Wnt antagonist gene polymorphisms and renal cancer

Authors

  • Hiroshi Hirata MD, PhD,

    1. Department of Urology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    Search for more papers by this author
  • Yuji Hinoda MD, PhD,

    1. Department of Oncology and Laboratory Medicine, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan
    Search for more papers by this author
  • Koichi Nakajima MD,

    1. Department of Urology, Toho University Faculty of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Nobuyuki Kikuno MD, PhD,

    1. Department of Urology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    Search for more papers by this author
  • Soichiro Yamamura PhD,

    1. Department of Urology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    Search for more papers by this author
  • Kazumori Kawakami MD, PhD,

    1. Department of Urology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    Search for more papers by this author
  • Yutaka Suehiro MD, PhD,

    1. Department of Oncology and Laboratory Medicine, Yamaguchi University Graduate School of Medicine, Yamaguchi, Japan
    Search for more papers by this author
  • Z. Laura Tabatabai MD,

    1. Department of Pathology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    Search for more papers by this author
  • Nobuhisa Ishii MD, PhD,

    1. Department of Urology, Toho University Faculty of Medicine, Tokyo, Japan
    Search for more papers by this author
  • Rajvir Dahiya PhD

    Corresponding author
    1. Department of Urology, San Francisco Veterans Affairs Medical Center, University of California at San Francisco, San Francisco, California
    • Urology Research Center (112F), Veterans Affairs Medical Center and University of California at San Francisco, 4150 Clement Street, San Francisco, CA 94121
    Search for more papers by this author
    • Fax: (415) 750-6639


Abstract

BACKGROUND:

Epigenetic silencing of several wingless-type mouse mammary tumor virus integration site (Wnt) pathway-related genes has been reported in renal cancer. Except for the T-cell factor 4 gene TCF4, there are no reports regarding Wnt pathway gene polymorphisms in renal cancer. Therefore, the authors of this report hypothesized that the polymorphisms in Wnt signaling genes may be risk factors for renal cancer.

METHODS:

In total, 210 patients (145 men and 65 women) with pathologically confirmed renal cell carcinoma (RCC) and 200 age-matched and sex-matched control individuals were enrolled in this study. We genotyped 14 single nucleotide polymorphisms (SNPs) in 6 genes. including Dickkopf 2 (DKK2) (reference SNP identification number 17037102 [rs17037102], rs419558, and rs447372), DKK3 (rs3206824, rs11022095, rs1472189, rs7396187, and rs2291599), DKK4 (rs2073664), secreted frizzled-related protein 4 (sFRP4) (rs1802073 and rs1802074), mothers against decapentaplegic homolog (SMAD) family member 7 or SMAD7 (rs12953717), and disheveled associated activator of morphogenesis 2 or DAAM2 (rs6937133 and rs2504106) using polymerase chain reaction-restriction fragment length polymorphism analysis and direct sequencing in the patients with RCC and in the healthy, age-matched control group. The relations also were tested between these polymorphisms and clinicopathologic data, including sex, tumor grade, tumor stage, lymph node involvement, distant metastasis, and overall survival.

RESULTS:

A significant decrease in the frequency of the guanine/adenine (G/A) + A/A genotypes in the DKK3 codon 335 rs3206824 was observed in the patients with RCC compared with the control group. The frequency of the rs3206824 (G/A) A-rs7396187 (guanine/cytosine [G/C]) C haplotype was significantly lower in patients with RCC compared with other haplotypes. In addition, DKK3 rs1472189 cytosine/thymine (C/T) was associated with distant metastasis, and, DKK2 rs17037102 G-homozygous patients had a decreased risk for death in multivariate Cox regression analysis.

CONCLUSIONS:

To the authors' knowledge, this is the first report documenting that DKK3 polymorphisms are associated with RCC and that the DKK2 rs17037102 polymorphism may be a predictor for survival in patients with RCC after radical nephrectomy. Cancer 2009. © 2009 American Cancer Society.

Renal cell carcinoma (RCC) is the third leading cause of death among urologic tumors, accounting for approximately 2% of adult malignancies.1 Although the rate of detection of RCC has increased with improved diagnostic techniques, metastatic lesions still are identified at diagnosis in approximately 25% of patients with RCC. Moreover, in patients with RCC, distant metastases sometimes are identified long after surgical removal of the primary tumor. After detecting these metastases, the 5-year survival rate generally is <10%.2 The standard treatment for localized renal cancer is surgical removal,2 whereas immunotherapy is used for metastatic disease because of its multidrug resistance. Interleukin 2 (IL-2) is the most common immunotherapy for RCC but is effective in only 10% to 15% of patients.3

Wingless-type mouse mammary tumor virus integration site (Wnt)/β-catenin signaling is involved in numerous processes in development, is strongly implicated in tumorigenesis, and is highly related to tumor invasion and metastasis.4 Wnt signals transduced by the canonical pathway play a role in determining cell fate, and signals transduced by the noncanonical pathway are important for the control of cell movement and tissue polarity.5

Canonical Wnt ligands bind to frizzled (FZD) family receptors and the low-density lipoprotein receptor-related protein 5 (LRP5)/LRP6 coreceptors, which stabilize β-catenin. Subsequently, β-catenin interacts with members of the lymphoid enhancer factor 1/T-cell factor (LEF1/TCF) family, resulting in the generation of a functional transcription factor complex and the expression of downstream target genes.5 Noncanonical Wnt ligands bind to FZD family receptors and to the receptor tyrosine kinase-like orphan receptor 2 (ROR2) and receptor related to tyrosine kinase (RYK) coreceptors.5-7 This signaling is involved mainly in cytoskeletal reorganization during cancer cell invasion and metastasis.6, 7 Currently, 5 Wnt antagonist families have been described, namely, the secreted frizzled-related protein (sFRP), Wnt inhibitory factor 1 (Wif-1), Xenopus Cerberus, Wise, and Dickkopf (DKK) families.8

The sFRP family (sFRP1-sFRP5) and Wif-1 are involved in inhibiting Wnt signaling by directly binding to Wnt molecules.8 The DKK family (DKK1-DKK4) inhibits Wnt signaling by binding to the LRP5/LRP6 component of the Wnt receptor complex.8

Transforming growth factor β (TGF-β) also regulates cell fate and proliferation cooperatively with Wnt proteins.9, 10 The mothers against decapentaplegic homolog (Smad) proteins transduce signaling from the activated TGF-β/TGF receptor complex.11 Smad proteins, together with β-catenin, act as transcriptional regulators in the TGF-β and Wnt signaling pathways.11

Epigenetic silencing of several Wnt pathway genes has been reported in renal cancer.12-15 However, to our knowledge, there have been no reports regarding Wnt pathway-related gene polymorphisms in renal cancer except for the TCF4 gene.16

Therefore, with this evidence, we hypothesized that gene polymorphisms associated with Wnt (canonical and noncanonical) and TGF-β cascades may be associated with renal cancer risk. To test this hypothesis, we conducted a case-control study with respect to 14 single nucleotide polymorphisms (SNPs), including DKK2 (reference SNP identification number 17037102 [rs17037102], rs419558, and rs447372), DKK3 (rs3206824, rs11022095, rs1472189, rs7396187, and rs2291599), DKK4 (rs2073664), sFRP4 (rs1802073 and rs1802074), disheveled associated activator of morphogenesis 2 (DAAM2) (rs6937133 and rs2504106), and SMAD7 (rs12953717). We selected these polymorphic sites based on previous reports and HapMap data (available at: http://www.hapmap.org/ Accessed on January 3, 2009),17, 18 which were composed of possibly functional SNPs (nonsynonymous and 5′ or 3′ untranslated region SNPs) or disease-associated SNPs. Then, we tested the relation between these polymorphisms and clinicopathologic data, including sex, grade, tumor stage, lymph node involvement, distant metastasis, and overall survival. We also investigated the relation between Wnt antagonist gene polymorphisms and the expression of β-catenin, which is 1 of the downstream targets of Wnt signaling.

MATERIALS AND METHODS

Samples

In total, 210 patients (145 men and 65 women) with pathologically confirmed, conventional RCC and a group of 200 healthy, age-matched and sex-matched controls were enrolled in this study. Genomic DNA was extracted from the peripheral blood of 154 patients and 200 healthy individuals (Shimane University Hospital, Izumo, Japan) and from paraffin-embedded, noncancerous kidney tissues of 56 patients (Toho University Hospital, Tokyo, Japan). A DNA minikit (Qiagen, Valencia, Calif) was used to extract DNA from normal tissue and peripheral blood according to the manufacturer's protocols. The mean ages of the patient and control groups were 62 years and 61 years, respectively (Table 1) (P = .35). All patients (n = 210) who were tested were diagnosed with RCC on the basis of histopathologic findings. They were classified according to World Health Organization criteria and were staged according to the tumor-lymph node-metastasis (TNM) classification. Healthy controls consisted of volunteers with no apparent abnormal findings upon medical examination at Shimane University Hospital. These samples are the same as those reported previously.19

Table 1. Characteristics of Patients with Renal Cell Carcinoma and Controls
CharacteristicNo. (%)P
Cases, n=210Controls, n=200
  1. SD indicates standard deviation; pT, pathologic tumor classification; pN, lymph node invasion; pM, distant metastasis.

Age: Mean±SD, y62.1±1260.1±16.35
Sex   
 Men145 (69)151 (75).23
 Women65 (31)49 (25) 
Grade   
 155 (26)  
 2130 (62)  
 3+425 (12)  
pT classification   
 pT1114 (54)  
 pT245 (21)  
 pT348 (23)  
 pT43 (2)  
pN status   
 Negative200 (95)  
 Positive10 (5)  
pM status   
 Negative191 (91)  
 Positive19 (9)  
Pathology   
 Clear cell carcinoma198 (94)  
 Granular cell carcinoma10 (5)  
 Chromophobe cell carcinoma2 (1)  

To ascertain that volunteers were healthy and free of cancer, they all underwent various tests, including physical examinations, questionnaires concerning their health and history, chest x-rays, blood and urine tests for various tumor markers, and abdominal ultrasound, gastric endoscopy, and colon enema. Peripheral blood samples were obtained from the patients and controls after written informed consent was obtained at Shimane University Hospital and Toho University Hospital.

Genotyping

Polymorphic sites with minor allele frequencies of >0.1 in Japanese populations were selected from HapMap data. Information regarding functional polymorphisms (nonsynonymous, synonymous, and 5′-untranslated and 3′-untranslated regions) are provided in Table 2, and diagrams of the DKK2, DKK3, and sFRP4 genes illustrating their functional domains and polymorphic sites are provided in Figure 1. Polymorphisms were analyzed by using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis. The genotyping methods, primer sets, and annealing temperatures that were used for RFLP are listed in Table 2. Each PCR reaction was carried out in a total volume of 20 μL that consisted of 0.3 μL of a 10 μmol/L solution of each primer, 1.5 mmol/L MgCl2, 0.8 mmol/L deoxynucleotide triphosphate, 0.5 unit RedTaq DNA polymerase (Sigma Chemical Company, St. Louis, Mo), 1 μL of genomic DNA (30 ng/μL), and 15.6 μL H2O using a PTC 200 Thermal Cycler (MJ Research, Waltham, Mass). All reactions were subjected to 2 rounds of amplification using a nested primer approach. The first and second PCR annealing temperatures and PCR cycles were 52°C for 48 cycles and 58°C for 45 cycles, respectively. The second PCR products were digested with each restriction enzyme and temperature (New England BioLabs, Waltham, Mass) for 3 hours, separated on 2% agarose gels, and subsequently stained with ethidium bromide.

Figure 1.

These are diagrams of the genes (1) Dickkopf 2 (DKK2), (2) DKK3, and (3) secreted frizzled-related protein (sFRP4) and their polymorphic sites. rs Indicates reference single nucleotide polymorphism number; Arg, arginine; Glu, glutamic acid; Gly, glycine; Pro, proline; Thr, threonine; Lys, lysine.

Table 2. Information Regarding Single Nucleotide Polymorphism Identification Number, Function Variation, Primers Sequence, Product Size, Polymerase Chain Reaction Conditions, and Restriction Enzymes for the Target Genes
Gene, SNP IDFunction VariationPrimer SequenceProduct Size, bpAnnealing Temperature (No. of Cycles)Restriction Enzyme
  1. SNP ID indicates single nucleotide polymorphism identification number (rs); bp, base pairs; DKK2, Dickkopf 2; F, forward; T, thymine; G, guanine; C, cytosine; A, adenine; R, reverse; Arg, arginine; Glu, glutamic acid; Gly, glycine; UTR, untranslated region; V, valine; P, proline; R, arginine; K, lysine; sFRP, secreted frizzled-related protein; SMAD7, mothers against decapentaplegic homolog (SMAD) family member 7; DAAM2, disheveled associated activator of morphogenesis 2.

DKK2, rs17037102NonsynonymousFirst: F5′-TGGCTTCATATTTCACATCAAGA-3′; R5′-TGTGTGGTCTTCCTAGATTCTGC-3′ 52 (48) 
 Exon 3 Arg146GluSecond: F5′-TGATCATCTCCAGGCATCTG-3′; R5′-ATTCTGCCATCCCAAGTCAT-3′13258 (40)DdeI
DKK2, rs4195583′ UTRFirst: F5′-TACGGACACAGGACCTCACA-3′; R5′-TCCAATATATGTGGGAAAAGAGC-3′ 52 (48) 
  Second: F5′-TCCCCTGGTTTCAAAGATGA-3′; R5′-TGGGAAAAGAGCTAACAGAGAGA-3′25158 (40)AflII
DKK2, rs447372Intron 1First: F5′-TGCTCAAAAAGCAGCATCTC-3′; R5′-AGCTAACGGTCACTCAACCAA-3′ 52 (48) 
  Second: F5′-TCAAAAAGCAGCATCTCAGG-3′; R5′-CTCCAAATTTTGCCCTACCA-3′24958 (40)EcoRV
DKK3, rs3206824NonsynonymousFirst: F5′-GAGGTCCCCGATGAGTATGA-3′; R5′-TAGGAAGAAGCCTGGTCAGC-3′ 52 (48) 
 Exon 7 Arg335Gly    
  Second: F5′-GGTCCCCGATGAGTATGAAG-3′; R5′-AGCACACACCTGGGGAAATA-3′20958 (40)DdeI
DKK3, rs11022095Intron 6 A/GFirst: F5′-TCAAGCAATCCTCCTGCATT-3′; R5′-AGTGGATGGTCCATGGAGAG-3′ 52 (48) 
  Second: F5′-TTACAGGTGTGAGCCACTGC-3′; R5′-CTCCCCAGACCTGTCACAAT-3′28858 (40)BssSI
DKK3, rs14721893′ Near gene C/TFirst: F5′-TGGGGGACCTAGTTTTATCTCA-3′; R5′-ACCATCAGGACCACCCTACA-3′ 52 (48) 
  Second: F5′-CTTCCTCAAGCTTTCCTTGC-3′; R5′-TACCTTGAAGGCATCCCAGT-3′29958 (40)BslI
DKK3, rs7396187Intron 4 G/CFirst: F5′-TTCCTTAGGTCCCTAGGTCCA-3′; R5′-AGGGCAAAGGAGACTCTTCA-3′ 52 (48) 
  Second: F5′-ACAGGGCATGGCAGTTAGAG-3′; R5′-CTCTTCACCCAACAGGCATT-3′24558 (40)Fnu4HI
DKK3, rs2291599Intron 4 C/TFirst: F5′-CAGAGGACATGGGGTGGAT-3′; R5′-CTGCTGCTCTGCTCTCCATT-3′ 52 (48) 
  Second: F5′-CTTGTCCTCCAGGAGTCAGC-3′; R5′-CATCATCGACGAGGACTGTG-3′25058 (40)AvaI
DKK4, rs2073664SynonymousFirst: F5′-GCCATGGCATTACTGCTTTT-3′; R5′-ATTGCTGGTCAATTGGCTTC-3′ 52 (48) 
 Exon 4 V169VSecond: F5′-CTGCGTGCTGTGTCTGTTTT-3′; R5′-AACGCTGGAAGATTTCTGGA-3′29258 (40)EcoNI
sFRP4, rs1802073NonsynonymousFirst: F5′-GAGCACCATAAAGGGGTGAG-3′; R5′-GGGCACATGGCCTTACATAG-3′ 52 (48) 
 Exon 6 P320TSecond: F5′-ACAGCGGAGAACAGTTCAGG-3′; R5′-TGGCCTTACATAGGCTGTCC-3′24658 (40)HinfI
sFRP4, rs1802074NonsynonymousFirst: F5′-AAGAGAGGCTGCAGGAACAG-3′; R5′-TCTGTACCAAAGGGCAAACC-3′ 52 (48) 
 Exon 6 R340KSecond: F5′-AGAGCGGAGAACAGTTCAGG-3′; R5′-TGGCCTTACATAGGCTGTCC-3′24658 (40)EarI
SMAD7, rs12953717Intron 3 C/TFirst: F5′-GTGCCACAGGGTCTCCTTC-3′; R5′-GGATGTGGAGACAATCAGGAA-3′ 52 (48) 
  Second: F5′-GCTTCGTTTCCACCCCTTAG-3′; R5′-AACCCAGGAGCCTCAGAGAT-3′27258 (40)Fnu4HI
DAAM2, rs6937133Intron 1 A/GFirst: F5′-GCACTGGTGCTCACTCCTCT-3′; R5′-TTCCTGCACAGCTGAGTGTC-3′ 52 (48) 
  Second: F5′-TTGCAGTTAAACCTGGGTGA-3′; R5′-TGTTCGCTGCCCAGTTTAGT-3′17558 (40)TaqI
DAAM2, rs2504106Intron 1 C/TFirst: F5′-TGGCCTCTCATACATCATGC-3′; R5′-CCCTCTTTCCTCCTTTTCCA-3′ 52 (48) 
  Second: F5′-GCCCAGAAAAGCCTCAAATA-3′; R5′-GGTCATTTCCCAAATGGTCA-3′21558 (40)SfcI

All assays were conducted blindly without the knowledge of case or control status. Two researchers performed RFLP and reading of the gels. All samples were retested, and the results were 100% concordant. To confirm the genotype ascribed by PCR-RFLP, approximately 50% of PCR sample products were selected randomly and subjected to direct sequencing using an ABI PRISM 377 DNA sequencer (Applied Biosystems, Inc., Foster City, Calif). There were no discrepancies in the results.

Immunohistochemical Study

We performed immunohistochemistry of the Wnt downstream target β-catenin in formalin-fixed, paraffin-embedded specimens using rabbit polyclonal antibody against human β-catenin (no. 9562; Cell Signaling Technology, Beverly, Mass). The staining procedure used was according to a commercial kit (Lab Vision, Fremont, Calif). We investigated the relation between Wnt antagonist gene genotypes and β-catenin expression. The sections were counterstained with Harris hematoxylin. A typical representative staining is shown in Figure 2.

Figure 2.

(a,b) These are representative immunohistochemical stains of β-catenin in renal cell carcinoma (RCC) tissues. RCC cells had mainly cytoplasmic and membranous expression (original magnification, ×200).

Statistical Analysis

The common homozygote was used as a reference to calculate the genotype-specific odds ratio. Hazards ratios (HRs) and 95% confidence intervals (CIs) were calculated from the proportional hazards assumption of the Cox regression model, including multivariate analysis. The probability of overall survival was estimated using Kaplan-Meier plots and log-rank tests. Hardy-Weinberg equilibrium and haplotype analysis were evaluated by SNPAlyze version 2.2 (DYNACOM, Tokyo, Japan), using an electron microscopic method. The chi-square test was used to compare the genotype frequency between patients and controls. All statistical analyses were performed using StatView software (version 5; SAS Institute Inc., Cary, NC).

We adopted false discovery rate according to Hochberg and Benjamini20, 21 and set the statistically significant level at P < .05. A false discovery rate of 0.05 was used as a critical value for the assessment whether the obtained P value was significant. The genotype frequencies of the polymorphisms in control samples (n = 200) and in case samples (n = 210) did not deviate from Hardy-Weinberg equilibrium (P > .05).

RESULTS

Comparison of Genotype Distribution Between Patients With Renal Cell Carcinoma and Controls

The genotype distributions of the DKK2 (rs17037102, rs419558, rs447372), DKK3 (rs3206824, rs11022095, rs1472189, rs7396187, rs2291599), DKK4 (rs2073664), sFRP4 (rs1802073, rs1802074), SMAD7 (rs12953717), and DAAM2 (rs6937133, rs2504106) polymorphisms between patients with RCC and healthy controls are listed in Table 3. A significant decrease in the frequency of the guanine/adenine (GA)+AA genotypes of DKK3 rs3206824 (nonsynonymous arginine 335 glycine [Arg335Gly]) was observed in patients with RCC compared with controls (OR, 0.43; 95% CI, 0.29-0.65) (Table 3).

Table 3. Association Between Polymorphisms in Wingless-Type Mouse Mammary Tumor Virus Integration Site (Wnt) Signaling Pathway Genes and Renal Cancer
Gene Polymorphism/GenotypeNo. (%)OR [95% CI]PFDR-Adjusted P*
Renal Cancer, n=210Control, n=200
  • OR indicates odds ratio; 95% CI, 95% confidence interval; FDA, false discovery rate; DKK2, Dickkopf 2; rs, single nucleotide polymorphism identification number; G, guanine; Ref, reference category; A, adenine; C, cytosine; T, thymine; sFRP, secreted frizzled-related protein; SMAD7, mothers against decapentaplegic homolog (SMAD) family member 7; DAAM2, disheveled associated activator of morphogenesis 2.

  • *

    FDR-adjusted P values were calculated using the Benjamini-Hochberg method, and values <.05 were considered statistically significant.

  • Remained significant after FDR adjustment.

DKK2, rs17037102     
 GG105 (50)106 (53)1 [Ref]  
 GA92 (44)82 (41)1.13 [0.76-1.69].54 
 AA12 (6)12 (06)1.01 [0.43-2.34].98 
 GA+AA104 (50)94 (47)1.11 [0.76-1.65].57.61
DKK2, rs419558     
 CC119 (57)88 (44)1 [Ref]  
 CT63 (30)88 (44)0.53 [0.34-0.81].003 
 TT28 (13)24 (12)0.86 [0.45-1.58].64 
 CT+TT91 (43)112 (56)0.60 [0.41-0.88].01.028
DKK2, rs447372     
 GG117 (56)88 (44)1 [Ref]  
 GA65 (31)88 (44)0.56 [0.36-0.84].0063 
 AA28 (13)24 (12)0.88 [0.47-1.62].67 
 GA+AA93 (44)112 (56)0.62 [0.42-0.92].01.0238
DKK3, rs3206824     
 GG150 (71)104 (52)1 [Ref]  
 GA53 (25)82 (41)0.45 [0.29-0.69].0001 
 AA7 (4)14 (7)0.35 [0.14-0.89].02 
 GA+AA60 (29)96 (48)0.43 [0.29-0.65].0001.0014
DKK3, rs11022095     
 GG28 (13)23 (11)1 [Ref]  
 GA88 (42)87 (44)0.83 [0.44-1.55].56 
 AA94 (45)90 (45)0.85 [0.46-1.59].63 
 GA+AA182 (87)177 (89)0.84 [0.47-1.52].57.57
DKK3, rs1472189     
 CC155 (74)134 (67)1 [Ref]  
 CT51 (24)59 (30)0.74 [0.48-1.16].19 
 TT4 (2)7 (3)0.49 [0.14-1.72].26 
 CT+TT55 (26)66 (33)0.72 [0.47-1.10].13.18
DKK3, rs7396187     
 GG138 (66)102 (51)1 [Ref]  
 GC58 (27)75 (38)0.57 [0.37-0.88].01 
 CC14 (7)23 (11)0.45 [0.22-0.92].02 
 GC+CC72 (34)98 (49)0.54 [0.37-0.81].002.01
DKK3, rs2291599     
 CC101 (48)113 (57)1 [Ref]  
 CT81 (39)62 (31)1.46 [0.95-2.24].08 
 TT28 (13)25 (12)1.25 [0.69-2.29].46 
 CT+TT109 (52)87 (43)1.41 [0.95-2.07].08.14
DKK4, rs2073664     
 CC144 (68)145 (72)1 [Ref]  
 CT40 (19)46 (23)0.88 [0.54-1.42].58 
 TT26 (13)9 (5)2.91 [1.32-6.42].006 
 CT+TT66 (32)55 (28)1.21 [0.79-1.85].38.48
sFRP4, rs1802073     
 CC51 (24)36 (18)1 [Ref]  
 CA82 (39)112 (56)0.52 [0.31-0.86].01 
 AA77 (37)52 (26)1.04 [0.61-1.82].87 
 CA+AA159 (76)164 (82)0.68 [0.42-1.11].12.19
sFRP4, rs1802074     
 AA36 (17)18 (9)1 [Ref]  
 AG69 (33)69 (34)0.51 [0.26-0.96].04 
 GG105 (50)113 (56)0.46 [0.25-0.87].01 
 AG+GG174 (83)182 (91)0.48 [0.26-0.87].01.04
SMAD7, rs12953717     
 CC125 (59)139 (70)1 [Ref]  
 CT68 (32)48 (24)1.57 [1.01-2.45].04 
 TT17 (9)13 (6)1.45 [0.68-3.11].33 
 CT+TT85 (41)61 (30)1.55 [1.03-2.33].03.06
DAAM2, rs6937133     
 GG30 (14)33 (16)1 [Ref]  
 GA65 (31)65 (33)1.1 [0.61-2.01].75 
 AA115 (55)102 (51)1.24 [0.71-2.17].45 
 GA+AA180 (86)167 (84)1.18 [0.69-2.03].53.62
DAAM2, rs2504106     
 TT72 (34)37 (17)1 [Ref]  
 TC84 (40)118 (59)0.36 [0.23-0.59].0001 
 CC54 (26)45 (23)0.62 [0.35-1.08].08 
 TC+CC138 (66)163 (82)0.44 [0.28-0.69].0002.0014

We also observed a significant decrease in the frequency of guanine/cytosine (GC)+CC genotypes of DKK3 rs7396187 in patients with RCC (OR, 0.54; 95% CI, 0.37-0.81) (Table 3). There also was a significantly increased frequency of the AA genotype of the rs1802074 SNP in the sFRP4 gene (OR, 2.15; 95% CI, 1.15-4.02), an increased frequency of the thymine/thymine (TT) genotype of rs2073664 in DKK4 (OR, 2.91; 95% CI, 1.32-6.42), and a significant increase in the TC+CC genotypes of DAAM2 rs2504106 (OR, 0.44; 95% CI, 0.28-0.69) in patients with RCC (Table 3). However, no significant difference was observed in the genotype distribution of other genotypes between patients and controls (Table 3).

Next, we investigated gene-gene interaction using DKK2 rs419558, DKK2 rs447372, DKK3 rs3206824, DKK3 rs7396187, SMAD7 rs12953717, sFRP4 rs1802074, and DAAM2 rs2504106, because we observed a significant difference in the genotype distribution between cases and controls. Among these combinations, when the combined effect of 2 polymorphisms (DKK3 rs3206824 and sFRP4 rs1802074) was evaluated, a decreased risk of RCC was observed for only the DKK3 rs3206824 GA+AA and sFRP4 rs1802074 AG+GG genotypes (OR, 0.19; 95% CI, 0.09-0.45 [P < .0001]) (Table 4).

Table 4. Gene-Gene Interaction Analysis of 2 Polymorphisms (Dickkopf 3 Single Nucleotide Polymorphism No. 3206824 [rs3206824] and Secreted Frizzled-Related Protein 4 rs802074)
DKK3 (GG/GA+AA)-sFRP4 (AA/AG+GG)*No. (%)OR [95% CI]P
Renal Cancer, n=210Control, n=200
  • DKK3 indicates Dickkopf 3; G, guanine; A, adenine; sFRP4, secreted frizzled-related protein 4; OR, odds ratio; CI, confidence interval.

  • *

    DKK3 rs3206824 and sFRP4 rs802074.

GG-AA27 (13)9 (4.5)1 [Reference] 
GG-AG/GG123 (58)95 (47.5)0.43 [0.19-0.96].04
GA/AA-AA9 (5)9 (5)0.33 [0.10-0.20].06
GA/AA-AG/GG51 (24)87 (43.5)0.19 [0.09-0.45]<.0001

Linkage Disequilibrium and Haplotype Analysis in DKK3 and DKK2 Gene Polymorphisms

The DKK3 rs3206824 polymorphism was in linkage disequilibrium (LD) with rs7396187 (D = 0.1627). Therefore, the frequency of each haplotype, including rs3206824 (GA) and rs7396187 (GC), was calculated between RCC patients and controls. The frequency of the A/C haplotype was significantly lower in RCC (P < .0001) compared with other haplotypes (Table 5).

Table 5. Haplotype Analysis and Renal Cancer Risk
HaplotypeOverall*Cases*Controls*P
  • DKK3 indicates Dickkopf 3; rs, reference single nucleotide polymorphism identification number; G, guanine; A, adenine; C, cytosine; T, thymine.

  • *

    Values are n (%).

DKK3 rs3206824 (G/A)-rs7396187(G/C)    
 G-G71.976.667.0<.001
 A-C18.713.224.5<.001
 G-C6.57.45.6.3155
 A-G2.92.82.9.9451
DKK2 rs419558-rs447372    
 C-G68.771.266.0.109
 T-A31.128.334.0.0796
 C-A0.20.50.167

The DKK2 rs419558 polymorphism was in LD with rs447372 (D = 0.2244). However, we did not observe a significant difference between cases and controls in the haplotype analysis using rs419558 or rs447372 (Table 5).

Relation Between Genotype Distribution and Clinicopathologic Characteristics

We investigated the effect of the 12 SNPs on clinicopathologic factors, including sex, grade, pathologic tumor classification (pT), pathologic lymph node status (pN), and pathologic metastasis status (pM). Regarding sex, pT, and pN, there were no significant effects of SNPs (Table 6). We observed a higher frequency of CT+TT genotypes in DKK3 rs1472189 among patients who had distant metastasis (pM positive).

Table 6. Comparison Between Gene Genotypes and Clinical Parameters*
Gene/GenotypeSexGradeStageLymph Node InvasionDistant Metastasis
Women, n=65Men, n=145P1/2, n=1853/4, n=25PpT1/pT2, n=159pT3/pT4, n=51PNegative, n=200Positive, n=10PNegative, n=191Positive, n=19P
  • pT indicates pathologic tumor classification; DKK2, Dickkopf 2; rs, single nucleotide polymorphism identification number; G, guanine; Ref, reference category; A, adenine; C, cytosine; T, thymine; sFRP, secreted frizzled-related protein; SMAD7, mothers against decapentaplegic homolog (SMAD) family member 7; DAAM2, disheveled associated activator of morphogenesis 2.

  • *

    Considering multiple comparisons.

  • P values <.05 were considered statistically significant.

DKK2, rs17037102              
 GG3570Ref9213Ref7926Ref996Ref9411Ref
 GA2468.278210.747022.89893.41857.49
 AA67.36112.76103.89121.78121.75
 GA+AA3075.459312.838025871014.52978.47
DKK2, rs419558              
 CC3980Ref10712Ref9128Ref1181Ref1136Ref
 CT1746.42567.834815.96576.004549.03
 TT919.94226.11208.58253.004244.08
 CT+TT2665.517813.356823.77829.0027813.02
DKK2, rs447372              
 GG3978Ref10512Ref9027Ref1161Ref1116Ref
 GA1748.31587.914916.82596.005569.04
 AA919.91226.11208.54253.004244.09
 GA+AA2667.418013.416924.65849.0038013.03
DKK3, rs3206824              
 GG50100Ref13614Ref11634Ref1437Ref13812Ref
 GA1340.234310.063716.28503.77467.26
 AA25.7961.6661.6170 70 
 GA+AA1545.244911.074317.39573.92537.41
DKK3, rs11022095              
 GG1117Ref253Ref217Ref271Ref235Ref
 GA2860.477612.686721.91835.61808.21
 AA2668.248410.997123.95904.87886.06
 GA+AA54128.3116022.8313844.931739.7516814.08
DKK3, rs1472189              
 CC47108Ref13817Ref11936Ref1505Ref1469Ref
 CT1635.89465.823813.74465.054110.003
 TT22.4113.000122.2140 40 
 CT+TT1837.74478.484015.55505.074510.005
DKK3, rs7396187              
 GG4692Ref12711Ref10830Ref1326Ref12711Ref
 GC1642.434612.014018.11544.46517.36
 CC311.36122.42113.92140 131.92
 GC+CC1953.315814.0145121.14684.69648.45
DKK3, rs2291599              
 CC3467Ref938Ref7625Ref956Ref9011Ref
 CT2655.827011.216219.84774.77747.61
 TT523.11226.04217.98280 271.24
 CT+TT3178.419217088326.881054.441018.37
DKK4, rs2073664              
 CC40104Ref12717Ref10737Ref1377Ref12915Ref
 CT1822.03355.913010.92373.52364.94
 TT719.93233.96224.26260 260 
 CT+TT2541.14588.955214.48633.92624.31
sFRP4, rs1802073              
 CC1536Ref429Ref3714Ref474Ref437Ref
 CA2557.897012.646517.37793.29766.21
 AA2552.71734.025720.85743.34716.26
 CA+AA50109.7814316.1512237.541536.2414712.17
sFRP4, rs1802074              
 AA1521Ref324Ref288Ref351Ref351Ref
 AG1554.03654.335316.92672.97645.35
 GG3570.368817.097827.68987.389213.09
 AG+GG50124.1315321.8713143.751659.5415618.15
SMAD7, rs12953717              
 CC4679Ref10817Ref9431Ref1196Ref11114Ref
 CT1355.01617.515117.98653.91653.11
 TT611.91161.37143.52161.85152.95
 CT+TT1966.03778.36652083814.97805.19
DAAM2, rs6937133              
 GG1218Ref264Ref228Ref291Ref264Ref
 GA2045.37596.544817.96623.77569.95
 AA3382.2310015.978926.641096.671096.12
 GA+AA53127.2515921.7913743.741719.6916515.38
DAAM2, rs2504106              
 TT1953Ref648Ref5715Ref684Ref666Ref
 TC2955.277311.716222.43804.837311.34
 CC1737.53486.994014.51522.63522.29
 TC+CC4692.3112117.7910236.391326.6912513.79

Relation Between Various Wnt-Antagonist Gene Genotypes and Expression of β-Catenin

We investigated the relation between Wnt-antagonist gene genotypes and β-catenin expression. However, we did not observe any correlation between these polymorphisms and β-catenin expression (Table 7).

Table 7. Relation Between Distribution of Wingless-Type Mouse Mammary Tumor Virus Integration Site (Wnt) Antagonist Genes and Expression of β-Catenin
Polymorphism/Genotypeβ-Catenin IHC Expression: No. (%)P
No, n=25Yes, n=31
  1. IHC indicates immunohistochemistry; DKK2, Dickkopf 2; rs, reference single nucleotide polymorphism identification number; G, guanine; A, adenine; C, cytosine; T, thymine; sFRP4, secreted frizzled-related protein 4.

DKK2, rs17037102   
 GG18 (72)21 (68)Reference
 GA5 (20)9 (29).49
 AA2 (8)1 (3).49
 GA+AA7 (28)10 (32).73
DKK2, rs419558)   
 CC16 (64)22 (71)Reference
 CT6 (24)3 (10).18
 TT3 (12)6 (19).63
 CT+TT9 (36)9 (29).58
DKK2, rs447372   
 GG16 (64)22 (71)Reference
 GA7 (28)3 (10).12
 AA2 (8)6 (19).37
 GA+AA9 (36)9 (29).58
DKK3, rs3206824   
 GG21 (84)24 (77)Reference
 GA4 (16)6 (19).72
 AA0 (0)1 (4) 
 GA+AA4 (16)7 (23).54
DKK3, rs11022095   
 GG5 (20)4 (13)Reference
 GA7 (35)10 (32).48
 AA13 (45)17 (55).52
 GA+AA20 (80)27 (87).47
DKK3, rs1472189   
 CC23 (92)23 (74)Reference
 CT2 (8)8 (26).08
 TT0 (0)0 
 CT+TT2 (8)8 (26).08
DKK3, rs7396187   
 GG21 (84)23 (74)Reference
 GC4 (16)5 (16).85
 CC0 (0)3 (10) 
 GC+CC4 (16)8 (26).38
DKK3, rs2291599   
 CC11 (44)15 (48)Reference
 CT9 (36)11 (35).86
 TT5 (20)5 (16).67
 CT+TT14 (56)16 (52).74
DKK4, rs2073664   
 CC16 (64)21 (68)Reference
 CT6 (24)3 (10).21
 TT3 (12)7 (22).45
 CT+TT9 (36)10 (32).77
sFRP4, rs1802073   
 CC6 (24)10 (32)Reference
 CA8 (32)11 (35).78
 AA11 (44)10 (32).37
 CA+AA19 (76)21 (68).49
sFRP4, rs1802074   
 AA8 (32)9 (29)Reference
 AG12 (48)11 (35).75
 GG5 (20)11 (35).35
 AG+GG17 (68)22 (71).82

Multivariate Cox Proportional Hazards Analysis of Overall Survival in Patients With Renal Cell Carcinoma

The prognostic values for overall survival using parameters, such as sex, age at diagnosis, tumor grade, and pTNM, in 14 SNPs were analyzed using Kaplan-Meier survival curves, and we observed that higher grade (grade 3 or 4), pT3 or pT4 tumors, positive pN status, positive pM status, and the DKK2 rs17037102 GA+AA genotype were associated with shorter survival (Fig. 3). In addition to pT3 or pT4 tumors, positive pN status, and positive pM status, the DKK2 rs17037102 GG genotype was identified as an independent, favorable factor for survival in multivariate analysis (OR, 0.407; 95% CI, 0.156-1.062 [P = .0489]) (Table 8).

Figure 3.

These Kaplan-Meier survival curves illustrate cumulative survival (Cum Survival) in 160 patients with renal cell carcinoma (RCC) according to (1) grade, (2) pathologic tumor classification (pT), (3) pathologic lymph node classification (pN), (4) pathologic metastasis classification (pM), and (5) the Dickkopf 2 (DKK2) single nucleotide polymorphism (no. rs17037102). Higher grade (grade 3 and;4), pT3 and pT4 tumor status, positive pN status (+), positive pM status, and the DKK2 rs17037102 guanine/adenine (G/A)+A/A genotypes were independent predictors of overall survival.

Table 8. Multivariate Cox Proportional Hazards Analysis of Overall Survival After Radical Nephrectomy (n=160)
ParametersHR (95% CI)P
  1. HR indicates hazards ratio; 95% CI, 95% confidence interval; pT, pathologic tumor classification; pN, lymph node invasion; pM, distant metastasis; DKK2, Dickkopf 2; rs, reference single nucleotide polymorphism identification number; G, guanine; A, adenine.

Grade 3 vs grade 1/21.515 (0.411-5.584).5324
pT: pT3/pT4 vs pT1/pT24.017 (1.469-10.986).0067
pN: Positive vs Negative8.885 (1.368-57.713).0221
pM: Positive vs Negative7.312 (2.003-26.692).0026
DKK2 rs17037102: G/G vs G/A+A/A0.407 (0.156-1.062).0489

DISCUSSION

In this study, we observed a significant decrease in the frequency of the GA+AA genotypes of DKK3 rs3206824 (nonsynonymous Arg335Gly) in cases (patients with RCC) compared with controls (OR, 0.43; 95% CI, 0.29-0.65). The functions of DKK1 and DKK3 have been investigated, and various studies have indicated that they are tumor suppressor genes.12, 22-32

DKK3 is a member of the Dickkopf family and regulates cell proliferation and apoptosis as a tumor suppressor gene both in vitro and in vivo.22, 23 DKK3 messenger RNA (mRNA) and protein levels were decreased significantly in human renal cancer tissues compared with normal kidney tissues.22 Regarding the mechanism of down-regulation of DKK3, Kobayashi et al detected hypermethylation in the promoter region in human cancer cell lines in which the expression of DKK3 was decreased.25 Those authors also detected a codon 335 SNP (rs3206824) in the DKK3 gene in mutation analysis and compared the distribution of genotypes between 200 healthy controls and 56 patients with lung cancer. However there was no significant difference between those 2 groups.25 Except for that previous report, to our knowledge, there have been no significant findings regarding codon 335 in SNP studies.

As a next step, we selected 4 polymorphic sites in the DKK3 gene based on HapMap data and conducted LD and haplotype analyses. The DKK3 rs3206824 (nonsynonymous Arg335Gly) was in LD with rs7396187, and the frequency of each haplotype, including rs3206824 (GA) and rs7396187 (GC), was calculated between cases and controls. The frequency of the A-C haplotype was significantly lower in RCC (P < .0001) compared with other haplotypes. An apparent added effect also was observed in the haplotype analysis. To our knowledge, this is the first report indicating that there is an association between the DKK3 gene haplotype and renal cancer.

We also selected 3 polymorphic sites in the DKK2 gene, and DKK2 rs419558 was in LD with rs447372, but there was no significant difference in the frequency of each haplotype, including these 2 polymorphisms between cases and controls. It is noteworthy that the DKK2 rs17037102 (nonsynonymous arginine 146 glutamic acid [Arg146Glu]) GA+AA genotype was associated with shorter survival in all patients. In addition, the DKK2 rs17037102 GG genotype was a favorable factor for survival in multivariate Cox regression analysis, suggesting that the DKK2 rs17037102 GA+AA genotypes may be useful parameters for identifying patients who have high-risk RCC.

However, this polymorphism was not associated with other clinical variables, including grade, stage, lymph node status, or metastases. Prognostic factors usually are associated with some clinicopathologic features. However, it has been demonstrated that functions of the Wnt signaling pathway are extremely diverse,33 suggesting that, currently it is not known whether conventional clinicopathologic factors reflect all of these functions. This will require further study. However, a recent report has indicated that an SNP can be associated with metastasis but not with prognosis.34

To the best of our knowledge, there are no reports regarding the DKK2 gene polymorphism in various cancers, including renal cancer; therefore, we could not compare our results with the findings from other studies. Regarding DKK2 function, Kremen proteins, which are Dickkopf receptors, modulate DKK2 activity during Wnt/LRP6 signaling.35, 36 DKK2 can either activate or inhibit the Wnt/β-catenin pathway, depending on cellular context.4 The role of DKK2 itself on deregulation in cancer is not well understood, because there are no other reports regarding DKK2 SNPs and cancer susceptibility. There was a marginal significance to the increased frequency of the AA genotype of rs1802074 (sFRP4) in patients with renal cancer compared with the control group (Table 3).

We tested for gene-gene interactions using DKK2 rs419558, DKK2 rs447372, DKK3 rs3206824, DKK3 rs7396187, DKK4 rs2073664, sFRP4 rs1802074, and DAAM2 rs2504106, and we observed significant differences the distribution of genotypes between cases and controls. There was a strong correlation, however, between DKK3 rs3206824 GA+AA and sFRP4 rs1802074 AG+GG (Table 4) suggesting that these 2 polymorphisms are related to RCC susceptibility.

It commonly is reported that the sFRP family is down-regulated by epigenetic inactivation in various cancers.25, 37-39 Urakami et al also investigated the methylation frequency of the sFRP family (sFRP1-sFRP5) in renal cancer tissue and adjacent normal kidney tissues and observed that the methylation level of sFRP1 was significantly higher in renal cancer tissues.12 In other SNP studies of the sFRP family, Caldwell et al reported 1 polymorphic site on exon 1 in sFRP1 but no significant association with the development of colorectal cancer.38 The sFRP4 protein directly binds Wnt7a to inhibit activation of β-catenin/canonical signaling40, 41 and was reportedly was down-regulated in renal cancer tissues by epigenetic mechanisms.12 However, currently, there are no reports investigating the potential effect of sFRP4 gene polymorphisms on RCC. Recently, Lee et al reported that DKK3 was a negative regulator of β-catenin.42

To investigate the effect of Wnt-antagonist polymorphisms on Wnt signaling, we did an immunohistochemical analysis of β-catenin expression and compared the relation between β-catenin expression levels and Wnt-antagonist polymorphisms. However, we did not observe any correlation between β-catenin expression and the polymorphisms. The detailed molecular mechanisms involved in how these polymorphisms have an effect on renal cancer are unclear. Nonsynonymous SNPs (nsSNPs) introduce amino acid changes and may affect protein function.43 Therefore, it is believed generally that nsSNPs may be associated with cancer susceptibility.43 PolyPhen (available at: http://genetics.bwh.harvard.edu/pph/Accessed on January 3, 2009) is a computer program that to predicts the effect of coding nsSNPs on protein structure and function. When we used this program, DKK3 Arg335Gly (rs3206824), DKK2 Arg146Glu (rs17037102), and sFRP4 arginine 340 lysine (Arg340Lys) (rs1802074) were judged benign, whereas sFRP4 proline 320 threonine (Pro320Thr) (rs1802073) had a high PolyPhen score (1.577) and was judged “possibly damaging.” Also, the x-ray repair complementing defective repair in Chinese hamster cells 1 gene (XRCC1) arginine 299 glutamine (Arg399Gln) reportedly was associated with the survival and prognosis of various cancers yet was judged benign by the PolyPhen program.44, 45 In addition, the synonymous SNPs can alter mRNA folding and can reduce mRNA stability, thereby altering translation through structural changes in the RNA.46 There is accumulating evidence regarding the functional effects of synonymous mutations.47, 48

The polymorphisms associated with the “odds” of RCC are not associated with clinical or pathologic factors or with survival. Similarly, the polymorphisms associated with clinical and pathologic features are not associated with survival. It is reasonable to consider that the functional role of an SNP as a risk factor is not always the same as that of a prognostic factor, because a risk SNP may contribute to the early stage of carcinogenesis of nearly normal cells, whereas a prognostic SNP may be involved in the progression of fully transformed cells. Indeed, there have been many examples that a risk SNP is of no significance as a prognostic SNP, and vice versa.49, 50

In conclusion, to our knowledge, this is the first report documenting that the DKK2 rs17037102 polymorphism may be a predictor for survival after radical nephrectomy. Although further studies with a larger sample size are necessary, the current findings contribute to an understanding of individual survival differences after nephrectomy.

Acknowledgements

We thank Dr. Roger Erickson for his support and assistance with the preparation of the article.

Conflict of Interest Disclosures

Supported by grants RO1CA130860, RO1CA111470, RO1CA108612, and T32-DK07790 from the National Institutes of Health; by the Veterans Affairs Research Enhancement Award Program; Merit Review grants; and the Yamada Science Foundation.

Ancillary