BACKGROUND

DNA damage is an important biomarker, in relation to human health and disease. It is primarily associated with cancer, as DNA damage is the precursor for mutations and chromosomal alterations, and they in turn can lead to cell transformation and tumorigenesis. However, the link is not straightforward, since almost all incident damage is repaired, and mutations and chromosome changes are effective only if they influence those genes - oncogenes, tumour suppressor genes and susceptibility genes - that are involved in the pathway to carcinogenesis. DNA damage is elevated in various diseases apart from cancer, whether as cause or effect.

A common factor in disease aetiology and progression is oxidative stress, caused by overproduction of ROS or deficient antioxidant defences. The level of oxidation of bases in DNA - typically measured in PBMN cells, but also sometimes in tissues, or buccal epithelial cells, or cells retrieved from alveolar lavage - serves as a convenient marker for general oxidative damage, and thus for oxidative stress. The method of choice for measuring DNA damage in human biomonitoring is the comet assay, or single cell gel electrophoresis (Figure 1).

Figure 1. The comet assay, as used to measure DNA base damage (such as oxidised guanine)

Briefly, cells are embedded in agarose on a microscope slide and lysed in a buffer containing detergents and high salt which together dissolve cell and nuclear membranes, release soluble constituents, and strip histones from the DNA, leaving the DNA attached at intervals to a nuclear matrix, effectively as a series of supercoiled loops. Alkaline incubation follows, converting alkali-labile sites to strand breaks (SBs). During electrophoresis at high pH, DNA loops that have lost their supercoiling because of a SB are drawn towards the anode, forming a comet-like image when viewed by fluorescence microscopy. The % of DNA in the comet tail reflects the frequency of DNA breaks (calibrated using ionising radiation which introduces a known break frequency per Gray). The lysis step can be followed by digestion of DNA with an enzyme that recognises specific lesions and converts them to DNA breaks. The most commonly used enzyme is formamidopyrimidine DNA glycosylase (FPG), which recognises 8- oxoguanine (and some other altered guanines). Thus the comet assay can be used to measure not just SBs but oxidised bases - a far more specific indicator of ROS attack and oxidative stress.

Recently developed high throughput versions of the comet assay, with automated scoring, make it far more amenable for use as a biomonitoring tool, since epidemiological studies demand large numbers of samples. The comet assay has been further modified to measure the capacity of cells to repair DNA damage - an essential element in our defences against cancer. In an in vitro assay, a cell extract is incubated with substrate DNA containing specific lesions; oxidised bases (to measure base excision repair, BER), or UV induced pyrimidine dimers (to measure nucleotide excision repair, NER). The DNA breaks induced by repair enzymes in the extract are measured with the comet assay.

hCOMET [human Comet], is centred on the use of the comet assay in human studies, and most of what follows will be exclusively about this application.

Biomonitoring in environmental and occupational exposure scenarios

The comet assay has been employed in around 140 studies of occupational exposure to xenobiotics, and approximately 40 investigations of environmental exposure of humans. Levels of DNA damage have been measured in populations exposed to environmental pollution, including polycyclic aromatic hydrocarbons, particulate matter, benzene and heavy metals, and in addition environmental tobacco smoke. Effects of exposure to ionising radiation following the Chernobyl accident, indoor radon, and high natural radiation levels have also been investigated.

Occupational exposure studies have focused on volatile organic compounds such as styrene, benzene, toluene, vinyl chloride; metals such as welding fumes, chromium, lead, mercury; pesticides; asbestos and mineral fibres; and anaesthetic gases, chemotherapy drugs and ionising radiation in medical personnel. (It is worth noting that the assay is also used in ecological biomonitoring, and much of the methodological discussion here applies equally to that field).

Nutritional epidemiology

The influence of nutritional factors on DNA damage levels (and by inference on disease incidence, notably cancer) has been studied with the comet assay using standard epidemiological approaches, in particular observational or ecological studies, and dietary interventions. Analyses of DNA damage levels in combination with information about dietary intake or measurements of micronutrients in the blood have shown, for instance, a negative correlation between carotenoids and DNA base oxidation, and a reduction of DNA damage in subjects on vegetarian or high vitamin content diets. Such correlations do not necessarily indicate causal relationships, but might simply be associations, brought about by other, undefined common factors. Intervention studies give more conclusive evidence of cause and effect. Many such studies have been carried out, with antioxidant or other phytochemical supplements, addition of specific foods (usually fruits) to the diet, or major changes in the diet (e.g. including 600 grams per day of vegetables or fruits).

To be reliable, trials should be placebo-controlled, though 'placebo' is a misnomer where real foods are concerned.

Crossover studies are popular, in which each subject takes placebo or supplement in two phases separated by a washout period.

With DNA damage as an endpoint, roughly half the studies have shown a reduction in SBs or oxidised bases and half have shown no effect. Individual antioxidant status can be assessed by treating white blood cells in vitro with an oxidising chemical (H2O2); a low yield of SBs indicates a high antioxidant status. The value of decreasing oxidative base damage in healthy individuals is debatable, since a certain amount of oxidative stress is inevitable and even desirable; for example, the inflammatory response deploys ROS as a defence mechanism, and ROS are involved in crucial cell signalling pathways.

Other endpoints are also amenable to study, notably DNA repair - shown in several studies to be enhanced by fruit consumption.

DNA damage as an indicator of disease

Oxidised DNA bases and/or SBs are detected at elevated levels in PBMN cells from patients with diabetes, inflammatory disorders (arthritis, ankylosing spondylitis), various cancers (breast, cervix, lung, oesophagus, prostate), neurodegenerative disorders (Alzheimer's and Parkinson's diseases), and cardiovascular diseases; the DNA oxidation damage is often accompanied by depressed antioxidant status.

To determine whether DNA damage is a cause or an effect of disease, prospective studies are necessary, but have not yet been carried out; a large cohort of healthy individuals would have their DNA damage analysed, and would be monitored for years so that disease incidence and mortality could be retrospectively linked to the earlier burden of DNA damage.

A similar approach is needed to establish whether intrinsic ability to repair DNA is a factor in determining susceptibility to disease (particularly cancer). 

DNA damage and aging

It is commonly believed that a major player in the process of aging is the accumulation of oxidative damage to biomolecules, including DNA. Such an accumulation might result from increased exposure to ROS (for instance, if mitochondria become 'leaky') or from depressed defences, such as a decline in DNA repair capacity.

Most studies to date have been carried out in PBMN cells, and results are not consistent, with only a weak positive correlation overall. DNA repair capacity, also, has been variously reported as not affected, decreased or even increased with age.

Pros and cons of the comet assay for measuring DNA damage: comparison with other methods

DNA damage can be detected with antibodies, by 32P-postlabelling, by diverse chromatographic techniques, or with the comet assay and related methods. Antibodies have been isolated with activity against 8-oxoguanine, for example; they are generally found to lack sufficient specificity to be used for quantitative measurements. 32P-postlabelling is valuable for measuring adducts in DNA (for instance, polycyclic aromatic hydrocarbons) with high sensitivity. Chromatographic techniques include gas chromatography-mass spectroscopy (GC-MS), high performance liquid chromatography with electrochemical detection (HPLC-ECD), and liquid chromatography-tandem mass spectroscopy (LCMS/MS).

In addition to the comet assay, two older established methods - alkaline unwinding and alkaline elution - are capable of measuring low levels of DNA breaks, and both have been adapted to incorporate FPG for detecting oxidised purines. The EC-funded Concerted Action ESCODD (European Standards Committee on Oxidative DNA Damage) was set up in the late 1990s to investigate the serious discrepancies in estimation of background levels of DNA oxidation (specifically, 8-oxoguanine) in humans. Depending on the technique used, estimates varied by orders of magnitude. By means of ring studies with samples containing different levels of experimentally induced 8-oxoguanine, distributed to partners using different techniques, it was established that while HPLC was very precise when measuring a range of concentrations of induced 8-oxoguanine lesions, it gave highly variable estimates of background levels, apparently because of oxidation occurring during sample preparation and giving rise to a serious artefact. (Mass spectroscopy methods tested by ESCODD were incapable of detecting the induced 8-oxoguanine dose response.) It was impossible to control the adventitious oxidation reliably. In contrast, the comet assay, alkaline unwinding and alkaline elution in combination with FPG - all relatively gentle procedures - were not prone to this oxidation and so are more accurate than chromatographic methods. The consensus in ESCODD was that background 8-oxoguanine levels are much lower than previously estimated, at between 0.3 and 4 per 106 guanines.

Progress beyond the state-of-the-art

FPG-based methods are not perfect. While the comet assay lpus FPG came out well in the ESCODD project in terms of accuracy, it lacks precision, compared with analytical chemical methods. Within a laboratory, intra-experimental variation (expressed as coefficient of variation for identical cell samples in one experiment) can be 10% or less, given well controlled experimental conditions, while inter-experimental variation (identical cell samples in different experiments) tends to be rather higher (typically around 14%). Problems arise when results are compared between laboratories; eight ESCODD partners were given identical samples to analyse for 8-oxoguanine, ostensibly using similar protocols, and the resulting CV was 80%. Since ESCODD, attempts have been made to reduce variation, and to identify its causes, by the ECVAG consortium, and by individual laboratories. Critical factors are agarose concentration (lower % tail DNA with higher agarose concentration), period of alkaline incubation (% tail DNA increasing with time), and electrophoresis conditions (% tail DNA increasing with voltage and with time, and also sensitive to temperature). Adoption of standard conditions should reduce inter-laboratory variation - a long-term aim. Scoring methods vary between laboratories, and this introduces variability that is more difficult to control.

Approaches to controlling variation

One solution to the problem of high variability is to employ reference standards. Internal standards would be ideal, with reference cells (containing a known amount of damage) included in the same gel as sample cells. However, it is necessary to identify the two classes of comets (sample and standard) after electrophoresis, and while different approaches have given promising preliminary results, they are not sufficiently practicable to be widely adopted.

Reference standards can, however, be useful, if run in parallel with samples but in separate gels. It is then possible (a) to check the performance of the assay (anomalous results from the standard cells indicating a technical problem); and (b) to 'normalise' sample results by applying a correction factor based on the value of % tail DNA seen in the reference standard. Thus, interexperimental variation can be controlled, and this is useful in the case of biomonitoring when typically hundreds of samples are collected and run in a series of experiments over a period of weeks or months. Inter-laboratory variation is also controllable.

In the ESCODD project, seven partners collected PBMN cells from volunteers, and measured 8-oxoguanine (FPG-sensitive sites); the overall mean value was 0.34 per 106 guanines, with a coefficient of variation (CV) of 43%. This could indicate real differences between the population samples (in six different countries). However, the laboratories also analysed 8-oxoguanine in identical reference standard cells sent by the coordinating laboratory. When the PBMN results were 'normalised' by simply dividing by the reference standard result for each laboratory, the CV (omitting one obvious outlier) was reduced to 14% - suggesting that there were no real differences in levels of DNA oxidation among the countries tested.

Consolidation of comet assay data

At the International Comet Assay Workshop in Kusadasi, Turkey, in September 2011, the project known as ComNet was initiated to address the need for better coordination of biomonitoring studies using the comet assay. To date, the project has been run without core funding, relying on individual researchers' resources. Under this project, laboratories publishing papers on human studies with the comet assay were invited to contribute to a collaborative effort, and over 100 laboratories worldwide registered on a dedicated website. A questionnaire was distributed to these laboratories, to establish the extent of epidemiological data available, and to collect information about the technical protocols that are in use. Responses have been received from over 50 laboratories, agreeing to donate their data. The geographical distribution of laboratories expressing interest is mostly European, but with significant participation from the US, India, China, and Latin America. Studies include cohort/prospective, cross-sectional and intervention studies, case-control studies of various diseases, investigations of occupational and environmental exposure to xenobiotics, nutritional studies, and examinations of age-related effects, sex differences, etc. The overall number of individual subject data sets that could be contributed by these laboratories is about 19,000. hCOMET is built on the foundations of ComNet.

A principal purpose of ComNet (and therefore of hCOMET) is to extract the maximum amount of scientifically valuable information from the pooled analysis of this wealth of data. This will establish baseline damage levels (SBs and oxidised bases) for future reference and will define the associations between comet assay measurements of DNA damage and repair, and factors such as sex, age, smoking status, nutrition and lifestyle.