When a pathologist looks at the biopsy of a suspicious lump, she checks for the telltale signs of cancer—the clumping of cells and distortion in their normal architecture—that reveal a malignancy. From there, she may test for biological markers that categorize the cancer, hoping to guide oncologists as precisely as possible in making treatment choices for the patient. That’s the hope, anyway. All too often, though, one cancer looks and tests very much like another. Yet chemotherapy will melt away one patient’s tumor, while having no effect on another, seemingly identical cancer. At that point, it may be too late to pursue other therapies, including experimental treatments that might have helped if used earlier on.

Genetic oncologists are beginning to understand why cancers that look the same react so differently to treatment. A range of new and related molecular technologies has begun to reveal a surprising fact about cancer: melanoma, lymphoma and virtually all other major cancer groupings have genetic subtypes, some of which resist treatment or may respond better to different therapies. Although the symptoms and the cells appear the same, they are effectively different diseases.

The effort to define these subtypes begins with a map of the genome of the cancer cell. Once armed with that knowledge, pathologists will know whether one therapy or another will be more likely to work for the specific cancer subtype. It also provides tools for seeking out new and better treatments.

That is why Vincent T. DeVita Jr., M.D., HS ’66, director of the Yale Cancer Center, and colleagues decided in February 2000 to launch the Yale Tissue Microarray Facility. Together with the Keck Biotechnology DNA Microarray Resource, which was expanded in 1999 with funds from the National Cancer Institute (NCI), it provides scientists at Yale and other research centers around the world with a powerful set of tools for experiments to get at the genetic underpinnings of cancer. “I don’t know of a single cancer that doesn’t have an identifiable genetic abnormality,” says DeVita, a former NCI director under Presidents Carter and Reagan and author of the leading oncology textbook. Once known, that abnormality becomes a potential marker for early detection and a precise target at which to shoot chemical compounds for a possible cure.

The Tissue Microarray Facility and the DNA Microarray Resource are part of a series of initiatives at the Cancer Center and elsewhere at Yale to apply the emerging technology of microarrays to the study of basic biology, cancer and other diseases. Most importantly, microarrays hold the potential for rapid advances in the design of new treatments. “You can address real and important questions with arrays,” says DeVita. “They’re going to be the tools for the future. I don’t have a crystal ball, but if I were a betting man, I’d bet that new treatments will happen faster than anyone expected previously because of them.” In fact DeVita has made some big bets at Yale that those technologies are key to coming up with better ways to find and treat cancer.

Chipping Away at Cancer

Microarrays, sometimes referred to as biochips, are created by positioning minute amounts of biological material, such as tiny slices of tissue or portions of genes or proteins, on a glass microscope slide or a computer chip. Robots drip nanoliter quantities through tiny tubes—in some instances spraying spots out like an inkjet printer and in others placing tiny droplets like a quill pen—at precisely known positions to create a matrix of spots on the chip’s or slide’s surface.

There are several ways of capturing data from microarrays, which use either electrical current or biochemical reactions to measure gene and protein expression or the properties of tissue. The arrayed material—chemically tagged with a fluorescent compound—can be treated with chemicals, complementary DNA, messenger RNA or other proteins, with which it interacts. The fluorescent tags light up during the reactions, and laser scanners pick out the glowing spots. Computers read and display the results quantitatively or graphically. The green, red, and yellow colors of gene expression arrays, indicating genes that are active in a cell, have become something of a signature image for microarray technology.

Where previous methods allowed one gene, protein or tissue sample to be studied at a time, a single microarray can be used to look simultaneously at virtually every gene in a cell or at tissue samples from hundreds of patients. A single glass slide DNA array can hold segments of more than 18,000 genes, and a few slides provide sufficient capacity to study every known gene in the human genome or all the genes in model organisms such as yeast or certain bacteria.

Experiments result in the quick generation of massive amounts of information—a mountain of data whose interpretation requires a combination of biology and computing skills that is currently in very short supply. Though there are many other bottlenecks in the ongoing development of technologies, microarrays already allow investigators to make advances far beyond what was possible with previous study tools. “An array gives you an all-inclusive opportunity to look at genes or tissue and to get a complete and much more accurate answer. There are hundreds of questions that could be addressed this way,” DeVita says. At least a dozen laboratories at Yale in departments including Genetics; Molecular, Cellular and Developmental Biology; and Pathology and the Section of Immunobiology are now using biochips to unlock the mechanisms of diseases, especially cancer.

Yale is home to one of 24 DNA microarray centers set up in 1999 at cancer centers around the country with NCI funds. The Howard Hughes Medical Institute supplemented those funds at Yale’s Keck Foundation Biotechnology Resource Laboratory. Archibald S. Perkins, M.D., Ph.D., associate professor of pathology and of molecular, cellular and developmental biology, co-directs the DNA Microarray Resource, which produces biochips for laboratories at Yale and around the world. Perkins uses microarrays in his own research to study how leukemia changes cells at the molecular level, in hopes of coming up with new targets for drug therapies. Microarrays permit him to study all the genes that are turned on in leukemia cells at various stages of their development. Before the advent of microarrays, that was impossible. “It’s really a fantastic achievement,” he says.

According to Co-Director Janet Hager, Ph.D., the capacity of the arrays has doubled in the last year, to a current maximum of 18,432 spots. “We now offer five different arrays for human and mouse,” said Hager, “some representing as many as 16,000 unique genes in the form of either cDNA or oligonucleotides.”

Much of the data produced remains undeciphered, and some laboratories specialize in studying raw data that gets posted on public websites. According to Perkins, “There’s almost a community effort to make sense of the results.” A major initiative is under way to create the Yale Microarray Database through the collaborative efforts of a group of investigators and the Yale Center for Medical Informatics.

Though it is still an emerging technology, the data flood has begun to resolve a number of long-standing puzzles, such as why certain cancers that appear to be identical under the pathologist’s microscope respond so differently to treatment. Small genetic differences are the key. When coupled with all the gene sequence data from the Human Genome Project, microarrays are giving scientists tools for understanding the genetic changes that take place when a normal cell becomes cancerous. Knowing the genetic defects specific to a cancer offers a potential diagnostic tool and gives pharmacologists a starting point for drug discovery. “We can now aim for a unique cancer target,” says DeVita, adding that such designer drugs hold great promise against previously untreatable cancers. One successful example is Gleevec, a pharmaceutical designed to cure certain forms of previously untreatable leukemia and a rare stomach cancer. “Gleevec is the most effective targeted treatment ever for leukemia,” he says. “It puts us in the era of specifically designed therapies. We’re going to see lots of them from now on.”

Scoring a Diagnosis

In 1999, NCI scientists developed the tools to make tissue microarrays for two reasons: to stretch out limited supplies of tissue and to standardize conditions for studying and comparing multiple samples. A single array can hold as many as 800 half-millimeter cross sections of tissue and can complement other types of large-scale studies in a variety of ways. Unlike tissue arrays, gene chips cannot be used with large patient populations because of their complexity and the huge amount of data that even a single DNA array generates. Laboratory results still require validation in larger, statistically verifiable studies, especially before expensive and complex clinical studies can begin. Studies of hundreds of patients can be done with a single tissue microarray—with minimal tissue expended—as a way of validating what was seen in other types of pilot experiments.

Yale is exceptionally well placed to take advantage of the new technologies because of its huge collection of cancer tissue samples. Early last century, Yale pathologists began to store tumor sections from cancer patients at what was then New Haven Hospital. The standard practice at the time, as it still is in most hospitals, was to keep pathology samples around for a few years and then discard them. Yale kept many of the tissue samples on hand as a study and teaching collection, and they have contributed to what is now known as the Yale Tissue Archive.

The pieces of tissue are, for the most part, plugs smaller than a thimble. They are embedded in paraffin blocks and kept in labeled cabinets in the basements beneath the medical school and in the hospital. In 1935 the state of Connecticut established the Connecticut Tumor Registry, requiring all hospitals and clinics in the state to report every case of cancer, along with follow-up, treatment and survival data. It was the country’s first such database. With the establishment of the Tumor Registry, pathologists began to archive patient information along with tissue at Yale.

In the early 1980s, Jon S. Morrow, Ph.D. ’74, M.D. ’76, HS ’77, who is now chair and the Raymond Yesner Professor of Pathology, headed a team that developed a clinical database for tracking tissue pathology and patient information. (That system, known as CoPath, is now in wide use around the world.) With support from Yale-New Haven Hospital, available records for tissue in the Yale archive were entered into the CoPath database, which today covers more than 3 million separate tissue samples taken from about 1 million patients. At Yale, one can locate virtually any form of cancerous tissue or group of related tissues, compare treatments and their effects and follow the course of the disease. “If we go looking for a specific type of cancer, it’s there,” says David L. Rimm, M.D., Ph.D., HS ’91, associate professor of pathology and director of the Tissue Microarray Facility. “The samples are usually thick, and we have instant 20 to 40 years of follow-up.”

The combination of massive tissue archive and computerized database may also have a big impact on the future practice of medicine. Robert L. Camp, M.D., Ph.D., an associate research scientist in Rimm’s laboratory, has developed software that can automatically “score” tissue arrays—that is, compare a cell’s features to those of other cancer or normal cells. The software, called Aqua, could eventually be used to measure tissue properties to determine the likelihood of response to therapy. Recognizing its commercial potential, the Office of Cooperative Research has been working with Rimm, Camp and outside investors to build a new company to develop it into a marketable diagnostic tool.

“Aqua,” says Rimm, “could replace some of the tasks currently done by a pathologist. There may ultimately be a day when pathologists run machines as opposed to making a subjective determination of a diagnosis. It’s going to happen sooner rather than later.”

Arrays for the Future

The various microarray tools could eventually find their way into clinical use. Some predict that the first biochips for diagnostic purposes will be seen in hospitals within two years. Within the next decade, chips could indicate to pathologists that telltale genetic markers for early forms of cancer are present, opening the door to earlier treatment before the malignancy spreads.

All the new technologies must be developed further, however, if they are to prove more than tools for the laboratory. “You want something that is inexpensive, simple and fast,” says Associate Professor of Pathology Paul M. Lizardi, Ph.D., who has invented several array-based technologies, including rolling circle amplification and whole genome amplification, which rapidly amplify genetic material for easier use of microarrays as diagnostic tools. Two New Haven companies, Molecular Staging Inc. and Agilix, have licensed technologies he developed at Yale for improved detection of DNA using microarrays. Says Lizardi, “Microarray technology is too expensive today, and it has not yet been optimized for clinical use. There’s a real need for microarray-based tools that are cheaper and better.”

Many Yale scientists are confident that as microarray technology becomes refined it will lead to progress at an unprecedented pace. In his laboratory DeVita hopes to find genes common to different cancers that resist standard chemotherapy. “If you find one gene that triggers a family of genes associated with drug resistance,” he says, “you could use the knowledge to solve a universal problem. That would be a tremendous advance.”

According to DeVita, all the new tools will soon provide patients with ever-more-precise diagnoses and effective treatments for cancer. “In the not-too-distant future,” he says, “cancer will be a chronic disease treated on an ongoing basis like diabetes or hypertension. This is the real payoff. It is what we’ve spent $42 billion for in the war on cancer.” YM