A dipstick test for breast cancer?
by Nancy Fliesler
Marsha Moses, PhD, of Children's Hospital Boston's Vascular Biology Program, wasn't expecting media attention when she published a report in the Journal of Biological Chemistry about a protein called ADAM 12. But attention came, inquiries poured in from around the world and unsolicited urine samples arrived in her lab.
The source of the excitement was a small pilot study in which Moses and postdoctoral fellow Roopali Roy, PhD, found that ADAM 12, when detected in urine, is an early sentinel of breast cancer.
When Moses and Roy tested urine samples donated by 71 breast cancer patients, 67 of them, or 94 percent, contained ADAM 12. In contrast, only seven of 46 samples from healthy controls—15 percent—had detectable ADAM 12, and only at very low levels. The likelihood of testing positive, and the levels of protein in the urine, were lowest in women with very early breast cancer and highest in women with advanced, metastatic breast cancer. [See study data in Table 1.1]
ADAM 12 testing isn't commercially available and needs further validation before it can be used clinically. Nevertheless, Children's is entertaining offers from companies interested in developing it as a diagnostic tool to complement mammography. While mammography can detect breast tumors at an early stage and has greatly improved survival rates, many women don't get mammograms at regular intervals and miss the window for early detection. Conversely, when mammograms are done, they often can't distinguish between cancerous and non-malignant breast lesions, which can lead to unnecessary biopsies. Moses's data suggest that ADAM 12 may be detectable in urine before a breast cancer is visible on a mammogram, and that its absence may help rule out breast cancer, reducing the false-positive findings that plague mammography.
ADAM 12 is just one of many so-called urinary biomarkers under study as part of the Moses Lab's Urinary Proteomics Initiative. The first-of-its kind, decade-old program is systematically looking for proteins that are easily detectable in urine and correlate with disease. "Urine turns out to be an incredibly rich resource that provides a noninvasive way of screening," Moses says. "Our patients have given us a real gift."
Moses's first generation of biomarkers grew out of her interest in a family of enzymes called the matrix metalloproteinases, or MMPs. In 1990, as a postdoctoral fellow, she was the first to show that MMPs are required for angiogenesis (the growth of new blood vessels), a critical step in a cancer's expansion. Without its own blood supply, a tumor can't grow beyond a certain size. The MMPs (and sister protein ADAM 12) clear the way for new blood vessels to sprout by breaking down the extracellular matrix, which is the dense, supportive, egg-carton-like structure that surrounds cells. Moses showed that angiogenesis is spurred by a variety of growth factors, but it can't proceed unless this physical barrier is breached. Once new vessels form, MMPs clear a path for them to penetrate the growing tumor, and as the cancer progresses, MMPs enable individual cancer cells to chew their way through the matrix, enter the bloodstream and metastasize to distant sites in the body.
Moses has published a series of studies linking urinary MMP levels with
cancer progression, and, with Children's Intellectual Property Office, has licensed the technology to GMP Companies, Inc. GMP has in turn incorporated several MMPs into a simple urine test that will soon enter formal clinical trials in adults with breast, prostate, bladder, lung and colon cancer, examining how well MMPs predict disease status over time.
With the right biomarkers—and more are in the pipeline—Moses envisions cancer someday being managed as a chronic illness, like diabetes or cardiovascular disease. Just as patients routinely have their glucose and cholesterol levels checked, they could be screened for cancer during routine medical visits (or even with home urine tests) and have that cancer treated before it becomes a serious threat. Urine biomarkers could help doctors determine the cancer's aggressiveness, choose an appropriate therapy and track whether it's working.
"People with diabetes are living long, healthy lives because they monitor their glucose levels," Moses points out. "The idea would be to catch cancer before it's a disease. Here's my fantasy: a doctor saying to a patient, ëYou've got things in your urine that suggest you may have some cancer activity; we'd better check it out further.'"
Moses's program director, cancer pioneer Judah Folkman, MD, shares and champions this vision. In the 1960s, he was the first to speculate that angiogenesis is required for a cancer to grow and progress. Today, he cites autopsy data from people who died from trauma without ever having had cancer. The autopsies revealed microscopic prostate tumors in over 40 percent of men in their 60s, microscopic breast tumors in more than one-third of women in their 40s and microscopic thyroid tumors in virtually everyone aged 50 to 70. Yet the rates of clinically diagnosed prostate, breast and thyroid cancers in these age groups are 1 percent or less.
In short, Folkman says, most people have a tumor somewhere in their body, but it remains microscopic and dormant until it flips a switch that triggers angiogenesis. Using animals as models, his lab has tracked when different types of tumors begin attracting a blood supply, and has found that angiogenesis often kicks in at a predictable time for a given tumor type. Moses's team has been comparing urine samples taken before and after the switch, searching for telltale biomarkers.
Both Moses and Folkman envision a day when patients start taking angiogenesis inhibitors as soon as urine tests indicate that the switch has happened or is about to occur. Angiogenesis inhibitors are generally nontoxic, Folkman points out, because they don't kill cells like chemo-therapy, but merely inhibit growth. What's more, tumors appear not to become resistant to angiogenesis inhibitors. Chemo-therapy drugs target cancer cells, whose high mutation rate allows them to quickly acquire biological defenses against the drugs. In contrast, angiogenesis inhibitors target blood-vessel-lining-endothelial cells that rarely mutate and are less apt to become drug-resistant.
A number of existing drugs for other diseases, such as the antibiotic doxycycline and rosiglitazone for type 2 diabetes, have anti-angiogenic properties. Folkman believes such drugs are natural candidates for preventive treatment of cancer.
With these relatively benign inhibitors, doctors could treat a tumor without even knowing its location—just as they prescribe cholesterol-lowering drugs without knowing which vessels have plaque buildup. "You don't wait until someone has a heart attack to prescribe statins," says Folkman. "You begin treatment when cholesterol is elevated. You treat the biomarker."
Near Moses's office is a "bank" with some 5,000 frozen urine samples from all types of cancer patients, both children and adults, and from healthy people matched to the cancer patients by sex and age. When Moses started the bank in 1994, there was little support for it. No one was much interested in cancer biomarkers, and researchers who were pursuing them were looking in blood. It was assumed that proteins, except a few very small ones, don't show up in urine.
"Urine was the poor cousin of body fluids," Moses recalls. "No one would give us a penny in funding."
A break came in 1995, when seed money arrived from the Breast Cancer Research Foundation of America and CaPCURE (now the Prostate Cancer Foundation). Meanwhile, protein detection systems improved and urine began to gain respect as a diagnostic and prognostic tool. Today, the Urinary Proteomics Initiative gets funding from the National Institutes of Health, and cancer research-ers routinely incorporate urine biomarker testing in their clinical trials.
"Originally, it was thought that larger proteins wouldn't get filtered out by the kidneys," explains Roy, who leads the urinary proteomics effort in the Moses Lab. "But we and others are finding that's really not true. Over the last few years, more sensitive methods have been developed that can show the
presence of large proteins at minute, nanogram levels."
In fact, she and Moses have helped show that many medium and some large proteins get into urine, albeit in small amounts, even in healthy people. The MMPs, for example, are all larger than 40 kilodaltons, which was thought to be the maximum size that could exit the kidney.
A few doors down from the urine bank, Roy begins the process of isolating these proteins. She loads urine samples into a series of chromatography columns, which are long, vertical glass or metal tubes filled with different resins. One column sorts proteins by size: proteins float downward through the resin, which is "seeded" with tiny, cage-like beads that temporarily trap the smaller proteins, so they're last to get to the bottom. Larger proteins float past the beads and reach the bottom first. Another column is loaded with chemically-charged resins that separate the proteins by binding to them with varying degrees of tenacity. Out at the lab bench, technician Alexis Exarhopoulos teases a sample's proteins apart even further by running a few drops through a zymogramóa flat gel-filled panel roughly the size of an index cardóthrough which an electric current is run.
When a sample contains an unknown protein like ADAM 12, Roy makes a
positive I.D. with mass spectroscopy, which uses a burst of energy to fragment the sample's proteins, then analyzes the fragments' charge and size to identify each protein. With this powerful tool, protein-finding has accelerated rapidly, enabling the urine bank to yield real dividends. (For an interactive feature on proteomics at Children's, visit www.childrenshospital.org/research/proteomics.)
Once proteins show the potential of being linked to cancer, hundreds of urine samples can be screened for them using a special plateóagain, not much larger than an index cardóloaded with antibodies to up to 96 different proteins. "If your protein is here, the antibody binds to it and it lights up in a different color," Roy says. Truly X marks the spot for disease diagnosis.
Moses's lab is now getting ready to report a host of new cancer-linked proteins. Some are specific to tumor type; others show up in all kinds of cancers; some appear just as angiogenesis is switched on; some turn on only when the cancer becomes metastatic. In the end, Moses expects to develop panels of highly reliable markers for each of the most important human cancers in their different stages of progression.
"We're going to discover as many key biomarkers as we can, then look to see how they can best be combined," Moses says. "We're not necessarily trying to replace existing tests, but to give clinicians more data so they can make informed decisions."
To learn more about supporting angiogenesis research at
Children's Hospital Boston,
contact Joan Romanition
in the Children's Hospital Trust at (617) 355-2429
or joan.romanition@chtrust.org.
