Molecular epidemiology
Integrating molecular techniques into conventional epidemiology is yielding a panoply of exciting new human data, with potential to replace some experiments on animals.
Classical epidemiological research is essentially a ‘black box’ approach. Exposure to causes of ill-health is correlated with outcome in populations, but without detailed exploration of the biological mechanisms, or the individual data that could firmly prove cause and effect.
Consequently, population research has often been supplemented with animal experiments aimed at providing the missing evidence. These experiments cause suffering and, as animal models are often imperfect [1-2], are far from ideal. Today, much sharper tools borrowed from molecular biology are prising open the ‘human black box’.
Gene genie out of the bottle
Dramatic strides have been made in genetic epidemiology by cross-fertilisation between disciplines. The first gene to be associated with asthma was identified as ADAM 33 on chromosome 20. The hunt started with a large genome-wide screen involving UK and US families with asthma [3]. Analysis of genetic linkage provided a statistical view of the co-inheritance of DNA, to see if a particular chromosomal region contained a gene related to the disease. This was followed by fine mapping of the likely chromosomal region, and led to identification of the gene by positional cloning. Further work will focus on the mechanisms by which ADAM 33 contributes to bronchial hyperreactivity.
Identical-twin studies reveal information about the impact of nature and nurture, and the roles of specific genes, in common diseases. The St Thomas’ UK Twin Registry, the largest in the world, contains clinical and genome-scan data for several thousand twins and has boosted research into cardiovascular, musculoskeletal, metabolic, dermatological and ophthalmological diseases.
Combined tactics have been used to probe how variations in DNA sequences might affect the regulation of baseline gene expression. Microarray analysis of immortalised peripheral blood lymphocytes from 14 families, combined with genome-wide linkage analysis, showed for the first time that levels of gene expression are amenable to analysis in the same way that other traits are [4]. This opens the way to a new and important suite of human genetic studies.
Tissue microarrays, developed in 1997, use thin sections from a paraffin block injected with tiny tissue cores; one section can contain tissues from hundreds of patients. Tissue microarrays are being used particularly for gene marker validation and also for multi-marker discovery [5]. The technique’s power is that it allows links to be forged between gene expression, pathology and detailed clinical data.
Hunting for biomarkers
Identifying new human biomarkers for population studies is a fast-moving field. Biomarkers can reveal exposure (e.g. to chemicals or cancer-causing viruses); assess dose or internal dose (e.g. formation of DNA adducts); identify altered structure or function (such as chromosomal aberrations or organ toxicity); and predict individual prognoses (e.g. from polymorphisms in drug-metabolism genes).
Ideally, biomarkers are measured non-invasively in the blood, breath or urine of volunteers. An intriguing example is metabonomics, a systems approach to examining hundreds or thousands of metabolites in tissues or body fluids.
A recent study of the severity of coronary heart disease used serum from human volunteers [6]. Nuclear magnetic resonance spectroscopy of serum samples, combined with novel pattern-recognition techniques, produced a highly sensitive and specific ‘fingerprint’ of biochemical changes characteristic of severe blockage of the coronary arteries. Further analysis of the molecular basis of the spectral differences should provide insights into underlying biological mechanisms.
Chemical toxicology
The predictivity of animal toxicity tests is limited by species variations and inherent uncertainties. Conventional epidemiology, lacking individual exposure indicators and being prone to confounding, has sometimes produced conflicting data. Thus for many chemicals the risks to human health remain unclear.
Organochlorine compounds have been linked with colorectal cancer, but occupational studies have given mixed results. A case-control study overcame the classical limitations by studying a population using serum levels of organochlorines as a marker for individual exposure [7].
Several organochlorines were measured in blood samples from colorectal cancer patients and controls. Point mutations in K-ras and p53 genes were assessed in tissue samples, and the expression of p53 protein was measured by immunohistochemical methods. Some polychlorinated biphenyls were associated with higher risk of colorectal cancer and with mutations in K-ras and p53 genes. The results strongly indicate a causal link.
In another example, the study population comprised workers occupationally exposed to cobalt and tungsten [8]. Exhaled breath condensate provided a means to assess individual doses in the lungs, and levels of malondialdehyde were a biomarker for pulmonary oxidative stress. Results suggested that exhaled breath condensate will be useful for integrating biomonitoring and health surveillance procedures among workers.
Recent technical developments (such as fluorescence in situ hybridisation painting of chromosomes) are transforming population-level assays for chromosomal aberrations and micronuclei. These sensitive biomarkers of exposure to genotoxic substances have proved their worth in studies of occupational and environmental hazards, and prevention policies can be expected to benefit.
Traditional epidemiology has already underpinned many life-saving public health policies. The new molecular toolbox is dramatically improving epidemiology’s ability to clarify links between exposure and effect, whether in terms of disease, treatments or toxicity, in the species of interest — as well as replacing experiments on animals.
References
1. Linazasoro G (2004). Recent failures of new potential symptomatic treatments for Parkinson’s Disease: causes and solutions. Movement Dis 19:743-754.
2. Sloviter RS (2005). The neurobiology of temporal lobe epilepsy: too much information, not enough knowledge. C R Biologies 328:143-153.
3. Holgate ST, Davies DE, Rorke S et al (2004). ADAM 33 and its association with airway remodeling and hyperresponsiveness in asthma. Clin Rev Allergy Immunol 27:23-34.
4. Morley M, Molony CM, Weber TM et al (2004). Genetic analysis of genome-wide variation in human gene expression. Nature 430:743-747.
5. Harding A (2005). Tissue microarrays go mainstream. The Scientist, 11 April, 26-28.
6. Brindle JT, Antti H, Holmes E et al (2002). Rapid and noninvasive diagnosis of the presence and severity of coronary heart disease using 1H-NMR-based metabonomics. Nat Med 8:1439-1444.
7. Howsam M, Grimalt JO, Guino E et al (2004). Organochlorine exposure and colorectal cancer risk. Environ Health Perspect 112:1460-1466.
8. Goldoni M, Catalani S, De Palma G et al (2004). Exhaled breath condensate as a suitable matrix to assess lung dose and effects in workers exposed to cobalt and tungsten. Environ Health Perspect 112:1293-1298.
(Originally published in the Dr Hadwen Trust’s Science Review 2005)