Tissue Microarrays In Prostate Cancer Research

Many of the initial studies that use tissue microarrays have been performed on prostate cancer tissue (12,19-39), and, therefore, much has been learned regarding optimizing tissue microarray production and use for prostate cancer research. This is reflected in the large number of prostate cancer tissue microarrays in production or already developed. In each case, the tissue microarrays address different scientific questions related to prostate cancer, and, therefore, have been designed to answer such questions. During the preparation and use of tissue microarrays for prostate cancer, specific issues related to this tumor type have been identified and addressed. For example, the architectural features of prostate cancer and the tendency to infiltrate between normal prostate gland tissues results in most cores being composed of both normal and cancer cells. This also makes it difficult for the technician to take cores from an area of tumor and guarantee that tumor is present on each small tissue core. They may have inadvertently taken an area of normal prostate within the tumor nodule, or the portion of prostate cancer may disappear after multiple slides have been prepared from the tissue core. This has led some investigators to question the value of tissue microarrays, in particular for biomarkers that require quantitative data, such as p53 staining (40). However, others have clearly demonstrated that well-designed tissue microarrays can be used for quantitative biomarker studies (41-44). These issues have been addressed by two approaches:

1. Larger tissue cores of up to 2.0 mm are taken from the tumor nodule.

2. Multiple small cores of 0.6 mm are taken from a single tumor nodule.

In both cases, the specimens on the tissue microarray slide become more representative of the original specimen. During staining, some core pieces may be lost from the slide, therefore most tissue microarrays use multiple tissue cores from the same patient tumor on the array. Initial studies have suggested that four to five cores from a single patient specimen may provide sufficient representation of a patient prostate cancer (43). To further improve the chances that tissues from a single patient are not completely lost during slide preparation and staining, the multiple tissue cores are spread over different tissue microarray blocks, thus, guaranteeing that a completed experiment will have sufficient material for evaluation. Quality control is performed, with selected slides (often every 10th or 20th) stained and evaluated for the presence of tissue and tumor (45,46). Using these techniques, up to 200 high-quality tissue microarray slides can be produced from limited specimen stores. Examples of different types of arrays are presented in the following sections.


For most prostate cancer research, the initial question is often "does my gene/protein of interest have any significance in cancer?" This is, in essence, two questions:

1. Where is this gene/protein found?

2. Is this gene/protein involved in cancer?

To answer the first question, one will often wish to examine the biomarker in a series of different tissues to understand the specificity of expression. This is the ideal use of a tissue microarray, which allows one to survey a large number of tissues on a single slide. Similarly, the use of a general cancer tissue microarray that is composed of many different types of cancers will provide the biomedical researcher with a sense of the types of tumors in which the biomarker may be present, thus, providing tumor specificity. This has been effectively used in numerous publications (1,47,48). No associated patient data is needed, because the simple question of biomarker presence or absence in tissues is being addressed. Both of these types of arrays are readily available through commercial vendors (see Subheading 12.). Data analysis is usually limited to the qualitative or semiquantitative expression and correlation with various tissue or tumor types. Statistical analysis is usually limited to simple x2 or Fisher's exact tests because of the small sample size. However, from these studies, a biomedical researcher can get an idea of the degree to which their biomarker is tumor or tissue specific.


There has been a strong scientific interest in the stepwise progression of prostate cancer from precancerous high-grade prostatic intraepithelial neoplasia to invasive prostate cancer to metastatic prostate cancer. Biomarkers that can be shown to mark various stages of tumor progression are of interest both for prostate cancer research and to predict patient progression and possibly determine the timing of treatment intervention. Tissue microarrays have been designed for the validation of biomarkers of prostate cancer disease progression, centering around two general formats:

1. The normal to high-grade prostatic intraepithelial neoplasia to invasive cancer progression.

2. The normal to cancer to metastatic cancer progression.

The first type contains samples from nonneoplastic prostate, high-grade prostatic intraepithelial neoplasia, and invasive prostate cancer (confined to the prostate), and examines the transition to invasive tumor. This allows for the identification of biomarkers that predict the earliest shift to invasive tumor, and, thus, could provide biomarkers for use in chemopreventive studies. Examples of these types of arrays have been extensively used in prostate cancer studies (21,33-35,49). The second type of progression array tracks tumors from invasive clinically localized cancers to metastatic prostate cancer. Biomarkers tested in this format focus on genes that predict tumor spread, and, thus, would help target aggressive therapy in patients likely to die of their disease. Because of the difficulty in obtaining specimens from tumor metastases, these tissue microarrays are more difficult to construct, but have been used in prostate cancer studies (12,32,50,51). In both cases, the tissue microarrays may come with associated patient outcomes data and can be analyzed using qualitative or semiquantitative methods and simple statistical associations. The identification of a statistical association with disease progression is taken as transitive evidence of biomarker correlation with disease progression and outcome. This is because the advanced disease state is a known marker of poor outcome. Yet, when one identifies a biomarker that is associated with specific stages of disease progression, this does not indicate that the biomarker is an independent predictor of patient outcome, and, therefore, additional studies are still needed to validate that biomarker.


Once a biomarker is associated with prostate cancer, the ability of the biomarker to provide additional prognostic data is often desired. In this situation, the biomarkers are tested on tissue microarrays that contain a large number of prostate cancers with associated pathological and outcomes data. Using an outcomes-based tissue microarray, one compares the expression of a marker against the patient's final clinical outcome (prostate-specific antigen recurrence, development of metastatic disease, or death from tumor). Alternatively, a biomarker can detect patients with high probability of early or rapid recurrence, even in a disease in which most patients have a recurrence, such as hepatocellular or pancreatic carcinoma. In the literature, most outcomes-based prostate cancer tissue microarrays are mixed outcomes and progression arrays and typically contain a series of prostate cancers and associated groups of nonmalignant prostate tissue. These mixed arrays are the most commonly used tissue microarrays (19,23-26,28,31,38,40,50,52-56). There are different types of nonmalignant prostate tissues, which may range from nodular hyperplasia (benign prostatic hypertrophy) to histologically normal prostate tissue from prostates isolated from organ donors. Some tissue microarrays also provide positive control cores that are composed of formalin-fixed pellets of human prostate cancer cell lines (46). The tumors cores are typically taken from radical prostatectomy specimens and represent the dominant tumor nodule from the prostate. These tissue microarrays typically contain 190 to 550 patients and the associated controls. The most important component is the associated patient data, which can vary from simple Gleason score to detailed patient clinical, demographic, pathological, and outcomes information. These data are usually supplied in a Microsoft Excel file format, although a recently released xml file format for tissue microarray data exchange is gaining popularity for data standardization and transfer (57,58). General prostate cancer tissue microarrays are currently available through collaboration with the National Cancer Institute's (NCI) Prostate Specialized Program in Research Excellence (SPORE) program or the NCI Cooperative Prostate Cancer Tissue Resource. Similar to general tissue and tumor arrays, data analysis is limited to qualitative and semiquantitative means, and simple statistical correlations can be made to the associated patient data. Although these results will provide a general idea of the role of a given biomarker in prostate cancer, the use of a general prostate cancer tissue microarray is best for the generation of clinically significant hypotheses and to provide initial frequency data for a subsequent large-scale case-control or cohort prospective validation study.


Once the general importance of a biomarker has been established for prostate cancer, there is often interest in further analyzing the biomarker in subsets of prostate cancer patients. For example, if an association is found between a biomarker and the Gleason grade of prostate cancer, further studies may be performed using tissue microarrays that contain specific Gleason grade tumors, thus, further defining the relationship between the biomarker and Gleason grade. Alternatively, there may be a significant scientific effort to identify biomarkers that relate to a clinically significant issue in prostate cancer (metastatic tumor potential or hormone response, for example), and, thus, only biomarkers that can be validated for that hypothesis are studied. In these cases, a carefully designed case-control tissue microarray can be used to support or refute the previous findings. Examples of such specific tissue microarrays that have been made or are in production are listed in Table 1. In many of the specialty arrays, a case-control design has been used with attempts made to maximize the statistical power of the array while working within the constraints of specimen availability and space on the array block.

Table 1

Specialized Prostate Cancer Tissue Microarrays

Array type


Metastatic prostate cancer array Gleason grade array Androgen status array Ethnicity array Perineural invasion

Was this article helpful?

0 0

Post a comment