Me and My Metabolism: Personalized medicine takes new direction

Physicians may someday predict a drug’s toxic effects in individual patients on the basis of their metabolisms, a proof-of-principle study in rats suggests. The finding could lead to a major shift in expectations for personalized medicine, which scientists generally have presumed would center on genetics.

Since people can vary widely in how they react to a particular medication type or dosage, many doctors consider personalized treatment to be one of medicine’s loftiest goals. Efforts to reach this end have focused mostly on pharmacogenomics, the study of how a person’s unique pattern of genes affects how he or she responds to any given drug.

However, notes biochemist Jeremy K. Nicholson of Imperial College London, genes can tell only so much about a body’s functions. Other factors, such as age, weight, emotional state, and gut bacteria, can have an enormous influence on how a patient processes medications. “Things that affect our lives quite a lot aren’t reflected in our genomes,” says Nicholson.

Since these factors influence metabolism, he and his colleagues wondered whether they could use individuals’ metabolic profiles before they receive medication to predict how patients might react to drugs. The scientists have named this approach pharmaco-metabonomics.

To test their idea, Nicholson’s team worked with 75 rats that belonged to an inbred strain and thus had closely matching genomes. The scientists began their work by collecting urine from all the animals. The researchers then ran all the samples through a machine that measured hundreds of molecules. The results provided a metabolic signature that varied slightly from rat to rat.

Next, the researchers fed each animal acetaminophen (Tylenol) in a single dose known to cause liver damage without killing a rat. Liver damage varied from animal to animal, despite the rats’ genetic similarities.

In half the rats, Nicholson and his colleagues examined whether the metabolic signatures correlated with the extent of the animals’ liver damage. Sure enough, the researchers found a striking relationship between the rats’ unique patterns of urine molecules and the toxic effects of the drug. Using this information, the scientists predicted with about 85 percent accuracy the liver damage in a second group of animals. Nicholson’s team reports these results in the April 20 Nature.

“This could be a very important advance in the study of personalized medicine,” says Richard Cote, a cancer researcher at the University of Southern California in Los Angeles. He adds that this approach eventually may give physicians a sense of a drug’s efficacy, as well as its toxicity, in an individual. Such information could prevent them from wasting time and money on ineffective treatments.

However, says David Jones of the Massachusetts Institute of Technology, researchers shouldn’t hastily give up on pharmacogenomics. Day-to-day variations in a patient’s routine could necessitate constant metabolic testing to make sure treatments are on target. On the other hand, he adds, “with pharmacogenomics, the answer you get is good for life.”