Return to story and artwork

Medical Futures
by Adam Rogers '92

As a science, medicine is a wonderful art form. Most of its practitioners admit that they have more questions than answers about what goes wrong with the human body and how to fix it. Like us patients, they're waiting for the day when they can wave that little sparkly salt shaker from Star Trek, see what's wrong on a Palm Pilot and deliver a cure via a needle-less injection that makes a shusshing sound. Been to the doctor lately? We ain't even close.

But we're getting there. Medicine, chemistry, biology, physics, computer science--a flight of fields called biomedicine when taken together--are taking hesitant steps toward the sparkly salt shaker. Perhaps more importantly, biomedical researchers are beginning to see how what they know now might extend into the future, a sort of daydreamy bootstrapping. Pomona alumni turn out to be at the forefront of research efforts with the potential to transform the way physicians heal, from basic research to clinical medicine to understanding the health of entire populations. The future is coming.

The fundamentals have to get worked out first, of course. Every biological organism is built out of cells. In complicated, multicelled animals like people, cells are microscopic factories with an array of specialized functions. Nerve cells conduct information, muscle cells move the body, blood cells carry oxygen, and so on. Inside just about every animal cell is a nucleus, and inside that, chromosomes. These are tightly packed coils of deoxyribonucleic acid--DNA, the storage medium for genetic information. Since DNA was first described in 1953, figuring out how human beings get from DNA to a miraculously complicated body, and what can go wrong in the process, has been prime research territory.

Today, that's basic research with many possible applications. Jennifer Doudna '85 didn't initially know that's where she wanted to stake her claim. As a grad student she got interested in RNA--ribonucleic acid, the "cheap copy" of DNA that cells actually read and use to make proteins. In an animal, if it's not bone or fat it's probably a protein. Weirdly, RNA on its own is also capable of making more RNA. This is software acting like hardware, as if a copy of Microsoft Word could also make print-outs. Doudna started her career by working out the structure of self-replicating RNA, using a technique called X-ray crystallography. If you can figure out the shape of a biological molecule, you can often figure out what it does or how it does it. "Just seeing this molecule for the first time, I almost can't communicate how profound the feeling of discovery was," Doudna says.

Many of the properties Doudna found for that curl of RNA turn out to hold true for other functional RNAs, as well. That's where the potential for new drug therapies comes in. For example, Doudna's research mostly concerns the virus that causes hepatitis C, which infects people worldwide and has no known treatment. "Part of the RNA in the virus forms a structure that is able somehow to hijack the host cell's protein synthesis machinery and direct it to making viral proteins," Doudna says. Nobody knows how it works, but if you stop the RNA, you stop the virus. "A number of these new drug targets are turning out to be RNA," Doudna says. "This hasn't been a focus until recently of pharmaceutical research." There aren't chemicals that interfere with RNA pathways the way thousands of so-called "small molecules" do with proteins. But many natural antibiotics interfere with RNA's role in protein synthesis, so the potential for new drugs is there.

Those small molecules, and their interactions with proteins, may yet provide a pharmacological gold mine. That's the research focus of J. Andrew McCammon '69. He's a theoretical chemist, but while at Pomona, he started working with early computers to design drugs. While RNA may have some unusual functions, the real action in cells, biochemically speaking, centers on proteins. Genes are fundamentally just lists that describe sequences of amino acids. Stuck together, those form proteins. But proteins aren't linear; atomic interactions among the amino acids twist and fold them into three dimensions, and it's in the pockets and grooves of these structures that the chemistry of life takes place.

Here's an example: HIV, the human immunodeficiency virus that causes AIDS, has three enzymes, proteins that have an active biological function. Ten years ago the only AIDS therapy attacked an enzyme called reverse transcriptase, a protein that converts the virus' RNA into DNA, ready for inserting into a human host cell. But another of the enzymes is a protease, which eats other proteins. McCammon studied the structure of the protease enzyme, converted it into a computer model and generated potential shapes for small molecules that would bind to the protease and disrupt its function. Chemists then synthesized these protease inhibitors, which are now the basis for the combination therapy that has been most successful at treating AIDS. "We've learned enough in the last 30 or 40 years about the structure of enzymes that we can approach the design of small-molecule inhibitors almost the way somebody would approach the design of a bridge or a building," says McCammon.

Where does it lead? McCammon's group at the University of California at San Diego is tackling the third HIV enzyme, which integrates viral genetic material into the human genome. And more powerful computers mean possible solutions to basic problems in protein structure. Like: proteins turn out to be more flexible than biochemists thought. "So you have to take into account the fact that these molecules might deform somewhat in binding to one another," McCammon says. In fact, proteins seem to fluctuate through many possible shapes, with some areas of stability and some that are atomically "fuzzy." Nobody knows how that works, but it could mean that interactions between small molecules and proteins rely not just on fitting hand-into-glove but on electrostatic forces between atoms, or even quantum mechanical effects. Quantum mechanics is so weird that biomedical researchers tell stories about it to scare each other, so we'll just move on.

The more precisely known the structures of proteins are, the easier it will be to create drugs that act on them. But protein structures are determined by genes, which in a nutshell is why the Human Genome Project seemed like such a good idea. "You don't discover genes anymore," says Sean Tavtigian '84. "You just associate them with their function." He should know. He bailed on a Harvard postdoc to go to work for an upstart company called Myriad Genetics, which gave him a front-row seat for the discoveries of the breast cancer genes BRCA 1 and 2 in the mid-1990s. Today he's Myriad's head of cancer research.

The problem Tavtigian's talking about is: What do you do with genes once you've got them? Endless lists of As, Ts, Gs and Cs, the basic alphabet of the genetic sequence, don't tell you anything about how genes go wrong. Sure, for simpler diseases like the rarer forms of breast cancer associated with the BRCA genes, researchers were able to develop diagnostic tests. But "diagnostics are traditionally a very low-profit industry," says Tavtigian, "and pharmaceuticals are traditionally a very high-profit industry." As a result, many genetics firms--Myriad included--are transforming themselves into drug discovery companies, trying to figure out how to do what McCammon does but faster, cheaper and for thousands of potential targets.

One reason genetics start-ups think they can be drug companies has to do with a field of study called pharmacogenomics. Basically, everyone responds to a given drug differently. While most people respond normally, some people will metabolize it so quickly that it has no effect, and a few others will have intolerable side effects. Those responses are genetic. "One can find drug response associations which would lead to genotyping tests where a patient could basically know how they'll respond to a drug," says Tavtigian. Now, that's not necessarily as great as it sounds. Drug companies may not want to know if some segment of their market can't take the drugs targeted to them. On the other hand, knowing who can take what might also allow drug companies to find markets for drugs that are dangerous to some people but help others. Either way, setting up a program to figure it out would cost big money--$500 million by Tavtigian's estimation. And nobody's sure who's going to pay.

Another way to go at the problem of what to do with the nearly complete human genome is to look at proteins. The 34,000 or so genes in a person make up the genome; so too do the million or so proteins those genes encode make up a "proteome." Proteomics, the science of studying them, is wicked complicated. Every gene isn't turned on, making protein, all the time. Time of day, temperature, mood, nutrition and a host of other environmental and biological factors regulate the process. Some proteins are combinations of subunits made by different genes, and those subunits can be put together in different ways. But those variations, those genetic-environmental interactions, are at the heart of just about every disease that isn't caused by a germ. That's especially true for the ones that affect human beings more as they age, like heart disease or Alzheimer's.

Large-scale proteomics is too new to have medical success stories, but the concepts behind it are leading researchers in interesting directions. Grover Bagby '64 directs the Oregon Cancer Institute at Oregon Health Sciences University. His leukemia research deals with a family of genes called Fanconi anemia genes. They show up in stem cells, the cells in bone marrow that make blood cells, and when mutated they lead to leukemia. His lab is examining the proteins these genes make, both mutated and normal, to figure out the differences that can lead to leukemia later on.

Bagby is looking toward microarray technologies. Developed to do high-speed genetics work, a microarray is essentially a silicon chip with millions of strands of DNA attached, each a perfect match for part of a human gene. By passing a sample of genetic material across the chip's surface and seeing where it sticks, a researcher can get a very fast read on which mutations are present in specific genes and which genes are switched on or off. "I'd say 10 years from now a patient with leukemia might come into a specialty treatment center and have their bone marrow looked at under a microscope," Bagby says. "A microarray would then tell the hemotologist what specific kind of leukemia that patient had, which chromosomal abnormalities it mapped to, and which form of treatment would be best."

That's not the only way to treat leukemia, of course. Cladd Stevens '63, Bagby's classmate in medical school, has a different approach. She's the medical director of the Placental Blood Program at the New York Blood Center, where they collect blood from the umbilical cords of newborn babies. That blood contains hematopoetic stem cells, the precursor cells to red blood cells, white blood cells and platelets --the cellular components of blood.

Leukemia is blood cancer, basically, so treatment often involves the destruction of a person's bone marrow, where blood is made. But replacing that diseased bone marrow with a transplant is a dicey proposition. The marrow has to be very precisely matched to the recipient's immune system. "Apparently the cord blood, because it's coming from a baby, is not well-educated, immunologically," Stevens says. "We've done transplants that are mismatched for tissue type and that wouldn't even be dreamed of with bone marrow." The hematopoetic stem cells seem to be more forgiving, and even more amazing, they home right in on the bone marrow, where they're most needed. And cord blood may have other uses. For example, it also contains osteoclasts, cells that break down old bone to make room for new tissue. Babies with a rare genetic disorder called osteopetrosis keep developing new bone without tearing down the old, so eventually, they die of severe brain damage caused by the abnormal growth. A cord blood transplant cures these kids outright.

At the same time that biomedical researchers are hoping genetic and proteomics work might lead to new treatments for specific patients, they're also learning how to apply this knowledge to entire populations. In banking cord blood, Stevens is creating a library of tissue with ultimate uses that aren't yet totally understood. At a practical level, population research helps parse out what actually goes wrong in human beings, and why. "Why is it that some people get heart disease and other people don't? Is it their stress level, something they eat, where they live?" asks Julie Buring '71. She heads the Physicians' Health Study and the Women's Health Study at Harvard Medical School, which follow thousands of participants for decades, looking at genes and lifestyles. "As much as we're making progress on treating people, it would be far better if we could prevent disease from ever occurring," Buring says. It's cheaper, and ultimately improves quality of life.

There's a continuum among these techniques. In classic genetics, researchers try to look at families with some inherited condition, and then look at their chromosomes to try to find similarities, to narrow down the location of the genes that might be the cause. In the old days of biochemistry, researchers could look at rising and falling levels of various proteins in the blood, but there was no way to tell whether those changing numbers caused a disease or were its consequences. Today, population-based research methods look at health trends among thousands of people, analyze their proteins to find differences among the sick and the healthy and then find the genes that underlie the proteins. "We took a blood specimen from everybody at the beginning of our study, not even knowing what to look for," says Buring. "Now we are beginning to know."

Michael Gorin '74 has roughly the same idea. An ophthalmologist and geneticist at the University of Pittsburgh, Gorin works with smaller populations trying to track down the causes of age-related macular degeneration, a progressive blindness. "The goal of my research is to look at what environmental or therapeutic interventions can reduce the risk," Gorin says. It's not a well-understood disease--lots of genes are involved, and so are environmental factors. A combination of genetics and population methods promises to tease them apart.

But for all this to work, researchers will need more powerful statistics. Laurel Beckett '68 heads the Division of Biostatistics at the University of California at Davis' medical school. She's collaborating with neurologists at Rush-Presbyterian St. Luke's Medical Center in Chicago on a study of Alzheimer's Disease among 900 Catholic priests, brothers and nuns. The subjects get annual neurological and physical exams, including extensive cognitive testing, and, when they die, they donate their brains for pathologists to count neurons, brain proteins and, if they're present, the classic plaques and tangles of the disease. Beckett's job is to figure out how to tell the difference between what's genetic, what's environmental and what's related to the disease's progress. "People have this idea of statistics as driven by means and averages, and it's really not true," she says. "The action is in variability, and trying to understand it."

The new biomedical landscape is covered in a torrent of data. Long-term, prospective studies like the ones Buring works on generate gigabytes of numbers. Geneticists looking for individual differences use sample sizes of only a few dozen people, but look at thousands of genes. With a database that size, it's easy to get misled by artifacts. "How much of that is the result of, when you look at ten thousand anything, something's going to pop up?" says Beckett. "We have to think of how to find pattern and meaning in a great deal of data." Eventually new statistical methods capable of handling huge streams of data will help set public health policy. Those stats will be the only way to tell if population-based preventive measures actually make a difference.

These futures aren't that far off--10 years, maybe 25. At greater distances it's possible to imagine replacement organs grown from spare cells, genetically engineered to avoid rejection by the immune system. Miniaturization technologies could mean pinhead-sized robots that can sweep clean a congested artery.

Amid all these sci-fi images, however, there is one cautionary note. Ironically, at the dawn of what promises to be a revolution in biomedical marvels, nearly half of all Americans today have turned, at least on occasion, to some kind of alternative medicine--acupuncture, chiropractics, herbalism--even though Western medicine usually works better. They turn to alternatives because of the skyrocketing cost and complexity of high-tech medicine, because they don't get any face-time with their doctor and because they like the feeling of being taken care of, not merely treated. It's a lesson Western medicine is struggling to assimilate. And it's a reminder that there's more to our medical future than that sparkly salt shaker.

That's something Grover Bagby, the leukemia researcher from Oregon, has known since medical school. On a hematology residency, he treated a woman who came to the hospital with acute leukemia. She had a 26-year-old husband and three children, and she was dead in 10 days. "I became almost a monk devoted to going after this disease," Bagby says. "The idea that the best doctor is the one that's most removed? That's crap. It's a way that some protect themselves. But the patients I've followed for many years, I feel like they're my brothers and sisters."

--Adam Rogers '92 is a reporter, sometimes focusing on science, for Newsweek.