Massachusetts Institute of Technology professor Bob Langer has spent 25 years trying to solve the puzzle of drug delivery. Every drug has a desirable therapeutic range. Above it, the drug can be toxic. Below, the medication can be ineffective.
Just how do you get the right dose to a patient at the right time? To find the answer, Langer has customized polymers that allow drugs to seep out into the body, experimented with magnets as a way to control polymer drug release, and worked on patches that slowly pass drugs into the body through the skin. To date, Langer holds 408 patents and has published 711 articles and 13 books. He has also licensed products to more than 75 companies. So when Langer has an idea, it matters.
His latest brainchild combines silicon chips with advanced medical-device technology to make drug delivery more intelligent. "This is a paradigm shift," he says. "We're putting pharmacies on a chip." It could also turn out to be big business. The U.S. drug delivery industry produces $38 billion to $40 billion in revenues each year. Mahesh Chaubal, founder of drug-delivery information service Drugdel.com, estimates that the market for computerized delivery systems could be worth $5 billion by 2012, vs. zero today.
DRUG WELLS. The concept of a computerized drug-delivery system came to Langer while he was watching a TV program about chip-giant Intel more than 10 years ago. But it wasn't until John Santini arrived as an MIT summer student in 1993 that the project started to take shape. After completing his PhD at MIT, Santini founded MicroCHIPS, a Cambridge (Mass.) startup that's designing chip-based drug-delivery technology. To date, it has attracted two rounds of venture funding totaling $17.25 million.
MicroCHIPS' product is a titanium device, about two-thirds the size of a standard pacemaker -- which itself is the size of a box of Chiclets gum. The device would be inserted in a patient's abdomen during an outpatient procedure. Inside is a silicon chip with as many as 400 tiny "wells" where drugs can be stored. Each well is sealed with a thin gold membrane, which prevents the drug from leaking out of the chip.
Releasing a unit of the drug involves sending a small electric voltage that dissolves the cap and allows the drug to diffuse into the tissue. MicroCHIPS is working with a major, though still unnamed, pharmaceutical company and plans to have the device in clinical trials within 18 months. It's likely that the device will be used for large-molecule protein drugs, steroids, or horomones.
IT NEVER FORGETS. A computerized drug-delivery device has several advantages. First, it allows for precise control of drug release, solving the key problem of delivering too much or not enough. Because it has a powerful chip inside, the device can be preprogrammed to release a prescribed dosage at precise intervals. It also can be manually controlled via a wireless device that looks like a supermarket scanner.
A system like MicroCHIPS also solves problems that are inherent in complex, combination drug therapies, where a patient needs to take specific dosages of several drugs at once. An implanted device releases drugs on schedule, eliminating the need for patients to remember to inject or swallow the right meds.
Finally, it's feasible: Technologists understand how silicon chips and implanted devices work. Cutting-edge gene therapies, where scientists try to coax the body into producing proteins or hormones to cure disease, are exponentially more complex because it isn't always clear what cascading, harmful effect they might have.
SKEPTICAL VIEW. "We're not reinventing the wheel," says Santini. "And that's really important because you don't want a technology where there are eight different new things you have to figure out before you can get a product onto the market. We are building on what has been done in the cardiac area, the wireless area, and the microchip area. We're integrating a lot of known things."
Even so, some drug-delivery analysts are skeptical of this approach. Tracy de Gregorio, a director at pharmaceutical research firm Decision Resources in Waltham, Mass., thinks computerized drug delivery could be a niche market at best. She says patients and HMOs may not embrace expensive, invasive procedures just to avoid needle injections, even for large-molecule drugs like proteins that can only be delivered this way.
Besides, "the overwhelming preference has always been and will always be oral delivery," she adds -- so that's where research money will be spent. "Using silicon chips will be applicable to a small percentage of protein and biologic drugs," De Gregorio says. "And its success will depend on the success of those therapies." She points out that the major pharmaceutical companies haven't put much muscle behind this type of innovation, leaving it to academic labs.
"ARTIFICIAL PANCREAS." In one area, though, computing technology already is making its way into medical devices: insulin delivery for diabetics. The U.S. market for glucose-monitoring devices alone totals $3.5 billion annually. Now, Medtronics MiniMed (MDT) in Northridge, Calif., is testing an "artificial pancreas" that includes an implantable glucose sensor and an insulin pump.
The sensor is inserted into one of the major veins in the heart, the superior vena cava, where it continuously records glucose levels by generating electrical current frequency changes proportional to the level of blood glucose. Using sophisticated software and a small microprocessor, the system then calculates the amount of insulin required using a mathematical algorithm and instructs the pump, which is implanted in the lower abdomen, to diffuse the correct amount of insulin.
Some 160,000 U.S. diabetes patients already use external insulin pumps and sensors. But putting this technology in the body could broaden the market, according to MiniMed Communications Director Deanne McLaughlin, since a successful artificial pancreas would change the lives of most diabetes patients. Today, many of them prick a finger to draw blood four or more times a day to measure their blood glucose.
OBVIOUS IMPROVEMENT? Even with that many tests, it's often difficult to tell whether the glucose level is changing, how fast, and in what direction. Computer technology would close the loop for diabetics, making it possible to adjust their glucose levels with a minimum of human intervention.
More than 300 patients in Europe are now testing MiniMed's implantable pump under supervision. On June 15, Dr. Eric Renard, a professor of medicine at Lapeyronie Hospital in Montpellier, France, presented a paper at a conference of the American Diabetes Assn. conference showing that the implanted sensor and pump worked well without any human intervention in a two-day test.
In retrospect, the idea of using silicon chips to make drugs smarter seems obvious. After all, over the past 20 years advanced microprocessors have found their way into nearly every other kind of device. Now if researchers such as Langer and Santini have their way, delivering custom doses of medicine will soon be a task for computing as well. By Jane Black in New York