NY Prostate Cancer Conference - C.L. Sawyers - Session 1: Gene copy number analysis as a prognostic tool
Integrated Prostate Cancer Oncogenome Dataset<br />218 tumors (181 primaries, 37 metastases)<br />13 prostate cancer cell lines and xenografts<br />Data collection:<br />arrayCGH<br />mRNA transcriptome (exon array)<br />microRNAtranscriptome<br />exonresequencing/mutation detection (157 genes)<br />linked to clinical outcome (5 year median follow-up)<br />William Gerald<br />
Copy Number Alterations in Prostate Cancer Genomes<br />8-21% of primaries<br />NCOA2<br />MYC<br />8q<br />n = 194 tumors<br />Barry Taylor, Chris Sander<br />
NCOA2 amplifies AR pathway output in prostate cancer<br />LNCaP cells<br />Human tumors<br />Vivek Arora, Haley Hieronymus<br />
Overview of AR pathway alterations<br /> (expression outliers)<br />Niki Schultz, Chris Sander<br />
Cooperativity betweeenTMPRSS2-ERG and other genetic alterations<br /><ul><li>ERG overexpression alone does not cause prostate cancer in transgenic mouse models
ERG may cooperate with other common alterations to drive prostate oncogenesis
Can we identify genetic events that cooperate with TMPRSS2-ERG to promote prostate oncogenesis?
aCGH analysis of prostate tumor set (n=121) to find copy number alterations associated with TMPRSS2-ERG fusion</li></li></ul><li>TMPRSS2-ERG positive tumors enriched for three regions of copy number loss<br />Amplification: No association of TMPRSS2-ERG and regions of copy number amplifications<br />Deletions associated with TMPRSS2-ERG:<br /><ul><li>PTEN loss: Chr10 copy number deletion spanning PTEN (p= 5e-4)
Correlation between copy number loss and expression decreases for ERG-associated genes<br />
Sequencing revealed SHQ1 point mutation,in addition to copy number truncation<br />
Do genomes define different subtypes<br />of prostate cancer?<br />-breast cancer is now divided into three major subgroups based on mRNA analysis and these women are treated differently<br /><ul><li>can we do the same thing with prostate cancer?
mRNA doesn’t work as well in prostate cancer. What about DNA changes?</li></li></ul><li>Do genomes define different subtypes<br />of prostate cancer?<br />-breast cancer is now divided into three major subgroups, defined in part by their transcriptomes<br /><ul><li>similarly robust subgroups have not emerged from prostate cancer transcriptomes
TMPRSS2-ERG fusion is present in 50% of prostate cancers but these tumors are not obviously distinct clinically</li></li></ul><li>Unsupervised Clustering of Tumors by Copy Number Alterations<br />Defines Robust Subgroups <br />
Copy Number Alteration at Diagnosis<br /> and Time to Biochemical Relapse<br />Barry Taylor, Brett Carver<br />
This finding raises the possibility of knowing who need aggressive therapy right away (surgery, radiation, etc) and who can be followed with active surveillance.<br />Next steps<br />evaluate the gene copy number hypothesis on 1000 tumors with 10 year follow-up<br />miniaturize the gene copy number test to work on needle biopsies and on circulating tumor cells<br />To determine whether CNAs predict disease-specific death in prostate cancer patients managed conservatively<br />