Abstract
Metabolomics was included among the ten emerging technologies
in an article recently published in the Technology Review of the
Boston MIT.
Metabolomics describes the chemical profile in terms of low molecular
weight metabolites present in the cells, tissues, organs and biological
fluids. Its components (metabolites) may be viewed as end products
of gene expression or of protein activity (enzymes), thus defining
the biochemical phenotype of an integral biological system, including
man.
While genomics and transcriptomics suggest a possible functioning
of t e system, metabolomics represents the real state of the system.
Over the next few years the metabolomic approach, combined with
functional genomics and proteomics, will have a strong impact
on drug manufacturing research, diagnostics, treatment and a broad
range of biotechnologies. In this article the theoretical and
experimental basis of the metabolomic analysis will be briefly
shown, as well as some fall-out in cancer research.
Alexander Bogdanov in his "Tektology: Universal organization
Science" (1913-1922) postulated that all the physical, biological
and human sciences could be unified by treating them as a set
of relations, highlighting the fact that the same organization
principles underlie all systems.
Living systems function by means of a vast range of reactions,
mostly catalyzed by enzymes and transport processes, which, through
the transformation of thousands of substances, governed by the
same laws of thermodynamics and kinetics as those governing the
inanimate world, determine the material and energy flows required
for their life and functioning. This set of flows represents the
metabolism.
Many of the metaphors used to describe a unifying principle that
would allow the functioning of a biological system to be interpreted
have over time been linked to the emerging technology, considered
as the "most powerful and advanced" at any given historical
time.
Thus in the 18th century, biological systems were compared to
a clockwork mechanism, in the 19th century to thermodynamic machines,
in the 20th century to digital computers controlled by logic ports,
down to the present time, in the century dominated by the world-wideweb
(www), where biological systems are represented as "networks".
Until only a few years ago, the dominant paradigm in representing
the functioning of a biological system was determined by a hierarchical
top-down view, in which the control of the system was generated
by the genome down to the hierarchically lower levels represented
by the physiological and functional aspects. A hierarchy of levels
was thus defined: gene ? transcriptions ? proteins ? metabolites
(fig.1), by means of which it was postulated
that the identification of the gene sequence of a biological system
would be sufficient on its own to predict the principal functional
characteristics.
 |
Fig.1
Biological information flux |
However, in the wake of the rapid development of biomolecular
technology, which allows the complete sequencing of the genome
of several living organisms, including part of that of human beings,
it became clear that such a approach merely offered the possibility
of evaluating and predicting the evolution of processes induced
by a disturbance due either to genetic modifications or to external
chemical and/or physical stimuli in terms of response by the entire
organism.
One of the main limits inherent in this approach consists in the
impossibility of interpreting and understanding the functioning
of an integral living system like the cell itself on the basis
of a reconstruction involving the use of a knowledge of the separate
characteristics and properties of the individual components of
the system (reductionist approach).
Indeed a living system is constantly being modified on the basis
of its physiological state (e.g. cell proliferation during the
stages of the cell cycle) and of interactions with the external
environment as a response to chemical and/or physical stimuli,
gene modifications, or by determining the variation in - or the
onset of - new structures permitted by the plasticity of the system,
thus allowing new strategies of functioning to take place including
the onset of a pathological state. As a complex system, a biological
system can come up with alternative pathways for processes determining
material and energy flows, while still respecting the laws of
thermodynamics and kinetics. In other words a complex system,
such as a cell, exerts control over its functioning, not by means
of a rigid hierarchical structure from the genes, to the proteins
and the metabolites, but by means of an organized structure equipped
with global and flexible interconnections among the gene, protein
and metabolic complements (fig.2). In recent
years it has thus been realized that it is necessary to make a
complementary appreciation of the innumerable data emerging from
the omic disciplines, such as genomics, proteomics and metabolomics.
In particular, the term metabolome describes the chemical profile
in terms of low molecular weight metabolites present in the cells,
tissues, organs and biological fluids. Its components (metabolites)
may be viewed as end products of gene expression or of protein
activity (enzymes), thus defining the biochemical phenotype of
an integral biological system, including man. It follows that
metabolomics provides an actual and not just a potential response
as far as functioning in a global fashion interconnected with
genomics and proteomics is concerned.
Likewise the metabolome, assessed via the biological fluids of
a human being, for example, reflects his/her history, including
his/her age, gender, life style, nutritional status, interactions
with the environment, possible pathological conditions and the
effect of drug therapies.
Although still in the early stages, the results obtained so far
appear important not only from the standpoint of basic science,
but also of applied science, with knock-on effects at the economic
and practical level.
For instance, the simultaneous qualitative and quantitative and
time-related evaluation of a large number of metabolites, such
as those that may be determined by means of NMR or mass spectrometry
in biological fluids affords a description with a reasonable degree
of probability of the current biochemical state of an organism,
providing information about the links between the various metabolic
processes defining such a state. It is thus possible to use the
study of biological fluids (such as plasma, urine, bile and cephalo-rachidian
liquid) to determine new criteria to define the state of health
or of disease on the basis of an integrated evaluation of the
variations in the level of the metabolites and the systemic metabolic
parameters, thus allowing the physio-pathological status to be
defined in systemic or organ-specific terms.
In this way, rather than take into consideration one or a few
metabolites with their related metabolic processes, metabolomics
examines the entire metabolic profile resulting from the interconnection
among all the different processes. For this purpose, Nuclear Magnetic
Resonance spectroscopy and Mass Spectrometry, with the use of
multivariate mathematical analysis methods, represent the most
powerful and most widely used methods for analysing the metabolome,
thus allowing both functional genomics and proteomics to open
up new research and applications pathways in the field of medicine,
pharmacology, food science and biotechnology in general.
 |
Fig.2
A 1H NMR spectrum of biological fluid divided
in three different
chemical shift region |
Nuclear magnetic resonance in metabolomics.
One of the experimental techniques best suited to acquiring the
required data for metabolomic analysis is Nuclear Magnetic Resonance
(NMR) spectroscopy (1).
Multinuclear NMR spectroscopy allows hundreds of metabolites to
be determined non destructively on a single sample, with or without
minimal pre-treatment of the sample itself, and without any a
priori knowledge of the compounds to be determined (metabolites).
For instance, from the acquisition of 1H NMR spectra from biological
samples, such as cells or culture media, tissue extract or the
biological fluids of living organisms, it is possible simultaneously
to obtain quali-quantitative information referring to more than
60 metabolites. If the analysis if performed as a function of
time it is possible to determine the direction of the observed
variations which depend on the disturbances to which we subject
the biological system considered.
By applying chemiometric and bioinformatic methods such as multivariate
data analysis, it is possible to make an integrated evaluation
of the specific variations observed in terms of changes and relations
among complex variables describing the effect of the perturbation
in that system, thus showing up the relative interconnections
and interrelations.
Another important aspect is the possibility of determining the
isotopomeric distribution of intermediate metabolites in different
metabolic pathways.
Using compounds enriched with stable isotopes (13C o 2H ), NMR
single and two-dimensional spectroscopy yields a description of
the isotope atoms in the skeleton of the metabolites involved
in the intra and inter-cellular metabolic flows.
 |
Fig.3
An esample of 13C isotopomer production starting from [U
13C]oleic acid in hepatcytes. |
Metabolomics in cancer research.
In 1980, with the advent of the molecular bases of medicine, it
was postulated that cancer was caused by the failure to regulate
several oncogenes or suppressor genes. It was thought that the
identification of these genes would make it possible to prevent
or treat cancer. It later became clear that tumors are much more
complex and heterogeneous than was previously believed and are
caused by changes in numerous processes and factors acting at
different levels.
The factors causing particular effects or alterations in a biological
system may be strongly dependent on the context and are regulated
by the activity of numerous components interacting in space and
time. It follows that in addition to the considerable interest
inherent in genomics directed towards the understanding the role
played by genes and their products, such as proteins, there is
growing interest in defining and understanding how the metabolism
can affect the genetic and proteic networks of particular tumor
phenotypes. As we have seen, the hierarchy derived by molecular
biology with its platonic top-down view in which a control is
postulated to exist from the genome to the transcriptomics, proteomics
and lastly metabolomics, is ultimately replaced by a bottom-up
view having a horizontal structure in which the dominant relative
aspect is provided by the interaction with the other components
of the system.
We have also seen how the metabolome may be defined as the quantitative
profile of all the low molecular weight molecules present in a
biological system in a particular dynamic state. Although it is
a known fact that metabolic networks vary in the way the substrates
are used and the flow distributed, thus reflecting the cellular
function and the phenotype, their control nodes have been satisfactorily
preserved in the course of the evolutionary processes and represent
points that are decisive for the interaction with pharmacologically
active molecules (drugs) also considering the limited number of
enzymatic isoforms, as well as of alternative pathways for the
metabolic processes, most of which are known. What apparently
distinguishes metabolomics from previous metabolism studies is
its focus on complete metabolic profiles and associated pathways
of a sample rather than one or more metabolites.
These profiles are represented by NMR or mass spectra and are
compared using multivariate statistical analysis of the data that
yield a complete overview of the molecular pathways inside the
biological system investigated. Therefore, while genomics and
transcriptomics suggest a possible functioning of the system,
metabolomics represents the real state of the system. This method
can provide information concerning the present state of the biological
system, by defining the real phenotype. Tumor cells possess a
number of mechanisms that can initiate and sustain the following
phenotypes: proliferative, differentiated, transformed, cell cycle
arrest, necrotic, apoptotic.
The common phenotype of advanced tumor cells is characterized
by a high degree of proliferation, low differentiation and high
transformation, as well as resistance to drugs and apoptosis.
These cells also display a high uptake of glucose, which is utilized
as primary substrate. The proliferation process is closely linked
to the de novo synthesis of macromolecules such as RNA, DNA, aminoacids
and fatty acids produced by low molecular weight substrates such
as glucose, short chain fatty acids and aminoacids through complex
and interrelated metabolic networks. Cancer may be viewed as a
“robust” system made up of tumor cells, the proliferation
of which is the property that must be maintained even though the
conditions of the microenvironment are not all favourable to growth
or the perturbations induced by antitumoral drugs inhibit it.
The "robustness" of cancer is defined in terms of system
and not of the individual tumor cell. Indeed the robustness of
the tumoral system is expressed by the functional redundancy generated
by the cellular heterogenicityheterogeneity of the tumor itself
and by the feed-back control system operating under extreme conditions
(for example, hypoxia and antitumor drugs) (2). One particular
tumor type very often corresponds to the inhibition of a specific
target, although it is commonly found that other tumor types do
not respond to such inhibition.
On the other hand it must be considered that the robustness of
a system is always relative and that the system itself may prove
to be fragile when it has to cope with perturbations of a different
kind. The aim is to identify the different sites and their interrelations
which determine the system’s fragility and find a method
to induce it systematically. Tumor phenotypes display differences
in their metabolome. For example, breast cancer or therapy-resistant
cells display differences in their metabolic processes compared
with therapy-sensitive pancreas cancer cells, and these differences
in the metabolic phenotype are reflected in the capacity to respond
to pro-apoptotic drugs. The main differences were identified by
analysing the metabolic profile using 13C enriched substrates,
in the rate of synthesis, elongation and desaturation of long-chain
fatty acids with reference to the different metabolic pathway
in the pentose cycle.
Therapy-resistant tumor cells also possess activities of desaturation
of fatty acid chains that involve the further oxidation of the
NADPH and allow the oxidative branch of the pentose phosphates
to operate in DNA synthesis even under treatment with inhibiting
drugs of the non oxidative pathway of the pentose cycle (3). Metabolomics
allows the biological characteristics and metabolic networks to
be linked with the objective to determine multiple enzymatic control
sites inside a network owing to the interconnection of the metabolic
networks with alternative synthesis pathways. On the basis of
long-established facts it may be postulated that inhibitors of
the pentose phosphate cycle effectively act on tumors that possess
a limited de novo synthesis of fatty acids, while tumors that
have a high turnover rate of fatty acids may be treated using
a combined approach involving fatty acid synthase, and elongase
and desaturase, with conventional drugs that act on the pentose
phosphate cycle, the production of nucleic acids from the backbone
of sugars, RNA synthesis, DNA replication and consequently cell
proliferation. The metabolic profile thus leads to the discovery
of new targets that may be less flexible and less variable than
genetic and proteic targets.
Pharmaceutical fallout
At the industrial level it is important to report the establishment
of a Consortium among British universities and multinational drug
companies (COMET project, 2001) to evaluate and validate metabolomic
analysis in drug toxicology via the creation of databases and
predictive expert systems.
In particular, COMET project objectives are to make a preclinical
evaluation of the toxicity and potential side effects of drugs
at the experimental stage, preclinical stage or in the early clinical
trial stage, which have often been found to be so strong as to
oblige the drug companies to withdraw the products from the market
or be held to pay large amounts in compensation. It should be
noted that in an article recently published in the Technology
Review of the Boston MIT metabolomics was included among the ten
emerging technologies with the number of publications in the sector
in quality international scientific journals having increased
exponentially between 2000 and 2005.
The results obtained so far, as well as increasing the interest
of drug manufacturing companies in the toxicity of drugs, are
focused on the search for new integrated metabolic indicators
in the diagnosis of human pathologies, for instance, coronary
pathologies, insulin resistance, degenerative diseases of the
central nervous system. Above all, an evaluation is being made
of the effect of diet and intestinal bacterial flora on the state
of health. In this field, in our laboratories, a data base is
being set up that is based on RMN 1H spectra obtained using plasma
taken from healthy volunteers and combined with anamnestic clinical
data and hematoclinal data. This database, today comprising 300
data obtained from 300 subjects, has been used in an ongoing investigation
of the metabolic variations caused by kidney insufficiency of
variable severity and the effect of dialysis and integrative therapies.
Future of metabolomics.
It is easy to predict that over the next few years the metabolomic
approach, combined with functional genetics genomics and proteomics,
will have a strong impact on drug manufacturing research, diagnostics,
treatment and a broad range of biotechnologies.In particular,
it could be applied to toxicity studies of and enhance the efficacy
of studies to develop drugs in the preclinical and clinical stage,
in screening and staging of patients with the potential to discover
new markers or a set of metabolic diagnostic markers. As well
as reducing the time and costs of research and experimentation
of new drugs, metabolomics combined with functional genetics genomics
can give indications for drugs already in production about the
possibility of use in treatment in the case of new pathologies
as well as reappraising drugs, the patents of which are due to
expire. Lastly, in the longer term, it could make a fundamental
contribution in the field of prevention but especially in the
personalization of treatment and diets associated with it.
Prof. Filippo Conti
Professor of Physical Chemistry
in Biotechnology Department of Chemistry,
”La Sapienza” University of Rome.
