BACKGROUND: The recent discovery that many human prostate cancers contain a recurrent genomic rearrangement involving an ETS family oncogene (usually ERG) has introduced a new class of biomarkers that may provide information about disease prognosis (Tomlins et al. Science 310:644-8, 2005). These recurrent mutations are generated early in the emergence of a prostate cancer clone by a rearrangement involving chromosome 21q22.2-3; in most cases these rearrangements place ERG expression under the control of an androgen regulated promoter from the TMPRSS2 gene. Several groups have reported that TMPRSS2-ERG fusions show considerable patient to patient variability, and Wang et al (Cancer Res 66:8347-51, 2006) have proposed that the structure of the specific TMPRSS2-ERG mRNA transcript(s) that are produced by a prostate cancer may be directly associated with the disease aggressiveness. We have analyzed TMPRSS2-ERG gene fusions at both the mRNA and chromosome level in a prospectively collected series of prostate needle biopsies (12 cores/patient). In this report, we focus on findings from the patients in that biopsy series who were diagnosed with prostate cancer.
METHODS: Patients scheduled for prostate biopsy for the usual indications (elevated serum PSA and/or abnormal DRE) were enrolled in a study protocol in which tissue print micropeel “touch preps” were collected from each biopsy core. Tissue cores were processed as usual for pathology diagnosis. After a diagnosis had been established, additional sections were obtained for fluorescent in situ hybridization (FISH) analysis of chromosome 21q22-3 rearrangements. Tissue print micropeels (one for each core) were snap frozen upon collection; purified RNA and DNA fractions were prepared from each print for molecular analysis. rt-PCR and direct sequencing were used to determine the characteristics of specific TMPRSS2-ERG transcripts.
RESULTS: Our work confirmed previous reports of patient-to-patient heterogeneity in this class of tumor biomarkers. Most interestingly, we found that in many patients with multiple TMPRSS2-ERG transcripts, the pattern of gene fusion expression was strikingly variable from one area of tumor to another within a single prostate gland. Alternate splicing of fusion transcripts probably accounts for some of this intra-gland heterogeneity, but in some cases sequence analysis showed more than one chromosome 21q22-3 lesion, consistent with tumor genetic heterogeneity arising from multiple clones (multi-clonal/muli-focal cancer) or from progressive genetic changes within a single malignant focus (clonal evolution). Interestingly, we found prostate cancers with EZH2 overexpression (a marker of relatively aggressive local disease that may arise from gains in chromosome 7q35-36) in both TMRRSS2-ERG positive and negative tumors.
CONCLUSIONS: The characteristics of the TMPRSS2-ERG gene fusions found in a patients prostate cancer at the time of biopsy diagnosis may provide new insights into the specific molecular and cellular events that have occurred during the emergence and progression of that malignancy. We hypothesize that patients with a high degree of TMPRSS2-ERG gene fusion heterogeneity in their prostate cancer(s) are showing evidence of genomic instability that may be associated with more aggressive disease. We will test this hypothesis in follow-on studies of our biopsy patient cohort.