Skip to main content
Premium Trial:

Request an Annual Quote

Imperial College s John Lindon on Using Metabolomics to Predict Drug Response


Name: John Lindon

Position: Professor, Imperial College Section of Biological Chemistry — 1998 to the present

Background: Head of spectroscopy department, Burroughs-Wellcome, 1976 to 1995

Education: : PhD in nuclear magnetic resonance spectroscopy from the University of Birmingham, UK — 1969

John Lindon and colleagues at Imperial College London and Pfizer have completed a proof-of-concept study for predicting response to the drug acetaminophen, possibly paving the way for the least developed of biomarker technologies — metabolomics — to enter clinics in the form of treatment-guiding diagnostics.

That day is still far off, though. The recent acetaminophen study, which was published in last week's Nature, reports the researchers' use of metabolomic profiles in predicting the extent of liver damage in rats.

The pharmacogenomic approach to predicting drug response fails to take into account the influence of drug absorption, distribution, metabolism, and excretion, the authors wrote. Metabolomic technology, however, accounts for a number of environmental factors, such as nutritional status, the gut microbiota, age, disease, and other drugs.

Pharmacogenomics Reporter spoke to Lindon this week about the future of metabolomic profiling for human medicine.

What sort of application do you see for this technology in humans?

Well, the aim would be that if you could do this in humans — and we are now working on a human study as well — you would have a facility for predicting how a human would tolerate a particular drug.

So, in a situation, for example, where there are lots of drugs on the market [for a specific condition], the doctor would try you on one, and if doesn't work, try you on another one. This might give you a way of focusing on the right drug for the right patients.

Can you tell me about the larger project you're working on in humans?

Not really, no. It's in collaboration with Pfizer, and we really don't want to reveal the results yet.

But what I can say is, 'They're promising.' Of course, we can't give humans a toxic dose of acetaminophen. We can only give them the normal therapeutic dose, but at least we can predict how they will metabolize the drug.

How many people are enrolled in the study?

A substantial number.

When will it be done?

The study is done. We're currently analyzing the data in more detail. Later this year I think we'll be in a position to publish that.

What is the advantage of metabonomics or metabolomics over pharmacogenomics?

['Metabonomics'] is how we describe it. There's a slight difference in definition between the two, [but] it doesn't matter.

Since we're looking at the metabolite level, we're able to look at the whole aspect of systems biology. Most diseases are a complicated mixture of gene effects versus environmental effects. So metabolism captures the whole lot, where the genes capture only the gene part. Of course, by looking in urine, for example — we could also have looked in blood plasma, but we looked in urine — it turns out that human biology and metabolism is intimately linked with the gut microflora, the bugs that we have in our intestine.

And so, by looking at the urine, we're looking at the whole system, we're looking at this complicated interaction between the human genome and all of these thousand different bug genomes. So if you just look at the genome of the host, you're missing a big part of the picture.

Are you looking at the host genome in your studies?

Well, of course. The metabolic results we see, the profile we see of hundreds and hundreds of small-metabolite levels reflects both the host genome and the bug genome as well.

How common is the use of metabolomics within pharma?

We have within the past 15 years collaborated with all the big pharma, doing these studies. And we're currently collaborating with a number now, and this work was done in collaboration with Pfizer, who have a big team of people doing it in-house. I know GSK are doing it in quite a large way — mainly in toxicity, I have to say there. But all the big pharma are doing this in-house now, and we have lots of collaborations with them.

How will this prove useful in humans?

What you might do, to every human being on the planet, you might actually profile their metabolism periodically and see if it's changed, for example, that would give you some information on diagnostics or whether their homeostasis has changed. Well, you could use that, and screen it against a statistical model you've prepared on a human population — see if it's normal, see if they need the drug, and then see which drug they need.

So is a person's propensity for toxic reactions to a drug likely to change?

It's not just toxicity — we're talking beneficial effects of the drug. We're trying to predict whether the drug would be beneficial to patients, not merely toxic.

So, for example, you've got 15 statins on the market for lowering cholesterol — which one is the right one? Both in terms of minimizing adverse side effects, but maximizing beneficial therapy.

Most people have pretty stable microflora. But different populations have different microflora. Therefore they will handle different drugs differently — they have different diets, which has a big effect on drug action. All of this is encapsulated in the metabolome.

Will acetaminophen be the first drug in which metabonomics will be useful?

Oh no, not at all. This is proof of concept, and it's something we know a lot about. We know a lot about the toxicity, we know a lot about how it's metabolized in the body — both in animals and in humans. So, this is just a good proof-of-concept study, which we could use and test for cross-species comparisons as well.

The drug industry tests drugs in animals and wants to test them in humans, and they need surrogate markers of efficacy and adverse effects, and metabolism offers the best platform, if you like, for getting the those trans-species biomarkers, and so [acetaminophen] was a good case to see when we come to do the humans, whether we're going to see any trans-species biomarkers.

You can take those biomarker combinations out of the rat or out of the monkey — whatever it is you're doing preclinically — and when you come to do clinical trials, then you're looking for those surrogate markers in the human population to see whether the drug's working or not.

If you take a complicated disease like many of the neurological diseases — schizophrenia, Alzheimer's disease, autism — all of those complicated diseases where it's difficult to work out whether the drug is working by conventional endpoints. The drug industry [is] looking for surrogate markers; biomarkers. And if your animal model of Alzheimer's disease is a mouse model, you might want to find some surrogate markers in the mouse that transfer to humans.

Is there a class of drug or disease area where metabolomics should first prove useful?

There are lots of studies where metabolic profiling has been useful for diagnosis, and we published one in Nature about three years ago for atherosclerosis based upon a blood plasma test for predicting the degree of severity of coronary artery stenosis, for example.

So there are lots of cases like that. This is a specific thing [that] is a new concept.

But no area or drug seems best suited?

No. I'm not a clinician, so it's difficult to comment. But you could imagine cancer chemotherapy, for example, where it's a constant balance between beneficial effect and adverse effect. You might want to tune your chemotherapy to match the patient.

There's an example, but I have no reason to believe it will be the first.

Filed under

The Scan

Octopus Brain Complexity Linked to MicroRNA Expansions

Investigators saw microRNA gene expansions coinciding with complex brains when they analyzed certain cephalopod transcriptomes, as they report in Science Advances.

Study Tracks Outcomes in Children Born to Zika Virus-Infected Mothers

By following pregnancy outcomes for women with RT-PCR-confirmed Zika virus infections, researchers saw in Lancet Regional Health congenital abnormalities in roughly one-third of live-born children.

Team Presents Benchmark Study of RNA Classification Tools

With more than 135 transcriptomic datasets, researchers tested two dozen coding and non-coding RNA classification tools, establishing a set of potentially misclassified transcripts, as they report in Nucleic Acids Research.

Breast Cancer Risk Related to Pathogenic BRCA1 Mutation May Be Modified by Repeats

Several variable number tandem repeats appear to impact breast cancer risk and age at diagnosis in almost 350 individuals carrying a risky Ashkenazi Jewish BRCA1 founder mutation.