Simple and reliable enumeration of micronucleated reticulocytes with a single-laser flow cytometer

Genetic Toxicology

ELSEVIER

Mutation Research 371 (1996) 283-292

Simple and reliable enumeration of micronucleated reticulocytes with a single-laser flow cytometer Stephen D. Dertinger *, Dorothea K. Torous, Kenneth R. Tometsko Litron Laboratories, 1351 Mount Hope Avenue, Rochester. NY 14620, USA

Received 28 May 1996; revised 3 September 1996; accepted 3 September 1996

Abstract A flow cytometric procedure for scoring micronuclei in mouse peripheral blood erythrocytes, especially reticulocytes, is described. The methods reported herein were developed in an effort to simplify the techniques and to reduce the equipment requirements associated with automated micronucleus analyses. With this procedure, fluorescein-conjugated monoclonal antibodies which bind to the CD71-defined antigen (the transferrin receptor) are used to label reticulocytes. The nucleic acid dye propidium iodide is used to identify cells with micronuclei. Given 488 nm excitation, four populations of erythrocytes are clearly resolved: normochromatic erythrocytes with and without micronuclei, and reticulocytes with and without micronuclei. Since the method is capable of simultaneously providing the incidence of micronuclei in both mature and immature erythrocyte populations, it is compatible with either chronic or acute treatment regimens. To demonstrate cell handling and flow cytometric procedures for quantitatively analyzing peripheral blood micronuclei, an experiment with the model clastogen methyl methanesulfonate is described. Additionally, a reconstruction experiment was performed whereby three mouse blood samples were spiked with successively greater volumes of blood from a clastogen-treated animal so each preparation differed slightly, but definitely, in micronucleus content. Each sample was scored six times by conventional microscopy and by flow cytometry so that the two methods could be directly compared. Collectively, the results from the methyl methanesulfonate experiment and the reconstruction study demonstrate the accuracy and reliability of the flow cytometric method. Furthermore, advantages associated with objective, high throughput scoring methodology are clearly indicated. Keywords: Micronucleus; Flow cytometry; Clastogenicity; Genotoxicity testing; Automated analysis; Anti-transferrinreceptor; CD71

antigen

1. Introduction The mouse micronucleus test has gained widespread use as a short-term system to screen chemicals for genotoxic activity (Matter and Schmid,

* Corresponding author. Tel.: (716)442-0930; Fax: (716)4420930.

1971; Heddle, 1973; Schmid, 1975). The test is based on the observation that mitotic cells with chromatid breaks or dysfunctional mitotic apparatus exhibit disturbances in the anaphase distribution of their chromatin. After telophase, this displaced chromatin can be excluded from the nuclei of the daughter cells and is found in the cytoplasm as a micronucleus (MN). Erythrocytes are particularly well suited

0165-1218/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. PII S0165- 121 8(96)001 30-9

284

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

for evaluating MN events since the nucleus of the erythroblast is expelled a few hours after the last mitosis. Consequently, micronuclei are particularly apparent in this population which is otherwise devoid of DNA. An important advance came with the observation that micronucleated erytbrocytes are not cleared from the blood of mice, thus allowing the analysis to be carried out more readily with peripheral blood samples (MacGregor et al., 1980, 1983). Even so, the low throughput capacity of microscopic scoring procedures makes the assay labor intensive and time consuming. It has been realized for some time that the next evolution of the assay should involve automated methods for objectively and accurately scoring larger numbers of micronucleated cells and thereby improve assay sensitivity and reliability (Heddle et al., 1991). This laboratory and others have reported flow cytometric (FCM) methods for scoring micronucleated erythrocytes in the total peripheral blood erythrocyte pool (Hayashi et al., 1992; Tometsko et al., 1993a,b,c; Tometsko et al., 1995). In these systems, DNA-specific dyes such as Hoechst or 4',6-diamidino-2-phenylindole (DAPI) are used to distinguish erythrocytes with or without micronuclei. Since the MN-inducing effect of clastogen exposure is diluted by erythrocytes pre-existing in circulation, these protocols require extended dosing regimens so that the effect of treatment sufficiently impacts the total blood pool. Given subchronic or chronic dosing protocols, these systems represent very sensitive and efficient means for measuring clastogenic activity (Dertinger et al., 1996). A considerable challenge has been to develop reliable automated methods for quantitating MN events in immature erythrocyte populations. The advantage of restricting the analysis to these newly formed cells is that this population can highlight genotoxic action resulting from acute exposures. A significant advance in this area came out of the laboratory of Grawe et al. (1992, 1993) who reported an FCM procedure for quantitating MN events in the peripheral blood reticulocyte (RET) population. In their system, a dual dye combination consisting of thiazole orange and Hoechst 33342 is employed. Thiazole orange stains the RNA component of the RET population, and Hoechst is used to label MN. The method, however, has high requirements in terms

of expertise and equipment relating to the dual-laser configuration of the flow cytometer needed to excite the two dyes. In this report, we describe a single-laser flow cytometric method for simultaneously quantitating MN events in the mature normochromatic erythrocyte (NCE) population and the immature RET population. Throughout the course of this research effort, a high priority was given to the development of methods which would provide reliable data with relative technical ease and modest equipment requirements - characteristics which are important for the widespread use of any automated scoring procedure. We believe that the method which was ultimately devised meets all of these criteria. It is based on the observation that RETs express the transferrin receptor (i.e., the CD71-defined antigen), and fluorescent antibodies directed against these surface markers can differentially label immature erythrocytes (Frazier et al., 1982; Seligman et al., 1983; Serke and Huhn, 1992). When fixed blood cells are treated with ribonuclease and stained with propidium iodide and anti-transferrin receptor antibody (FITCconjugate), NCEs and RETs with and without MN are clearly resolved. An experiment with the model clastogen methyl metbanesulfonate is described which demonstrates the MN scoring procedure outlined above. Additionally, a reconstruction study was performed so that MN-RET measurements obtained through manual scoring and flow cytometry could be directly compared. The data reported herein demonstrate the accuracy and reliability of flow cytometric MN data.

2. Materials and methods

2. I. Animals and blood cell fixation Adult male and female B A L B / c mice from a randomly bred closed colony were purchased from Charles River Laboratory, Inc., Wilmington, MA. The animals (16-18 g) were acclimated for 1 week before these experiments were initiated. Purina Mills' Lab Diet 5001 and water were available to the mice ad libitum. The peripheral blood samples collected for these studies were obtained from the tail vein after a brief warming period under a heat lamp.

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

Unless otherwise specified, approximately 50 txl was collected from each animal per harvest into a tube containing 250 Ixl anticoagulant solution (500 USP units heparin/ml saline). Blood samples were maintained at room temperature for no more than 2 h before being fixed. Fixation was achieved by forcefully delivering a 180 Ixl aliquot of each blood suspension from a pipettor into a polypropylene centrifuge tube containing 2 ml - 7 0 ° C methanol. The tubes were struck sharply several times to break up any aggregates, and stored at - 7 0 ° C for at least 24 h. To prepare cells for analysis, the tubes of fixed blood cells were struck sharply several times and 8 ml bicarbonate-buffered saline (0.9% NaC1 + 5.3 mM sodium bicarbonate, pH 7.5) was added. The cells were isolated by centrifugation and the cell pellets were stored at 4°C until analysis. 2.2. Anti-transferrin receptor staining

Subsequent to optimizing an immunofluorescent labeling procedure for resolving mouse reticulocytes, an experiment was performed to demonstate the reliability of the method as it relates to the micronucleus assay. In this experiment, one animal was treated in a manner which is known to stimulate erythropoiesis, and another animal was exposed to a drug which inhibits hemapoietic function. Specifically, one female mouse was bled extensively at time 0 h (approximately 300 p~l) in order to induce RBC production. Subsequent blood samples were collected at 24 and 48 h to track the influx of immature erythrocytes into the peripheral blood pool. A second female mouse was also bled at 0, 24 and 48 h, although not as extensively. To impair erythropoiesis, this animal was treated with methotrexate via intraperitoneal injection at time 0 and 24 h (50 mg methotrexate/kg bw; Sigma Chemicals; Cas. No. 59-05-2). The frequency of immature erythrocytes was determined for each of these six blood samples using two methods. The first method employed the nucleic acid dye propidium iodide (PI; Sigma Chemical; Cas. No. 25535-16-4). Similar to new methylene blue or acridine orange, PI differentially stains the immature erythrocytes based on their RNA content (Wallen et al., 1980). For this analysis, 20 p,l of fixed blood cells were transferred to tubes containing

285

1 ml bicarbonate-buffered saline with 1.25 I-tg P I / m l . The samples were analyzed with a flow cytometer providing 488 nm excitation. For each measurement, 500 000 total erythrocytes were interrogated, and the population of cells expressing a high red fluorescent signal (FL2 > 150) were scored as RNA-positive RETs. The second method utilized commercially available fluorescein isothiocyanate conjugated antitransferrin receptor antibody (FITC-ATR-Ab; Sigma Chemical, catalog no. F-2652). This monoclonal antibody reagent was used to differentially label and score immature erythrocytes. For this analysis, 20 txl aliquots of fixed blood cells were added to tubes containing 80 ~1 working FITC-ATR-Ab solution (10 Ixl stock FITC-ATR-Ab per ml bicarbonatebuffered saline). The cells were placed at 4°C for 30 min, resuspended with 1 ml cold bicarbonate-buffered saline, and analyzed with 488 nm excitation. As with the PI analyses, 500000 erythrocytes were interrogated per blood sample. The population of erythrocytes expressing a high green fluorescent signal (FLI > 150) were scored as RETs. 2.3. Model clastogen exposure

An experiment was performed to demonstrate the reliability of the FCM-based method to score clastogen-induced MN. For this experiment, five male B A L B / c mice were treated with 100 mg methyl methanesulfonate/kg bw via intraperitoneal injection at time 0 h (MMS; Sigma Chemical; Cas No. 66-27-3). Clastogen was delivered in a volume of 25 m l / k g bw. Blood samples were obtained at time 0, 24, 40, 48 and 72 h. Blood samples were fixed and stored at - 70°C until completion of the experiment. Fixed cells were washed out of methanol as described and stored at 4°C. For FCM analysis, 20 Ixl aliquots of fixed blood cells were added to tubes containing 80 Ixl working FITC-ATR-Ab solution with lmg ribonuclease A / m l (RNAse A; Sigma Chemical; Cat. No. R-5250). This solution prepared the cells for scoring by simultaneously labeling RETs with FITC-ATR-Ab and eliminating RNA content. After 30 min at 4°C, 1 ml cold propidium iodide staining solution was added (1.25 Ixg P I / m l bicarbonate buffer). Note that degradation of reticulocytes' RNA had to be achieved so that PI fluorescence would represent a DNA (micronuclei) specific

286

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

signal. Following the addition of the PI staining solution, cells were kept at 4°C until analysis.

2.4. Reconstruction experiment To facilitate a direct comparison between MNRET frequencies derived from manual and FCM methods, and to eliminate the influence of animal to animal variation, a reconstruction experiment was performed. For this study, aliquots of each 0-h and 40-h blood sample from the MMS experiment above were transferred to two separate tubes. The cells were vortexed and 500 Ixl of the pooled 0-h preparation was transferred to each of three tubes labeled 'control', 'spiked low' and 'spiked high'. The control tube was set aside, while 40 txl of pooled MMS blood was added to the 'low' tube, and 80 txl was transferred to the tube labeled 'high'. In this manner, preparations with modestly different MN-RET frequencies were constructed. Each preparation was scored blind six times using acridine orange-coated slides (Hayashi et al., 1990), or flow cytometrically based on the FITC-ATR-Ab and PI staining procedure described above. For manual scoring, 1000 RETs were analyzed for MN per replicate. FCM measurements were based on an evaluation of 10 000 RETs per replicate.

2.5. Flow cytometric analysis The FCM analyses outlined above were carried out with a FacStar Plus flow cytometer (Becton Dickenson). The laser was tuned to provide 488 nm excitation. Cells were passed through the laser at an average rate of 2500 erythrocytes/second. The FacStar Plus is equipped with four photomultiplier tubes (PMTs) which are used to sense forward light scatter, side light scatter, red and green fluorescence signals. Erythrocytes were isolated by gating on the light scatter parameters. Filters were placed before the green and red PMTs such that the green PMT registered fluorescence emission between 420 nm and 560 nm, whereas the red PMT measured emission greater than 580 nm. The combination of PI and FITC labeling is widely employed to simultaneously label nucleic acids and cell surface antigens (Kruth et al., 1981; Montecucco et al., 1985). It is well appreciated by those in the field that electronic com-

pensation which eliminates the longer wavelength emissions of the FITC signal enhances the resolution of PI and FITC-labeled cells. To optimally resolve the MN-RET population, we set FL2-%FL1 and FLI-%FL2 compensation to 51% and 0.6%, respectively (Consort 30 software, Becton-Dickenson). The dimensions of the MN analysis windows were guided by analyses of fixed erythrocytes from a malaria (P. berghei) infected mouse. Malaria is an excellent biological model, as it endows the erythrocyte population with an MN-like event, but has the advantage of being homogeneous in DNA content and more plentiful in peripheral blood (Tometsko et al., 1993a). Optics a n d / o r PMT voltages were adjusted to set this population to consistent red and green fluorescent channels. In this manner, the same analysis windows could be used to quantitate the number of NCEs, MN-NCEs, RETs and MN-RETs between samples, and even between experiments. These measurements were accomplished by analyzing at least 500000 total erythrocytes per blood sample. When quantitatively analyzing rare events such as the MN-RET population, it is critical that the flow cytometer's sample tubing is clean and free of debris which may interfere with highly accurate scoring. For these experiments, a particle-free solution consisting of 1% Clorox TM bleach with 50 mM NaOH in d H 2 0 was passed through the sample line for approximately 1 min between each sample. A similar solution has been reported to facilitate FCM analysis of ultrarare events (Gross et al., 1993), and represents a critical component of this application.

3. Results

3.1. Relevance of anti-transferrin staining Bleeding and methotrexate administration were employed to induce changes in the frequency of peripheral blood RETs. As expected, the number of RETs in the heavily bled animal was observed to rise in response to stimulated erythropoietic function. The administration of high dose methotrexate was observed to diminish RBC production. In both the erythropoiesis stimulated and inhibited animals, the immunofluorescent labeling procedure is found to

S.D. Dertinger et al. / Mutation Research 3 71 (1996) 283-292

closely parallel the nucleic acid staining method of P1 (see Fig. 1). Note that the absolute count of RETs is slightly lower for the FITC-ATR-Ab procedure. This is in agreement with other reports which suggest that the method labels approximately the youngest 70% of the RET population normally scored with supravital stains (Serke and Huhn, 1992). These data demonstrate that CD71 positive erythrocytes are extremely well suited as a target population for the peripheral blood micronucleus assay. Given appropriate blood sampling times, MN events in CD71 positive cells can be attributed to a chemical treatment administered during a recent cell cycle.

6]Propidium Iodide Method

3•

Melh

2"

1"

0

•

i 12

•

i 24 Time

7" FITC-ATR-Ab

•

g 36

-

i 48

•

i 48

•

60

(hours)

Method

6" Heavily Bled

4" %

RET 3"

2" 1"

0

i 12

•

i 24

•

i 36

•

60

Time (hours) F i g . 1. T h e e f f e c t o f h e a v y b l e e d i n g a n d m e t h o t r e x a t e a d m i n i s t r a t i o n on the p e r c e n t a g e o f p e r i p h e r a l b l o o d r e t i c u l o c y t e s is g r a p h e d as a f u n c t i o n o f t i m e . N o t e t h a t the i m m u n o f l u o r e s c e n t t e c h n i q u e ( b o t t o m ) c l o s e l y p a r a l l e l s the r e s p o n s e m e a s u r e d b y p r o p i d i u m iodide staining.

287

3.2. M M S response

MMS served as a model clastogen, having been studied carefully in other mouse micronucleus systems (Tsuyoshi et al., 1989; Sugiyama et al., 1992). To closely evaluate the effectiveness and reliability of the automated scoring method, an experiment was designed to track the incidence of MN in the peripheral blood of mice over time after a single exposure to clastogen. Quantitative FCM analysis was performed and the number of NCEs, MN-NCEs, RETs and MN-RETs was determined. A representative bivariate graph which illustrates the resolution of these four erythrocyte populations and the dimensions of the analysis regions is provided (Fig. 2). The complete data set is presented in Table 1. The mean MN frequency in the mature NCE population was found to rise gradually over the course of the experiment: 0.25% initially and 0.34% by 72 h. This modest effect on the mature erythrocyte population is expected, since MMS-induced micronuclei are diluted by the vast pool of pre-existing cells. In contrast, the frequency of MN events in the RET population was observed to rise sharply over the course of the first 40 h, reaching a maximal level of 4.20%. The incidence of MN-RET proceeds to fall to nearly background frequencies by 72 h (0.41%). The highly temporal effect of acute MMS exposure on MN frequency in the short-lived RET population is expected. The clastogenic action exerted by MMS reaches a maximum level and then quickly subsides as the hydrophilic compound is metabolized and excreted. The potency and kinetics of MMS-induced MN formation are in good agreement with the observations of Sugiyama et al. (1992). While Sugiyama et al. did not obtain a 40 h sample, they did observe a maximum response between their 24 and 48 h sampling times (i.e., 36 h). Aside from comparing MN frequencies over time, it is informative to study the frequency of MN events in the set of 0 h samples. As indicated by Table 1, the mean frequency of MN-NCEs and MN-RETs are very similar at time 0 h (0.25% versus 0.27%, respectively). We would expect these values to be approximately equal, since mice do not effectively clear MN from circulation (MacGregor et al., 1980). Note that before optimal staining and FCM operating procedures were realized, these values often diverged

288

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

these

f r e q u e n c y d e t e r m i n a t i o n s c o u l d be c o l l e c t e d o v e r

n u m b e r s was one useful tool w h i c h p r o v i d e d evi-

significantly.

The

correspondence between

time. This e x p e r i m e n t was p e r f o r m e d with b l o o d

d e n c e that M N - N C E and M N - R E T p o p u l a t i o n s w e r e

s a m p l e s o b t a i n e d f r o m M M S m o u s e No. 1. F o r this

b e i n g s c o r e d reliably and with precision.

investigation, e a c h o f the a n i m a l ' s 5 b l o o d s a m p l e s

W e e x t e n d e d the M M S e x p e r i m e n t to include an

w e r e p r e p a r e d for F C M analysis and s c o r e d 1 day

evaluation o f the c o n s i s t e n c y with w h i c h M N - R E T

and again 7 d a y s after the initial F C M m e a s u r e m e n t s

Table 1 Flow cytometric analysis of MMS-induced micronuclei Time

Mouse No.

No. NCE

No. MN-NCE

Freq. (%) MN-NCE

No. RET

No. MN-RET

Freq. (%) MN-RET

0

1 2 3 4 5 Average SD

490631 490648 489526 491259 493156

1339 1240 1111 1150 1390

0.27 0.25 0.23 0.23 0.28 0.25 0.02

8471 8541 9825 7883 6030

26 23 24 21 16

0.31 0.27 0.25 0.27 0.27 0.27 0.02

24

1 2 3 4 5 Average SD

490949 490112 488089 487462 490011

1244 1287 1145 1254 1391

0.25 0.26 0.23 0.26 0.28 0.26 0.02

8531 9390 11532 12234 9391

46 88 80 162 55

0.54 0.93 0.69 1.31 0.58 0.81 0.32

40

1 2 3 4 5 Average SD

490511 485301 486302 484237 492350

1405 1285 tl78 1445 1458

0.29 0.26 0.24 0.30 0.30 0.28 0.03

8373 13570 12679 14533 6785

386 592 652 581 262

4.41 4.18 4.89 3.81 3.72 4.20 0.48

48

1 2 3 4 5 Average SD

487164 481651 480648 476364 490250

1515 1538 1298 1517 1339

0.31 0.32 0.27 0.32 0.27 0.30 0.03

11794 17187 18484 22687 8990

355 366 396 431 211

2.92 2.09 2.10 1.86 2.29 2.25 0.40

72

1 2 3 4 5 Average SD

473052 462539 450096 461882 475141

1593 1573 1607 1628 1575

0.34 0.34 0.36 0.35 0.33 0.34 0.01

25838 36327 48913 36940 23744

112 147 224 150 82

0.43 0.40 0.46 0.40 0.34 0.41 0.04

(hrs)

Abbreviations: NCE, normochromatic erythrocytes; MN-NCE, micronucleated normochromatic erythrocytes; RET, reticulocytes; MN-RET, micronucleated reticulocytes; SD, standard deviation.

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

3

289

scoring reliability, and (ii) the relatively simple cell handling procedures and the use of a malaria model to guide instrument set-up makes it possible to compare data obtained on different days or from different experiments.

4

•!I! :ii i!iiii i i 2

3.3. Reconstruction

......

I

. . . . . . .

Propidium Iodide Fluorescence

Fig. 2. A representative bivariate graph illustrating the resolution of the various erythrocyte populations is presented. Window No. 1 corresponds to cells which are low in green and red fluorescence, i.e., NCEs; window No. 2: cells high in red fluorescence, i.e., MN-NCEs; window No. 3: cells high in green fluorescence, i.e., RETs; and window No. 4: cells high in both red and green fluorescence, i.e., MN-RETs.

reported in Table 1. These data are graphically presented in Fig. 3. The high reproducibility of the numbers suggest that (i) fixed samples can be stored at 4°C for at least 1 week without appreciable loss in

The six FCM and six manual MN-RET measurements accumulated for each of three mouse blood samples are presented in Table 2. As expected, the frequency of MN-RET is observed to increase as greater volumes of high-incidence MN blood (from an MMS-treated animal) were added to control preparations. Note that whereas the averages are very similar between the two methods, the amount of variation in the data is considerably lower with the FCM method. The low scoring error associated with the automated procedure greatly improved assay sensitivity, as evidenced by the high significance with which the two spiked samples were distinguished from the control preparation (one tailed t-tests, Table 2).

5.0 • ------o.--- day0 4.5- ~ dayl

Table 2 Manual and flow cytometric data compared

/ I

4.0-

3.5["

3.0-

~

2.5

AO Slides/Microscopic scoring

2.0 1.5 1.0-

Sample

6 replicate measurements (MN per 1000 RET)

Avg. %MNRET _+SD.

p-value ~

Untreated Spiked low Spiked high

2, 4, 2, 0, 2, 5 8, 5, 3, 5, 2, 4 6, 8, 6, 9, 3, 7

0.25 -+ 0.18 0.45 + 0.21 0.65_+0.21

0.0509 0.0024

0.5

Flow cytometric scoring

0.0

•

0

i

8

•

i

•

i

•

i

•

i

•

i

•

i

-

i

•

i

•

16 24 32 40 48 56 64 72 80

Sample

6 replicate measurements (MN per 10000 RET)

Avg. %MNRET _+SD

p-value a

Untreated Spiked low Spiked high

22, 21, 20, 24, 28, 24 57, 41, 48, 51, 47, 46 63, 48, 73, 66, 72, 55

0.23_+0.03 0.48 _+0.05 0.63_+0.10

0.0001 0.0001

Time (hours) Fig. 3. In this experiment, a mouse was injected with methyl methanesulfonate and the incidence of MN-RET was tracked in the peripheral blood pool over 72 h. The 5 blood samples were analyzed the day they were washed out of methanol (day 0), and subsequently on days 1 and 7. The high reproducibility of these data suggest that fixed and washed blood samples are stable at 4°C for at least 1 week, and that these flow cytometric scoring procedures provide reliable and accurate micronucleated cell measurements.

a One tailed t-tests were used to compare the mean frequency of manually scored MN-RET in the control sample relative to each of the spiked preparations• The same analyses were performed for the FCM-scored samples. SD, standard deviation.

290

S.D. Dertinger et al. / Mutation Research 371 (1996) 283-292

4. Discussion We have found that reliable and consistent fixation is a critical aspect of this FCM method. The fixing procedure must provide cells with the following characteristics: (1) in suspension and free of aggregates; (2) permeable to propidium iodide and RNAse; (3) recognizable CD71 antigen; and (4) low autofluorescence. A variety of fixing techniques were evaluated, and the ultracold methanol procedure described herein was considered optimal. We have found that the method is technically simple and highly reproducible. Importantly, the method does not generate undesirable autofluorescence as is the case with glutaraldehyde, a fixative which limits the resolution of FITC-labeled RETs. The most critical parameter of the methanol procedure is temperature. Cells must be added to - 70 to - 90°C methanol. At this stage, the erythrocytes are stable at - 7 0 ° C indefinitely. After cells are washed with bicarbonate-buffered saline, it is important to maintain them at 4°C. Prolonged periods of time at room temperature results in unacceptable degradation of light scatter and fluorescent staining characteristics. Washed cells maintained at 4°C are stable for at least l week. Beyond a reliable fixing procedure, a dye or dye combination was needed to adequately resolve the target MN-RET population. Classically, RETs are divided into five populations which are defined by the staining pattern observed in the presence of RNA-precipitating dyes (Vander et al., 1963). Stains such as thiazole orange (Lee et al., 1986) and acridine orange (Seligman et al., 1983; Hayashi et al., 1990) are widely employed. However, in regards to an FCM-based micronucleus assay, these and other RNA dyes are problematic. Since RNA dyes actually bind to DNA as well, overlapping signals tend to limit the resolution of MN-RETs from MN-NCEs. Furthermore, the dissimilar wavelengths necessary for the excitation of DNA and RNA dyes necessitates the use of dual-laser flow cytometers (Grawe et al., 1992). For an automated MN scoring protocol to gain wide acceptance and use, high resolution and modest equipment requirements are needed. We have found that significant advantages are realized when fluorescent antibodies directed against the transferrin receptor are used to label the RET population. As

considerable flexibility is available in terms of a fluorescent tag, it is possible to choose RET and MN-specific labels which are excited by a similar wavelength but exhibit significantly different emission spectra. For this application, we used an FITCsignal to differentiate the RET population, and propidium iodide to resolve the MN population. As Fig. 2 indicates, proper filter sets and electronic compensation results in well defined resolution of four cell populations (NCE, RET, MN-NCE and MN-RET). Fluorescent resolution is an important starting point for highly accurate quantitative analyses, and the simple FITC and PI labeling procedure described herein provides these conditions with single-laser excitation. The data reported for the MMS experiment suggest that the analysis windows used to define the MN-NCE and MN-RET populations are highly appropriate. Over the 72 h experimental time-frame, the incidence of MN in the NCE population rose slightly, and the frequency of MN-RETs was observed to rise and fall quickly. Given the persistence of these populations in the peripheral blood pool of mice, these profiles are expected. Furthermore, the high reproducibility found when blood samples from an MMS-treated mouse were re-analyzed over the course of a week suggests that the analysis windows can be used to compare FCM data obtained on different days. We believe that the use of malaria-infected erythrocytes to set instrument parameters each day and to monitor optics and PMT voltages over the course of a day is an important element of this application. With these controls in place, inter-experimental comparisons can be made with confidence. The micronucleus endpoint continues to grow in popularity as a genotoxic screen to identify chromosome-damaging agents. The high throughput capabilities of FCM technology represents a means by which these MN data could be significantly improved in terms of accuracy and reliability. Whereas traditional manual methods typically score the incidence of MN in 1000-2000 reticulocytes, state-ofthe-art flow cytometers are capable of analyzing hundreds of thousands of cells in just minutes. The experiment in which blood samples were scored both flow cytometrically and microscopically with acridine orange-coated slides demonstrates the advan-

S.D+ Dertinger et al. /Mutation Research 371 (1996) 283-292

tages associated with objective high throughput analyses. While the absolute MN-RET values obtained by the two methods are very similar, the degree of scoring error differs considerably. Given the low amount of variation and the very high confidence intervals achieved with the FCM method relative to manual scoring, significant improvements to assay sensitivity are realized. Furthermore, when FCM scoring precision is coupled with the important benefits associated with peripheral blood sampling (MacGregor et al., 1980), clear advantages in terms of assay reliability, efficiency, turn-around-time and costs are also likely. In addition to measuring MN in mouse peripheral blood, it is conceivable that methods similar to those described herein could be applied to studies with bone marrow samples. Preliminary studies with mouse bone marrow are very promising, and suggest that the high specificity with which the immunochemical reagent FITC-ATR-Ab labels RETs may help overcome the presence of debris which is appreciable in bone marrow preparations (Romagna and Staniforth, 1989). Experiments are planned to investigate this possibility.

Acknowledgements This research was supported in part by Grant Number 1 R43 ES07707-01 from the National Institutes of Health.

References Derringer, S., D. Torous and K. Tometsko (1996) Induction of micronuclei by low doses of azidothymidine (AZT). Mutation Res., 368, 301-307. Frazier, J., J. Caskey, M. Yoffee and P. Seligman (1982) Studies of the transferrin receptor on both human reticulocytes and nucleated human cells in culture: Comparisons of factors regulating receptor density. J. Clin. Invest., 69, 853-865. Grawe J., G. Zetterberg and H. Amneus (1992) Flow-cytometric enumeration of micronucleated polychromatic erythrocytes in mouse peripheral blood. Cytometry, 13, 750-758. Grawe, J., G. Zetterberg and H. Amneus (1993) Effects of extended low-dose-rate exposure to ~37Cs detected by flow-cytometric enumeration of micronucleated erythrocytes in mouse peripheral blood. Int. J. Radiat. Biol., 63, 339-347. Gross, H.+ B. Verwer, D. Houck and D. Recktenwald (1993)

291

Detection of rare cells at a frequency of one per million by flow cytometry. Cytometry, 14, 519-526. Hayashi, M., T. Morita, Y. Kodama, T. Sofuni and M. Ishidate Jr. (1990) The micronucleus assay with mouse peripheral blood reticulocytes using acridine orange-coated slides. Mutation Res.+ 245, 245-249. Hayashi, M., H. Norppa, T. Sofuni and M. Ishidate Jr. (1992) Flow cytometric micronucleus test with mouse peripheral erythrocytes. Mutagenesis, 7, 257-264. Heddle, J. (1973) A rapid in vivo test for chromosome damage. Mutation Res., 18, 187-190. Heddle, J., M. Cimino, M. Hayashi, F. Romagna, M. Shelby, J. Tucker, P. Vanparys and J. MacGregor (1991) Micronuclei as an index of cytogenetic damage: past, present and future. Environ. Mol. Mutagen., 18, 277-291. Kruth, H. R. Braylau, N. Benson and V. Nourse (1981) Simultaneous analysis of DNA and cell surface immunoglobulin in human B-cell lymphomas by flow cytometry. Cancer Res.+ 41, 4895-4899. Lee, L., C. Chen and L. Chiu 0986) Thiazole-orange: a new dye for reticulocyte analysis. Cytometry, 7, 508-516. MacGregor, J., C. Wehr and G. Gould (1980) Clastogen-induced micronuclei in peripheral blood erythrocytes: the basis of an improved micronucleus test. Environ. Mutagen., 2, 509-514. MacGregor, J., R. Schlegal, W. Choy and C. Wehr (1983) Micronuclei in circulating erythrocytes: a rapid screen for chromosomal damage during routine toxicity testing in mice, in: A.W. Hayes, R.C. Schnell and T.S. Miya (Eds+), Developments in the Science and Practice of Toxicology, Elsevier, Amsterdam, pp. 555-558. Matter, B. and W. Schmid (1971) Trenimon-induced chromosomal damage in bone-marrow cells of six mammalian species evaluated by the micronucleus test. Mutation Res., 12, 417425. Montecucco, C., A. Riccardi, G. Mazzini, M. Danova, G. Ucci and P. Giordano (1985) DNA content analysis of peripheral blood B-lymphocytes in plasma cell malignancies. Basic Appl. Histochem.+ 29, 275-282. Romagna, F and C. Stan ifortb (1989) The automated bone marrow micronucleus test. Mutation Res., 213, 91-104. Schmid, W. (1975) The micronucleus test. Mutation Res., 31, 9-15. Seligman, P., R. Allen, S. Kirchanski and P. Natale (1983) Automated analysis of reticulocytes using fuorescent staining with both acridine orange and an immunofluorescence technique. Am. J. Hematol., 14, 57-66. Serke, S. and D. Huhn (1992) Identification of CD71 (transferrin receptor) expressing erythrocytes by multiparameter-flow-cytometry (MP-FCM): correlation to the quantitation of reticulocytes as determined by conventional microscopy and by MP-FCM using a RNA-staining dye. Br. J. Haematol., 81, 432-439. Sugiyama, C., Y. Miyamae, H. Kobayashi, Y. Fujino, M. Mori and K. Ohara (1992) The micronucleus test of methyl methanesulfonate with mouse peripheral blood reticulocytes using acridine orange-coated slides. Mutation Res., 278, 117120.

292

S.D. Dertinger et aL / Mutation Research 371 (1996) 283-292

Tometsko, A., D. Torous and S. Dertinger (1993a) Analysis of micronucleated cells by flow cytometry. 1. Achieving high resolution with a malaria model. Mutation Res., 292, 129-135. Tometsko, A., S. Dertinger and D. Torous (1993b) Analysis of micronucleated cells by flow cytometry. 2. Evaluating the accuracy of high-speed scoring. Mutation Res., 292, 137-143. Tometsko, A., D. Torous and S. Derringer (1993c) Analysis of micronucleated cells by flow cytometry. 3. Advanced technology for detecting clastogenic activity. Mutation Res., 292, 145-153. Tometsko, A., S. Dertinger and D. Torous (1995) Analysis of micronucleated cells by flow cytometry. 4. Kinetic analysis of cytogenetic damage in blood. Mutation Res., 334, 9-18.

Tsuyoshi, T., M. Takeuchi, H. Hirano and Y. Masamoto (1989) Micronucleus test with methyl methanesulfonate administered by intraperitoneal injection and oral gavage. Mutation Res., 223, 383-386. Vander, J., C. Harris and S. Ellis (1963) Reticulocyte counts by means of fluorescent microscopy. J. Lab. Clin. Med., 62, 132-140. Wallen, C., R. Higashikubo and L. Dethlefsen (1980)Comparison of two flow cytometric assays for cellular RNA: acridine orange and propidium iodide. Cytometry, 3, 155.