Mutagenesis Advance Access originally published online on September 27, 2005
Mutagenesis 2006 21(1):11-13; doi:10.1093/mutage/gei053
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Performance of flow cytometric analysis for the micronucleus assay—a reconstruction model using serial dilutions of malaria-infected cells with normal mouse peripheral blood
Litron Laboratories, 200 Canal View Boulevard, Rochester, NY 14623, USA, 1Toxicological Research Center, Nitto Denko Corporation, 1-1-2 Shimohozumi, Ibaraki, Osaka 567-8680, Japan, 2Division of Safety Information on Drug, Food and Chemicals and 3Division of Genetics and Mutagenesis, National Institute of Health Sciences, 1-18-1 Kamiyoga, Setagaya-ku, Tokyo 158-8501, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
To confirm the performance and statistical power of a flow cytometric method for scoring micronucleated erythrocytes, reconstruction experiments were performed. For these investigations, peripheral blood erythrocytes from untreated mice, with a micronucleated erythrocyte frequency of
0.1% were combined with known quantities of Plasmodium berghei (malaria) infected mouse erythrocytes. These cells had an infected erythrocyte frequency of 0.7%, and mimic the DNA content of micronuclei (MN). For an initial experiment, samples with a range of MN/malaria (Mal) content were constructed and analysed in triplicate by flow cytometry until 2000, 20 000 and 200 000 total erythrocytes were acquired. In a second experiment, each specimen was analysed in triplicate until 2000, 20 000, 200 000 and 1 000 000 erythrocytes were acquired. As expected, the sensitivity of the assay to detect small changes in rare erythrocyte sub-population frequencies was directly related to the number of cells analysed. For example, when 2000 cells were scored, increases in MN/Mal frequencies of 3.9- or 2.7-fold were detected as statistically significant. When 200 000 cells were analysed, a 1.2-fold increase was detected. These data have implications for the experimental design and interpretation of micronucleus assays that are based on automated scoring procedures, since previously unattainable numbers of cells can now be readily scored. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
From a statistical point of view, in order to achieve a higher power of detection, sample size should be increased. For many experimental situations, it is not always feasible to increase the number of subjects studied. When the event under consideration is rare as to cause appreciable scoring error, then an alternative would be to enhance the precision of each measurement. For example, in the rodent erythrocyte micronucleus assay, the evaluation of 2000 immature erythrocytes per animal and 5 animals per dose group represents commonly cited minimum values. Owing to the rarity of micronucleated cells, even this minimal assay design results in tedious and time-consuming efforts. The use of flow cytometry (1
–3) realizes the ability to evaluate high numbers of erythrocytes, something that is impossible to achieve by manual microscopy. By reducing scoring error in this manner, flow cytometry has the potential to increase statistical power.
In the present study, we evaluated the relationship between statistical power to detect a rare erythrocyte sub-population, i.e. micronucleated or malaria-infected erythrocytes (MN/Mal), and the total number of erythrocytes analysed. These experiments were accomplished using a reconstruction model whereby known quantities of malaria-infected erythrocytes were added to blood from an untreated mouse. Malaria is a known model for micronucleated erythrocytes, as they endow the target cells of interest with a micronucleus-like DNA content (4,5). The samples were analysed by flow cytometry to measure the MN/Mal frequency through the interrogation of 2000 (2k), 20 000 (20k), 200 000 (200k) and 1 000 000 (1m) erythrocytes. The results presented here show the capability of flow cytometric technology to reduce scoring error, and also the extent to which this affects the ability to detect small changes to baseline micronucleus frequencies.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Staining of blood specimens
Methanol-fixed blood from untreated and malaria-infected mice used in this study were two ‘biological standards’ which accompany the Mouse MicroFlow®PLUS kits (Litron Laboratories, NY). MicroFlow PLUS kits were the source of these specimens.
Before analysis, malaria-infected specimens and untreated mouse specimens were washed out of fixative with 12 ml Hank's Balanced Salt Solution. Procedures for the 3-colour labelling technique which appear in the MicroFlow®PLUS instruction manual (version 031230) were scaled up 7-fold in order to provide at least 10 ml each of control and malaria blood in a cell density range that is recommended for this assay (between 2000 and 6000 events/s). Anti-CD71-FITC, anti-CD61-PE and all other flow cytometry reagents were also supplied in the kits. After the labelling procedures were accomplished, the cell density of the malaria-infected sample was adjusted so that it was equal to that of the control blood sample. Initial cell densities were measured with a Coulter Counter, model ZM. After adjustment with additional propidium iodide staining solution, equal cell densities were confirmed by Coulter Counter measurements. Normalization of cell densities was an important experimental design consideration, as this allowed us to calculate the expected MN/Mal frequencies in the diluted samples once the frequencies of the original control (0.10 and 0.09% for Experiments 1 and 2, respectively) and malaria-infected (0.67 and 0.70% for Experiments 1 and 2, respectively) samples were determined with high precision (i.e. control and malaria-infected %MN/Mal frequencies are the mean value of triplicate analyses with 1m erythrocytes per analysis).
Dilution of malaria blood specimen
Malaria-infected blood (Sample H) was diluted with control blood (Sample A) in the following ratios (v/v): 1:1 (Sample G), 1:3 (Sample F), 1:7 (Sample E), 1:15 (Sample D), 1:31 (Sample C) and 1:63 (Sample B). These blood specimens were stored at 4°C until flow cytometric analysis, which occurred on the same day. Each sample was analysed three times to evaluate reproducibility.
Flow cytometric analysis
All samples were analysed according to the MicroFlow® PLUS 3-colour technique. One deviation to the kit-supplied data acquisition and analysis template was that the frequency of erythrocytes with malaria or micronuclei was determined without restriction to CD71-expression level. That is, the Mal and MN frequencies measured and reported here are based on total peripheral blood erythrocytes. A second deviation from standard practices is that the default stop mode of 20 000 reticulocytes was not utilized. Rather, each specimen was analysed until the following number of erythrocytes were acquired: 2k, 20k and 200k erythrocytes in the first experiment and 2k, 20k, 200k and 1m erythrocytes in the second experiment.
Statistical analysis
The average of triplicate MN/Mal measurements associated with the control blood sample were compared with those associated with each of the other seven specimens by the Fisher's exact method. A P-value of 0.05 divided by 7 (number of sample groups) was considered evidence of a statistically significant difference. Expected versus observed MN/Mal frequencies were graphed for each measurement performed in the second experiment. Microsoft Excel (Microsoft Corp., Seattle, Washington) was used to determine a best-fit line. The associated equations and r2 values were determined.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Data from Experiments 1 and 2 are summarized in Table I and include the expected and observed MN/Mal frequencies. The MN/Mal frequencies shown are the average of triplicate analyses. As shown in Table I, for measurements based on 2k erythrocytes, samples with expected MN/Mal frequencies of 0.39 and 0.24% were found to be significantly different from control samples, in Experiments 1 and 2, respectively. These values correspond to fold increases of 3.9 and 2.7 for the first and second experiment, respectively. As more erythrocytes were analysed per sample, the detection limit was improved. For instance, measurements based on the evaluation of 200k erythrocytes per analysis show statistical significance for expected MN/Mal samples of 0.12 and 0.11%. These values correspond to an increase of
1.2-fold. In fact for the second experiment, when a stop mode of 1m erythrocytes was investigated, statistical significance was observed between the control blood sample (0.09% MN/Mal) and the specimen with the lowest frequency of malaria (0.10% MN/Mal; P = 0.00005).
|
As an aid for visualizing the performance characteristics associated with the various number of cells analysed, scattergrams showing %MN/Mal measurement are presented (Fig. 1). Best-fit lines and equations are included with these graphs, and illustrate the degree to which the experimentally derived data agree with the linear relationship that is known to exist among MN/Mal frequencies for these specimens.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
To evaluate the performance and statistical power of a flow cytometric approach to score micronucleated erythrocytes, we performed a reconstruction model experiment by the serial dilution of malaria-infected mouse blood with normal mouse blood. As expected, the present results illustrate that the power of rare event detection is directly related to the number of cells analysed per specimen. By analysing 3m (triplicate of 1m) cells per group, 0.10% is significantly different (P = 0.00005) when compared with 0.09%. Even so, it must be appreciated that the biological significance of minute changes must be considered in addition to statistical significance.
Previously, we have shown that individual differences were negligible in the mouse micronucleus assay when 1000 cells per animal and 5 or 6 mice per dose group were analysed (6–8) and the statistical unit for the evaluation can be assigned to a cell but not to an animal. According to the present results and also results by Asano et al. (9), the variability of the data was high when 2k cells were analysed. Under these circumstances, the difference among animals is not apparent, as they are likely to be smaller than the scoring error. While, in the case of the present malaria dilution experiments, when 200k or 1m cells per sample were analysed, the scoring error decreased and converged to a value. This, however, is not true in the case of the actual micronucleus assays using model chemicals (9). When 200k or 1m immature erythrocytes were analysed, differences between individual animals became apparent and there was data variability within each dose group. Therefore, even if the experimental size in the animal experiments is increased, we cannot expect the same increment of detecting power. This finding suggests that optimizing the statistical procedure also includes evaluating individual differences.
Based on the present results, we confirm the accuracy and high performance of the micronucleus assay system using flow cytometry and we propose that the number of reticulocytes analysed for the micronucleus assay using flow cytometry be a minimum of 20k. We suggest that the analysis of 20k reticulocytes is approximately equivalent to the manual microscopic analysis according to test guideline OECD 474 (9,10). We anticipate that the experimental size of the MN assay will be recommended and set by expert committees based on the evaluated data. In addition to statistical sensitivity, biological variability between animals and as a consequence of treatment should also be considered. There appears to be diminishing value to analyses based on 200k or even 1m per animal. These may be useful in certain special circumstances, for instance when looking for evidence of threshold or practical threshold effects (9).
|
Notes |
|---|
* To whom correspondence should be addressed. Email: dtorous{at}litronlabs.com
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
-
1. Grawé,J., Zetterberg,G. and Amnéus,H. (1992) Flow-cytometric enumeration of micronucleated polychromatic erythrocytes in mouse peripheral blood. Cytometry, 13, 750–758.[CrossRef][Web of Science][Medline]
2. Dertinger,S.D., Torous,D.K. and Tometsko,K.R. (1996) Simple and reliable enumeration of micronucleated reticulocytes with a single-laser flow cytometer. Mutat. Res., 371, 283–292.[CrossRef][Web of Science][Medline]
3. Torous,D.K., Hall,N.E., Illi-Love,A.H. et al. (2005) Interlaboratory validation of a CD71-based flow cytometric method (Microflow) for the scoring of micronucleated reticulocytes in mouse peripheral blood. Environ. Mol. Mutagen., 45, 44–55.[CrossRef][Web of Science][Medline]
4. Tometsko,A.M., Torous,D.K. and Dertinger,S.D. (1993) Analysis of micronucleated cells by flow cytometry. 1. Achieving high resolution with a malaria model. Mutat. Res., 292, 129–135.[CrossRef][Web of Science][Medline]
5. Dertinger,S.D., Torous,D.K., Hall,N.E., Tometsko,C.R. and Gasiewicz,T.A. (2000) Malaria-infected erythrocytes serve as biological standards to ensure reliable and consistent scoring of micronucleated erythrocytes by flow cytometry. Mutat. Res., 464, 195–200.[Web of Science][Medline]
6. Hayashi,M., Yoshimura,I. Sofuni,T. and Ishidate,M. (1989) A procedure for data analysis of the rodent micronucleus test involving a historical control. Environ. Mol. Mutagen., 13, 347–356.[Medline]
7. Hayashi,M., Hashimoto,S., Sakamoto,Y., Hamada,C., Sofuni,T. and Yoshimura,I. (1994) Statistical analysis of data in mutagenicity assays: rodent micronucleus assay. Environ. Health Perspect., 102 (Suppl. 1), 49–52.
8. Adler,I.-D., Bootman,J., Favor,J., Hook,G., Schriever-Schwemmer,G., Welzl,G., Whorton,E., Yoshimura,I. and Hayashi,M. (1998) Recommendations for statistical designs of in vivo mutagenicity test with regard to subsequent statistical analysis. Mutat. Res., 417, 19–30.[Medline]
9. Asano,N., Torous,D.K., Tometsko,C.R., Dertinger,S.D., Morita,T. and Hayashi,M. (2005) Practical threshold for micronucleated reticulocyte induction observed for low doses of mitomycin C, Ara-C, and colchicine. Mutagenesis, 21, 15–20.
10. OECD (1997) Guideline for the Testing of Chemicals: Mammalian Erythrocyte Micronucleus Test. Guideline 474. Organisation for Economic Cooperation and Development, Paris, France.
Received on May 23, 2005; revised on July 21, 2005; accepted on July 22, 2005.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Zeiger, L. Recio, T. R. Fennell, J. K. Haseman, R. W. Snyder, and M. Friedman Investigation of the Low-Dose Response in the In Vivo Induction of Micronuclei and Adducts by Acrylamide Toxicol. Sci., January 1, 2009; 107(1): 247 - 257. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. D. Dertinger, M. E. Bishop, J. P. McNamee, M. Hayashi, T. Suzuki, N. Asano, M. Nakajima, J. Saito, M. Moore, D. K. Torous, et al. Flow Cytometric Analysis of Micronuclei in Peripheral Blood Reticulocytes: I. Intra- and Interlaboratory Comparison with Microscopic Scoring Toxicol. Sci., November 1, 2006; 94(1): 83 - 91. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||