Dr. Toby Athersuch explains the delicate art of Metabolic Profiling

Metabonomics: Seeking Meaningful Signatures In Your Biological Fluids

In the first of a new interview series for minimolecule.com, I caught up with Dr. Toby Athersuch to understand the rapidly growing field of metabolic profiling. This research is allowing scientists to analyse blood, urine, plasma and other biological fluids to help characterise disease states and advise on potential treatment and prognosis. It is also critical in developing new non-intrusive diagnostic techniques and may make the science-fiction-eque rapid diagnosis using a handheld bioanalytic instrument a reality!

What is metabolic profiling?

In broad terms, metabolic profiling (also known as metabonomics) is the
study of small molecules (metabolites) in biological systems. We are
interested in finding out what they do and why they change. Typically, we
use spectroscopic analysis to provide direct information of the identity
and relative concentration of many metabolites at a time, in different
complex biofluids/tissues (e.g. urine, blood, saliva) that we can sample
easily. Each spectroscopic technique we can use (and there are many)
reports on a subset of the whole, which we term the metabolome.
Investigating the metabolome can be very informative, and report on a
large number of processes in biological systems; in humans, the metabolome
is influenced by our own genetic makeup, environment, lifestyle,
occupation, and health status. Finding out which of these factors
influence the metabolome is a big part of profiling research; in many
cases, it helps to be broad-minded – if we approach our analysis with no
preconceptions (an agnostic, or untargeted approach) then we may find
signals related to our factor of interest (disease etc) that we have not
yet encountered. In this way we can start working out what these things
are, and why they might be interesting. My own research at Imperial is
involved in research aimed at characterising how metabolism is influenced
by environmental exposures, and what consequences this has on our risk of
chronic disease.

What determines a good biomarker?

Foremost, high specificity for what it represents, as this helps us be
sure that we are observing relevant changes, and that they are not
confounded by other processes. If I had a wishlist for other attributes, I
think I would opt for the following: presence in an easily accessible
biofluid/tissue (easy to get), easily detected in that biofluid/tissue
(easy to measure), a large response relating to the factor of interest
compared to normal physiological/inter-individual variation (easy to see a
response), and good translation between different biological systems (easy
to relate results at the lab bench to those in the clinic). It¹s usually
not that straightforward!

NB: These issues were recently addressed in relation to drug development
by the Predictive Safety Testing Consortium (PSTC).

Does the the metabolite profile in urine and other biofluids reflect a
disease state well?

It depends. How well one can relate changes in the metabolome to a

particular disease or treatment will depend on several factors. In
general, because metabolites mediate and regulate so many biological
processes, finding changes in one or more of them can not only tell us
about disease state, but also about the biological mechanisms that are
responsible. Other techniques can obviously be very complimentary to this,
and help get a better picture of the system we are interested with as a
whole. Additionally, I would note that different biofluids report on
different things; urine offers a good idea of the pooled excretion of
metabolites over a period of time, whereas blood is effectively an
instantaneous picture of circulating metabolites. Choosing a suitable
biofluid to sample is obviously a good start to any investigation of this
nature!

Do you feel metabonomics has a big role to play in developing more
targeted therapies and have there been any success cases so far?

The tailoring of healthcare to the individual (personalised healthcare
and precision medicine) will benefit a lot from the involvement of
metabolite profiling approaches. It has already been shown by researchers
at Imperial that the metabolism of a drug can partially be predicted by a
urine sample provided before administration ­ pharmacometabonomics
recently reviewed in [1]. Developing this idea to relate to predicting
drug efficacy, surgical success, and therefore determine a more optimal
course of treatment is a priority for researchers in the field of medicine
at present. There has been considerable success in applying metabonomics
within the field of cancer research over recent years; cancer cells
typically have several metabolic traits that distinguish them from normal,
healthy cells, and these can be exploited from several angles including
diagnosis, patient stratification and drug efficacy monitoring.

What is the most significant challenge in your field of research and
how do you see the field progressing over the next decade?

The sensitivity, resolution and throughput of spectroscopic
instrumentation that we use continues to rapidly increase, so we can
characterise an ever-increasing proportion of the metabolome.  Making
sense of the vast quantities of data that can now be generated will
require development of advanced analysis tools and databasing solutions,
and is probably the most significant challenge at present. Advancing the
application of metabonomics in the two main areas of healthcare provision
(e.g. the Imperial NIHR Biomedical Research Centre [2]) and molecular
epidemiology studies (e.g. EU FP7 EXPOsOMICS project [3]) present
additional challenges that result from the timescales involved for
analysis, and the size of the sample sets involved.

[1]
http://acb.sagepub.com/content/early/2013/07/25/0004563213497929.abstract
[2]
http://imperialbrc.org/our-impact/case-studies/intelligent-knife-surgery
[3]
http://www3.imperial.ac.uk/newsandeventspggrp/imperialcollege/newssummary/news_19-11-2012-10-22-31

Note: TJA is a contributor to the EU FP7 EXPOsoMICS project.