BioWorld International Correspondent

LONDON - The era of personalized drug therapy is inching closer, but the investigations to determine which is the best drug is likely to include a urine test, as well as your genotype.

Researchers in Europe have shown that it is possible to predict an animal's response to a drug from the metabolic "signature" of its urine. The scientists' preliminary investigations suggest that human urine, too, holds hitherto untold information about how drugs are metabolized in individuals, based on their intrinsic metabolic profiles.

Jeremy Nicholson, professor of biological chemistry at Imperial College London, told BioWorld International: "We have shown a proof of principle that a urinary metabolite profile has the potential to predict whether or not you are likely to have good or bad effects from drugs. If this approach can be fully developed, urine tests of this kind could be used on a huge scale to predict drug metabolism and drug fate, and possibly many other things about a person's health."

The same method could help predict how people will respond to a change in their diet, or even to events such as dealing with a bacterial infection.

"It's potentially applicable to every area of human and even animal medicine," Nicholson added. "It's a completely new way of looking at personalized health care for the future - a new way of screening people."

Nicholson has been working on metabolic profiling for the past 25 years. He calls the approach "pharmaco-metabonomics" to distinguish it from pharmacogenomics.

The latter is the study of how genomic variations affect the ways in which a drug is metabolized, and whether it is toxic. For example, there are some well-known polymorphisms in genes encoding enzymes that influence how well people can metabolize certain types of drugs. People who cannot metabolize those drugs well may suffer toxic side effects.

But genetic influences on drug metabolism can be very complex, and they only are part of the story.

"Pharmacogenomics cannot capture the environmental element," Nicholson said.

The environment includes, for example, the millions of microbes that live on and in humans. Many have enzymes that can metabolize drugs, and they dose the human body with drug-like secondary metabolites, which switch human enzymes on and off. Earlier work by Nicholson's group showed that changing gut microbes caused drug metabolism to change, too.

To address those issues, Nicholson and his team decided to focus on the animal's metabolic profile, as seen in the urine, using nuclear magnetic resonance (NMR) spectroscopy.

The strategy involves analyzing the NMR spectra of urine before and after an animal is given a drug, and using the pre-dose metabolic signature to predict the outcome of the drug dose.

Experiments carried out by the group demonstrated that the strategy can work. In one experiment, the researchers conducted NMR spectroscopy on urine samples from 10 rats. They then gave the animals galactosamine. It is a natural metabolite, not a drug, but given in large doses, some rats will suffer liver damage.

Among the 10 rats, six animals developed liver damage and four didn't. Analysis of the pattern of the NMR spectra showed that the animals that fell sick all had a similar type of NMR pattern, while those that remained healthy had a different NMR pattern.

A second experiment involved dosing rats with paracetamol at a level that sometimes causes liver damage. The researchers analyzed the pre-dose NMR patterns of the animals' urine, gave the drug and then asked a histopathologist to score the level of liver damage.

That approach provided them with numerical values for both the NMR spectra and the degree of liver damage. "You build a mathematical model that connects all the variables from the NMR spectra - the signature - to the outcome data on drug metabolism and drug toxicity," Nicholson said.

"We found we could take the urine and predict, in a quantitative way, the metabolic fate of the drug and the quantitative level of liver damage," he added.

The experiments are reported in the April 20, 2006, issue of Nature in a paper titled "Pharmaco-metabonomic phenotyping and personalized drug treatment."

Nicholson said preliminary tests using human urine, aimed at predicting the quantitative metabolism of paracetamol, have had "very interesting" results.

"It looks as if we will be able to apply it to paracetamol," he said. "Future work will aim to find out how many different drug classes can be predicted in this way."