View Report Details

Molecular Biomarkers of Toxicity from Dietary Chemicals

Project Code: T01025

30/09/2005

Section of Experimental Medicine and Toxicology, Imperial College London
Boobis, A ; Edwards, R; Taylor, G

Background

In assessing the risk from dietary chemicals, either data from epidemiological studies of exposed individuals are used or, more often, results from studies in experimental animals are extrapolated to humans. The problems of extrapolating results from experimental animals to humans are well established, and whilst the current paradigm has served well, the ability to obtain information from human subjects would enable the risk assessment to be refined, and would reduce some of the uncertainty in the process. Further, some toxicity is idiosyncratic, occurring in only a very small number of exposed subjects, due to an unknown combination of genetic and environmental factors. It is not practical and indeed may not even be feasible, to establish an experimental model with which to screen for such toxicity. Hence, studies in subjects exposed to normal levels of the chemicals in their diet are necessary. For risk assessment, epidemiological studies require an estimate of exposure and of outcome. As exposure may occur over prolonged periods of time and serious outcomes in large numbers of subjects are very unlikely, it is difficult, if not impossible to use morbidity as an effective measure of effect in subjects exposed to the low levels of most chemicals that occur in the diet. Some sensitive, early indication of effect is required. Ideally, a sensitive specific biomarker of effect would be available, with which to monitor large numbers of exposed subjects. If such a biomarker were causally linked to the event chain ultimately leading to morbidity or mortality, it could be used to detect effects of exposure much sooner and at lower exposure levels than otherwise would be possible.

Rationale and Objectives

Many proteins are released from cells, either as a consequence of normal function or chemically-induced dysfunction. These will be present in various body fluids, depending on their site of release and their normal fate. Examples include plasma, urine, saliva, sweat and broncho-alveolar lavage fluid (and cerebrospinal fluid). Accessible tissues/cells that can be analysed include: peripheral lymphocytes, granulocytes, monocytes, buccal mucosal cells, nasal mucosal cells, scalp hair follicles, sputum cells, cells from bronchoalveolar lavage, exfoliated colon cells, cervical epithelial cells, exfoliated urothelial cells and male germ cells. Modern clinical chemistry is in part based on the concept of selective release of soluble proteins from affected tissues/organs. Examples include serum transaminases from damaged hepatocytes, troponin from cardiac muscle and gamma-glutamyl transferase, from proximal tubular cells.

Proteomics offers the opportunity to perform a broad-based screen for changes in releasable proteins or peptides from cells in response to chemical insult. Whilst toxicity often only ensues after a period of continuous exposure, perhaps months or even years, it is now apparent that even during early exposure, critical events are initiated. Toxicity is a consequence of how the system adapts and responds to such effects. It should be possible to monitor these early changes with a suitable biomarker, which would then be reflective of longer term responses.

One of the areas in which most difficulty in assessment is encountered is neurotoxicity. The effects of neurotoxicants may be subtle, particularly at low doses, with little or no apparent morphological change. Interpretation of the results of tests of neurobehaviour in experimental animals is notoriously difficult. In humans, objective measures of neurological function are complex and often not possible. On the other hand, reliance on subjective estimates of behaviour is extremely unreliable, confounded by subject and, perhaps, experimenter bias. Hence, biomarkers would be particularly valuable in the study of potential neurotoxicity. Many compounds found in the diet are proven or suspected neurotoxicants. These include naturals components, such as cycasin (methylazoxymethanol) and cyanide, pesticides such as the organophosphates and pyrethrins, contaminants such as heavy metals (e.g. aluminium, lead, methylmercury), PCBs and dioxins, food additives such as aspartame, contaminating mould products such as fumonosins and 3-nitropropionic acid, and supplements such as vitamin B6.

The aim of this project was to establish proof of principle of a novel strategy for identifying molecular biomarkers of toxicity that may be used to assess early changes occurring on exposure to potential neurotoxicants.

Approach

Both in vivo and in vitro approaches were employed to identify protein biomarkers responsive to the effects of the peripherally-acting neurotoxicant, acrylamide, and the centrally-acting neurotoxicant, methylmercury. In the in vivo approach groups of rats were treated with these neurotoxicants and the release of proteins into serum, cerebral spinal fluid, urine and lung lavage fluid was assessed. Initial studies were performed with groups of rats dosed with a range of concentrations of acrylamide or methylmercury for up to 3 weeks. Samples of serum, cerebral spinal fluid, urine, and lung lavage fluid were collected at the end of the treatment period. Potential biomarkers in these biological fluids were sought using surface-enhanced laser desorption ionisation time of flight (SELDI-TOF) mass spectrometry by comparing protein profiles in samples from treated rats with those in controls. Based on these results a second study was then conducted in which groups of rats were treated with a fixed concentration of each compound, selected from the results of the first study, for up to 10 weeks. In this study analysis was restricted to samples of serum and urine. The in vitro approach employed the use of both rat and human cultured neuronal cells. The cells were treated with acrylamide and methylmercury and then analysed for biomarkers released into the medium as a result of this treatment using SELDI-TOF mass spectrometry as described above.

The results obtained from the in vitro studies were compared with those from in vivo studies in order to determine if there was coincident detection of biomarkers from both approaches and thus allow a ready validation of potential biomarkers and establish the utility of the approach for the use of human cells.

Outcome/Key Results Obtained

The protein profiling approach explored here using SELDI-TOF MS successfully demonstrated the potential of this approach for the discovery of candidate biomarkers of neurotoxicity. It appeared that the levels of a number of protein ions detected in both serum and urine of rats varied in response to the treatment of rats with both acrylamide and methylmercury. Also, it appeared possible to differentiate between the effects of the two neurotoxicants, which may reflect their different proposed mechanisms of toxicity on the neuronal system. Similarly, experiments performed in vitro using cultured rat and human neuronal cells also indicated a number of changes in secreted proteins and that these could be classified into responding to exposure to acrylamide, methylmercury, or both neurotoxicants. There appeared to be some relationship between changes in the protein profiles found in vivo and in vitro suggesting that biomarkers discovered using the approach explored here could be used for developing biomarkers for use in surveillance studies in humans.

The SELDI-TOF MS approach proved highly effective at finding differences in protein expression following treatment both in vivo and in vitro. However, the specific identification of the proteins involved proved beyond the technical capability of the instrument, which we were aware of at the outset. We attempted to purify candidate biomarkers by micro-purification techniques and made some progress with this. However, we were limited by the low amounts of the proteins of interest present in the relatively small samples available. This meant that the requirement to produce sufficient material in a partially purified form to allow proteomic identification to be performed could not be achieved. In other studies that we have performed in our laboratory we have been able to successfully purify candidate biomarker proteins from human plasma, but this required about 12 ml of plasma each time. In this study we only had only small volumes of serum and this proved to be insufficient. We are in the process of developing more sensitive techniques to address this problem so that in future studies results such as those produced in this project can be translated more efficiently to human studies.

What it means and why it is important

The main aim of this project was to establish proof of the principle that potentially novel protein biomarkers responsive to the effects of neurotoxicants can be detected in accessible biofluids, such as blood or urine. This was successfully achieved. This is important because at present there is no ready way of assessing many of the effects of neurotoxicants objectively in living subjects or experimental animals and this is a first step in addressing this deficiency. As neurotoxicants are widely present in the environment and humans are exposed to such compounds through a variety of routes it is important to develop the means to assess their risk to human health. The development of validated biomarkers of neurotoxicity would represent a significant achievement and would allow appropriate epidemiological studies to be performed.

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: