Human prostate cancers frequently show loss of heterozygosity (LOH) at loci on the long arm of chromosome 16 (16q). In this study, we analyzed prostate cancer specimens from 48 patients (Stage B, 20 cases; Stage C, 10 cases; cancer death, 18 cases) for allelic loss on 16q, using either restriction fragment length polymorphism (RFLP)- or polymerase chain reaction (PCR)-based methods. Allelic losses were observed in 20 (42%) of 48 cases, all of which were informative with at least one locus. Detailed deletion mapping identified three distinct commonly deleted regions on this chromosome arm: q22.1-q22.3, q23.2-q24.1, and q24.3-qter. On the basis of a published sex-averaged framework map, the estimated sizes of the commonly deleted regions were 4.7 (16q22.1-q22.3), 17.2 (16q23.2-q24.1) and 8.4 cM (16q24.3-qter). Allelic losses on 16q were observed more frequently in the cancer-death cases (11 of 18; 61%) than in early-stage tumor cases (9 of 30; 30%; P < 0.05). In 7 of 11 patients from whom DNA was available from metastatic cancers as well as from normal tissues and primary tumors, the primary cancer foci had no detectable abnormality of 16q, but the metastatic tumors showed LOH. These results suggest that inactivation of tumor suppressor genes on 16q plays an important role in the progression of prostate cancer. We also analyzed exons 5-8 of the E-cadherin gene, located at 16q22.1, in tumor DNA by means of PCR-single strand conformation polymorphism and direct sequencing, but we detected no somatic mutations in this candidate gene.

DNA samples from tumors and paired normal tissues from 48 patients with prostate cancer (stage B, 16 cases; stage C, 14 cases; stage D, 18 cases) were examined with 26 polymorphic markers spanning chromosome 10. Allelic losses were observed in 17 of the 46 cases (37%) that were informative with at least one of the markers. Detailed deletion mapping identified two distict commonly deleted regions on the long arm of chromosome 10 (10q22-q24:7 cM and 10q25.1:17 cM) and one on 10p, suggesting that at least three tumor suppressor genes associated with prostate cancer are present on this chromosome. We observed loss of heterozygosity more frequently in tumors from fatal cases (stage D, 8/16, 50%) than in localized tumors (stage B, 0/16, 0%; P = 0.001 or stage B + C, 5/30, 17%; P = 0.02 Fisher's exact test). All metastatic tissues showed allelic loss at one or more loci on 10q. In five of the nine patients from whom DNAs were available from both metastatic and primary tumors, the primary cancer foci had no detectable abnormality of chromosome 10, while the metastatic foci showed allelic loss on chromosome 10. These results suggested that inactivation of one or more tumor suppressor genes on chromosome 10 plays an important role in late stages of prostate cancer.

Frequent allelic losses on chromosome arm 13q are observed in carcinomas of the head and neck, breast, ovary, and pituitary gland. We analyzed 59 primary prostate tumors (stage B, 18 patients; C, 12 patients; D1, 4 patients; and endocrine therapy-resistant cancer death, 25 patients), as well as 18 metastatic tissues from 14 of the 25 cancer death patients for loss of heterozygosity (LOH) using 35 microsatellite markers on chromosome arm 13q. Of the 59 primary tumors, 31 (53%) showed LOH involving at least one locus. Detailed deletion mapping identified a distinct commonly deleted region in the I-cM interval flanked by D13S153 and D13S273 on 13q14 and this region overlapped a part of the RB1 gene. Paired DNAs were available from both primary and metastatic tumors in the 14 cases of cancer death; among those pairs, we detected LOH on 13q in seven (50%) primary tumors, and in all metastatic foci (P = 0.0029). Moreover, the regions lost in metastatic tissues were more extensive than those seen in the corresponding primary tumors. These results suggest that inactivation of a putative tumor suppressor gene(s) including the RB1 gene on 13q14 plays an important role in human prostate cancer.