Allelic Variants of KLK2 Gene Predict Presence of Prostate Cancer at Biopsy

Single nucleotide polymorphisms (SNPs) are the most abundant form of human genetic variations and a resource for mapping complex genetic traits [1,2]. Several eforts, for example the HapMap project, have been made to document the frequencies of various SNPs in diferent ethnic groups and human races and to evaluate their associations with diseases [3]. Currently, there are no clinical tests to evaluate genetic predisposition to prostate cancer risk in men with or without elevated prostate specifc antigen (PSA), abnormal digital rectal examination (DRE), or both. Reports by Nam et al. showed that two SNPs in KLK2 gene: rs2664155 (AA or AG variants) and rs198977 (TT or TC variants) were strongly associated with the presence of prostate cancer at biopsy [4]. Genome-wide association studies have also identifed several groups of SNPs (haplotypes) in diferent genes that are linked to prostate cancer risk [5,6]. However, the SNPs in KLK2 gene warrant further investigation because the protein product of KLK2 gene (hK2) is only secreted by the prostate and its serum levels correlate with prostate cancer development [3]. Biologically, the hK2 activates the PSA, which is involved in the liquefaction of the seminal fuid thereby aiding sperm motility [7].


Introduction
Single nucleotide polymorphisms (SNPs) are the most abundant form of human genetic variations and a resource for mapping complex genetic traits [1,2]. Several eforts, for example the HapMap project, have been made to document the frequencies of various SNPs in diferent ethnic groups and human races and to evaluate their associations with diseases [3]. Currently, there are no clinical tests to evaluate genetic predisposition to prostate cancer risk in men with or without elevated prostate specifc antigen (PSA), abnormal digital rectal examination (DRE), or both. Reports by Nam et al. showed that two SNPs in KLK2 gene: rs2664155 (AA or AG variants) and rs198977 (TT or TC variants) were strongly associated with the presence of prostate cancer at biopsy [4]. Genome-wide association studies have also identifed several groups of SNPs (haplotypes) in diferent genes that are linked to prostate cancer risk [5,6]. However, the SNPs in KLK2 gene warrant further investigation because the protein product of KLK2 gene (hK2) is only secreted by the prostate and its serum levels correlate with prostate cancer development [3]. Biologically, the hK2 activates the PSA, which is involved in the liquefaction of the seminal fuid thereby aiding sperm motility [7].

Objective
Te objective of this study was to fnd out if any of the SNPs in the KLK2 gene could predict the presence of prostate cancer at biopsy. Te second objective was to fnd out if any associated SNP has a correlation with the disease phenotype (tumour grade and/ or pathological grade). Te third objective was to evaluate the performance of a TaqMan SNP genotyping assay for any of the associated SNP and to use the assay for genotyping DNA from archived formalin fxed, parafn-embedded (FFPE) tissue sections.

Subjects
Sixty consecutive patients who consented to the study (afer favourable ethical approval by the Bedfordshire Research Ethics Committee) were sampled for 5 ml of peripheral blood before they had transrectal ultrasound (TRUS) guided prostate biopsy. Most of the patients had elevated PSA level (≥ 4.0 ng/ml) and/ or abnormal DRE at the time of referral for biopsy. Some patients were referred to the clinic due to persistent urinary symptoms. No patient was a known case of prostate cancer. Te patients were Caucasian whites (British). Te blood samples were centrifuged at 1840 g for 5 minutes and plasma removed for PSA testing. Te corpuscular components were lysed in a cold red cell lysis bufer (1.55 M Ammonium chloride, 0.01 M EDTA and 0.1 M Potassium bicarbonate; adjusted pH 7.4 using 10 M HCl). Afer two washes in the lysis bufer at 10 minutes intervals, white cell pellets were fnally washed in phosphate bufered saline (PBS) before lysis in 1 ml of guanidine isothiocyanate (GITC) bufer. Genomic DNA was extracted from 200 μl of the lysates using QiaAmp DNA kit and the Qiacube automated extraction system (Qiagen UK).

FFPE Prostate tissue sections
Archived FFPE prostate tissue blocks (n=138) were retrieved from the Royal Gloucestershire tissue store following favourable ethical approved by the Royal Gloucestershire Research Ethics Committee. Two 25 micrometre thick sections were asceptically cut from each tissue block and picked into a 2 ml tube. Te tubes were briefy centrifuged, deparafnised in two washes of 1 ml xylene for 10 minutes each; the xylene decanted and the tissue rehydrated by two washes in 1 ml of 100% ethanol before allowing the pellets to dry for 5 minutes on a dry heat block kept at 37°C. Te tissue pellets were digested overnight in 540 μl of ATL tissue lysis bufer and 60 μl of proteinase K (Qiagen, UK). Te digest was centrifuged at 800 g for 5 minutes and the supernatants (containing nucleic acids) were collected. A 200 μl of the supernatant was used for DNA extraction. Another 300 μl of the supernatant was used for RNA extraction; and subsequent complementary DNA synthesis using random hexamer priming and the Molony murine leukaemic virus reverse transcriptase (M-MLV RT) in a fnal 50 μl reaction volume according to manufacturer's recommendation (Promega, UK). Te RNA elute was 30 μl and the quality was checked by automated electrophoresis using BioRad Experion system. Quality of gDNA was checked by the amplifcation of a 150 bp G6PD from each sample using 2 μl of the genomic DNA, 300 nM each of forward and reverse primers in a fnal 20 μl reaction volume using a standard polymerase chain reaction (PCR). Te primer sequences and thermal protocol are available on request.

Direct sequencing
Te KLK2 gene was sequenced in four parts covering nearly the entire KLK2 gene. Te Big Dye X terminator purifcation system and the ABI 3130 genetic analyzer (Applied biosystems) were used for the sequencing according the manufacturer's instructions. Te nucleotide sequences for primers have been described in details elsewhere [4]. Te sequences were aligned using the SeqMan component of the DNA star sofware (DNASTAR Inc, USA). Te sequences were correlated with canonical sequences from the UCSC genome browser (http://genome. ucsc.edu/) and Entrez SNP database (http://www.ncbi.nlm.nih.gov/).

Allelic discrimination assays
Using the same genomic DNA from blood samples, predesigned SNP genotyping assays, based on TaqMan VIC-and FAMfuorescent labelled minor groove binding (MGB) probes for the two alleles and specifc primers (Applied biosystems, UK) were used in genotyping SNPs rs198972 and rs198977: the assay identities were C_8705643 and C_736084 respectively. Te ABI 7900HT sequence detection system (Applied biosystems, UK) was used according to manufacturer's instructions. Genomic DNA from the FFPE tissue materials was also genotyped by this method. For each reaction, 50 to 100 ng (1-2 μl) of sample gDNA was used in setting up the real time PCR reaction, for allelic discrimination, in a fnal 25 μl reaction volume using thermal protocol as recommended by the manufacturer (Applied Biosystems).

Relative Quantifcation of KLK2 transcript
Te KLK2, G6PD and ABL1 transcripts in the samples were measured by a triplex real time PCR in a 20 μl reaction volume using 2 μl of cDNA, TaqMan gene expression master mix and TaqMan probes (FAM for ABL1 and TET for KLK2 using a black hole quencher BHQ and CY5 for G6PD with a BHQ). Primer and probe sequences are shown in Table 1. Te thermal profle was 50°C for 2 minutes, initial denaturation at 95°C for 10 minutes; cycle denaturation of 95°C for 15 seconds, annealing and extension at 60oC for 1 minute for 40 cycles. Data was acquired at all stages. Te geometric mean of G6PD and ABL1 genes were used in normalizing the gene expression of KLK2. Te normalized relative quantity (NRQ) of KLK2 was calculated using the formula NRQ=E Ct KLK2 /E Gm of Ct ABL and G6PD where E is the PCR efciency for each target, Gm is the geometric mean of Ct values of ABL1 and G6PD, the endogenous control genes [8]. Figure 1a shows a representative SYBR safe stain of 2% agarose gel of 150 bp G6PD quality control check on gDNA from FFPE tissue materials. Figure 1b shows a representative SYBR safe stain of PCR products prior to setting up Big Dye terminator reaction. Single intense bands yielded good electropherograms during sequencing. Figure 1c shows the normalized relative quantity of KLK2 transcript in benign, non-involved and prostate cancer cases. KLK2 mRNA (normalized) was signifcantly reduced in the prostate cancer cases compared to BPH cases. Table 2 contains a summary of all the 18 known SNPs identifed in the KLK2 gene. Eleven SNPs (61%) of the 18 SNPs identifed in the 60 patients' cohort through direct DNA sequencing had 100% frequency of the reference (ancestral) allele in the study population, therefore were not investigated further. Te remaining 7 SNPs had a heterozygosity of 0.07 to 0.40 in the sample population (n=60). Four of them were intronic SNPs; three were coding sequence (CDS) SNPs. Te CDS located SNPs are shown in Figure 2. SNPs rs10422897 and rs198872 did not show any signifcant association with any prostate disease (P values=0.14 and 0.22 respectively, Chi Square test). Te codons in both SNPs are for Arginine (R) and Leucine (L) respectively. However, the SNP rs198977 had a signifcant association with prostate cancer (P=0.0037, Chi Square test). Te T/T allele of the rs198977 occurred only in prostate cancer cases. Tere were 12 prostate cancer cases in the 60 patients' cohort (representing 20% cancer detection rate by biopsy in this cohort). Four (33%) of the 12 prostate cancer cases had T/T allele of rs198977. At the time of sampling their histology status was not known. Te benign hyperplasia cases were 38 (about 63%) of the cohort. Prostatic intraepithelial neoplasia (PIN) was present in 2 cases (about 3% of the cohort) and there were 8 non-involved cases (cases of chronic infammation, no dysplasia, no carcinoma; about 13% of the cohort).

Results
None of the four intronic SNPs had any signifcant association with prostate cancer (P>0.05, Chi Square test). However, the A/A allele of the SNP rs2664155 occurred only in benignhyperplasia patients    Overall, there were 7 cases of T/T allele of rs198977 identifed in this study using direct DNA resequencing and TaqMan SNP genotyping assay of 198 samples in total. Two of the seven cases had Gleason score 8-10; four cases had Gleason Score 6 and one case had Gleason score 7. Tere were no low grade cancers (Gleason score 4-5) identifed with the T/T allele, indicating a possible association with high tumour grade.

Discussion
Of the 18 known SNPs identifed in the KLK2 gene through direct DNA re-sequencing of 60 patients, only two SNPs showed a signifcant association with prostate diseases. Te T/T allele of rs198977 was signifcantly associated with presence of prostate cancer at biopsy. Te sample size in this study was relatively small. However, in many SNP association studies, categories of prostate diseases such BPH, PIN and chronic infammation are usually all grouped into 'no tumour category' and are compared with 'tumour group' . Tis practice could skew results because the detection rate of prostate cancer by histology in a pool of patients attending TRUS biopsy is at best 30% [9], and in this cohort the cancer detection rate was even lower (20%). Tere was a higher frequency of BPH in prostate lesions. Terefore, larger population studies incorporating these sub-groups are necessary to confrm the association of SNPs with prostate diseases. Contrary to the report by Nam et al., our study showed that the rs2664155 was associated with risk of benign hyperplasia rather than prostate cancer. In addition, our study showed that the occurrence of T/T allele of rs198977 was associated with high tumour grade. Larger studies are still required to confrm these associations. Tis study did not evaluate the association of these SNPs in haplotypes. In addition, genotyping SNPs from archived tissue samples could be problematic since possible somatic mutations were diferentiated from germ line variations.
Te normalized quantifcation of KLK2 showed that the transcript level was signifcantly reduced in prostate cancer cases compared to benign cases (P<0.05). Although the protein product of KLK2 is known to be increased in prostate cancer progression, our study, in agreement with some other studies [10,11], showed that the normalized mRNA level was reduced in prostate cancer. It is therefore unclear how the KLK2 gene is up-regulated during prostate cancer progression.
Te TaqMan SNP genotyping is a cheap, fexibly high throughput and was found suitable for genotyping gDNA from archived parafn tissue materials and blood samples. Both the rs198977 and rs2664155 could be determined in the same plate reaction for each patient sample using the TaqMan SNP assay. However, known allelic variants (calibrators for the three possible variants determined by re-sequencing) must always be included in each run for quality control purposes. Again the amount of gDNA used in the TaqMan real time PCR for genotyping is very important (our optimization assay showed that 50 to 100ng amount of gDNA was required for good quality base calling).
SNPs could provide additional genetic information to help stratify patients into risk groups. It might help identify males who will beneft from prostate cancer annual screening programmes, which are, in many countries, strongly debated [12]. Males who have the T/T allele of rs198977, for example, could be enrolled into annual screening programme before the age of 50 (the recommended age in some countries). SNPs could also inform prognosis, and may help in choice of treatment options. For example, if it is confrmed that the T/T allele of rs198977 was associated with higher tumour grade, this can infuence patients with localised prostate cancers to choose a radical treatment option earlier on. Although the rs198977 SNP is located in the CDS

SNP ID
Molecular type Genotypes and their frequency (n = 60) Reference allele  exons of KLK2 gene, the resulting codon is a 'coding synonymous' , which means there is no change in peptide sequence for whichever allele of this particular SNP the patient has. It is therefore unclear how allelic variation afects the function of the KLK2 gene product. Te NRQ of the KLK2 gene also showed that it was down regulated in prostate cancer cases compared to benign cases.

Conclusion
Two SNPs in the KLK2 gene: rs198977 and rs2664155 were signifcantly associated with prostate cancer risk and benign hyperplasia respectively. Te T/T allele of rs198977 occurred only in prostate cancer cases and was also associated with high tumour grade. Te A/A allele of rs2664155 was, contrary to previous reports, associated with benign nodular hyperplasia. Te normalized mRNA level of KLK2 was signifcantly reduced in prostate cancer tissues compared to benign cases. Te TaqMan SNP genotyping assay provided a cheap and fexibly high throughput assay for clinical allelic discrimination in samples including gDNA from archived FFPE prostate tissue blocks and blood samples.