Single N ucleotide P olymorphism o f t he H uman K allikrein-2 Gene H ighly C orrelates W ith S erum H uman K allikrein-2 Levels a nd i n C ombination E nhances P rostate Cancer D etection

Purpose: We examined the relationship between a mutant (T) for wild-type (C) allele substitution of the human kallikrein-2 gene (KLK2), circulating human kallikrein-2 (hK2) levels and prostate cancer risk. Patients and Methods: We studied 1,287 consecutive men who underwent prostate biopsies because of an abnormal prostate-specific antigen level. Serum and DNA were obtained before biopsy. Cases were patients with cancer, and controls were patients with no cancer. The mutant and wild-type alleles of the KLK2 gene were designated as the T and C alleles, respectively. Results: Of the 1,287 men, 616 had cancer, and 671 had no cancer. The overall distribution of the CC, CT, and TT KLK2 genotypes was 55.1%, 38.2%, and 6.8%, respectively. The median hK2 levels for men with the CC, CT, and TT genotypes were 0.24, 0.18, and 0.062 ng/mL and correlated with the genotypes, respectively (P .0001). The adjusted odds ratios for prostate cancer for patients with the TT and CT genotypes compared with patients with the CC genotype, were 2.13 (95% confidence interval [CI], 1.3 to 3.5; P .004) and 1.51 (95% CI, 1.2 to 2.0; P .002), respectively. The adjusted odds ratio for prostate cancer for patients in the fourth quartile of hK2 compared with the first quartile was 4.33 (95% CI, 2.9 to 6.4; P .0001). When combined, the adjusted odds ratio for having prostate cancer was 13.92 (95% CI, 6.6 to 29.2; P .0001) for patients with high hK2 levels and at least one T allele. Conclusion: The C/T polymorphism of the KLK2 gene and circulating levels of hK2 are correlated and, in combination, are highly predictive for prostate cancer. J Clin Oncol 21:2312-2319. © 2003 by American Society of Clinical Oncology.