technology
by david steen
Bioinformatics
Replacing the Human Guinea Pig
The news about specific drugs is often positive—for example, a new leukemia drug might have remarkable effectiveness and minimal side effects. But news about a drug can be negative, too, such as with Rezulin, an antidiabetic agent recalled last year due to drug-induced liver failures. How do we end up with more good news and less bad news? Bioinformatics is an important tool to achieve that goal.
Bioinformatics is a term born of convenience to describe a wide range of activities. It’s easier to say than: modern processes involving automated, precise measurement of biological reactions, with results evaluated by a series of methods including complex software and statistical comparisons, intended to predict the safety and effectiveness of a chemical when humans swallow, inject, inhale or apply it.
In bioinformatics, scientists use specialized cells, tissues and probes as human body substitutes, expose them to “candidate” compounds, and compare the results to other drugs and chemicals. We want to minimize the time, risk, and expense of a drug being developed, “trialed,” and released, if its safety or effectiveness ultimately proves unacceptable.
The FDA-mandated process for human trials of drugs requires three phases. Typically, about 25 healthy people participate in a Phase 1 trial (usually young males). They receive a drug for a specified period and are tested extensively to ensure that the drug doesn’t produce an adverse reaction.
In Phase 2 trials, the drug is given to perhaps 100 people who have the condition the drug is intended to treat, and the drug is evaluated for safety and effectiveness. In Phase 3 clinical trials, 4,000 or more people participate. These folks, who also have the “indicated” condition, represent a cross–section of the population that would receive the drug. Half receive the candidate, the other half a current therapy or the famous sugar pill (“placebo”). If things go well, the drug is approved.
Bioinformatics precedes the Phase I stage in order to catch problems before the drug moves to human testing because we can now predict likely human response in ways that were unknown only a few years ago.
For instance, it is desirable for some drugs to penetrate the infamous “blood-brain” barrier, because treating the particular condition requires altering chemistry in the brain. But we don’t want that to happen with other drugs—a good example is older antihistamines, which caused drowsiness because they penetrated that barrier and affected the brain’s chemistry. Systematic procedures now determine the likelihood of penetration of the blood-brain barrier before any person or creature receives a compound. As with all other bioinformatics systems, any dosage level of the compound can be evaluated.
XenoTech has collaborated with Kansas University Medical Center to develop special probes that accurately, inexpensively and rapidly determine the amount of RNA that livers will produce when exposed to a drug. Virtually all drugs and chemicals pass through the liver, and some production of liver RNA is needed for normal metabolism. Excessive stimulation of liver RNA can be a harbinger of significant risk; we can only wonder if that knowledge might have resulted in a different outcome when Rezulin was being developed.
We are a long way from absolute accuracy; the task is complicated by the enormous diversity among humans. For instance, genetic factors cause some people to lack key enzymes used to metabolize certain drugs, and it’s not currently possible to identify them in advance. But bioinformatics is moving us rapidly toward reducing the amount of bad news, and increasing the incidence of good news.
David Steen is Executive Director at XenoTech LLC in Kansas
City, Kan. He may be reached by phone at 913.588.7530 or by e-mail at
dsteen@kumc.edu.