Background: In spite of increasing knowledge about the tumor biology of prostate cancer (PC), molecular events involved in tumor progression are not well characterized. There is evidence that a number of genetic alterations play a role in tumor progression and in addition, angiogenesis also contributes. In this study, comparative genomic hybridization (CGH), a sensitive method for detecting regional DNA copy number abnormalities, and microsatellite analysis was used to identify frequent genome changes in PC. Correlation of these data with microvessel density (MVD) and clinical follow-up data was performed to determine genetic alterations that are associated with angiogenesis and subsequent tumor progression.
Methods: Fifty-seven paraffin embedded radical prostatectomy (RP) specimens were microdissected. DNA from the microdissected PC tissue was amplified by degenerate oligonucleoitide primed (DOP)-polymerase chain reaction (PCR), and CGH was performed on the PCR product. Quantitative analyses of the CGH profiles were performed using a t-statistic. Additionally, a microsatellite analysis of chromosome 13q was performed on a subgroup of 31 of the tumors. Using a polyclonal antibody against factor VIII, MVD was determined for all RP specimens. The results of CGH and microsatellite analysis were correlated with the clinical data of the patients and with MVD.
Results: Forty-two of the tumors (75%) showed one or more gains while 39 (70%) showed one or more losses per tumor. The most frequent DNA copy number gains were on chromosome 3, 4, 7, 8, 10, 11, 12, 13, and X. The most frequent losses were on chromosomes 2, 5, 6, 8, 10, 13, 15, and 16. Cancer recurrence occurred in 15 patients. The total number of DNA copy number losses was significantly higher in patients with this progression (86%) than without (52%) (P < 0.001). There was no significant difference in the number of gains in patients with or without progression. Contingency table analysis showed a significant correlation between progression and losses in regions of chromosomes 6q and 13q and a gain of chromosome 7q. In multivariate analysis, only loss of chromosome 6 was independently prognostic. The gains that correlated most closely with MVD > 35 were on chromosomes 2q, 7q, and Xq, while the losses most closely associated with MVD > 35 were on chromosomes 8q, 10q, and 13q. However, only the association between loss of chromosome 13q and MVD > 35 was statistically significant. Microsatellite analysis revealed a statistically significant correlation between MVD and instability of locus 171.
Conclusions: This study indicates that the frequency of genetic alterations in PC as detected by CGH correlates with clinical outcome, and that losses of DNA from chromosomes 6q and 13q are important events that correlate with tumor progression, with loss of 13q, especially instability of locus 171, also associated with angiogenesis.
Copyright 2004 Wiley-Liss, Inc.