Download COX-2 promoter polymorphisms and the association with prostate

Carcinogenesis vol.29 no.12 pp.2347–2350, 2008
doi:10.1093/carcin/bgn245
Advance Access publication October 28, 2008
COX-2 promoter polymorphisms and the association with prostate cancer risk in
South African men
Pedro Fernandez1,, Pieter M.de Beer1,2, Lize van der
Merwe3,4 and Chris F.Heyns1,2
1
Department of Urology, Stellenbosch University, PO Box 19063, Tygerberg
7505, South Africa, 2Department of Urology, Tygerberg Hospital, Tygerberg
7505, South Africa, 3Biostatistics Unit, Medical Research Council, PO Box
17090, Tygerberg, 7505, South Africa and 4Department of Statistics, University
of the Western Cape, Private Bag X17, Bellville, 7535, South Africa
To whom correspondence should be addressed. Tel: +27 21 938 9288;
Fax: þ27 21 933 8010;
Email: [email protected]
Cyclooxygenase-2 (COX-2) converts arachidonic acid to prostaglandins, which are important mediators of cell proliferation and
inflammation. Evidence indicates that COX-2 plays a role in carcinogenesis and that it is over-expressed in prostate tumours. We
investigated the role of COX-2 variants in prostate cancer in
a case–control study of South African Coloured men, consisting
of 151 cases and 134 controls. The genotype frequencies of four
single-nucleotide polymorphisms (SNPs) in the COX-2 promoter
were initially determined in 50 control subjects. One SNP, rs20417
(2899G>C), was monomorphic and excluded from further investigation. Three SNPs, rs3918304 (21285A>G), rs20415
(21265C>T) and rs5270 (2297C>G), were genotyped in all the
case patients and control subjects. The 21285 G-allele and 21265
T-allele were associated with increased risk of prostate cancer
[odds ratio (OR) 5 3.53; confidence interval (CI) 5 2.14–5.90;
P < 0.0001 and OR 5 3.01; CI 5 1.82–5.02; P < 0.0001] after
adjusting for age. Haplotype GTC conferred increased risk of
prostate cancer in South African Coloured men (OR 5 3.54 versus ACC; CI 5 2.12–5.92; P < 0.0001). These findings in conjunction with findings in other populations of African descent might
suggest a common causal variant for prostate cancer in COX-2, or
a variant in a nearby gene.
Introduction
Cyclooxygenase-2 (COX-2) is an inducible enzyme that converts
arachidonic acid to prostaglandins, which play a role in cell proliferation and are potent mediators of inflammation (1). Furthermore,
COX-2 has been postulated to influence carcinogenesis by inhibiting
apoptosis (2), inducing angiogenesis (3) and by chronic activation of
immune responses (4). Several studies have shown increased expression of COX-2 in various cancers (5–7). Increased COX-2 expression
has also been observed in prostate tumours (8,9), although it has been
suggested that the over-expression is confined to proliferative inflammatory atrophy (10), a precursor of prostate cancer.
Even though a number of single-nucleotide polymorphisms (SNPs)
have been reported in COX-2, only a few SNPs have been identified in
the coding regions, suggesting evolutionary selection in favour of
maintaining the enzyme’s role in body homeostasis (11). Genetic
associations between allelic variations in the COX-2 non-coding, promoter and 3# untranslated region and risk of developing different
forms of cancer have been demonstrated (12,13). Moreover, progressively more associations between COX-2 SNPs and risk of developing
prostate cancer have been reported. Panguluri et al. (14) described an
increased risk of prostate cancer in African American men and variants in the COX-2 promoter that alter putatively functional transcription factor-binding sites. They additionally identified a COX-2
Abbreviations: C/EBP, CCAAT/enhancer-binding protein; CI, confidence interval; COX-2, cyclooxygenase-2; OR, odds ratio; SNP, single-nucleotide
polymorphism.
promoter haplotype that conferred increased risk of prostate
cancer in African American and Nigerian men (14). Subsequent studies described a decreased risk of developing prostate cancer in Swedish men with the intronic COX-2 polymorphisms rs20432
(þ3100T.G) and rs689470 (þ8365G.A) (15) and North American
Caucasian men with rs2745557 (þ202G.A) and rs2206593
(þ6994C.T) (16). Cheng et al. (16) additionally reported a higher
risk of prostate cancer in African American men with the AA genotype of the COX-2 þ8365G.A polymorphism, but no association in
Caucasian men. More recent findings from a large case–control study
in North American non-Hispanic Caucasian men did not support
associations between five SNPs and prostate cancer risk (17), which
included SNPs where associations had been described in other investigations (15,16).
In the present study, our aim was to examine the role of COX-2
polymorphisms and the risk of developing prostate cancer in a case–
control study of South African Coloured men (Coloured refers to
a unique South African population group of mixed racial origin).
We describe a genetic association between promoter SNPs and a haplotype that confer increased prostate cancer risk. Our findings are
to a degree similar to a previously described association between
COX-2 promoter polymorphisms and haplotype identified in African
American and Nigerian men.
Materials and methods
Study population
Unrelated male subjects from the South African Coloured ethnic group (18)
were enrolled in the case–control study to determine genetic risk factors for
prostate cancer that was conducted between the years 2004 to 2007. The study
population comprising 151 cases [mean (range) age: 69 (47–88) years] with
histologically confirmed prostate cancer were all from the Western Cape Province of South Africa and had undergone radical prostatectomy, transurethral
resection of the prostate or prostatic biopsy at the Department of Urology,
Tygerberg Hospital (Cape Town, South Africa). Controls were selected among
subjects admitted to the same hospital during the same period and comprised of
134 individuals that were matched for ethnicity and age [mean (range) age:
65 (52–91) years] and were from the same geographical region. The selection
criteria of age-, ethnic-, geographical- and institutional-matching increase the
likelihood that the controls were representative of the source population of
cases.
Blood samples were collected from each subject. Clinical characteristics
including prostate-specific antigen, Gleason grade, tumour node metastasis
stage, age at diagnosis and family history were obtained from medical records.
All controls had prostate-specific antigen levels 2.5 ng/ml and a normal
digital rectal examination. All subjects were informed and gave written consent
to participate in the study and allow their biological samples to be genetically
analyzed, according to the Helsinki declaration. The study was approved by the
Stellenbosch University, Faculty of Health Sciences, Committee for Human
Research.
SNP genotyping
Genomic DNA was extracted from whole blood using a QIAamp DNA Kit
(Qiagen, GmbH, Hilden, Germany). We used ABI 3100 Mutiplex SNaPshotTM
Primer extension analysis (Applied Biosystems, Foster City) to screen SNPs in
an 1156 bp region in the COX-2 promoter that starts 125 bp upstream of the
transcription start site. The outer primer set for amplification of the 1156 bp
fragment were COX-2prmF 5#-CTCACATGCTCCTCCCTGAG-3# and COX2prmR 5#-CTCCTCTCCCCTTAAAAAAATTGCG-3#. Our aim was to replicate the study by Panguluri et al. (14) and therefore we selected the SNPs
rs3918304 (1285A.G), rs20415 (1265G.A), rs20417 (899G.C) and
rs5270 (297C.G). The respective primers for the primer extension analysis
were 1285 F 5#-CCATCTAATTAATTCCTCATCCAACT-3#, 1265 R 5#GGCTTATTGGGGCTAATTTT-3# (reverse extension direction in this study),
899 F 5#-TTAGGACCAGTATTATGAGGAGAATTTACCTTTCCC-3# and
297 F 5#-CAGCTTCCTGGGTTTCCGATTTTCTCATTT-3#. Initially, we
determined the allele frequency for each SNP in a panel of 50 control subjects.
Ó The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]
2347
P.Fernandez et al.
Among our controls, SNP rs20417 was monomorphic and it was subsequently
removed from further genotype analysis.
To determine whether population stratification might be a confounding factor leading to spurious associations (19), genotype analysis was performed in
a group of 65 cases and 65 controls for 24 independent polymorphic SNPs that
were not in linkage disequilibrium with COX-2.
Statistical analyses
Genotype and allele frequencies were calculated for each SNP. Each polymorphism was tested in the controls to confirm that they were in Hardy–
Weinberg equilibrium and linkage disequilibrium between polymorphisms
was assessed. For each SNP, logistic regression was used to estimate prostate
cancer–genotype odds ratios (ORs), 95% confidence intervals (CIs) and corresponding P-values adjusted for age. Logistic regression was combined with
an expectation maximization algorithm to infer haplotype frequencies and
assess the association between prostate cancer and specific haplotypes. Haplotype analysis inferred a probability distribution of possible haplotypes for
each individual. From these, we estimated haplotype frequencies for the group
(pooled), as well as separate case and control frequencies. Prostate cancer ORs
(and 95% CIs) for specific haplotypes compared with the reference (highest
frequency) haplotype were estimated (20). P-values for prostate cancer–
haplotype association were also determined. The OR and P-values are all based
on a single model, so that they are all adjusted for each other. Analyses were
done in R, a language and environment for statistical computing, freely available from http://www.R-project.org. The R packages genetics, LDheatmap and
haplo.stats were used.
controls, among the 24 SNPs, was 1.39 (P 5 0.3557), resulting in
an inflation factor of k 5 1.39/0.4549 5 3.06 (23). All our results
remained significant when we inflated our P-values by multiplying
them with the inflation factor (data not shown). Additionally, the
P-values we obtained for our significant results were much smaller
than any of those found for our unlinked independent panel of 24
SNPs, further indication that our results are probably true.
Three haplotypes of 1285A.G, 1265C.T and 297C.G had
inferred frequencies of 5% in the study groups (Table II). Five
haplotypes with individual frequencies ,5% in the study groups were
grouped together for the analysis (Table II). The global P-value for the
fit of the logistic regression model of prostate cancer case–control
status was highly significant (P , 0.0001). Haplotype GTC, compared with the haplotype ACC, was associated with prostate cancer
in South African Coloured men (OR 5 3.54; CI 5 2.12–5.92)
(Table II). Highly significant (P , 0.0001 after inflating) differences
in haplotype frequencies between cases and controls were found for
both haplotypes ACC (52% in cases and 73% in controls) and GTC
(31% in cases and 16% in controls) (Table II).
Results
All three SNPs were in Hardy–Weinberg equilibrium among the controls (P . 0.05). A significant association with prostate cancer was
observed for the 1285 AG/GG (OR 5 3.53; CI 5 2.14–5.90;
P , 0.0001) and 1265 CT/TT genotype (OR 5 3.01; CI 5 1.82–
5.02; P , 0.0001) after adjusting for age (Table I). The 297 CG
genotype showed an increased risk of prostate cancer (OR 5 2.56;
CI 5 0.98–7.58), although this result was not significant after adjustment for age (P 5 0.0673) (Table I). No 297 GG genotype was
observed in any of the subjects screened in the present study (Table I).
As expected, there was highly significant linkage disequilibrium between the SNP pair 1285/1265 (Figure 1).
Our study panel is considered to be a present-day homogeneous
population (21), despite historically having received genetic contributions from different ancestral populations (18). We genotyped a subgroup of 65 cases and 65 controls for a panel of 24 unlinked SNPs in
an attempt to address the issue of possible population stratification as
a confounding factor leading to spurious associations. To test for
stratification across the 24 loci, we calculated the sum of the 24 allelic
association v2 statistics (22), which gave a total of 54.0 at 24 degrees
of freedom (P 5 0.0004), indicating that significant stratification exists between the cases and controls included in our study. The median
v2 value for comparison of allele frequencies between cases and
Fig. 1. Pairwise linkage disequilibrium analysis using D#.
Table I. Genotype frequencies and ORs for prostate cancer among COX-2 promoter SNPs in South African Coloured men
SNP
rs3918304 (1285A.G)
rs20415 (1265C.T)
rs5270 (297C.G)
Genotype
AA
AG
GG
AG and GG
CC
CT
TT
CT and TT
CC
CG
GG
OR (95% CI)a
Number of subjects (%)
Cases (%)
Controls (%)
48 (32)
91 (60)
12 (8)
103 (68)
47 (31)
94 (62)
10 (7)
104 (69)
135 (89)
16 (11)
0
85 (63)
46 (34)
3 (2)
49 (36)
76 (57)
52 (39)
6 (4)
58 (43)
128 (96)
6 (4)
0
1.00 (reference)
3.39 (2.04–5.73)
5.52 (1.59–25.72)
3.53 (2.14–5.90)
1.00 (reference)
3.15 (1.89–5.34)
1.88 (0.59–6.36)
3.01 (1.82–5.02)
1.00 (reference)
2.56 (0.98–7.58)
ND
ND, not determined; OR, 95% CI and P-value could not be estimated because the observed counts in both cases and controls are zero.
a
Adjusted for age.
2348
P-valuea
,0.0001
0.0130
,0.0001
,0.0001
0.2920
,0.0001
0.0673
ND
COX-2 association with prostate cancer
Table II. Inferred haplotype frequencies, joint ORs and joint P-values for tests of association with prostate cancer for each inferred haplotype in South African
Coloured men
1285A.G
A
A
G
Rarea
1265C.T
C
T
T
297C.G
C
C
C
Haplotype frequency
Cases
Controls
OR (95% CI)
P-value
0.52
0.05
0.31
0.12
0.73
0.06
0.16
0.05
1.00 (reference)
1.30 (0.64–2.64)
3.54 (2.12–5.92)
3.72 (1.82–7.61)
,0.0001
0.5714
,0.0001
Global P-value for model ,0.0001.
a
Haplotypes with joint/pooled frequencies ,5% were grouped together. No P-value is given for prostate cancer-rare haplotype association, because it is not a single
haplotype.
Discussion
The present study identified two SNPs and a haplotype in the COX-2
promoter that are associated with prostate cancer risk in South African
Coloured men. A previous study reported an increased association
between prostate cancer and having the 1265 A-allele and 899
C-allele and an inverse association with the 297 G-allele (14). In our
study, we found that the 1265 T-allele increased the risk of prostate
cancer, which confirms the association described by Panguluri et al.
(14). However, our study differs marginally because the 899G.C
SNP was excluded from this investigation, an increased risk of prostate cancer was identified with the 1285 G-allele and no statistically
significant association was observed with the 297 G-allele. Our
study demonstrated that the GTC haplotype increases the risk of developing prostate cancer by 3.54-fold in South African Coloured men
compared with the ACC haplotype. We also noted that the ACC
haplotype occurred with a significantly higher frequency in the controls than in the cases. This finding could possibly suggest that this
haplotype could be protective against developing prostate cancer in
South African Coloured men.
The promoter region upstream of the COX-2 transcriptional start
site contains multiple putative transcription factor-binding sites
(1,24). Several studies have described an elevated COX-2 expression
in prostate tumours (8,9,25,26), although one study suggested that
increased COX-2 levels may suppress tumour development (27). In
the present study, two promoter variants were shown to be associated
with prostate cancer. The 1285A.G variant does not alter a promoter transcriptional factor site, whereas the 1265 A-allele putatively eliminates a CCAAT/enhancer-binding protein (C/EBP) a site
and creates a Pit-1a- and Hb-binding site (14). Increased levels of
C/EBPa have been shown to down-regulate COX-2 expression in
mouse skin carcinoma cells (28), whereas another study demonstrated
that up-regulation of COX-2 is associated with decreased levels of
C/EBPa in rat hepatocytes (29). The third polymorphism analyzed
in this study (297C.G) has been shown to create a C/EBP delta
(d)- and a ribosomal DNA-binding site (14). Interestingly, increased
COX-2 expression in mouse skin carcinogenesis has been correlated
with decreased levels of C/EBPa and concomitant increased C/EBPd
expression (28). At present, it is not yet known whether the elimination of a C/EBPa site and creation of a C/EBPd site may independently or synergistically alter the expression of COX-2.
Numerous investigations have shown that sub-Saharan, West and
Central African and African American populations are related genetically (30,31), and suggestions have been made that common susceptibility alleles are probably to be present across different African
populations (32). Increased genetic risk between COX-2 variants
and prostate cancer has been shown in African American (14,16)
and Nigerian men (14); concomitantly, these studies reported reduced
risk between COX-2 variants and prostate cancer in Caucasian men.
High incidence rates of prostate cancer have been detected in South
African Coloured men (33), a unique ethnic group descendant predominantly of Khoi and San (African) inhabitants, with genetic contributions from European settlers (predominantly Dutch, German and
French) and Asian (Indonesian and Madagascan) migrants who inhabited the Western Cape Province of South Africa in the late 1600s
(18). Taken together, the genetic association described in this Southern
African population in conjunction with findings in other populations of
African descent (14,16) might suggest a common causal variant for
prostate cancer in COX-2 or a variant in a nearby gene.
Our study has a number of weaknesses and strengths. The sample
size was relatively small, although it was sufficient to detect small to
moderate associations. Our control group was closely matched to the
cases by age, ethnicity, geography and medical institution to reduce
the likelihood of spurious association due to population stratification.
When we tested a panel of cases and controls for 24 independent
SNPs, we observed significant stratification between the groups. However, even after genomic control inflation of our P-values, the results
remained highly significant. The genetic association was detected in
a male population of mixed African, European and Asian descent and
it should be researched in other South African ethnic groups (Black
African and Caucasian). A limitation of the study is that the SNPs that
were selected only provide a limited coverage of COX-2. Further
investigations should be undertaken on additional COX-2 SNPs, on
variants in genes flanking COX-2 and in other genes in the inflammatory pathway.
In conclusion, we have detected a significant genetic association
between variants in COX-2 and prostate cancer risk in South African
men. These findings provide supporting evidence to the link between
inflammation and prostate cancer. Genotype analyses of additional
gene variants are warranted and the influence of the COX-2 promoter
variants on gene expression requires investigation.
Funding
National Research Foundation, Thuthuka Programme; Medical
Research Council of South Africa; JH Hayes Fund.
Acknowledgements
We would like to thank all the participants of the study, Drs Sı̂an Hemmings
and Craig Kinnear (Division of Molecular Biology and Human Genetics) and
the urologists who facilitated the collection of samples. Even though the work
is supported by the Medical Research Council, the views and opinions expressed are not those of the Medical Research Council but of the authors of
the material produced or publicized.
Conflict of Interest Statement: None declared.
References
1. Smith,W.L. et al. (2000) Cyclooxygenases: structural, cellular, and molecular biology. Annu. Rev. Biochem., 69, 145–182.
2. Tsujii,M. et al. (1995) Alterations in cellular adhesion and apoptosis in
epithelial cells overexpressing prostaglandin endoperoxide synthase 2.
Cell, 83, 493–501.
3. Masferrer,J.L. et al. (2000) Antiangiogenic and antitumor activities of
cyclooxygenase-2 inhibitors. Cancer Res., 60, 1306–1311.
2349
P.Fernandez et al.
4. O’Byrne,K.J. et al. (2001) Chronic immune activation and inflammation as
the cause of malignancy. Br. J. Cancer, 85, 473–483.
5. Half,E. et al. (2002) Cyclooxygenase-2 expression in human breast cancers
and adjacent ductal carcinoma in situ. Cancer Res., 62, 1676–1681.
6. Shariat,S.F. et al. (2003) Cyclooxygenase-2 is highly expressed in carcinoma in situ and T1 transitional cell carcinoma of the bladder. J. Urol., 169,
938–942.
7. Yu,H.P. et al. (2003) Cyclooxygenase-2 expression in squamous dysplasia
and squamous cell carcinoma of the esophagus. Cancer Lett., 198,
193–201.
8. Gupta,S. et al. (2000) Over-expression of cyclooxygenase-2 in human
prostate adenocarcinoma. Prostate, 42, 73–78.
9. Kirschenbaum,A. et al. (2000) Expression of cyclooxygenase-1 and
cyclooxygenase-2 in the human prostate. Urology, 56, 671–676.
10. Zha,S. et al. (2001) Cyclooxygenase-2 is up-regulated in proliferative inflammatory atrophy of the prostate, but not in prostate carcinoma. Cancer
Res., 61, 8617–8623.
11. Fritsche,E. et al. (2001) Functional characterization of cyclooxygenase-2
polymorphisms. J. Pharmacol. Exp. Ther., 299, 468–476.
12. Campa,D. et al. (2004) Association of a common polymorphism in
the cyclooxygenase 2 gene with risk of non-small cell lung cancer.
Carcinogenesis, 25, 229–235.
13. Cox,D.G. et al. (2004) Polymorphisms in prostaglandin synthase 2/cyclooxygenase 2 (PTGS2/COX2) and risk of colorectal cancer. Br. J. Cancer,
91, 339–343.
14. Panguluri,R.C. et al. (2004) COX-2 gene promoter haplotypes and prostate
cancer risk. Carcinogenesis, 25, 961–966.
15. Shahedi,K. et al. (2006) Genetic variation in the COX-2 gene and the
association with prostate cancer risk. Int. J. Cancer, 119, 668–672.
16. Cheng,I. et al. (2007) COX2 genetic variation, NSAIDs, and advanced
prostate cancer risk. Br. J. Cancer, 97, 557–561.
17. Danforth,K.N. et al. (2008) Polymorphic variants in PTGS2 and prostate
cancer risk: results from two large nested case-control studies. Carcinogenesis, 29, 568–572.
18. Nurse,G.T. et al. (1985) The growth of hybrid communities. In Harrison,
G.A. (ed.) The Peoples of Southern Africa and Their Affinities. Clarendon
Press, Oxford, UK, pp. 219–224.
19. Cardon,L.R. et al. (2003) Population stratification and spurious allelic association. Lancet, 361, 598–604.
2350
20. Schaid,D.J. et al. (2002) Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am. J. Hum. Genet., 70, 425–434.
21. Barreiro,L.B. et al. (2006) Promoter variation in the DC-SIGN-encoding
gene CD209 is associated with tuberculosis. PLoS Med., 3, e20.
22. Ziegler,A. et al. (2006) Association analysis in unrelated people. In A
Statistical Approach to Genetic Epidemiology: Concepts and Applications.
Wiley-VCH, Weinheim, Germany, pp. 210–213.
23. Bacanu,S.A. et al. (2000) The power of genomic control. Am. J. Hum.
Genet., 66, 1933–1944.
24. Tazawa,R. et al. (1994) Characterization of the genomic structure, chromosomal location and promoter of human prostaglandin H synthase-2
gene. Biochem. Biophys. Res. Commun., 203, 190–199.
25. Uotila,P. et al. (2001) Increased expression of cyclooxygenase-2 and nitric
oxide synthase-2 in human prostate cancer. Urol. Res., 29, 23–28.
26. Denkert,C. et al. (2007) Overexpression of cyclooxygenase-2 in human
prostate carcinoma and prostatic intraepithelial neoplasia-association with
increased expression of Polo-like kinase-1. Prostate, 67, 361–369.
27. Bol,D.K. et al. (2002) Cyclooxygenase-2 overexpression in the skin of
transgenic mice results in suppression of tumor development. Cancer
Res., 62, 2516–2521.
28. Kim,Y. et al. (1998) Transcriptional regulation of cyclooxygenase-2 in
mouse skin carcinoma cells. Regulatory role of CCAAT/enhancer-binding
proteins in the differential expression of cyclooxygenase-2 in normal and
neoplastic tissues. J. Biol. Chem., 273, 27686–27694.
29. Callejas,N.A. et al. (2000) Regulation of cyclooxygenase 2 expression in
hepatocytes by CCAAT/enhancer-binding proteins. Gastroenterology, 119,
493–501.
30. Chen,Y.S. et al. (2000) mtDNA variation in the South African Kung and
Khwe-and their genetic relationships to other African populations. Am. J.
Hum. Genet., 66, 1362–1383.
31. Ely,B. et al. (2006) African-American mitochondrial DNAs often match
mtDNAs found in multiple African ethnic groups. BMC Biol., 4, 34.
32. Tishkoff,S.A. et al. (2002) Genetic analysis of African populations: human
evolution and complex disease. Nature Rev. Genet., 3, 611–621.
33. Heyns,C.F. et al. (2003) Problems with prostate specific antigen screening
for prostate cancer in the primary healthcare setting in South Africa. BJU
Int., 91, 785–788.
Received June 8, 2008; revised October 16, 2008; accepted October 23, 2008