Mutagenesis vol. 19 no. 5 © UK Environmental Mutagen Society 2004; all rights reserved.
An automated scoring procedure for the micronucleus test by image analysis
Department of Human Genetics, University of Ulm, 89069 Ulm, Germany, 1Metasystems, Altlussheim, Germany and 2Department of Gynaecology, Technische Universität, München, Germany
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
The micronucleus assay (MNT) in human lymphocytes is frequently used to assess chromosomal damage as a consequence of environmental mutagen exposure, to assess the effect of mutagens or to search for reduced DNA repair capacity after a mutagenic challenge. We have established an automated scoring procedure for the cytokinesis blocked MNT based on computerized image analysis (Metasystems Metafer 4 version 2.12). To evaluate the results we used the reproducibility of counts, established a dose–response curve for
-irradiation and used the ability of the system to differentiate between breast cancer patients and controls as a biological reference, a difference which we had observed before by visual counting. Blood cultures were irradiated with -rays (2 Gy) at the beginning and treated with cytochalasin B during the last 24 h. The slides were stained with Giemsa for visual counting and with DAPI for automated analysis. Our test sample consisted of 73 persons (27 with breast cancer and 26 female and 20 male controls). A comparison between visual counting (controls, mean MN frequency 313) and automated counting (mean MN frequency 106) in slides from the same culture revealed a large drop for the automated counts. However, the automated counts were as reproducible as the visual counts [coefficient of variation (CV) on the sample 
20%; CV on repeated counts of the same slides 5%] and both counts were highly correlated. Furthermore, the discrimination between cases and controls improved for automated counting of slides from the same cultures [visual odds rato (OR) 
4.0, P = 0.009; automated OR > 16, P < 0.0001], with a strong dependence on the set of parameters used. This improvement was confirmed in a validation sample of an additional 21 controls and 20 cases (OR = 11, P = 0.0018) performed as a prospective or diagnostic test. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
Micronuclei (MN) are derived from chromosome fragments arising from asymmetrical structural aberrations or represent whole chromosomes that are not incorporated into the nucleus at cell division. Acentric fragments are most often seen after irradiation of cells, whereas entire chromosomes are more frequent in spontaneously occurring MN or after induction by spindle poisons without any clastogenic treatment, as was demonstrated by anti-kinetochore antibody staining (Fenech and Morley, 1989
; Tucker and Eastmond, 1990). The most frequently used method for the micronucleus test (MNT) is to score MN in cells that have passed one mitosis but are prevented from undergoing cytokinesis (see, for example, Fenech and Morley, 1986; Kirsch-Volders et al., 1997; Fenech, 2000). This is achieved by the use of cytochalasin B. It is important to count MN in binucleated cells (BNC) for several reasons: (i) cells have to pass through one cell cycle and mitosis after irradiation in order to form MN; (ii) MN may be lost in the second to the third cycle; (iii) new MN may arise in the second to the third cycle. The MNT in BNC is a well-established assay, especially for mutagenicity testing (see, for example, Kalantzi et al., 2003; Palus et al., 2003) and for human population monitoring (Bonassi et al., 2003; Neri et al., 2003). It has also been used to investigate chromosomal instability in humans who have mutations in genes which are needed for the repair of DNA damage, as in the case of Fanconi anemia (Zunino et al., 2001) and ataxia teleangiectasia (AT) (Gutierrez-Enriquez and Hall, 2003). The observed chromosomal instability as a consequence of a limited capacity to repair DNA damage prompted a search for increased radiosensitivity in carriers of BRCA1 or BRCA2 mutations using the MNT, which was indeed observed (Rothfuss et al., 2000; Trenz et al., 2002). This phenotype is constitutive in carriers of chromosomal instability syndromes like AT or Fanconi anaemia and may also be present in other conditions predisposing to cancer. It may be possible to detect these persons by the MNT, a fact that has been shown for a proportion of sporadic breast cancer patients (Scott et al., 1998; Baeyens et al., 2002). There are a lot of factors which influence the frequency of MN. The most important and interesting one is the biological differences between individuals. Other points concern the culture time, the cell harvesting methodology and the fixation and preparation of the slides (Bonassi et al., 2001). Another source of variation is the scoring procedure for MN, which is usually done visually. The criteria for scoring have been standardized in order to minimize non-biological variation and to allow comparison between laboratories. Visual scoring of MN is very time consuming and the results depend on subjective interpretation of nuclei and MN. The question of which binucleated cells and micronuclei to accept results in differing MN frequencies between observers and laboratories (Fenech et al., 2003a). Two different types of automated MNT are presently used, flow cytometry (Schreiber et al., 1992; Nusse and Marx, 1997; Styles et al., 2001) and MN counting by image analysis (Verhaegen et al., 1994; Bocker et al., 1995, 1996). Both of these techniques have their advantages and disadvantages: In the case of flow cytometry results cannot be rechecked after measurement and no data for individual cells can be obtained (such as, for example, the number of MN). The basic problems of MN counting by image analysis are: identification of the cytoplasm and, hence, the assignment of nuclei and micronuclei to a cell, as well as identification of background signals. The advantage of both systems is the fast acquisition of results, which allows the analysis of large numbers of slides and the exclusion of subjective judgement and individual scoring skills. In the case of image analysis another advantage is the possibility of repeated scoring of the same slide.
We here report on a new and fast image analysis system for scoring of MN in BNC using fluorescent dyes specific for DNA. We determined the coefficient of variation (CV), which proved to be only slightly smaller than that obtained with visual counting, established a dose–response curve for -radiation and evaluated the ability of the system to discriminate between breast cancer patients and controls.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
Sample
The present sample comprised 114 persons from southern Germany and was derived from a larger study aimed at discrimination between breast cancer patients and controls, the results of which are to be published in detail elsewhere (D. Varga, S. Jainta, M. Hoehne, B. Patino-Garcia, I. Michel, R. Kreienberg, U. Schwarz-Boeger, M. Kiechle, T. Paiss, C. Maier and W. Vogel, in preparation). In a test sample 27 patients with breast cancer (25 before treatment), 26 female controls and 20 male controls were investigated, first by visual counting and afterwards by automated analysis in order to directly compare both procedures. Further details of the participants are given in Table I. In order to verify the ability of the automated system to detect the biological difference between breast cancer patients and controls, namely a higher radiosensitivity, an additional sample of 20 cases and 21 controls was analysed (validation sample, Table II).
|
|
Micronucleus test
Heparinized blood samples were stored at room temperature (18–25°C) until the start of culture. The time before start of culture ranged from 2 to a maximum of 48 h. An aliquot of 0.3 ml of blood was diluted 1:9 with chromosome medium 1 A (Gibco BRL) supplemented with 2% PHA-L (Gibco BRL). Cultures were exposed to a total dose of 2 Gy
-rays from a 137Cs source (Gammacell 2000; Molsgaard Medical, Heorsholm, Denmark) immediately after being set up. To establish a dose–response curve, 0, 1, 2 and 4 Gy were used. Thereafter, the cultures were incubated at 37°C for 44 h (after PHA stimulation) and then cytochalasin B (Sigma) was added to a final concentration of 6 µg/ml. After an additional 24 h growth, cultures were harvested by centrifugation, treated with 0.56% KCl solution and fixed with methanol/acetic acid (5:1) mixed with an equal amount of 0.9% NaCl. Afterwards, fixation with undiluted methanol/acetic acid (5:1) was performed three times. The slides were coded and the entire analysis was carried out blind. Air dried slides were stained with 7% Giemsa in phosphate buffer (pH 7.0) for visual counting. MN were counted under a light microscope in 500–1000 BNC. We performed multiple measurements in most probands by visual counting. Slides from the same blood cultures were later prepared for automated analysis. When establishing the automated analysis, a series of non-fluorescent dyes was tested (Table IV) because Giemsa was found not to be suitable. These other dyes mostly produced insufficient contrast or stained debris. For automated analysis we finally chose fluorescence staining (DAPI) because of the background in Giemsa stained slides, which led to false positive counts, debris being mistaken for MN. DAPI proved to be the most suitable of the fluorochromes among a series of dyes tested (details in Table V). All of these dyes were tested with appropriate filter settings. DAPI was chosen for two reasons: it is the dye with the highest known specificity for DNA and it has a high fluorescence yield. It is therefore expected to be best suited to detect small signals such as MN. DAPI actually yielded the best fluorescence signal for automated image analysis with our system and provided the clearest pictures without cytoplasmic background. After establishing the staining procedure, the slides were stained routinely with 1 µg/ml DAPI in 4x SSC for 15 min, rinsed in distilled water, air dried and embedded with Vectashield (Vector Laboratories, Burlingame, CA) to prevent loss of fluorescence and to provide a strong initial fluorescence.
|
|
Slide scanning principle
Automated slide scanning is generally performed by moving the slide with reference to the fixed objective lens of the microscope in a regular meander-like pattern, leaving no gaps between the image fields. Because of speed considerations, image acquisition is done at the lowest possible optical magnification that still allows resolution of the features of interest (to detect BNC and count MN a 10x objective, giving a final magnification of 100x, is used). Each field of view is captured and analysed for the presence of analysable objects (e.g. BNC). If cells of interest are detected within a field, they are further analysed and stored in an image gallery along with their position and feature data (e.g. no. of micronuclei). After the scan the image gallery can be used to review the detected cells, to reject unsuitable cells and to correct feature values if necessary. Since the coordinates of all cells detected were stored during the scan, any cell can automatically be relocated under the microscope for direct visual inspection.
Stage movement
To scan a slide or a part thereof all three axes in space have to be controlled by the software. A motorized stage covers the x- and y-axes, while movement of the specimen in the z-direction is done using the microscope's internal software controlling all motorized components, which can be driven either by the microscope controls or by external devices connected to the microscope.
Focusing
Precision and speed of autofocus are of the utmost importance for the quality of object detection and the slide scanning results. The plane of best focus is determined at a number of grid positions regularly distributed across the scan area. This is done by automatically moving the stage in the z-direction, capturing images in different focal planes and analysing the focus quality based on a local contrast criterion. Typically, 11 z-positions are analysed within 2 s. During the subsequent scan the slide is automatically kept within the plane of best focus by bilinear interpolation.
Image acquisition
As the fluorescence signal intensity varies significantly between different positions on the same slide, automatic exposure control is a must to ensure correct image quality and a wide dynamic range. The software estimates the correct exposure time based on the histogram (the intensity distribution) of an image captured without integration. Depending on the histogram shape, a reduced exposure time (using the built-in electronic shutter) or a longer time integration is used. This strategy allows for exposure times of from 1/10 000 to 30 s.
Central unit
The central unit of the Metafer system is a Dell microcomputer (Langen, Germany) equipped with an Intel® Pentium® III processor (1 GHz), 256 MB RAM memory and a 40 GB hard drive and running the Microsoft® Windows® 2000 operating system. For data archiving purposes the system is equipped with a magneto-optical disk drive (2.3 GB capacity).
Image acquisition hardware
For image acquisition a high resolution monochrome megapixel charge coupled device (CCD) camera (M1; JAI AS, Glostrup, Denmark) with a resolution of 1280 x 1024 pixels (2/3 inch CCD, pixel size 6.7 x 6.7 µm, signal-to-noise ratio 56 dB) is used. It is connected to a grey level digitizer board installed in the central unit, which provides real-time digitization of video signals at a resolution of 1280 x 1024 pixels with 256 grey levels. The camera is connected to the microscope via a standard 1.0x TV adapter (C-mount 60 C, 2/3 inch; Carl Zeiss, Göttingen, Germany).
Microscope
The computer is connected to a motorized Axioplan 2 Imaging E MOT microscope (Carl Zeiss) and uses the microscope components for automated focusing, light source adjustment (for bright field imaging) and fluorescence filter changes.
Scanning stage
Eight slides (3 x 1 inches) per scanning run can be loaded onto the motorized scanning stage (Märzhäuser, Wetzlar, Germany) with a total scanning area of 225 x 76 mm. The stage is connected to a 2-axis stepping motor controller board inside the central unit (PCSMOC3), providing a step size of 1 µm and a maximum step frequency of 72 000 Hz. For visual movement of the stage a trackball is connected to the system (Microspeed, El Cajon, USA).
Detection of binucleated cells and counting of micronuclei
A human observer identifies a BNC by the presence of two nuclei in the cytoplasm of a single cell and accepts it using standardized criteria (Fenech et al., 2003b). In the present automated system no information on the cytoplasm is acquired and, therefore, a BNC was defined by the occurrence of two similar (shape, size, etc.) nuclei, close to each other but completely separated. For automatic analysis, the slides were DAPI stained to minimize difficulties in MN counting due to cytoplasmic staining. To count MN a circular area with a diameter corresponding to the average size of a cell around the epicenter between the nuclei was analysed. Scanning for MN was performed using a 10x microscope objective (10x Fluar; Carl Zeiss) and a DAPI filter set (Zeiss/Chroma). After automatic image acquisition the nuclei are enhanced by image processing using a sharpening filter. Following this, the system automatically sets the object threshold used to separate objects from the background. The threshold algorithm takes the presence of discontinuities in the image background into account. Objects above this threshold are analysed for certain morphometric criteria:
- size (see Appendix and Figure 1) (within a size range specified in the classifier set-up);
- aspect ratio (the ratio of the longest to the shortest diameter of the object; see Figure 2);
- relative concavity depth, to exclude clusters (see Figure 2).
18 µm) were accepted for further analysis. As the nuclei of a BNC are considered to have approximately the same size, a maximum relative size difference (symmetry criterion) of the two nuclei was set by the system to reject nuclei of different cells that are close together. After an initial search for criteria we used two classifiers (A and B) differing only in the symmetry criterion (A, 75%; B, 90%). Finally, the object area of all other objects within a specified region of interest (the central point of the line connecting the centres of the two nuclei forms the centre of the region of interest circle with a specified radius of 40 µm) (Figure 1) is measured and the cell is rejected if one of these objects exceeds the maximum area specified.
|
|
In a second step the system analyses the region of interest for the presence of MN. After applying image processing operations to reduce the background and enhance the usually weaker MN, a threshold is set for MN in the same way as for the nuclei before. Objects above this threshold and within a defined size range (minimal size 3 pixels, maximal 46 pixels) are tested for their morphological features, which are principally the same as described above for the detection of nuclei. A MN is counted if it meets these criteria and additionally is located within a certain distance (Figure 1) from the centre of the region of interest (29 µm). For each BNC cell a gallery image (Figure 3) is generated which indicates the number of MN. The galleries of the counted slides are saved and only slides with >500 BNC are taken as data. All automatically counted slides were counted using both classifiers (see Appendix for exact specifications) whose parameters were defined by analysing thousands of appropriate cells and MN on the PC, gauging their dimensions and afterwards specifying the range for the morphological values required. The two classifiers use identical specifications except for the symmetry criterion for the two nuclei of a BNC. The concordant parameters of the two classifiers offered the most satisfying results in the visual re-examination of BNC. This applied to all parameters except the symmetry criterion. We soon realized that both symmetry criteria offered good results and used both of them to determine the influence of this criterion on counting of the slides.
|
The system for automated MN detection, Metafer MicroNuclei, automatically identifies BNC that are stained with a single nuclear/cytoplasmic stain (transmitted light or fluorescence). This detection of BNC is one of the difficulties for image analysis and herein lies one of the big differences to visual counting. A human observer identifies a cell and afterwards takes a close look to decide whether the cell is appropriate or not. Our system searches for nuclei and, if the parameter specifications (see Appendix) are correct, defines it as a BNC.
Statistical methods
All statistical analyses were carried out using StatView version 5.0.1 (SAS Institute Inc., Cary, NC). The 75th percentile of controls was used to recode the counts into a nominal variable for logistic regression. The results are depicted as box plots; the numbers are presented in Tables I–III. In the case of the visually counted slides we used the mean values from the different observers and compared these values with the automatically counted ones. In order to determine the variation inherent in counting by image analysis, one slide was counted four times and the CV calculated. This gives a CV per individual and was repeated for 10 different persons. An overall CV was then calculated using all 40 measurements. Furthermore, polynomial (2nd order) regression was used to determine a dose–response curve. We did not correct for effects such as age, storage time of the samples (2–48 h) nor, in cancer patients, for treated or untreated cases, because it does not affect a direct comparison between visual and automated counting (identical blood cultures) and because discrimination between cases and controls was not the primary goal of this study.
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
In the test sample MN frequencies were first determined visually. Then new slides were prepared from the same blood cultures and counted twice by image analysis, employing classifiers A and B, respectively. The results from these counts along with the CV of the different groups and the age of the probands are reported in Table I. The corresponding data for the validation sample are presented in Table II. There is a large difference between visual counting and automated scoring. Mean values for lymphocytes from controls and cancer patients ranged from 289.5 to 395.9 MN in visually counted slides (Table I), whereas automated counting resulted in 94.2–149.9 MNs per 1000 BNC, about half to one-third of the visually obtained values. The values from both scoring procedures are highly correlated despite the difference in absolute MN frequencies (Figure 4), as determined by regression analysis (y = 11 + 0.3 x x with r2 = 0.83). This high correlation is seen for all measurements, but is valid in particular also for the baseline values without induction (r2 = 0.71) (Figure 5). Both procedures yielded highly significant differences between cases and controls (Table III). The lower MN frequencies obtained by image analysis raised the question as to how specific these counts are and, therefore, we established a dose–response curve for
-irradiation from 0 to 4 Gy in 10 controls. These data show that MN frequencies obtained with our system of image analysis are strictly dose dependent with very little variation and that they can be used to detect exposure to -irradiation. The polynomial fit (Figure 6) gave: y = 22 + 58 x x + 4.8 x x2 with r2 = 0.86 and P < 0.0001. The CV for MN frequencies is considered another criterion for quality of counting and, therefore, has been calculated on our data (Tables I and II). The CV values on the different subgroups were in the range 17–32% which includes biological variability between the probands (Tables I and II). In order to assess the reproducibility of the automated system we prepared and stained one slide each from 10 subjects. The slides were counted and after 6 weeks three repeat counts of the same slide were made, i.e. four counts were performed for each person. From these four counts we calculated the CV of the MN frequencies for classifiers A (CV = 9.5%) and B (CV = 5.4%). These data demonstrate the reproducibility of the data obtained with the system. In order to confirm the ability of the automated scoring to detect a biological difference a second sample of patients and controls was analysed (validation sample) (Tables II and III). Both automated analyses were carried out running the two different classifiers (A and B) in parallel (Figures 7 and 8). The discriminatory ability of classifier B was ‘better’ with the test sample. The odds ratios for the test sample ranged between 4 (visual), 5 (A) and 28.03 (B) (Table III). Interestingly, in the validation sample it was lower, with an odds ratio of 11.5 for classifier A and 6.5 for classifier B (Table III, Figure 8).
|
|
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
The aim of our study was to evaluate the results obtained by automated image analysis and to compare them with those from visual counting. The major difference between image analysis and visual counting was a dramatic drop in the MN frequencies in automated analysis. Simultaneously we switched from Giemsa staining to DAPI. This drop in MN frequencies is difficult to explain, since visual re-examination of DAPI stained slides indicated similar sensitivities for image analysis and human observers. Obviously, automated scoring detects only a subset of the MN accepted by human observers. This may in part be explained by stricter scoring criteria imposed by the restrictions on morphometric parameters in image analysis, such as contact between MN and the nucleus (Figure 3). Since the slides were prepared from the same blood cultures, the only differences between the two analyses concern the scoring criteria (the parameter set of the system) and staining with DAPI, which is very specific for DNA. One might think that this high specificity of DAPI for DNA avoids counting artefacts, as may happen on Giemsa stained slides. However, it is hard to believe that nearly half of the MN detected by visual counting represent artefacts. In any case, the ability to discriminate between patients and controls by visual counting has been improved upon by using automated scoring (Table III and Figure 4).
Anyway, we decided to evaluate the system using (i) the stability of counting as determined by the CV, (ii) the ability to detect the increasing effects of an increasing dose of -irradiation and (iii) a biological difference, namely the higher radiosensitivity of breast cancer patients compared with controls, which was established in another study (Varga et al., in preparation). All three approaches to evaluate MN counting by image analysis showed equal or superior results compared with human observers counting Giemsa stained slides. Regardless of the question what the ‘true’ MN frequencies might be, we have demonstrated that the system provides highly reproducible results and that these are biologically relevant. The analysis of our data shows a difference in the MN rates between breast cancer patients and healthy persons. This discrimination has been reported by others (Scott et al., 1998; Rothfuss et al., 2000; Baeyens et al., 2002) confirmed by us in a larger unpublished study employing visual counting extended by automated analysis (D. Varga, S. Jainta, M. Hoehne, B. Patino-Garcia, I. Michel, R. Kreienberg, U. Schwarz-Boeger, M. Kiechle, T. Paiss, C. Maier and W. Vogel, in preparation). Using automated counting we were able to improve discrimination in terms of the observed odds ratios.
Concerning reproducibility, our system proved to be at least comparable with experienced human observers who had worked for several years in cytogenetics. It has the advantage, however, that the scoring criteria can be modified in a completely controlled way in small steps, which cannot be kept constant by a human observer. An example of the impact of this fine scaling emerged from the use of two classifiers (A and B) with a differing symmetry criterion (75 versus 90%) for the nuclei in the test and validation samples. Classifier B gave a much higher odds ratio on the test sample while classifier A performed better on the validation sample. We cannot explain this effect at the moment, but it may be related to the time the cells were kept in fixative (months versus days) and, in consequence, a difference in spreading of the cells on the slides may result. It is difficult for human observers to assess the difference between 90% and 75% symmetry of nuclei and its effect on discrimination may explain differing results between human observers, despite highly standardized scoring criteria (Fenech et al., 2003a). The HUMN project states ‘The results of these studies indicate clearly that even after standardizing culture and scoring conditions it will be necessary to calibrate scorers and laboratories if MN, MNed cell and nucleoplasmic bridge frequencies are to be reliably compared among laboratories and among populations’ (Fenech et al., 2003a). Laboratories equipped with automated scoring could cross-check their results more easily and identical classifiers could be used for joined analyses. The present status of computer hardware, recognition algorithms and the choice of appropriate staining allow fast (7 min/slide) and reliable analysis and offer a good chance to automate and standardize counting of MN in BNC.
|
Appendix. The parameter set (classifier) used to identify binucleated cells and micronuclei in the automated scoring process |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
Image processing used standard procedures like average and sharpen before determining object size and level above threshold (as implemented in the Metafer/Metacyte software packages). Two classifiers (A and B) were used in parallel with identical parameters except for the ‘nuclei maximum area asymmetry’ (75 and 90%).
Nuclei:
- sharpen(3,4)
- nuclei object threshold: 20
- nuclei minimum area: 4000 in 1/100 µm2
- nuclei maximum area : 20 000 in 1/100 µm2
- nuclei maximum relative concavity depth in 1/1000: 120
- nuclei maximum side ratio in 1/1000: 1500
- nuclei maximum distance in 1/10 µm: 180
- nuclei maximum area asymmetry: A, 75%; B, 90%
- nuclei region of interest radius in 1/10 µm: 400
- nuclei maximum object area in roi in 1/100 µm2: 2000
- nuclei object threshold: 20
Micronuclei:
- medianv(3), medianh(3), average(3,1), sharpen(5,5)
- micronuclei object threshold: 5%
- micronuclei minimum area in 1/100 µm2: 100
- micronuclei maximum area in 1/100 µm2: 2100
- micronuclei maximum relative concavity depth in 1/1000: 1000
- micronuclei maximum side ratio in 1/1000: 1720
- micronuclei maximum distance in 1/10 µm: 290
- micronuclei object threshold: 5%
|
Acknowledgments |
|---|
We are most grateful to all patients and controls for their donations of blood samples and the self-support groups for active collaboration. We would like to thank Guenter Speit for valuable discussions and for critical reading of the manuscript. The excellent technical assistance of Karina Eiwen is highly appreciated. This work was supported in part by the Deutsche Krebshilfe, grant no. 70-2680-VO2. B.P.G. was a fellow of the DAAD and D.V. of GRK 460. T.J. is an employee of Metasystems, makers of the image analysis system used in this study.
|
Notes |
|---|
3 To whom correspondence should be addressed. Tel: +49 731 50023430; Fax: +49 731 50023438; Email: walther.vogel{at}medizin.uni-ulm.de
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendix. The parameter set...References |
|---|
-
Baeyens,A., Thierens,H., Claes,K., Poppe,B., Messiaen,L., De Ridder,L. and Vral,A. (2002) Chromosomal radiosensitivity in breast cancer patients with a known or putative genetic predisposition. Br. J. Cancer, 87, 1379–1385.[CrossRef][ISI][Medline]
Bocker,W., Muller,W.U. and Streffer,C. (1995) Image processing algorithms for the automated micronucleus assay in binucleated human lymphocytes. Cytometry, 19, 283–294.[CrossRef][ISI][Medline]
Bocker,W., Streffer,C., Muller,W.U. and Yu,C. (1996) Automated scoring of micronuclei in binucleated human lymphocytes. Int. J. Radiat. Biol., 70, 529–537.[CrossRef][ISI][Medline]
Bonassi,S., Fenech,M., Lando,C. et al. (2001) HUman MicroNucleus project: international database comparison for results with the cytokinesis-block micronucleus assay in human lymphocytes: I. Effect of laboratory protocol, scoring criteria and host factors on the frequency of micronuclei. Environ. Mol. Mutagen., 37, 31–45.[CrossRef][ISI][Medline]
Bonassi,S., Neri,M., Lando,C. et al. (2003) Effect of smoking habit on the frequency of micronuclei in human lymphocytes: results from the Human MicroNucleus project. Mutat. Res., 543, 155–166.[CrossRef][ISI][Medline]
Fenech,M. (2000) The in vitro micronucleus technique. Mutat. Res., 455, 81–95.[ISI][Medline]
Fenech,M. and Morley,A.A. (1986) Cytokinesis-block micronucleus method in human lymphocytes: effect of in vivo ageing and low dose X-irradiation. Mutat. Res., 161, 193–198.[ISI][Medline]
Fenech,M. and Morley,A.A. (1989) Kinetochore detection in micronuclei: an alternative method for measuring chromosome loss. Mutagenesis, 4, 98–104.
Fenech,M., Bonassi,S., Turner,J. et al. (2003a) Intra- and inter-laboratory variation in the scoring of micronuclei and nucleoplasmic bridges in binucleated human lymphocytes. Results of an international slide-scoring exercise by the HUMN project. Mutat. Res., 534, 45–64.[ISI][Medline]
Fenech,M., Chang,W.P., Kirsch-Volders,M., Holland,N., Bonassi,S. and Zeiger,E. (2003b) HUMN project: detailed description of the scoring criteria for the cytokinesis-block micronucleus assay using isolated human lymphocyte cultures. Mutat. Res., 534, 65–75.[ISI][Medline]
Gutierrez-Enriquez,S. and Hall,J. (2003) Use of the cytokinesis-block micronucleus assay to measure radiation-induced chromosome damage in lymphoblastoid cell lines. Mutat. Res., 535, 1–13.[ISI][Medline]
Kalantzi,O.I., Hewitt,R., Ford,K.J. et al. (2003) Low-dose induction of micronuclei by lindane. Carcinogenesis, 25, 613–622.
Kirsch-Volders,M., Elhajouji,A., Cundari,E. and Van Hummelen,P. (1997) The in vitro micronucleus test: a multi-endpoint assay to detect simultaneously mitotic delay, apoptosis, chromosome breakage, chromosome loss and non-disjunction. Mutat. Res., 392, 19–30.[ISI][Medline]
Neri,M., Fucic,A., Knudsen,L.E., Lando,C., Merlo,F. and Bonassi,S. (2003) Micronuclei frequency in children exposed to environmental mutagens: a review. Mutat. Res., 544, 243–254.[CrossRef][ISI][Medline]
Nusse,M. and Marx,K. (1997) Flow cytometric analysis of micronuclei in cell cultures and human lymphocytes: advantages and disadvantages. Mutat. Res., 392, 109–115.[ISI][Medline]
Palus,J., Rydzynski,K., Dziubaltowska,E., Wyszynska,K., Natarajan,A.T. and Nilsson,R. (2003) Genotoxic effects of occupational exposure to lead and cadmium. Mutat. Res., 540, 19–28.[ISI][Medline]
Rothfuss,A., Schutz,P., Bochum,S., Volm,T., Eberhardt,E., Kreienberg,R., Vogel,W. and Speit,G. (2000) Induced micronucleus frequencies in peripheral lymphocytes as a screening test for carriers of a BRCA1 mutation in breast cancer families. Cancer Res., 60, 390–394.
Schreiber,G.A., Beisker,W., Braselmann,H., Bauchinger,M., Bogl,K.W. and Nusse,M. (1992) An automated flow cytometric micronucleus assay for human lymphocytes. Int. J. Radiat. Biol., 62, 695–709.[ISI][Medline]
Scott,D., Barber,J.B., Levine,E.L., Burrill,W. and Roberts,S.A. (1998) Radiation-induced micronucleus induction in lymphocytes identifies a high frequency of radiosensitive cases among breast cancer patients: a test for predisposition? Br. J. Cancer, 77, 614–620.[ISI][Medline]
Styles,J.A., Clark,H., Festing,M.F. and Rew,D.A. (2001) Automation of mouse micronucleus genotoxicity assay by laser scanning cytometry. Cytometry, 44, 153–155.[CrossRef][ISI][Medline]
Trenz,K., Rothfuss,A., Schutz,P. and Speit,G. (2002) Mutagen sensitivity of peripheral blood from women carrying a BRCA1 or BRCA2 mutation. Mutat. Res., 500, 89–96.[ISI][Medline]
Tucker,J.D. and Eastmond,D.A. (1990) Use of an antikinetochore antibody to discriminate between micronuclei induced by aneuploidogens and clastogens. Prog. Clin. Biol. Res., 340B, 275–284.[Medline]
Verhaegen,F., Vral,A., Seuntjens,J., Schipper,N.W., De Ridder,L. and Thierens,H. (1994) Scoring of radiation-induced micronuclei in cytokinesis-blocked human lymphocytes by automated image analysis. Cytometry, 17, 119–127.[CrossRef][ISI][Medline]
Zunino,A., Degan,P., Vigo,T. and Abbondandolo,A. (2001) Hydrogen peroxide: effects on DNA, chromosomes, cell cycle and apoptosis induction in Fanconi's anemia cell lines. Mutagenesis, 16, 283–288.
Received on April 29, 2004; accepted on July 12, 2004.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Varga, J. Hoegel, C. Maier, S. Jainta, M. Hoehne, B. Patino-Garcia, I. Michel, U. Schwarz-Boeger, M. Kiechle, R. Kreienberg, et al. On the difference of micronucleus frequencies in peripheral blood lymphocytes between breast cancer patients and controls Mutagenesis, September 1, 2006; 21(5): 313 - 320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Patino-Garcia, J. Hoegel, D. Varga, M. Hoehne, I. Michel, S. Jainta, R. Kreienberg, C. Maier, and W. Vogel Scoring variability of micronuclei in binucleated human lymphocytes in a case-control study Mutagenesis, May 1, 2006; 21(3): 191 - 197. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||