By Robert M. Frederickson June 15, 2003
| HIGH-THROUGHPUT SCIENCE HAS RESULTED in a flood of potential drug targets and biomarkers of human disease. Of course, these potential targets and biomarkers remain just that — potential — until validated by scrupulous analysis. The key is to correlate molecular findings with clinical information and tissue-based diagnoses made by a pathologist.
Traditional methods involving conventional microscopy and immunohistochemistry (IHC) are labor-intensive, slow, costly, and suffer from a high degree of subjectivity. Recently, high-throughput tissue analysis has emerged as an important new tool to automate histological screening. A number of young companies are pioneering the commercialization of tissue microarrays (TMAs) along with the imaging equipment and software tools necessary to automate their use and analysis. The advantages of TMAs are their high throughput and scale — saving both time and money — and the generation of a higher level of experimental standardization both within and between groups.
Creating TMAs requires access to archival tissue samples, preferably those associated with significant clinical information on the patient, disease diagnosis, and outcome. Hollow needles generate cylindrical tissue "cores" about 0.6 mm in diameter, from which 5-micron-thick slices are cut and placed onto slides. Arrays are often disease based and contain anywhere from 200 to 1,000 independent samples per slide.
"Few of the TMAs available from the commercial enterprises come with the depth of patient clinical data that would allow strong statistical analyses," says David B. Seligson, director of the University of California at Los Angeles Tissue Array Core Facility.
TMAs that come with richer clinical information tend to come from academic facilities. The Yale Cancer Center Tissue Microarray Facility, a leading example, has access to the Yale pathology archive, established in the 1930s, and the only such archive with fully computerized records. A strong advantage is that extensive molecular and clinical follow-up had been conducted on these patients. Yale has recently licensed its resources to a new company, Histometrix, which will provide access to the tissue arrays and the associated rich annotation for a fee.
"TMA analysis is multiparametric and inherently complex," says Dirk Soenksen, CEO of Aperio Technologies. "Unlike DNA microarrays, in which one or two fluorescent intensities are measured at each spot, a TMA image represents thousands of cells that can be captured and scored at every spot. TMA users must have an easy way to access the images and to associate that data with crucial spot-specific clinical information."
Sliding home: Aperio's ScanScope creates virtual TMA slides at resolutions about 200 times higher than a typical laser printer.
Traditionally, pathologists have manually scored staining using a crude intensity scale that ranges from 0 to 4. However, new imaging and informatics technology has allowed the data to be digitized. Software is also being developed to allow easy storage and image manipulation.
Aperio has developed a scanner, called the ScanScope, to rapidly digitize entire slides, including TMAs, and software tools that view and analyze the resulting images — known as "virtual slides." At 100,000 dpi, a slide containing 300 to 500 tissue spots can be captured in about 5 to 10 minutes.
ChromaVision Medical Systems offers the Automated Cellular Imaging System (ACIS). ACIS software scans IHC stained slides, and then processes the color and morphology information to display cells or tissues of interest for review by a pathologist. "It's very much a pathologist-driven tool," says Judy Kropp, director of marketing at ChromaVision.
TissueInformatics created its TissueAnalytics Arrayf(x) software using a process called "machine vision" to automate structure and pattern recognition and the quantification of tissue staining. Results can be automatically tabulated, graphically displayed, and made available for correlation with proteomic and genomic data sets.
"Using TissueInformatics' Arrayf(x), pathologists can perform complete statistical analysis of about 5,000 tissue cores in approximately one day," says Rajiv Dhir, director of the Paraffin Tissue Array Facility, University of Pittsburgh Medical Center.
Histometrix's AQUA software and microscope-based imaging platform use molecular co-localization rather than feature extraction. TMAs are stained with fluorescently labeled antibodies, and the co-localization of the fluorescence signals within pixel-based subcellular compartments is measured. In this way, one can determine whether a particular protein co-localizes with a membrane, nuclear, or other localized reference standard. David Rimm, a scientific founder of Histometrix, says the accuracy is comparable to an ELISA, "allowing us to see relationships undetected by conventional pathologist-based analysis."
Far from replacing the pathologist, these new tools promise to both speed and aid the transition of their field to the new high-throughput reality.
Robert M. Frederickson is a biotech writer based in Seattle. He can be reached at firstname.lastname@example.org.