Different chromosomal alterations correlate with gene expression in African American (AA) versus Caucasian American (CA) prostate cancer (PC) patients

Abstract

5000 Background: We used a genome wide approach to identify chromosomal and gene expression alterations that might explain the aggressive clinical behavior of PC in AA compared with CA patients. Methods: We frequency matched 20 AA with 20 CA patients for potential confounding factors (PSA, age, Gleason grade, pathologic stage). We utilized array-CGH (Spectral Genomics), to identify noteworthy genomic regions with significantly different chromosomal alterations between the 2 groups. RNA samples were available for 19 AA and 14 CA patients. Gene expressions were examined using Affymetrix HG-U133 Plus 2.0 arrays. Correlations were examined between the noteworthy altered chromosomal regions and gene expression. Results: Regions of chromosomal gain or loss in each sample were determined using the circular binary segmentation approach coupled with an outlier detection method to identify singleton markers exhibiting alterations. We identified 28 chromosomal regions that were significantly different between the 2 cohorts (Fisher’s exact test). 11 regions (3q28- 29, 3q25–26, 4p12–14, 4p13, 9q21, 10q11, 11q14, 12p13, 14q12, 16p11, 20p11–20q11) were more commonly altered in AA patients compared to CA patients (P<0.005). The altered chromosomal regions were associated with corresponding gene expression changes. Gene ontology and pathway analyses showed a relatively short list of affected pathways: insulin, STAT, TGF-beta signalings, proteosome degradation, integrin- mediated cell adhesion, adipose and calcium metabolisms. Hierarchical clustering of the 33 patients (19 AA and 14 CA) based on the shortlist of gene expressions showed two clusters, one consisting of CA and the other with AA patients, with only 3 AA patients showing similar gene expression to CA group and vice versa. Conclusions: Data demonstrate the presence of biologic differences between the AA and CA cohort of patients studied and suggest that the disproportionately unfavorable outcome of PC in AA patients cannot solely be explained by advanced tumor at presentation. Validation in an independent larger sample size is a prerequisite before further conclusions can be made. No significant financial relationships to disclose.