PerspectivesAre you interested in submitting a Perspective Article? Be sure to read The Science Advisory Board's Editorial Guides for Perspective Articles. Click here. Tissue Microarrays: Bringing Histology up to Speed by Jim Brady, Ph.D. Anyone who has conducted scientific research over the past couple of decades, particularly in the field of molecular biology, surely has noticed one inexorable trend: the growing popularity of commercial kits and the related tendency for lengthy and complicated procedures to evolve into routine methods. Twenty years ago, determining the DNA sequence of a small gene was an accomplishment worthy of publication in a respectable journal. Other procedures such as recombinant protein production, plasmid purification, and cDNA cloning used to be laborious tasks that kept graduate students, post-docs, and research associates occupied for days or even months on end. Now, all of these methodologies are performed with relative ease using optimized kits, labor saving instruments, and other customized products. However, there is still one area where automation and ready-to-use kits have not completely eliminated the need for tedious labor and skilled hands: the field of histology. Despite the availability of automated slide processors and specialized histochemistry reagents, informative tissue analysis still requires experienced histologists to generate large numbers of high quality sections one at a time. The difficulty of introducing automation and high throughput methodologies into histology has led to a blockage in the research pipeline. As increasing amounts of gene expression data are generated through the burgeoning application of DNA microarrays and other rapid genomic analysis techniques, researchers must identify more efficient ways to validate the vast amounts of potentially useful information that are being produced. Five years ago, scientists at the National Institutes of Health developed a new research tool with the potential to overcome many of the limitations that are inherent with standard histology methods. This tool is known as the tissue microarray (TMA), and its utility stems from the fact that it permits hundreds of tissue specimens to be analyzed simultaneously on one microscope slide. Small punches remove from up to 1,000 different tissues, are embedded in a single block, which is then sectioned to produce an array of circular samples derived from the original tissues. Although TMAs have great potential to expedite research, particularly in fields such as biomarker discovery and drug development, they also have problems that must be addressed before their usefulness can be fully appreciated. Once created, TMAs provide an efficient way to analyze many different samples in a single experiment. However, the time and effort needed to produce a TMA are far from negligible. To appreciate the effort that goes into generating a typical TMA using a manual microarray production instrument, take a look at the video posted on the National Cancer Institute's Tissue Array Research Program (TARP) web site, which depicts all of the steps that are involved in TMA production. Beecher Instruments claims that their ATA-70 automated machine can transfer up to 180 sample cores per hour, but even with automation, the process requires a considerable amount of practice. Dr. Stephen Hewitt, profiled in The Science Advisory Board, Member Spotlight, is a researcher at the TARP lab who is pioneering new protocols for generating TMAs from a wide range of different tissues. As an alternative to creating TMAs in-house, many companies offer pre-made TMAs representing a number of species and tissue types. Researchers can purchase arrays from a catalog or request custom arrays produced from specimens in tissue repositories. Prices for pre-made arrays from a commercial vendor range from about $100 to $250 per slide. Of course, outsourcing usually does not provide all of the options that a TMA user would like. In addition, the number of slides per order is small (usually less than five). In some ways, this situation mirrors the early days of DNA microarrays, when probe availability was limited and gene chip suppliers had narrow product lines. Now, there are many different commercial suppliers offering a wide range of DNA microarrays that are suited to a variety of specialized research needs, such as microbial genotyping or toxicogenomics. Similar product diversification may occur with TMAs. However, the narrowly focused requirements of TMA users, particularly in the area of tumor analysis, will make it harder for vendors to offer ready-made solutions comparable to those in the gene chip market. Another issue with TMAs relates to the small diameter of the individual array samples. Tumor biopsies are often heterogeneous, containing areas of normal tissue interspersed with cancerous cells. Moreover, cancerous cells in various regions of the biopsy may show different stages of advancement. Thus, a single core removed from a tissue block for TMA analysis may not accurately represent the pathological characteristics of the biopsy. In addition, it is not uncommon for tissues to be damaged or displaced during slide processing. This may not be a serious problem with standard tissues sections, but with TMAs, even minor damage can lead to the complete loss of multiple samples. Aside from the difficulties associated with producing and handling TMAs, data storage and analysis issues present another set of challenges. Interpreting the data from just one TMA slide often requires the use of sophisticated instrumentation and software to capture and process digital images from hundreds of samples. Moreover, comparing data between studies can be difficult if there is no standard format for information storage or file annotation. Pathology data are often free text and may reflect the idiosyncratic biases of the person who enters the information. Finally, privacy issues stemming from concerns over patient confidentiality add a further layer of complexity to the issues of data storage and information retrieval. Despite their limitations, TMAs offer significant advantages over standard tissue slides. One area where TMAs can be very useful is multicenter research studies. For example, an array of several hundred tumor samples collected in a preclinical study at one medical school can be turned into a series of identical slides which may be shipped to other research groups to assay for expression of various biomarker proteins or to look for genetic alterations. Previously, the logistics of producing and shipping several hundred individual slides to multiple labs made such studies hard to conduct, and the variability associated with many histology methods created difficulties in data interpretation. In addition, TMAs present an ideal method for quality control of diagnostic histology procedures. Different pathology laboratories can verify the accuracy of their results by using standard TMAs that were designed specifically as analytical controls. A search of the literature reveals that TMAs are becoming an increasingly common tool for cancer studies. However, it is surprising that there are very few published reports of TMAs being used for other scientific objectives. One could imagine, for instance, neuroscientists using TMAs to study amyloid plaque formation in Alzheimer's patients, or epidemiologists could utilize TMAs to characterize changes in tissue morphology caused by infectious disease agents. There are also many potential applications beyond the medical arena. For example, DNA microarray studies are providing vast amounts of gene expression data for a variety of organisms including plants, fungi, and invertebrates. To validate these data and to identify candidate genes for further study, TMAs representing a variety of strains, tissues, or species corresponding to the aforementioned organisms could serve as efficient screening tools. It will be interesting to see the extent to which TMAs will improve research productivity in the coming years. Clearly, there are still issues related to production efficiency, standardization, and data processing that need to be addressed before TMAs can reach their full utility. Also, TMAs are unlikely to create the same scientific impact as DNA microarrays, since TMAs are used mostly for validation or analysis of previously identified markers rather than for discovery of new targets. However, researchers or diagnosticians who analyze large numbers of tissue sections certainly will appreciate the potential for TMAs to change the way that they perform their jobs. ### Jim Brady, Ph.D. Senior Science Analyst The Science Advisory Board ### << Previous Next >> [ View All Perspectives ] |
|
