Gene expression profiling identifies clinically relevant subtypes of prostate cancer - PubMed (original) (raw)

. 2004 Jan 20;101(3):811-6.

doi: 10.1073/pnas.0304146101. Epub 2004 Jan 7.

Chunde Li, John P Higgins, Matt van de Rijn, Eric Bair, Kelli Montgomery, Michelle Ferrari, Lars Egevad, Walter Rayford, Ulf Bergerheim, Peter Ekman, Angelo M DeMarzo, Robert Tibshirani, David Botstein, Patrick O Brown, James D Brooks, Jonathan R Pollack

Affiliations

Gene expression profiling identifies clinically relevant subtypes of prostate cancer

Jacques Lapointe et al. Proc Natl Acad Sci U S A. 2004.

Abstract

Prostate cancer, a leading cause of cancer death, displays a broad range of clinical behavior from relatively indolent to aggressive metastatic disease. To explore potential molecular variation underlying this clinical heterogeneity, we profiled gene expression in 62 primary prostate tumors, as well as 41 normal prostate specimens and nine lymph node metastases, using cDNA microarrays containing approximately 26,000 genes. Unsupervised hierarchical clustering readily distinguished tumors from normal samples, and further identified three subclasses of prostate tumors based on distinct patterns of gene expression. High-grade and advanced stage tumors, as well as tumors associated with recurrence, were disproportionately represented among two of the three subtypes, one of which also included most lymph node metastases. To further characterize the clinical relevance of tumor subtypes, we evaluated as surrogate markers two genes differentially expressed among tumor subgroups by using immunohistochemistry on tissue microarrays representing an independent set of 225 prostate tumors. Positive staining for MUC1, a gene highly expressed in the subgroups with "aggressive" clinicopathological features, was associated with an elevated risk of recurrence (P = 0.003), whereas strong staining for AZGP1, a gene highly expressed in the other subgroup, was associated with a decreased risk of recurrence (P = 0.0008). In multivariate analysis, MUC1 and AZGP1 staining were strong predictors of tumor recurrence independent of tumor grade, stage, and preoperative prostate-specific antigen levels. Our results suggest that prostate tumors can be usefully classified according to their gene expression patterns, and these tumor subtypes may provide a basis for improved prognostication and treatment stratification.

PubMed Disclaimer

Figures

Fig. 1.

Fig. 1.

Hierarchical cluster analysis of prostate samples. (a) Thumbnail overview of the two-way hierarchical cluster of 112 prostate specimens (columns) and 5,153 variably expressed genes (rows). Mean-centered gene expression ratios are depicted by a log2 pseudocolor scale (ratio fold-change is indicated); gray denotes poorly measured data. The complete data set depicted here is available at

http://microarray-pubs.stanford.edu/prostateCA

. (b) Enlarged view of the sample dendrogram. Terminal branches for normal prostate samples are colored pink, and those for tumor samples are colored according to gene expression subgroups: III (purple), I (yellow), and II (dark blue). Two tumors clustering with normal samples (see text) are colored light blue. Clinicopathological features associated with individual tumor samples are indicated by black boxes below the dendrogram (asterisks indicate missing data). High grade indicates Gleason grade ≥4 + 3; advanced stage indicates pathological stage ≥T3; tumor recurrence indicates PSA rise after surgery or clinical metastasis. (c_–_l) Selected gene expression “features” extracted from cluster (locations indicated by vertical colored bars). Because of space limitations, only selected genes are indicated. Genes are annotated as indicated if associated in supervised analysis with high-grade (blue circles), advanced stage (green squares), short time to recurrence (red triangles), or long time to recurrence (red inverted triangles). Genes positively and negatively associated with epithelial cell content are indicated by colored text (dark blue and light blue, respectively; see Supporting Note 1). Genes characterized by immunohistochemistry are indicated with arrow. m, moving average (41-gene window) plots for the t test statistic (grade and stage) and Cox's proportional hazards partial likelihood score (recurrence-free survival) shown for the 5,153 genes in the cluster. Note that peaks (high grade, advanced stage, early recurrence) and valleys frequently correspond to gene expression features characterizing tumor subtypes.

Fig. 2.

Fig. 2.

Genes associated with high grade, advanced stage, and tumor recurrence. Genes identified in a supervised analysis using the significance analysis of microarrays (SAM) method (see Supporting Note 2) are ordered by rank value of their SAM score; samples are grouped by clinicopathological parameter and ordered by rank value within groups. Gene expression ratios are depicted by a log2 pseudocolor scale (ratio fold-change is indicated). (a) Forty-one genes (represented by 55 cDNAs), positively associated with high grade, with a FDR of 2%; note that, at this FDR, no negatively associated genes were identified. (b) Eleven genes (represented by 12 cDNAs) positively associated with advanced stage (FDR 8%); at this FDR, no negatively associated genes were identified. (c) Four genes positively and 19 genes negatively associated with short time interval to tumor recurrence (FDR 16%). Orange bars indicate samples and genes associated with high grade (a), advanced stage (b), or early tumor recurrence (c).

Fig. 3.

Fig. 3.

Expression of MUC1 and AZGP1 predict prostate tumor recurrence. (a and b) Immunohistochemical staining of prostate cancer tissue microarray. Representative positively and negatively staining cores are shown for MUC1 (a) and AZGP1 (b). Original magnifications are ×200 and ×400 (Inset). (c_–_e) Kaplan–Meier recurrence-free survival analysis based on immunostaining for MUC1 (c, 173 scoreable cases), AZGP1 (d, 170 scoreable cases), or both (e, 160 scoreable cases). MUC1 expression is stratified by positive vs. negative staining. AZGP1 expression is stratified by strong vs. weak/negative staining. P values (log rank test) are indicated.

Similar articles

Cited by

References

    1. Parkin, D. M., Bray, F. I. & Devesa, S. S. (2001) Eur. J. Cancer 37, S4–S66. - PubMed
    1. Hsing, A. W., Tsao, L. & Devesa, S. S. (2000) Int. J. Cancer 85, 60–67. - PubMed
    1. Jemal, A., Murray, T., Samuels, A., Ghafoor, A., Ward, E. & Thun, M. J. (2003) CA Cancer J. Clin. 53, 5–26. - PubMed
    1. Farkas, A., Schneider, D., Perrotti, M., Cummings, K. B. & Ward, W. S. (1998) Urology 52, 444–448; discussion, 448–449. - PubMed
    1. Han, M., Partin, A. W., Piantadosi, S., Epstein, J. I. & Walsh, P. C. (2001) J. Urolol. 166, 416–419. - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources