It is unlikely that personalized medicine will be enabled for a wide range of major diseases using genetic knowledge alone.

Much has been made of the potential of pharmacogenetics to identify the genetic variation underlying differences in drug responses between individuals and how this information can be used to improve drug safety and efficacy. However, a particular drug-response phenotype is not determined solely by the genotype — it is also influenced by environmental factors such as nutrition, concurrent medication, underlying disease and age. Jeremy Nicholson and colleagues speculate that it is therefore unlikely that personalized medicine will be enabled for a wide range of major diseases using genetic knowledge alone, and describe in Nature a proof-of-principle 'pharmaco-metabonomic' study of paracetamol (acetaminophen) metabolism and toxicity in which they were able to predict an aspect of paracetamol metabolism from the pre-dose metabolic signatures of rats.

Nicholson and colleagues originally came up with the concept of pharmaco-metabonomics — defined as “the prediction of the outcome of a drug or xenobiotic intervention in an individual based on a mathematical model of pre-intervention metabolite signatures” — after finding that rats given galactosamine hydrochloride fell into 'responder' and 'non-responder' groups, which could be distinguished by their pre-dose urinary metabolite profiles.

To investigate this concept further, the authors obtained pre- and post-dose urinary metabolite profiles and post-dose liver mean histology scores from 65 rats given a single toxic-threshold dose (600 mg per kg body weight) of paracetamol. The amounts of the urinary paracetamol metabolites were determined and the variation in these data was modelled in relation to the variation in the pre-dose metabolite profiles. The authors also modelled the variation in the mean histology scores relative to the variation in the pre-dose metabolite profiles in an attempt to predict histological outcome from the pre-dose metabonomic data.

The most convincingly predicted drug metabolite parameter was found to be the mole ratio of paracetamol glucuronide to paracetamol (G/P). The authors therefore built and validated a mathematical model (known as a projection to latent structure) that enabled them to predict expected G/P values from individual pre-dose metabolite profiles.

Although the authors did not produce a fully validated model for predicting post-dose histology, principal components analysis revealed a statistically significant association between the nature of the pre-dose metabolite profile and the extent of the induced liver damage, which was demonstrated by assigning rats into histological classes 1–3, with 3 representing the highest degree of liver damage. In particular, a higher pre-dose level of taurine was associated with a lower severity of liver injury, whereas a higher combined pre-dose level of trimethylamine-N-oxide (TMAO) and betaine was associated with a greater degree of liver damage.

More research is needed, but this initial study demonstrates relationships between the nature of the pre-dose metabolite profile and two independent post-dose parameters, and provides validation of the concept of pharmaco-metabonomics. In addition to being used in drug and dose selection in the clinic, where the aim would be to increase drug efficacy and to minimise adverse reactions, metabonomic signatures might be a more reliable method of extrapolating toxicity tests from animals to humans in preclinical development, and could aid the discovery of safety biomarkers for new drugs.