerom oxidatively damaged DNA.13,14 A Ser326Cysolymorphism in hOGG1 has been identified inung and esophageal cancers.15,16

The origin of bladder cancer is not fully ex-lained by a single risk factor. Thus, various en-ymes, cytokines, and the gene repair system maye involved in carcinogenesis, recurrence, and pro-ression of bladder cancer. In this study, we per-ormed molecular epidemiologic analyses to deter-

ine whether polymorphisms of TNF-�, VEGF,OGG1, GSTM1, and GSTT1 are risk factors forladder cancer development, recurrence, and pro-ression among Korean people.

MATERIAL AND METHODS

A hospital-based, age-matched and sex-matched case-con-rol investigation was conducted in 153 patients with bladderancer (Department of Urology, Chungbuk National Univer-ity Hospital, Cheongju, South Korea) and 153 control sub-ects. The inclusion criteria for the patients were primary blad-er cancer, pathologically proven transitional cell carcinoma,nd follow-up of more than 3 years after diagnosis. Controlsere recruited from the same department and included indi-iduals with benign disease. All subjects provided informedonsent, and the Institutional Review Board of Chungbuk Na-ional University College of Medicine approved the study.amples of blood (5 mL) from patients and controls wereransferred into 0.1-mL ethylenediaminetetraacetic acid-con-aining tubes. DNA extraction was performed according to therocedure of Lahiri and Nurnberger.17

Table I outlines the primer sets and polymerase chain reac-ion (PCR) conditions used for genotyping. The PCR mixtureor each primer set comprised 100 ng genomic DNA, 10 pM ofach primer, dNTP mix (each at 20 mM), 2.5 �L of 10�eaction buffer and Taq DNA polymerase (0.25 U) in a totalolume of 25 �L.

ENOTYPE OF TNF-� �308 PROMOTER REGIONFor detection of the G to A transition polymorphism at

osition �308 of the TNF-� gene, we used the PCR-restrictionragment length polymorphism (RFLP) technique. Substitu-ion of G with A at position �308 of TNF-� creates an Nco I

TABLE I. Sequences of pr

Factors Primer

NF-� S 5=-GCA ATA GGT GAG GGC CAT-3=AS 5.-TGG GGA CAC ACA AGC ATC AA-3=

EGF S 5=-ATA AGG GCC TTA GGA CAC CA-3=AS 5.-GCT ACT TCT CCA GGC TCA CA-3=

OGG1 S 5=-TCC TAC AGG TGC TGT TCA GT-3=AS 5.-TTG GAA CCC TTT CTG CGC TT-3=

STM1 S 5=-GAA CTC CCT GAA AAG CTA AAG C-3=AS 5.-GTT GGG CTC AAA TAT ACG GTG G-3

STT1 S 5=-TTC CTT ACT GGT CCT CAC ATC TC-3=AS 5.-TCA CCG GAT CAT GGC CAG CA-3=

EY: PCR � polymerase chain reaction; D � denaturation temperature; A � annealivascular endothelial growth factor; hOGG1 � human 8-oxoguanine DNA glycos

estriction site absent in the wild-type gene, which is detected m

ROLOGY 65 (1), 2005

y RFLP.18 A 150-bp PCR product was obtained using a DNAhermal Cycler (Perkin-Elmer, Fermont, Calif). Next, ampli-ed products were digested with 4 U of Nco I (Takara Korea,eoul, South Korea). Cleaved bands (130-bp and 20-bp frag-ents) correspond to the GG genotype, the non-cleaved (150-

p) band represents the AA genotype, and both 130-bp and50-bp bands signify the GA genotype.

ENOTYPE OF VEGF �2578 PROMOTER REGIONThe 422-bp PCR product was mixed with 4 U of Bgl II

Takara Korea) and reaction buffer, according to the manufac-urer’s instructions. The reaction was performed for 12 hourst 37°C and terminated by incubation for 20 minutes at 80°C.leaved bands (264-bp and 158-bp fragments) correspond to

he CC genotype, the non-cleaved (422-bp) band representshe AA genotype, and all three (422-bp, 264-bp, and 158-bp)ands represent the CA genotype.

NALYSIS OF Ser326Cys POLYMORPHISM OF

OGG1 GENEThe 82-bp PCR product (16 �L) was mixed with 1 U of Fau

(Bioneer, Chungbuk, South Korea) and 2 �L of 10� Accu-ut greenN reaction buffer in a total volume of 20 �L, accord-

ng to the manufacturer’s instructions. The reaction was per-ormed for 24 hours at 55°C and terminated by incubation for0 minutes at 65°C. Cleaved bands (47-bp and 35-bp frag-ents) signify the Ser326Ser genotype, the noncleaved band

82 bp) represents the Cys326Cys genotype, and the 82-bp,7-bp, and 35-bp bands correspond to the Ser326Cys geno-ype.

To confirm the accuracy of the PCR-RFLP assay, purifiedCR products of each target were sequenced using an auto-ated laser fluorescence sequencer (ABI PRISM 377 and 3100enetic Analyzer, Applied Biosystems, Foster City, Calif).

ULTIPLEX PCR TO DETECT PRESENCE OR ABSENCE OF

STM1 AND GSTT1 GENESMultiplex PCR was used to detect the presence or absence of

STM1 and GSTT1 genes, as described previously.1

TATISTICAL ANALYSISAssociations between genotypes and clinical parameters

ere measured using the adjusted odds ratio (AOR) with agend sex determined by unconditional logistic regression anal-sis because of frequency matching. Additionally, the influ-nce of five different genotypes on bladder cancer develop-

r sets and PCR conditionsLength

(bp) D A E Cycle

150 94°C 52°C 72°C 3515 s 30 s 30 s

422 95°C 60°C 72°C 3515 s 30 s 30 s

82 94°C 53°C 72°C 4030 s 30 s 30 s

210 94°C 62°C 74°C 3060 s 60 s 60 s

473 94°C 62°C 74°C 3060 s 60 s 60 s

perature; E � extension temperature; TNF-� � tumor necrosis factor-alpha; VEGF; GSTM1 � glutathione S-transferase-�; GSTT1 � glutathione S-transferase-�.

ime

=

ng tem

ent, recurrence, and progression was measured by multiple

71

lBaSciA8t

D

hcdfcisGo

s0ttavhSd2asm23

RS

T

S

AT

V

h

G

G

K*G†

7

ogistic analysis controlling for the effects of other factors.ecause TNF-� (GG versus GA and AA), VEGF (CC versus CAnd AA), and hOGG1 (Cys326Cys versus Ser326Ser ander326Cys) genotypes correlated significantly with bladderancer in univariate analyses, we assigned these two categoriesn each target as reference groups for multiple logistic analysis.ll analyses were performed with the SAS package (version.1, SAS Institute, Cary, NC), and data were considered statis-ically significant at P �0.05.

RESULTS

EVELOPMENT OF BLADDER CANCER

Table II shows the genotypes of TNF-�, VEGF,OGG1, GSTM1, and GSTT1 among patients andontrols. The genotypes in this study were in Har-y-Weinberg equilibrium. The GSTM1-negativerequencies were 60.1% for patients and 47.8% forontrols (P � 0.020, AOR 1.725, 95% confidencenterval [CI] 1.091 to 2.729), which represents atatistically significant risk of bladder cancer. TheSTT1-negative genotype was observed in 46.4%

TABLE II. Distribution of age, sex, an

CharacteristicControls

(n � 153)Patients

(n � 153)

ex (n)Male 132 129Female 21 24

ge (mean � SD) 60.7 � 11.8 62.9 � 11.8NF-� (n)GG 130 139GA 22 12AA 1 2P value* 0.211

EGF (n)CC 11 13CA 69 69AA 73 71P value 0.717

OGG1 (n)Cys326Cys 45 26Ser326Cys 70 90Ser326Ser 38 37P value 0.137STM1Positive 80 61Negative 73 92P value 0.020†

STT1Positive 64 82Negative 89 71P value 0.044†

EY: S � superficial; I � invasive; AOR � adjusted odds ratio; other abbreviationsP values represent statistically significant levels of variation in each genetic patterSTT1).Meaningful results (protective factor: AOR �1 and risk factor: AOR �1).

f patients with bladder cancer and 58.2% control q

2

ubjects, which was statistically significant (P �.044). The AOR and 95% CI of the GSTT1-nega-ive type was 0.625 and 0.395 to 0.988, suggestinghat individuals with the GSTT1-positive genotypere approximately at 1.6-fold increased risk of de-eloping bladder cancer. Compared with theOGG1 Cys326Cys genotype, the Ser326Ser ander326Cys genotypes displayed risk factors for theevelopment of bladder cancer (P � 0.012, AOR.028, 95% CI 1.171 to 3.521). On multiple logisticnalysis, the GSTM1 and hOGG1 genotypes wereignificantly related to bladder cancer develop-ent (P � 0.033, AOR 1.667, 95% CI 1.043 to

.663 and P � 0.026, AOR 1.912, 95% CI 1.080 to

.390, respectively; Table III).

ELATIONSHIP BETWEEN GENOTYPES AND CANCER

TAGE AND GRADE

Stage was significantly associated with theNF-� genotype, with the GG genotype more fre-

notypes among patients and controlsStage(n �141)

Grade(n � 139)

Recurrence(n � 93)

Progression(n � 141)

S I 1 2 3 No Yes No Yes

77 41 22 60 35 45 32 97 2116 7 3 13 6 10 6 16 7

87 40 22 66 37 51 36 103 244 8 2 6 4 3 1 8 42 0 1 1 0 1 1 2 00.208 0.526 0.872 0.630

7 6 1 5 7 6 1 12 139 25 10 33 21 20 19 47 1747 17 14 35 13 29 18 54 100.080 0.015† 0.827 0.713

14 8 6 10 4 12 2 19 358 24 13 44 25 32 26 68 1421 16 6 19 12 11 10 26 110.422 0.232 0.765 0.098

34 20 15 27 11 20 14 37 1759 28 26 46 14 35 24 76 110.562 0.598 0.963 0.006

46 30 24 37 14 24 22 61 1547 18 17 36 11 31 16 52 130.116 0.484 0.179 0.969

able I.ated by analysis of variance (TNF-�, VEGF, and hOGG1) or t test (GSTM1 and

d ge

as in Tn estim

uent in patients with superficial disease compared

UROLOGY 65 (1), 2005

w0ctcwA

RR

rpoms

hrfgt2(naC

wGct0ma0

A

U

M

Ka*CG†

TAB

LEIV

.R

elati

onsh

ipbet

wee

nca

ncer

stage,

gra

de,

recu

rren

ce,

pro

gre

ssio

n,and

gen

otyp

esSta

geG

rade

Rec

urre

nce

Pro

gres

sion

naly

sis

PVal

ueA

OR

(95

%C

I)P

Val

ueA

OR

(95

%C

I)P

Val

ueA

OR

(95

%C

I)P

Val

ueA

OR

(95

%C

I)

ivar

iate

NF-

�*

0.0

36

†0.2

93

†(0

.093–0

.925)†

0.7

37

1.2

43

(0.3

51

–4.4

03

)0

.85

80

.85

0(0

.141–5

.102)

0.2

01

2.3

31

(0.6

38–8

.547)

EGF

0.4

03

1.6

50

(0.5

10–5

.336)

0.0

98

0.3

58

(0.1

06

–1.2

07

)0

.17

44

.50

5(0

.514–4

0.0

00)

0.1

96

4.0

49

(0.4

86–3

3.3

33)

OG

G1

0.8

08

1.1

31

(0.4

28–2

.967)

0.2

63

2.0

00

(0.5

93

–6.7

57

)0

.03

3†

5.5

80

†(1

.145–2

7.1

83)†

0.2

50

2.2

73

(0.5

61–9

.207)

STM

10

.562

1.2

41

(0.5

98–2

.578)

0.8

27

1.0

93

(0.4

93

–2.4

22

)0

.82

70

.90

7(0

.378–2

.174)

0.0

09

†0.3

03

†(0

.123–0

.745)†

STT1

0.1

16

1.8

01

(0.8

65–3

.748)

0.3

03

0.6

63

(0.3

03

–1.4

50

)0

.13

70

.51

7(0

.219–1

.487)

0.9

28

0.9

61

(0.4

04–2

.288)

ltiv

aria

teN

F-�

0.0

69

2.9

87

(0.9

20–9

.705)

0.9

43

1.0

49

(0.2

85

–3.8

56

)0

.62

70

.61

8(0

.089–4

.291)

0.4

78

1.6

46

(0.4

16–6

.518)

EGF

0.1

08

0.5

23

(0.2

37–1

.153)

0.0

90

0.4

87

(0.2

11

–1.1

20

)0

.87

10

.92

8(0

.378–2

.281)

0.5

89

0.7

67

(0.2

93–2

.008)

OG

G1

0.4

41

0.6

61

(0.2

30–1

.894)

0.4

52

1.6

31

(0.4

55

–5.8

48

)0

.02

6†

6.4

94

†(1

.247–3

3.3

33)†

0.4

14

1.8

66

(0.4

18–8

.333)

STM

10

.717

0.8

68

(0.4

03–1

.869)

0.7

52

1.1

43

(0.5

00

–2.6

10

)0

.62

80

.79

2(0

.307–2

.038)

0.0

16

†0.3

23

†(0

.130–0

.807)†

STT1

0.1

38

0.5

64

(0.2

65–1

.201)

0.2

86

0.6

47

(0.2

91

–1.4

40

)0

.06

30

.41

6(0

.165–1

.048)

0.8

12

0.8

96

(0.3

62–2

.217)

revi

atio

nsas

inT

able

sI

and

III.

NF

-��

GG

vers

usG

A�

AA

;VE

GF

�C

Cve

rsus

CA

�A

A;h

OG

G1

�C

ys32

6Cys

vers

usSe

r326

Ser

�Se

r326

Cys

;GST

M1

�po

siti

veve

rsus

nega

tive

;GST

T1

�po

siti

veve

rsus

nega

tive

.ea

ning

fulr

esul

ts(p

rote

ctiv

efa

ctor

:AO

R�

1an

dri

skfa

ctor

:AO

R�

1).

U

ith the GA and AA genotypes (P � 0.036, AOR.293, 95% CI 0.093 to 0.925; Table IV). A notableorrelation was observed between the VEGF geno-ype and grade (P � 0.015; Table II). High-gradeancers were more frequently observed in patientsith the CC genotype compared with the CA andA genotypes.

ELATIONSHIP BETWEEN GENOTYPES AND CANCER

ECURRENCE AND PROGRESSION

In this study, we defined recurrence as tumorecurrence of superficial bladder cancer withoutrogression and progression as cancer progressionf both superficial bladder cancer to invasive oretastatic disease and invasive cancer to meta-

tatic disease after adequate treatment.In patients with superficial bladder cancer, the

OGG1 genotype was related to recurrence. Theecurrence of superficial bladder cancer was morerequent in patients with Ser326Ser and Ser326Cysenotypes compared with the Cys326Cys geno-ype (P � 0.033, AOR 5.580, 95% CI 1.145 to7.183). On multiple logistic analysis, the hOGG1Cys326Cys versus Ser326Ser and Ser326Cys) ge-otype was also statistically significantly associ-ted with recurrence (P � 0.026, AOR 6.494, 95%I 1.247 to 33.333; Table IV).Cancer progression was significantly associatedith the GSTM1 genotype. Patients with theSTM1-positive genotype were at a 3.3-fold in-

reased risk of cancer progression compared withhose with the GSTM1-negative genotype (P �.009, AOR 0.303, 95% CI 0.123 to 0.745). Onultiple logistic analysis, cancer progression was

lso associated with the GSTM1 genotype (P �

TABLE III. Statistical analysis for bladdercancer development

nalysis P Value AOR 95% CI

nivariateTNF-�* 0.156 0.597 0.293–1.217VEGF 0.729 0.862 0.371–2.000hOGG1 0.012† 2.028† 1.171–3.521†

GSTM1 0.020† 1.725† 1.091–2.729†

GSTT1 0.044† 0.625† 0.395–0.988†

ultivariateTNF-� 0.217 0.631 0.304–1.310VEGF 0.982 0.995 0.662–1.592hOGG1 0.026† 1.912† 1.080–3.390†

GSTM1 0.033† 1.667† 1.043–2.663†

GSTT1 0.145 0.703 0.438–1.129

EY: AOR � adjusted odds ratio; 95% CI � 95% confidence interval; other abbrevi-tions as in Table 1.TNF-� � GG versus GA � AA; VEGF � CC versus CA � AA; hOGG1 �ys326Cys versus Ser326Ser � Ser326Cys; GSTM1 � positive versus negative;STT1 � positive versus negative.Meaningful results (protective factor: AOR �1 and risk factor: AOR �1).

.016; Table IV).

ROLOGY 65 (1), 2005

A

Un T V h G G

Mu T V h G G

Abb

*T

†M

73

ttacg(w(obG

uuasgmssgSfwtptapbAwgeieSpb

tagisasw

chtcSt

swcicnliabt

Gtcagrsut

aga

a9

miO

lp3

lTe

ncs

p1

gtc

mH

tf

7

COMMENT

Molecular epidemiologic studies have revealedhat individuals with the GSTM1-negative geno-ype have a greater risk of bladder cancer.1,19,20 Ingreement with earlier data, our results have indi-ated an association between the GSTM-negativeenotype and an increased risk of bladder cancerAOR 1.725). In contrast, the bladder cancer riskas decreased with the GSTT1-negative genotype

AOR 0.625). Our findings are supported by thosef a Japanese investigation showing an associationetween a decreased risk of bladder cancer and theSTM1-positive/GSTT1-negative genotype.21

The repair effectiveness of hOGG1 may be mod-lated by genetic polymorphisms. However, pop-lation studies on the hOGG1 326 polymorphismnd cancer susceptibility have been inconclu-ive.22–26 The distribution of the hOGG1 codon 326enotypes varied in different populations, withuch less prevalence of the Cys allele in Western

ubjects.15 Among the Korean participants in ourtudy, the distribution of the hOGG1 codon 326enotypes (Cys326Cys versus Ser326Ser ander326Cys) of controls was significantly differentrom that of the patients with bladder cancer inhom the Cys326Cys genotype displayed a protec-

ive effect against developing bladder cancer com-ared with the Ser326Ser and Ser326Cys geno-ypes. Currently, no convincing explanation isvailable for our finding that Cys326Cys is a highlyrotective genotype. In any case, hOGG1 may note the only gene associated with oxidative damage.n alternative DNA oxidative damage repair path-ay to minimize the effects of 8-oxoguanine inenomes was recently reported.27 However, in gen-ral, additional studies are necessary to clarify thenvolvement of this gene in bladder cancer. In ourxperiments, the GSTM1-negative, and hOGG1er326Ser and Ser326Cys genotypes were inde-endent risk factors for bladder cancer, as revealedy both univariate and multivariate analyses.We showed that cancer stage and grade are sta-

istically significantly associated with the TNF-�nd VEGF genotypes. These cytokines have angio-enic function, which is possibly critical for cancernvasion and metastasis. High-grade and high-tage bladder cancers are more prone to invasionnd metastasis. Our results suggest that the aggres-iveness of the primary bladder cancer is associatedith the TNF-� and VEGF genotypes.Recurrence in patients with superficial bladder

ancer was significantly associated with theOGG1 genotype. The hOGG1 Cys326Cys geno-ype was protective against superficial bladder can-er recurrence compared with the Ser326Ser ander326Cys genotypes. The GSTM1-positive geno-

ype correlated significantly with cancer progres- 1

4

ion. In this study, the GSTM1-negative genotypeas a statistically significant risk factor for bladder

ancer development. In contrast, the GSTM1-pos-tive genotype was an independent risk factor forancer progression, suggesting that increased uri-ary excretion of unknown substances metabo-

ized by GSTM1 may promote cancer progressionn patients with bladder cancer. Additional studiesre needed to identify the relationship betweenladder cancer progression and the GSTM1-posi-ive genotype.

CONCLUSIONS

The results of our study revealed that theSTM1, GSTT1, and hOGG1 genotypes are related

o the development of bladder cancer. Bladder can-er aggressiveness was associated with the TNF-�nd VEGF genotypes, and the hOGG1 and GSTM1enotypes correlated closely with bladder cancerecurrence and progression. Our data collectivelyuggest that these genetic polymorphisms may beseful as prognostic markers for bladder cancer inhe clinical setting.

REFERENCES1. Kim WJ, Lee HL, Lee SC, et al: Polymorphisms of N-

cetyltransferase 2, glutathione S-transferase mu and thetaenes as risk factors of bladder cancer in relation to asthmand tuberculosis. J Urol 164: 209–213, 2000.

2. Kim WJ, Kim H, Kim CH, et al: GSTT1-null genotype isprotective factor against bladder cancer. Urology 60: 913–18, 2002.

3. Kohno T, Shinmura K, Tosaka M, et al: Genetic poly-orphisms and alternative splicing of the hOGG1 gene that is

nvolved in the repair of 8-hydroxyguanine in damaged DNA.ncogene 16: 3219–3225, 1998.

4. Gadducci A, Ferdeghini M, Fanucchi A, et al: Serumevels of soluble receptors for tumor necrosis factor (p55 and75 sTNFr) in endometrial cancer. Anticancer Res 16: 3125–128, 1996.

5. Gadducci A, Ferdeghini M, Castellani C, et al: Serumevels of tumor necrosis factor (TNF), soluble receptors forNF (55- and 75-kDa sTNFr), and soluble CD14 (sCD14) inpithelial ovarian cancer. Gynecol Oncol 58: 184–188, 1995.

6. Okamoto T, Watanabe N, Yamauchi N, et al: Endoge-ous tumor necrosis factor exerts its protective function intra-ellularly against the cytotoxicity of exogenous tumor necro-is factor. Cancer Res 52: 5278–5281, 1992.

7. Steiner MS: Review of peptide growth factors in benignrostatic hyperplasia and urological malignancy. J Urol 153:085–1096, 1995.

8. Sasaki M, Nakajima K, Perinchery G, et al: Frequentenotype changes at �308 of the human tumor necrosis fac-or-alpha promoter region in human uterine endometrial can-er. Oncol Rep 7: 369–373, 2000.

9. Warzocha K, Ribeiro P, Bienvenu J, et al: Genetic poly-orphisms in the tumor necrosis factor locus influence non-odgkin’s lymphoma outcome. Blood 91: 3574–3581, 1998.10. Oh BR, Sasaki M, Perinchery G, et al: Frequent geno-

ype changes at �308, and 488 regions of the tumor necrosisactor-� (TNF-�) gene in patients with prostate cancer. J Urol

63: 1584–1587, 2000.

UROLOGY 65 (1), 2005

pgn

lr

d8B

hi

SC

pC

mR

T1

cc(C

yfel

Sc

ilS

hC

pc8

pJ1

hmc

lm

U

11. Brown LF, Berse B, Jackman RW, et al: Increased ex-ression of vascular permeability factor (vascular endothelialrowth factor) and its receptors in kidney and bladder carci-omas. Am J Pathol 143: 1255–1262, 1993.12. Shahbazi M, Fryer AA, Pravica V, et al: Vascular endothe-

ial growth factor gene polymorphisms are associated with acuteenal allograft rejection. J Am Soc Nephrol 13: 260–264, 2002.

13. Lee YS, Lee HS, Park MK, et al: Identification of 8-hy-roxyguanine glycosylase activity in mammalian tissues using-hydroxyguanine specific monoclonal antibody. Biochemiophys Res Commun 196: 1545–1551, 1993.14. Tchou J, Kasai H, Shibutani S, et al: 8-Oxoguanine (8-

ydroxyguanine) DNA glycosylase and its substrate specific-ty. Proc Natl Acad Sci USA 88: 4690–4694, 1991.

15. Sugimura H, Kohno T, Wakai K, et al: hOGG1er326Cys polymorphism and lung cancer susceptibility.ancer Epidemiol Biomarkers Prev 8: 669–674, 1999.16. Xing DY, Tan W, Song N, et al: Ser326Cys polymor-

hism in hOGG1 gene and risk of esophageal cancer in ahinese population. Int J Cancer 95: 140–143, 2001.17. Lahiri DK, and Nurnberger JI Jr: A rapid non-enzymaticethod for the preparation of HMW DNA from blood forFLP studies. Nucleic Acids Res 19: 5444, 1991.18. Park KS, Mok JW, Rho SA, et al: Analysis of TNFB and

NFA NcoI RFLP in colorectal cancer. Mol Cells 8: 246–249,998.19. Bell DA, Taylor JA, Paulson DF, et al: Genetic risk and

arcinogen exposure: a common inherited defect of the car-inogen-metabolism gene glutathione S-transferase M1GSTM1) that increases susceptibility to bladder cancer. J Natl

ancer Inst 85: 1159–1164, 1993. 1

ROLOGY 65 (1), 2005

20. Brockmöller J, Cascorbi I, Kerb R, et al: Combined anal-sis of inherited polymorphisms in arylamine N-acetyltrans-erase 2, glutathione S-transferases M1 and T1, microsomalpoxide hydrolase, and cytochrome P450 enzymes as modu-ators of bladder cancer risk. Cancer Res 56: 3915–3925, 1996.

21. Katoh T, Inatomi H, Kim H, et al: Effects of glutathione-transferase (GST) M1 and GSTT1 genotypes on urothelialancer risk. Cancer Lett 132: 147–152, 1998.

22. Janssen K, Schlink K, Gotte W, et al: DNA repair activ-ty of 8-oxoguanine DNA glycosylase 1 (OGG1) in humanymphocytes is not dependent on genetic polymorphismer326/Cys326. Mutat Res 486: 207–216, 2001.

23. Xu J, Zheng SL, Turner A, et al: Associations betweenOGG1 sequence variants and prostate cancer susceptibility.ancer Res 62: 2253–2257, 2002.24. Wikman H, Risch A, Klimek F, et al: hOGG1 polymor-

hism and loss of heterozygosity (LOH): significance for lungancer susceptibility in a Caucasian population. Int J Cancer8: 932–937, 2000.25. Hanaoka T, Sugimura H, Nagura K, et al: hOGG1 exon7

olymorphism and gastric cancer in case-control studies ofapanese Brazilians and non-Japanese Brazilians. Cancer Lett70: 53–61, 2001.26. Hardie LJ, Briggs JA, Davidson LA, et al: The effect of

OGG1 and glutathione peroxidase I genotypes and 3p chro-osomal loss on 8-hydroxydeoxyguanosine levels in lung

ancer. Carcinogenesis 21: 167–172, 2000.27. Klungland A, Rosewell I, Hollenbach S, et al: Accumu-

ation of premutagenic DNA lesions in mice defective in re-oval of oxidative base damage. Proc Natl Acad Sci USA 96:

3300–13305, 1999.

75