Microarray Analysis and Prostate Cancer Research
By Samuel E. DePrimo, PhD, and James D. Brooks, MD
Recently emerged nucleic acid microarray technologies are accelerating the pace of biological research and are likely to lead to improvements in the clinical setting. The technology utilizes arrays consisting of hundreds or thousands of distinct DNA sequences immobilized on a solid surface, most typically on either glass slides or silicon chips.1,2 Arrays on silicon chips consist of oligonucleotide molecules, while the glass slide and other approaches use larger cDNA molecules. Hybridization between DNA molecules on the arrays and fluorescently labeled nucleic acid mixtures in the samples being examined is the key step in the approach. Fluorescent labeling allows for computer-assisted visualization and allows identification on a large-scale of elements, which are differentially abundant between control and case samples. Utility of microarray analysis has been demonstrated in fields of biomedical research such as genotyping, gene mapping, and most extensively thus far, analysis of gene expression.3 This type of "expression profiling" has yielded extensive information from yeast experimental systems, and similar studies with human samples are providing biologically relevant information on an unprecedented scale and, often, of an unexpected nature.
Cancer research is likely to benefit particularly from the novel insight provided by microarray-based investigations. Reports to date have focused on identifying collections of genes that are either over- or under-represented in tumorigenic cells as compared to normal. Analysis of array data identifies which messenger RNA species in a given sample are differentially expressed; the sequences which are immobilized on arrays are complementary DNA molecules (cDNAs) that represent either known genes or currently uncharacterized transcripts referred to as expresed sequence tags (ESTs). The inclusion of these ESTs extends the utility of the array approach, as extensive characterization of expression patterns of an EST can provide insight into the biological activity of its gene product. This insight can be enhanced through the application of statistical algorithms to comparisons between different experiments and the thousands of data points which are measured for each. Such analysis has been termed hierarchical clustering, and the methodologies are still evolving as several approaches are being developed and refined upon data validation.4
Utility of Hierarchial Clustering
Although the informatics-based analysis of microarray data is in its early days, a number of reports illustrate the potential of such analysis for understanding of cancer biology. In a study comparing gene expression in colon tumors to normal tissue, clustering was shown to be useful not only for grouping genes into functional groups but also for classifying tissue type (i.e., normal vs tumor) based on similarity of expression patterns.5 A related approach was applied to a study of breast tumors and similar conclusions were drawn.6 Clustering allowed for distinction between expression patterns of tumor cells and those of noncarcinoma cells present within the same tissue samples, such as stromal cells and B lymphocytes. Another recent study describes an approach to cancer classification, termed class discovery and class prediction, which is demonstrated to be able to distinguish between cases of acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL) based solely on gene expression monitoring.7 These and other ongoing studies set the stage for attempts to correlate expression pattern data with clinical cancer outcomes such as tumor grade, sensitivity to chemotherapeutics, and long-term survival.
Benefits for Prostate Cancer Research
Research on prostate cancer stands to benefit considerably from microarray approaches, as indicated by some recent studies. Prostate cancer is the second leading cause of cancer deaths among males.8 Tumors of the prostate are marked by an initial dependence on androgen stimulation for development and progression but are also characterized by almost invariable development of an androgen-independent phenotype following androgen deprivation therapy.9 Ongoing studies using microarray analysis are addressing both of these phenomena with the goal of identifying and characterizing key genes and proteins crucial to prostate cancer biology.
Bubendorf and associates used cDNA microarrays to compare expression patterns of 5184 genes in nude mouse xenograft tumors of hormone-refractory human cancer strain CWR22R and in xenografts of the parental androgen-sensitive strain CWR2210. Expression of 0.7% of the genes was increased more than two-fold in the hormone-refractory xenografts compared to parental, while expression of 2.6% of the genes was reduced by 50% or more. The two most consistently over-expressed genes were insulin-like growth factor binding protein 2 (IGFBP2) and the 27-kd heat-shock protein (HSP27). The authors used a method conceptually related to cDNA microarrays, that of tissue section microarrays, in combination with immunohistochemistry to demostrate that IGFBP2 and HSP27 proteins were indeed over-expressed in a much higher percentage of hormone-refractory clinical tumor samples as compared to primary tumors. It should be noted that the tissue microarray approach, in which hundreds of clinical biopsy specimens are arrayed into a single pariffin block, has also been used in a study of gene amplification in prostate cancer progression.11 Fluorescent in situ hybridization (FISH) analysis of several genes implicated in prostate cancer was applied in a high-throughput fashion to a large number of prostate tumor specimens. The further combination of the cDNA and tissue section microarray approaches should continue to yield insights into the nature of hormone-refractory prostate cancer.
The role of androgen stimulation in prostate cancer genesis, as well as in normal prostate biology, is another fundamental question amenable to microarray analysis. Androgen exerts its effects through activation of the androgen receptor, which functions as a transcription factor which activates or enhances expression of several target genes. One of these is the prostate-specific antigen (PSA), which is widely used as a marker for the early diagnosis and management of prostate cancer in patients. In an effort to identify other potential markers, Lin and colleagues profiled androgen-induced gene expression patterns in the well-characterized androgen-responsive human prostate cancer cell line LNCaP.12 The microarrays used here consisted of 1500 cDNAs derived from prostate tissue cDNA libraries. Stimulation with androgen induced expression of nine genes; in this report, the authors chose one of these for further analysis. This is the TMPRS22 gene, which encodes a serine protease. Other serine proteases include PSA and human glandular kallikrein 2 (hK2), both of which are prostate-specific. Expression of TMPRS22 is shown to be localized to normal prostate basal cells and to prostate carcinoma cells and its sequence indicates cell-surface localization of the protein. It is suggested that TMPRS22 might be exploited as a diagnostic or therapeutic target.
Conclusion
Further characterization of androgen-induced gene expression changes should reveal more insights into prostate function and tumorigenesis. In this laboratory, microarrays consisting of approximately 8000 cDNAs are being used in androgen-dependent expression profiling not only in LNCaP cells but also in three other androgen-sensitive cell lines (MDA Pca 2a, MDA Pca 2b, and LAPC-4) in a variety of experimental conditions. Preliminary data indicate an extensive list of genes for which expression is either repressed or induced.13 Clustering analysis will be applied in an effort to classify responsive genes into functional subgroups. Other experimental contexts under investigation in this laboratory focus on the effects of putative food-derived chemopreventative chemicals on gene expression in prostate cells. Examples of such agents include sulphoraphane, selenium, and vitamin D.14 These and many other studies are likely to benefit from the broad perspective afforded by the new microarray technologies, with potentially the greatest benefit resulting from translation of research findings into the treatment and prevention of cancer. (Samuel E. DePrimo, PhD, is a Postdoctoral Fellow in the laboratory of James D. Brooks, MD, in the Department of Urology, Stanford University School of Medicine, Stanford, CA.)
References
1. Schena M, Shalon D, Davis RW, et al. Science 1995;270: 467-470.
2. Lockhart DJ, Dong H, Byrne MC, et al. Nat Biotechnol 1996;14:1675-1680.
3. Brown PO, Botstein D. Nat Genet 1999;21(1 Suppl): 33-37.
4. Eisen MB, Spellman PT, Brown PO, et al. Proc Natl Acad Sci U S A 1998;95:14863-14868.
5. Alon U, Barkai N, Notterman DA, et al. Proc Natl Acad Sci U S A 1999;96:6745-50.
6. Perou CM, Jeffrey SS, van de Rijn M, et al. Proc Natl Acad Sci U S A 1999;96:9212-9217.
7. Golub TR, Slonim DK, Tamayo P, et al. Science 1999;286:531-537.
8. Landis SH, Murray T, Bolden S, et al. CA Cancer J Clin 1999;49:8-31.
9. Isaacs JT. Urol Clin North Am 1999;26:263-273.
10. Bubendorf L, Kolmer M, Kononen J, et al. J Natl Cancer Inst 1999;91:1758-1764.
11. Bubendorf L, Kononen J, Koivisto P, et al. Cancer Res 1999;59:803-806.
12. Lin B, Ferguson C, White JT, et al. Cancer Res 1999;59:4180-4184.
13. DePrimo SE, Nelson JB, Brown PO, et al. Microarray analysis of the transcriptional program activated by exposure of prostate cancer cells to androgen (abstract). Proc Am Assoc Cancer Res 2000;In press.
14. Brooks JD , Paton V. Potent Induction of Carcinogen Defense Enzymes with Sulforaphane, a Putative Prostate Cancer Chemopreventitive Agent (abstract). Prostate Can Prostatic Dis 1999;In press.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content