Editors' ChoiceCancer Genomics

Power in Numbers

See allHide authors and affiliations

Science Translational Medicine  28 Jul 2010:
Vol. 2, Issue 42, pp. 42ec120
DOI: 10.1126/scitranslmed.3001508

Among carcinomas, prostate cancer is unique in many ways, not the least of which is its extreme clinical heterogeneity. In many patients, the cancer can be successfully treated while it is still localized to the prostate and, even if it recurs, can be controlled for many years. For other patients, however, prostate cancer follows an aggressive course. Measures such as the Gleason score and blood PSA (prostate-specific antigen) concentrations can help forecast the clinical course, but much uncertainty remains. Could a detailed genetic analysis of the tumor better predict its future behavior? Could new mutations or pathways that drive the tumor be discovered?

To answer these questions, the group led by the late William Gerald, a renowned molecular pathologist, carefully assembled a collection of 218 samples from prostate cancer patients at various stages of disease. With detailed clinical information in hand, the team performed three sets of analyses on the tumors: expression profiling, array comparative genomic hybridization with normal tissue, and sequencing of the coding region of 157 genes hand-picked for their suspected roles in prostate cancer.

Unlike breast cancer, lymphoma, or glioblastoma, prostate cancer expression profiling alone did not yield any prognostic information. However, the profile of copy-number alterations did provide prognostic information beyond standard clinical measures. Many genes implicated in prostate and other cancers—such as PTEN, RB1,and TP53—showed abnormal copy numbers, but they rarely carried mutations. In fact, only one gene (the androgen receptor) was mutated in four patients, whereas other mutations affected three or fewer patients out of the 218 analyzed. The copy number variations and mutations collectively pointed to three pathways: PI3K, RAS, and the androgen receptor. Although the androgen receptor gene itself was mutated only in late-stage prostate cancer, a member of the androgen receptor pathway called NCOA2 was altered by mutation or copy number alteration in up to 20% of localized prostate cancers, pointing to a previously unsuspected role as the “first hit” in the androgen receptor pathway.

The data generated by the investigators are massive and have received accolades for their high quality. It is publicly available for anyone to mine and will probably drive prostate cancer research for years to come. Yet, it is only the tip of the iceberg. With better sequencing technology, efforts to sequence the entire genome of cancers—not just a hand-picked set of genes—are already on the way.

B. S. Taylor et al., Integrative genomic profiling of human prostate cancer. Cancer Cell 18, 11–22 (2010). [Abstract]

Navigate This Article