Towards unravelling biological mechanisms behind radiation-induced oral mucositis via mass spectrometry-based proteomics.

Objective: Head and neck cancer (HNC) accounts for almost 890,000 new cases per year. Radiotherapy (RT) is used to treat the majority of these patients. A common side-effect of RT is the onset of oral mucositis, which decreases the quality of life and represents the major dose-limiting factor in RT. To understand the origin of oral mucositis, the biological mechanisms post-ionizing radiation (IR) need to be clarified. Such knowledge is valuable to develop new treatment targets for oral mucositis and markers for the early identification of "at-risk" patients. Methods: Primary keratinocytes from healthy volunteers were biopsied, irradiated in vitro (0 and 6 Gy), and subjected to mass spectrometry-based analyses 96 h after irradiation. Web-based tools were used to predict triggered biological pathways. The results were validated in the OKF6 cell culture model. Immunoblotting and mRNA validation was performed and cytokines present in cell culture media post-IR were quantified. Results: Mass spectrometry-based proteomics identified 5879 proteins in primary keratinocytes and 4597 proteins in OKF6 cells. Amongst them, 212 proteins in primary keratinocytes and 169 proteins in OKF6 cells were differentially abundant 96 h after 6 Gy irradiation compared to sham-irradiated controls. In silico pathway enrichment analysis predicted interferon (IFN) response and DNA strand elongation pathways as mostly affected pathways in both cell systems. Immunoblot validations showed a decrease in minichromosome maintenance (MCM) complex proteins 2-7 and an increase in IFN-associated proteins STAT1 and ISG15. In line with affected IFN signalling, mRNA levels of IFNß and interleukin 6 (IL-6) increased significantly following irradiation and also levels of secreted IL-1ß, IL-6, IP-10, and ISG15 were elevated. Conclusion: This study has investigated biological mechanisms in keratinocytes post-in vitro ionizing radiation. A common radiation signature in keratinocytes was identified. The role of IFN response in keratinocytes along with increased levels of pro-inflammatory cytokines and proteins could hint towards a possible mechanism for oral mucositis.

The effect of data aggregation on dispersion estimates in count data models.

For the modelling of count data, aggregation of the raw data over certain subgroups or predictor configurations is common practice. This is, for instance, the case for count data biomarkers of radiation exposure. Under the Poisson law, count data can be aggregated without loss of information on the Poisson parameter, which remains true if the Poisson assumption is relaxed towards quasi-Poisson. However, in biodosimetry in particular, but also beyond, the question of how the dispersion estimates for quasi-Poisson models behave under data aggregation have received little attention. Indeed, for real data sets featuring unexplained heterogeneities, dispersion estimates can increase strongly after aggregation, an effect which we will demonstrate and quantify explicitly for some scenarios. The increase in dispersion estimates implies an inflation of the parameter standard errors, which, however, by comparison with random effect models, can be shown to serve a corrective purpose. The phenomena are illustrated by γ-H2AX foci data as used for instance in radiation biodosimetry for the calibration of dose-response curves.

Posterior subcapsular cataracts are a late effect after acute exposure to 0.5 Gy ionizing radiation in mice.

PURPOSE: The long-term effect of low and moderate doses of ionizing radiation on the lens is still a matter of debate and needs to be evaluated in more detail. MATERIAL AND METHODS: We conducted a detailed histological analysis of eyes from B6C3F1 mice cohorts after acute gamma irradiation (60Co source; 0.063 Gy/min) at young adult age of 10 weeks with doses of 0.063, 0.125, and 0.5 Gy. Sham irradiated (0 Gy) mice were used as controls. To test for genetic susceptibility heterozygous Ercc2 mutant mice were used and compared to wild-type mice of the same strain background. Mice of both sexes were included in all cohorts. Eyes were collected 4 h, 12, 18 and 24 months after irradiation. For a better understanding of the underlying mechanisms, metabolomics analyses were performed in lenses and plasma samples of the same mouse cohorts at 4 and 12 h as well as 12, 18 and 24 months after irradiation. For this purpose, a targeted analysis was chosen. RESULTS: This analysis revealed histological changes particularly in the posterior part of the lens that rarely can be observed by using Scheimpflug imaging, as we reported previously. We detected a significant increase of posterior subcapsular cataracts (PSCs) 18 and 24 months after irradiation with 0.5 Gy (odds ratio 9.3; 95% confidence interval 2.1-41.3) independent of sex and genotype. Doses below 0.5 Gy (i.e. 0.063 and 0.125 Gy) did not significantly increase the frequency of PSCs at any time point. In lenses, we observed a clear effect of sex and aging but not of irradiation or genotype. While metabolomics analyses of plasma from the same mice showed only a sex effect. CONCLUSIONS: This article demonstrates a significant radiation-induced increase in the incidence of PSCs, which could not be identified using Scheimpflug imaging as the only diagnostic tool.

Dose-dependent long-term effects of a single radiation event on behaviour and glial cells.

PURPOSE: The increasing use of low-dose ionizing radiation in medicine requires a systematic study of its long-term effects on the brain, behaviour and its possible association with neurodegenerative disease vulnerability. Therefore, we analysed the long-term effects of a single low-dose irradiation exposure at 10 weeks of age compared to medium and higher doses on locomotor, emotion-related and sensorimotor behaviour in mice as well as on hippocampal glial cell populations. MATERIALS AND METHODS: We determined the influence of radiation dose (0, 0.063, 0.125 or 0.5 Gy), time post-irradiation (4, 12 and 18 months p.i.), sex and genotype (wild type versus mice with Ercc2 DNA repair gene point mutation) on behaviour. RESULTS: The high dose (0.5 Gy) had early-onset adverse effects at 4 months p.i. on sensorimotor recruitment and late-onset negative locomotor effects at 12 and 18 months p.i. Notably, the low dose (0.063 Gy) produced no early effects but subtle late-onset (18 months) protective effects on sensorimotor recruitment and exploratory behaviour. Quantification and morphological characterization of the microglial and the astrocytic cells of the dentate gyrus 24 months p.i. indicated heightened immune activity after high dose irradiation (0.125 and 0.5 Gy) while conversely, low dose (0.063 Gy) induced more neuroprotective features. CONCLUSION: This is one of the first studies demonstrating such long-term and late-onset effects on brain and behaviour after a single radiation event in adulthood.

Double-strand breaks in lymphocyte DNA of humans exposed to [18F]fluorodeoxyglucose and the static magnetic field in PET/MRI.

BACKGROUND: Given the increasing clinical use of PET/MRI, potential risks to patients from simultaneous exposure to ionising radiation and (electro)magnetic fields should be thoroughly investigated as a precaution. With this aim, the genotoxic potential of 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) and a strong static magnetic field (SMF) were evaluated both in isolation and in combination using the γH2AX assay detecting double-strand breaks in lymphocyte DNA. METHODS: Thirty-two healthy young volunteers allocated to three study arms were exposed to [18F]FDG alone, to a 3-T SMF alone or to both combined over 60 min at a PET/CT or a PET/MRI system. Blood samples taken after in vivo exposure were incubated up to 60 min to extend the irradiation of blood by residual [18F]FDG within the samples and the time to monitor the γH2AX response. Absorbed doses to lymphocytes delivered in vivo and in vitro were estimated individually for each volunteer exposed to [18F]FDG. γH2AX foci were scored automatically by immunofluorescence microscopy. RESULTS: Absorbed doses to lymphocytes exposed over 60 to 120 min to [18F]FDG varied between 1.5 and 3.3 mGy. In this time interval, the radiotracer caused a significant median relative increase of 28% in the rate of lymphocytes with at least one γH2AX focus relative to the background rate (p = 0.01), but not the SMF alone (p = 0.47). Simultaneous application of both agents did not result in a significant synergistic or antagonistic outcome (p = 0.91). CONCLUSION: There is no evidence of a synergism between [18F]FDG and the SMF that may be of relevance for risk assessment of PET/MRI.

CREB Signaling Mediates Dose-Dependent Radiation Response in the Murine Hippocampus Two Years after Total Body Exposure.

The impact of low-dose ionizing radiation (IR) on the human brain has recently attracted attention due to the increased use of IR for diagnostic purposes. The aim of this study was to investigate low-dose radiation response in the hippocampus. Female B6C3F1 mice were exposed to total body irradiation with 0 (control), 0.063, 0.125, or 0.5 Gy. Quantitative label-free proteomic analysis of the hippocampus was performed after 24 months. CREB signaling and CREB-associated pathways were affected at all doses. The lower doses (0.063 and 0.125 Gy) induced the CREB pathway, whereas the exposure to 0.5 Gy deactivated CREB. Similarly, the lowest dose (0.063 Gy) was anti-inflammatory, reducing the number of activated microglia. In contrast, induction of activated microglia and reactive astroglia was found at 0.5 Gy, suggesting increased inflammation and astrogliosis, respectively. The apoptotic markers BAX and cleaved CASP-3 and oxidative stress markers were increased only at the highest dose. Since the activated CREB pathway plays a central role in learning and memory, these data suggest neuroprotection at the lowest dose (0.063 Gy) but neurodegeneration at 0.5 Gy. The response to 0.5 Gy resembles alterations found in healthy aging and thus may represent radiation-induced accelerated aging of the brain.

Automated scoring of dicentric chromosomes differentiates increased radiation sensitivity of young children after low dose CT exposure in vitro.

PURPOSE: Automated detection of dicentric chromosomes from a large number of cells was applied to study age-dependent radiosensitivity after in vitro CT exposure of blood from healthy donors. MATERIALS AND METHODS: Blood samples from newborns, children (2-5 years) and adults (20-50 years) were exposed in vitro to 0 mGy, 41 mGy and 978 mGy using a CT equipment. In this study, automated scoring based on 13,000-31,000 cells/dose point/age group was performed. Results for control and low dose points were validated by manually counting about 26,000 cells/dose point/age group. RESULTS: For all age groups, the high number of analyzed cells enabled the detection of a significant increase in the frequency of radiation induced dicentric chromosomes in cells exposed to 41 mGy as compared to control cells. Moreover, differences between the age groups could be resolved for the low dose: young donors showed significantly increased risk for induced dicentrics at 41 mGy compared to adults. CONCLUSIONS: The results very clearly demonstrate that the automated dicentric scoring method is capable of discerning radiation induced biomarkers in the low dose range (<100 mGy) and thus may open possibilities for large-scale molecular epidemiology studies in radiation protection.

Age-dependent differences in DNA damage after in vitro CT exposure.

PURPOSE: Age dependent radiation sensitivity for DNA damage after in vitro blood exposure by computer tomography (CT) was investigated. MATERIALS AND METHODS: Radiation biomarkers (dicentrics and gammaH2AX) in blood samples of newborns, children under five years and adults after sham exposure (0 mGy), low-dose (41 mGy) and high-dose (978 mGy) in vitro CT exposure were analyzed. RESULTS: Significantly higher levels of dicentric induction were found for the single and combined newborns/children group compared to adults, by a factor of 1.48 (95% CI 1.30-1.68), after exposure to 978 mGy. Although a significant dose response for damage induction and dose-dependent repair was found, the gammaH2AX assay did not show an age-dependent increase in DNA damage in newborns/children compared to adults. This was the case for the gammaH2AX levels after repair time intervals of 30 minutes and 24 hours, after correcting for the underlying background damage. For the low dose of 41 mGy, the power of the dicentric assay was also not sufficient to detect an age-dependent effect in the sample size investigated. CONCLUSION: A 1.5-fold increased level of dicentric aberrations is detected in newborns and children under five years after 1 Gy radiation exposure.

Lifetime study in mice after acute low-dose ionizing radiation: a multifactorial study with special focus on cataract risk.

Because of the increasing application of ionizing radiation in medicine, quantitative data on effects of low-dose radiation are needed to optimize radiation protection, particularly with respect to cataract development. Using mice as mammalian animal model, we applied a single dose of 0, 0.063, 0.125 and 0.5 Gy at 10 weeks of age, determined lens opacities for up to 2 years and compared it with overall survival, cytogenetic alterations and cancer development. The highest dose was significantly associated with increased body weight and reduced survival rate. Chromosomal aberrations in bone marrow cells showed a dose-dependent increase 12 months after irradiation. Pathological screening indicated a dose-dependent risk for several types of tumors. Scheimpflug imaging of the lens revealed a significant dose-dependent effect of 1% of lens opacity. Comparison of different biological end points demonstrated long-term effects of low-dose irradiation for several biological end points.

The second gamma-H2AX assay inter-comparison exercise carried out in the framework of the European biodosimetry network (RENEB).

PURPOSE: Within the EU RENEB project, seven laboratories have taken part in training and harmonisation activities to strengthen triage gamma-H2AX-based radiation exposure assessment. This has culminated in a second triage biodosimetry exercise. MATERIALS AND METHODS: Whole blood and separated lymphocyte samples were homogenously irradiated with 60Co gamma rays at 0.5, 2.5 (blind samples), 0 and 2 Gy (reference samples). Following post-exposure incubations of 4 and 24 h, 16 samples were shipped on ice packs to each partner. The samples were stained and scored for gamma-H2AX foci, using manual and/or automated fluorescence microscope scoring strategies. Dose estimates were obtained and used to assign triage categories to the samples. RESULTS: Average dose estimates across all the laboratories correlated well with true doses. The most accurate assignment of triage category was achieved by manual scoring of the 4-h blood and lymphocyte samples. Only three samples out of a total of 46 were miscategorized in a way that could have adversely effected the clinical management of a radiation casualty. CONCLUSIONS: This inter-comparison exercise has demonstrated that following a recent acute radiation exposure, the gamma-H2AX assay could be a useful triage tool that can be successfully applied across a network of laboratories.

Differences in DNA Repair Capacity, Cell Death and Transcriptional Response after Irradiation between a Radiosensitive and a Radioresistant Cell Line.

Normal tissue toxicity after radiotherapy shows variability between patients, indicating inter-individual differences in radiosensitivity. Genetic variation probably contributes to these differences. The aim of the present study was to determine if two cell lines, one radiosensitive (RS) and another radioresistant (RR), showed differences in DNA repair capacity, cell viability, cell cycle progression and, in turn, if this response could be characterised by a differential gene expression profile at different post-irradiation times. After irradiation, the RS cell line showed a slower rate of γ-H2AX foci disappearance, a higher frequency of incomplete chromosomal aberrations, a reduced cell viability and a longer disturbance of the cell cycle when compared to the RR cell line. Moreover, a greater and prolonged transcriptional response after irradiation was induced in the RS cell line. Functional analysis showed that 24 h after irradiation genes involved in "DNA damage response", "direct p53 effectors" and apoptosis were still differentially up-regulated in the RS cell line but not in the RR cell line. The two cell lines showed different response to IR and can be distinguished with cell-based assays and differential gene expression analysis. The results emphasise the importance to identify biomarkers of radiosensitivity for tailoring individualized radiotherapy protocols.

New mutation in the mouse Xpd/Ercc2 gene leads to recessive cataracts.

Cataracts are the major eye disorder and have been associated mainly with mutations in lens-specific genes, but cataracts are also frequently associated with complex syndromes. In a large-scale high-throughput ENU mutagenesis screen we analyzed the offspring of paternally treated C3HeB/FeJ mice for obvious dysmorphologies. We identified a mutant suffering from rough coat and small eyes only in homozygotes; homozygous females turned out to be sterile. The mutation was mapped to chromosome 7 between the markers 116J6.1 and D7Mit294;4 other markers within this interval did not show any recombination among 160 F2-mutants. The critical interval (8.6 Mb) contains 3 candidate genes (Apoe, Six5, Opa3); none of them showed a mutation. Using exome sequencing, we identified a c.2209T>C mutation in the Xpd/Ercc2 gene leading to a Ser737Pro exchange. During embryonic development, the mutant eyes did not show major changes. Postnatal histological analyses demonstrated small cortical vacuoles; later, cortical cataracts developed. Since XPD/ERCC2 is involved in DNA repair, we checked also for the presence of the repair-associated histone γH2AX in the lens. During the time, when primary lens fiber cell nuclei are degraded, γH2AX was strongly expressed in the cell nuclei; later, it demarcates clearly the border of the lens cortex to the organelle-free zone. Moreover, we analyzed also whether seemingly healthy heterozygotes might be less efficient in repair of DNA damage induced by ionizing radiation than wild types. Peripheral lymphocytes irradiated by 1Gy Cs137 showed 6 hrs after irradiation significantly more γH2AX foci in heterozygotes than in wild types. These findings demonstrate the importance of XPD/ERCC2 not only for lens fiber cell differentiation, but also for the sensitivity to ionizing radiation. Based upon these data, we hypothesize that variations in the human XPD/ERCC2 gene might increase the susceptibility for several disorders besides Xeroderma pigmentosum in heterozygotes under particular environmental conditions.

Genetic factors in individual radiation sensitivity.

Cancer risk and radiation sensitivity are often associated with alterations in DNA repair, cell cycle, or apoptotic pathways. Interindividual variability in mutagen or radiation sensitivity and in cancer susceptibility may also be traced back to polymorphisms of genes affecting e.g. DNA repair capacity. We studied possible associations between 70 polymorphisms of 12 DNA repair genes with basal and initial DNA damage and with repair thereof. We investigated DNA damage induced by ionizing radiation in lymphocytes isolated from 177 young lung cancer patients and 169 cancer-free controls. We also sought replication of our findings in an independent sample of 175 families (in total 798 individuals). DNA damage was assessed by the Olive tail moment (OTM) of the comet assay. DNA repair capacity (DRC) was determined for 10, 30 and, 60min of repair. Genes involved in the single-strand-repair pathway (SSR; like XRCC1 and MSH2) as well as genes involved in the double-strand-repair pathway (DSR; like RAD50, XRCC4, MRE11 and ATM) were found to be associated with DNA damage. The most significant association was observed for marker rs3213334 (p=0.005) of XRCC1 with basal DNA damage (B), in both cases and controls. A clear additive effect on the logarithm of OTM was identified for the marker rs1001581 of the same LD-block (p=0.039): BCC=-1.06 (95%-CI: -1.16 to -0.96), BCT=-1.02 (95%-CI: -1.11 to -0.93) and BTT=-0.85 (95%-CI: -1.01 to -0.68). In both cases and controls, we observed significantly higher DNA basal damage (p=0.007) for carriers of the genotype AA of marker rs2237060 of RAD50 (involved in DSR). However, this could not be replicated in the sample of families (p=0.781). An alteration to DRC after 30min of repair with respect to cases was observed as borderline significant for marker rs611646 of ATM (involved in DSR; p=0.055), but was the most significant finding in the sample of families (p=0.009). Our data indicate that gene variation impacts measurably on DNA damage and repair, suggesting at least a partial contribution to radiation sensitivity and lung cancer susceptibility.

Radiation-induced alterations of histone post-translational modification levels in lymphoblastoid cell lines.

BACKGROUND: Radiation-induced alterations in posttranslational histone modifications (PTMs) may affect the cellular response to radiation damage in the DNA. If not reverted appropriately, altered PTM patterns may cause long-term alterations in gene expression regulation and thus lead to cancer. It is therefore important to characterize radiation-induced alterations in PTM patterns and the factors affecting them. METHODS: A lymphoblastoid cell line established from a normal donor was used to screen for alterations in methylation levels at H3K4, H3K9, H3K27, and H4K20, as well as acetylation at H3K9, H3K56, H4K5, and H4K16, by quantitative Western Blot analysis at 15 min, 1 h and 24 h after irradiation with 2 Gy and 10 Gy. The variability of alterations in acetylation marks was in addition investigated in a panel of lymphoblastoid cell lines with differing radiosensitivity established from lung cancer patients. RESULTS: The screening procedure demonstrated consistent hypomethylation at H3K4me3 and hypoacetylation at all acetylation marks tested. In the panel of lymphoblastoid cell lines, however, a high degree of inter-individual variability became apparent. Radiosensitive cell lines showed more pronounced and longer lasting H4K16 hypoacetylation than radioresistant lines, which correlates with higher levels of residual γ-H2AX foci after 24 h. CONCLUSION: So far, the factors affecting extent and duration of radiation-induced histone alterations are poorly defined. The present work hints at a high degree of inter-individual variability and a potential correlation of DNA damage repair capacity and alterations in PTM levels.

Are mouse lens epithelial cells more sensitive to γ-irradiation than lymphocytes?

In this pilot study we compared for the first time the radiation sensitivity of mouse lens epithelial cells (LECs) and mouse lymphocytes. We freshly prepared LECs and lymphocytes and irradiated them with γ-rays ((137)Cs; doses ranging from 0.25 to 2 Gy). DNA damage and repair were evaluated by alkaline comet assay and γH2AX foci assay. Using the comet assay, we observed a dose-dependent increase in DNA damage in both cell types. The faster formation of single- and double-strand breaks in LECs of C57BL/6 mice at doses below 1 Gy needs to be confirmed in other mouse strains. Immunofluorescence for γH2AX foci showed a higher degree of lesions in LECs from C57BL/6J mice compared to those of JF1 mice and to lymphocytes of both strains. Correspondingly, repair of DNA damage proceeded faster in LECs of C57BL/6J mice compared to LECs of JF1 mice and lymphocytes of both strains. It is obvious that the lymphocytes of both strains repaired DNA lesions more slowly than the corresponding LECs. In conclusion, our results demonstrate that LECs of C57Bl/6 mice show a steeper dose-response than lymphocytes in both types of experiments. It shows that both test systems are able to be used also at doses below 0.25 Gy. The observed difference in DNA repair between the LECs from C57BL/6J mice compared to the LECs from JF1 mice and to the lymphocytes of both strains warrants further experiments to identify the underlying molecular mechanisms.

Evaluation of different biomarkers to predict individual radiosensitivity in an inter-laboratory comparison--lessons for future studies.

Radiotherapy is a powerful cure for several types of solid tumours, but its application is often limited because of severe side effects in individual patients. With the aim to find biomarkers capable of predicting normal tissue side reactions we analysed the radiation responses of cells from individual head and neck tumour and breast cancer patients of different clinical radiosensitivity in a multicentric study. Multiple parameters of cellular radiosensitivity were analysed in coded samples of peripheral blood lymphocytes (PBLs) and derived lymphoblastoid cell lines (LCLs) from 15 clinical radio-hypersensitive tumour patients and compared to age- and sex-matched non-radiosensitive patient controls and 15 lymphoblastoid cell lines from age- and sex- matched healthy controls of the KORA study. Experimental parameters included ionizing radiation (IR)-induced cell death (AnnexinV), induction and repair of DNA strand breaks (Comet assay), induction of yH2AX foci (as a result of DNA double strand breaks), and whole genome expression analyses. Considerable inter-individual differences in IR-induced DNA strand breaks and their repair and/or cell death could be detected in primary and immortalised cells with the applied assays. The group of clinically radiosensitive patients was not unequivocally distinguishable from normal responding patients nor were individual overreacting patients in the test system unambiguously identified by two different laboratories. Thus, the in vitro test systems investigated here seem not to be appropriate for a general prediction of clinical reactions during or after radiotherapy due to the experimental variability compared to the small effect of radiation sensitivity. Genome-wide expression analysis however revealed a set of 67 marker genes which were differentially induced 6 h after in vitro-irradiation in lymphocytes from radio-hypersensitive and non-radiosensitive patients. These results warrant future validation in larger cohorts in order to determine parameters potentially predictive for clinical radiosensitivity.

Heritability of radiation response in lung cancer families.

Radiation sensitivity is assumed to be a cancer susceptibility factor due to impaired DNA damage signalling and repair. Relevant genetic factors may also determine the observed familial aggregation of early onset lung cancer. We investigated the heritability of radiation sensitivity in families of 177 Caucasian cases of early onset lung cancer. In total 798 individuals were characterized for their radiation-induced DNA damage response. DNA damage analysis was performed by alkaline comet assay before and after in vitro irradiation of isolated lymphocytes. The cells were exposed to a dose of 4 Gy and allowed to repair induced DNA-damage up to 60 minutes. The primary outcome parameter Olive Tail Moment was the basis for heritability estimates. Heritability was highest for basal damage (without irradiation) 70% (95%-CI: 51%-88%) and initial damage (directly after irradiation) 65% (95%-CI: 47%-83%) and decreased to 20%-48% for the residual damage after different repair times. Hence our study supports the hypothesis that genomic instability represented by the basal DNA damage as well as radiation induced and repaired damage is highly heritable. Genes influencing genome instability and DNA repair are therefore of major interest for the etiology of lung cancer in the young. The comet assay represents a proper tool to investigate heritability of the radiation sensitive phenotype. Our results are in good agreement with other mutagen sensitivity assays.

Validation of a fully automated COMET assay: 1.75 million single cells measured over a 5 year period.

The COMET assay is recognized as a rapid and sensitive method in quantifying radiation induced DNA damage. We investigated the distorting influence of endogenous, assay-inherent factors onto base (single cell level) and primary outcome measures (experimental/slide level), such as olive tail moment (OTM) and percentage DNA in the tail (%tail-DNA). From 2003 to 2008, we performed the assay on lymphocytes isolated from the blood samples of 355 lung cancer patients, 170 controls, and 610 relatives, as well as one single reference individual, repeated 170 times. In total, the data from 10,016 single experiments containing around 1,750,000 cells have been included in this study. This is the first time that the endogenous variability of the COMET assay has been validated systematically on such a huge data set over a 5 year period. Assuming that the reference sample reflects assay specific white noise, we estimated a proportion of 7-95% of the variability of the outcome measures due to assay variation (white noise) depending on parameter, exposure level, and study group. The proportion of white noise was largest for the initial radiation damage. The specific endogenous factors considered attribute to 14.8% of the total variability in the primary outcome measurements of the OTM and 6.9% of the %tail-DNA. OTM turns out to be a sensitive parameter to detect variation, but is also more susceptible to disturbance caused by endogenous factors than %tail-DNA. To reduce the experimental variability in COMET assays, we recommend a highly standardized operation protocol as well as inspecting and/or adjusting the primary outcome measures according to endogenous factors before calculating secondary outcome measures, e.g. DNA repair capacity (DRC) or testing statistical inference. A human reference (HR) sample is also useful to inspect homogeneity in the temporal progression of long lasting investigations, but not for calibrating primary outcome measurements.