- Email: [email protected]
Analysis of micronucleated cells by flow cytometry. 4. Kinetic analysis of cytogenetic damage in blood
Environmental
ELSEVIER
Mutagenesis
Mutation Research 334 (1995) 9-18
Analysis of micronucleated cells by flow cytometry. 4. Kinetic analysis of cytogenetic damage in blood Andrew M. Tometsko *, Stephen D. Dertinger, Dorothea K. Torous Litron Laboratories, 1351 Mt. Hope Ave., Rochester, NY 14620, USA
Received 15 October 1993;revision received 7 March 1994; accepted 18 March 1994
Abstract Micronucleated cells (MN cells) generated spontaneously or by clastogen action accumulate in the peripheral blood of the mouse, and their presence reflects the level of chromosome damage. Traditionally, micronucleated cells have been scored by visual inspection. With the development of flow cytometry based scoring procedures, vast numbers of cells can be analyzed, making it possible to determine the change in the number of MN cells in the total peripheral blood pool. This report describes experiments whereby initial blood samples were obtained before dosing, providing mouse-specific controls for measuring subsequent changes in MN cells. Mice were then dosed with saline (solvent control), methyl methanesulfonate, cyclophosphamide or colchicine every 48 h and bled every 96 h for 12 days. For each blood sample, one million fixed erythrocytes (RBCs) were interrogated for the presence of micronuclei, and regression analysis was used to determine the rate of MN cell influx per day for each animal or sets of animals. To evaluate the effect of treatment on MN induction, the mean slopes of solvent and chemically treated animals were compared using t-tests. The results of these experiments indicate that the kinetics of MN induction continues near the background frequency for saline dosed mice, whereas clastogenic agents or spindle poisons cause a significant influx of MN events into the blood. The results suggest that some studies may benefit from a flow cytometry based analysis of multiple blood samples, especially when the number of mice is limited, or when a weak clastogen is being investigated. Keywords: Micronucleus; Flow cytometry; Clastogens; Chromosome damage; Kinetics
I. Introduction Since its introduction, the mouse micronucleus assay has been in a state of evolution with improvements to every aspect of the method. The assay (Matter and Schmid, 1971; Heddle, 1973; Schmid, 1975) is based on the fact that clastogens
* Corresponding author.
or spindle poisons cause genetic damage to erythroblasts. The resulting accumulation of chromosomes and chromosome fragments can be scored in the red blood cell population which is normally D N A deficient. This assay was originally restricted to bone marrow, where the number of micronucleated cells (MN ceils) could easily be counted by visual inspection of immature polychromatic erythrocytes (PCEs) (Salamone et al., 1980; Heddle et al., 1983; Salamone and Heddle,
0165-1161/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0165-1161(94)00026-L
10
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
1983; Mavournin et al., 1990). However, it was found that micronucleated erythrocytes are not cleared from the blood in the mouse (MacGregor et al., 1980; Schlegel and MacGregor, 1982), thus allowing the analysis to be carried out with blood samples (Mavournin et al., 1990; Collaborative Study Group for the Micronucleus Test, 1992). With manual scoring methods a limited number of cells are analyzed for each mouse (i.e. 10002000 total PCEs per blood or bone marrow sample). Since micronuclei are present as rare events at a frequency of about 0.2% at background levels, analyzing only 1000 cells adversely impacts on assay statistics and results in a large coefficient of variance (_+70%) (Tometsko et al., 1993b). It has been realized for some time that the next step in the evolution of the assay should involve automated methods for objectively and accurately scoring larger numbers of micronucleated erythrocytes to improve assay statistics (Tucker et al., 1989). Over the years, flow cytometry (FCM) has received considerable attention as a possible solution to objectively analyzing micronucleated cells, and a variety of methods have been proposed (Hutter and Stohr, 1982; Grawe et al., 1992; Hayashi et al., 1992a,b; Krishna et al., 1993; Tometsko et al., 1993a,b,c). These methods differ considerably in their approaches to scoring and analyzing micronucleated cells and should be studied carefully for resolution of MN cells and elimination of 'noise' during data acquisition (Heddle et al., 1991). With flow cytometry, the rate of data acquisition is increased by orders of magnitude over hand scoring, and processing speeds of 1 000000 cells per minute have been achieved (Tometsko et al., 1993b,c). The sheer volume of data provided by flow cytometry makes quality control and scoring accuracy important concerns. Therefore, chemical, biological and instrument parameters must be carefully optimized and proper control of the associated variables should be demonstrated as a prelude to a flow cytometry based analysis procedure. With this in mind, the authors focused previous reports on the resolution of MN cells from other cell populations and the development of methods for assessing scoring accuracy. It was only subsequent to
these resolution and accuracy studies that the analysis of a biological response due to chemical treatment was carried out (Tometsko et al., 1993a,b,c). This article describes the analysis of multiple blood samples by flow cytometry in order to determine the kinetics of MN cell influx into the peripheral blood. Multiple sample analysis had been previously reported in conjunction with hand scoring (Luke et al., 1988a,b). Although initial blood samples were obtained, the 0-h data were excluded from the analysis of the treatment groups where data for each mouse were averaged over time to yield a 'temporal average'. In the present method, the initial blood sample plays a central role in the analysis by serving as a mouse specific control (internal standard) for evaluating the net change in MN cells. The bleeding and dosing regimen described herein is based on previous work (Tometsko et al., 1993c) in which mice were dosed at 0 h and 48 h and were bled initially (0 h; before dosing) and at 96 h. For continuity, this previous bleeding and dosing protocol was extended through two additional cycles, terminating with a blood sample obtained on day 12. The results suggest that the analysis of multiple samples by flow cytometry may provide an important advantage when increased sensitivity is required in the micronueleus assay (e.g. studies involving a weak clastogen), or when the number of mice is limited (e.g. studies involving transgenic mice). The methods and procedures presented here are easily adaptable to subchronic studies by periodically obtaining blood samples.
2. Material and methods Biological and chemical considerations
Adult male and female mice (strain BALB/c) from a randomly bred closed colony were purchased from Charles River Breeding Laboratories, Inc., Wilmington, MA. A commercial diet and water were available to the mice ad libitum. An experiment with methyl methanesulfonate involved one treatment group consisting of five male mice, whereas the cyclophosphamide and colchicine studies included both solvent and
A.M. Tometskoet al. / Mutation Research334 (1995)9-18 chemical treatment groups with four male and four female mice in each set. On days 0, 2, 4, 6, 8 and 10, animals were dosed with saline (solvent control), cyclophosphamide (CP: CAS No. 605519-2: Sigma Chemical Co.; 100 /~g/g b.w.) (Tinwell et al., 1990; MacGregor et al., 1980; Salamone et al., 1980; Schlegel and MacGregor, 1982), methyl methanesulfonate (MMS: CAS No. 66-27-3: Eastman Kodak Co.; 100 ~ g / g b.w.) (Jenssen and Ramel, 1976, 1978; Tsuyoshi et al., 1989) or colchicine (CAS No. 64-86-8: Sigma Chemical Co.; 1/~g/g b.w.) (Hayashi et al., 1989). Each of the solutions was prepared in saline. The mice were bled initially (i.e. 0 h; before dosing) and on days 4, 8 and 12. Micronuclei were stained with Hoechst 33258 (HST: CAS No. 23491-44-3: Sigma Chemical Co.) which is specific for DNA (Latt et al., 1976). Propidium iodide (PI: CAS No. 25535-16-4) (Hudson et al., 1969; Wallen et al., 1982) was used to quantitate the PCE population.
Preparation of blood samples for FCM analysis In these studies, tail vein blood samples (5 drops; about 200 ~1)were collected in heparin (0.5 ml; 500 USP units/ml) initially and every 4 days. The blood cells were fixed by adding 180/~1 of a blood suspension to 2 ml of ultracold ( - 70°C) methanol (Tometsko et al., 1993a). All of the samples for each set of mice were stored at - 7 0 ° C until the completion of the dosing regimen. At ultralow temperatures, fixed blood samples are stable indefinitely. At the time of analysis, cells were diluted with sodium chloride/sodium bicarbonate buffer (8 ml) (sodium chloride (9 g) plus sodium bicarbonate (0.445 g) per liter). The cells were collected by centrifugation and were stored at 4°C until analysis. A sample of the pellet (5-10 /zl) was diluted into Tris-borateEDTA (TBE) buffer (1 ml) containing Hoechst 33258 (84 ng/ml TBE). Quantitative analysis of cell populations by flow cytometry The number of MN cells present per one million erythrocytes was determined for every blood sample in this study using a FACStar Plus flow cytometer (Becton Dickinson). The DNA-specific fluorochrome Hoechst 33258 was used to differ-
11
entially stain red blood cells with and without micronuclei. The cells were interrogated with a UV laser beam from a 5-W Innova 90 Argon Ion Laser (Coherent). The fluorescence emission from each cell was collected by photomultiplier tubes (PMTs) and the number of MN cells in the analysis area was tabulated. The FACStar Plus flow cytometer is equipped with four detectors, which are used to sense forward scatter, side scatter, blue fluorescence and yellow fluorescence signals. The blue PMT registered fluorescence emission between 420 nm and 555 nm, whereas the yellow PMT measured emission greater than 580 nm. Data were acquired in the dual parameter mode. The dimensions of the micronucleus analysis window were set using red blood cells containing the malaria parasite Plasmodium berghei as a model. By interrogating one million total blood cells per sample, a coefficient of variance of < 3% should be obtained for rare events present at a 0.2% frequency when analyzed under optimum conditions (Tometsko et al., 1993a,b). Since new MN events must pass through the PCE population, toxicity at the erythroblast level can be detected as a decrease in the number of PCEs present in the blood. FCM can provide a rapid assessment of toxicity as a prelude to micronucleus analysis. In these experiments, the P C E / N C E ratio was determined by staining blood samples with propidium iodide which causes PCEs to fluoresce red when excited by a UV laser beam. Erythroblast toxicity was minimal under the dosing regimens as described.
Slope analysis The experimental design of these studies makes use of initial and subsequent blood samples so that the induction of MN events is carefully monitored over time for each animal. With these multiple data points, it is possible to calculate a regression line for each animal or each set (Cricket-Graph T M Software) which reflects the influx of MN cells per million erythrocytes per day (i.e. the kinetics of MN formation). It should be apparent that a number of statistical methods may be employed to test whether the slopes of these lines are significantly different or whether they might be estimating the same population. For this
A.M. Tometsko et al. /Mutation Research 334 (1995) 9-18
12
study, we chose to analyze the effect of chemical treatment on mean slopes with one-tailed t-tests (Statview II). An additional analysis was performed on colchicine data where the response of each female mouse was transformed into a series of three slopes covering the range 0 - 4 days, 4 - 8 days, and 8-12 days. This treatment is based on experiments which indicate that the influx of new MN cells returns to the spontaneous background level 48 h after dosing. In one such experiment, the ability of the system to reset to background frequencies was evaluated with four male mice that were dosed with MMS ( 1 0 0 / x g / g b.w.) at 0 and 2 days only, and were bled at 0, 4 and 8 days. As expected, the influx was 353 _+ 85 MN cells/ day for the first 4-day cycle. However, the rate of influx was - 2 5 _+ 77 MN cells/day for the second 4-day cycle, showing that MN cell influx returned to background frequencies within 48 h of dosing. Therefore, each cycle represents a sep-
arate and independent experiment, and the starting point for a subsequent cycle can be reset each time a mouse is bled. By calculating the slope for each 4-day time interval as a separate experiment (as discussed for colchicine later), a three-fold increase in the number of independent readings was available for statistical evaluation (one-tailed t-test).
3. Results
As a result of cytogenetic damage, MN cells accumulate in the peripheral blood of mice where they can be analyzed relative to background levels. The initial blood sample (day 0) represents a baseline reading on the spontaneous background frequency of MN cells for each mouse. Subsequent samples therefore reflect any accumulation of MN cells resulting from clastogen exposure. Newly formed MN cells require about 48 h to
Table 1 Cyclophosphamide response Mouse
Sex
Treatment
MN cells
a/106 erythrocytes
Slope b
day0
day4
day 8
day l2
1 2 3 4 5 6 7 8 Average SD
Female Female Female Female Male Male Male Male
Saline Saline Saline Saline Saline Saline Saline Saline
2485 2718 2825 2360 2649 2359 2630 2485 2563 169
2813 2762 2463 2347 2665 2675 2709 2619 2632 155
2608 2734 2890 2479 2808 2717 2679 2731 2706 124
3002 2910 2781 2957 2805 2608 2674 2634 2796 150
34 14 7 48 15 20 3 14 19.4 14.8
9 10 11 12 13 14 15 16 Average SD
Female Female Female Female Male Male Male Male
CP c CP CP CP CP CP CP CP
3010 2375 2624 2420 2560 2539 2594 2569 2586 191
3361 3262 3412 3271 2741 3415 3398 3322 3273 223
5069 3983 4015 4540 4239 3893 4170 4051 4245 388
6210 5212 6335 5179 5760 4729 5246 5357 5504 552
283 231 293 239 277 176 218 227 243.0 39.3
MN cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c CP, cyclophosphamide (100 # g / g b.w.; i.p. injection). The difference between means (eight animals per treatment group) was compared with a one-tailed t-test: p value = 0.0001.
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
reach the bloodstream (MacGregor et al., 1980). In previous studies (Tometsko et al., 1993), mice were dosed with MMS at 48-h intervals and were bled initially and at 96 h. For this study the dosing/bleeding regimen was extended to 12 days (three cycles) during which time five male mice were treated with methyl methanesulfonate. Fig. 1 shows the response of each mouse to multiple doses of MMS. Since this regimen provided multiple blood samples per mouse, a measurement of the kinetics of MN cell influx into the peripheral blood could be readily obtained by regression analysis. For the MMS dosed mice shown in Fig. 1, the regression equation is: y = 1917 + 1126x R 2 = 0.94 This equation describes an influx rate of 1126
13
20000 18000 -
cycle 1
I
cycle 2
]
cycle 3
I
16000 14000
-
12000 MN
cells 10 0 0 0 8000 6000 4000 2000 0 0
i
i
2
4
•
i
,
6 Time
i
•
8
i
r
10
12
(Days)
Fig. 1. The n u m b e r of M N cells per million erythrocytes was plotted as a function of time for five male B A L B / c mice dosed with methyl methanesulfonate (100 t x g / g b.w.) every 48 h and bled every 4 days. Each 4-day period defined one cycle of the b l e e d i n g / d o s i n g regimen. Notice that each of the five mice displayed similar responses to MMS treatment.
Table 2 Colchicine response Mouse
Sex
Treatment
M N cells a/106 erythrocytes
Slope b
day 0
day 4
day 8
day 12
1 2 3 4 5 6 7 8 Average SD
Female Female Female Female Male Male Male Male
Saline Saline Saline Saline Saline Saline Saline Saline
2000 2111 1918 2082 2642 2594 2366 2438 2269 277
1991 2096 2136 2256 2549 2238 2599 2879 2343 303
2084 2287 1993 2401 2780 2376 2766 2414 2388 282
2570 2265 2262 2579 2731 2593 2773 2820 2574 213
45 16 22 41 12 3 35 17 23.9 14.9
9 10 11 12 13 14 15 16 Average SD
Female Female Female Female Male Male Male Male
Colchicine c Colchicine Colchicine Colchicine Colchicine Colchicine Colchicine Colchicine
2168 2221 1903 2212 2087 2236 2519 2423 2221 190
2477 2544 2345 2666 3166 3198 3602 2944 2868 433
3154 3420 3538 3650 3546 4194 4461 3101 3633 474
3970 3745 5307 4852 dead 4685 6233 3217 4572 1022
152 136 285 223 182 209 300 63 193.8 78.4
a M N cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c 1 / x g / g b.w., i.p. injection. T h e difference between m e a n s was compared with a one-tailed t-test: p value = 0.0001. The difference between m e a n s (four female mice per treatment group) was compared with a one-tailed t-test: p value = 0.0015; degrees of freedom = 6.
A.M. Tometsko et al. /Mutation Research 334 (1995) 9-18
14
MN cells per million erythrocytes per day during the MMS exposure time frame. In a second multiple dose experiment, mice were treated with saline (i.e. solvent) or CP, and the number of MN cells per million erythrocytes was obtained for each sample. The means reflect an increase due to CP treatment (see Table 1). The mean slopes obtained by regression analysis describe an influx rate of 19.4 MN cells/day and 243 MN cells/day for saline and CP treated mice, respectively. These results indicate that CP caused a 12.5-fold increase in the rate at which MN cells accumulated in the peripheral blood pool under these exposure conditions. In order to provide a graphic overview, the rate of influx (MN cells/million R B C s / d a y ) for each saline and CP dosed mouse was plotted as a bar graph in Fig. 2. This figure shows that the rate of micronucleus formation for all saline mice was low, whereas an elevated response was obtained for mice dosed with cyclophosphamide. This type of graphic presentation is useful for highlighting the biological variation (i.e. rate of MN cell in-
350
300
250
200 Rate MN
of Influx 15o
100
0 I
2
3
4
5
6
7
8
9
l0
11
12
13
14
15
16
Mouse N u m b e r
Fig. 2. In this experiment, mice were bled every 4 days and injected with 0.9% saline or cyclophosphamide (CP; 100 p~g/g b.w.) every 2 days for 12 total days. Regression analysis was used to determine the rate at which micronucleated cells entered the peripheral blood pool for each mouse over the experimental time frame. The rate of MN cell influx (i.e. MN cells/day) was obtained from the slope of the regression line and was plotted as a bar graph for each mouse, providing an overview of the biological variation for the saline and CP dosed mice.
5000 []
Saline
y - 2249.5 + 24.000x • .
45OO
=
.
R'2 - 0.910 .
~
=
.
.
40OO
Average Number o f MN cells
3500
3000
2500,
20o0
•
i 2
,
i
,
4
i
i
6 Time
8
10
12
(Days)
Fig. 3. The average n u m b e r of MN cells was plotted as a function of time for eight mice that were bled every 4 days and dosed with saline or colchicine (1 ~ g / g b.w.) every second day for 12 days. Each blood sample was fixed, stained with Hoechst 33258, and was quantitatively analyzed for the presence of MN cells by high speed flow cytometry. Regression analysis was performed on the mean MN cell values and the regression equations yielded slopes of 24 MN cells/day and 195 MN cells/day for saline and colchicine dosed mice, respectively.
flux) of individual mice due to clastogen treatment. In a spindle poison experiment, mice were also dosed with colchicine and the results are presented in Table 2. In this experiment regression analysis produced a mean slope for the saline animals of 24 MN cells/day while colchicine treatment resulted in a slope of 194 MN cells/day. Therefore, the effect of colchicine at this dose was an eight-fold increase in the influx of MN cells above the background level. The graphic view of this effect (Fig. 3) was obtained by plotting the average MN value per million RBCs for each treatment group versus time. This presentation shows that for this low level of genotoxic activity, a longer dosing and bleeding regimen is preferred since it allows the treatment profile to diverge from the background profile. While solvent controls show a slight rise over time, clastogen treatment results in a more rapid elevation (accumulation) of MN events in the blood. Additional analysis of the colchicine data was performed to illustrate how multiple sample analysis may offer a means for improving the sensitivity of the assay or possibly reducing the number
15
A.M. Tometsko et aL / Mutation Research 334 (1995) 9-18
of animals required. For this demonstration, it was necessary to reduce the number of animals per set from eight to four so that the p values would be higher and would be more sensitive to changes in the degrees of freedom. Accordingly, only data from the female mice (Table 2) were used for this analysis and each dosing/bleeding cycle was processed as a separate and discrete experiment, resulting in a three-fold increase in the degrees of freedom. With this adjustment, it was possible to obtain a p value with four mice that was obtained when the treatment group consisted of eight animals each. Operationally, the data for each of the four saline and colchicine animals were transformed into a series of three successive slopes corresponding to days 0-4, days 4 - 8 and days 8-12, respectively (see Table 3). Given the time frame in which clastogen induced MN cells enter the peripheral blood pool, each successive slope reflects a separate application of test material, and each analysis can be thought of as an independent observation. The statistics of the assay are affected considerably by the three-fold increase in statistical sample size. For example, when the slopes were compared for four saline dosed and four colchicine dosed females, a p value of 0.0015 was obtained (one-tailed t-test; degrees of freedom = 6). This is in contrast to a p value of 0.0001 when the data were processed as three
independent slope readings per mouse (degrees of freedom = 22). These results indicate that by increasing the degrees of freedom, multiple independent readings improve the confidence level of statistical analysis, even when small numbers of animals are used.
4. Discussion The purpose of this presentation is to demonstrate that FCM based methods and procedures can be used to assess the kinetics of MN cell influx into the peripheral blood. In these studies, an initial blood sample was obtained from each mouse and was used for subsequent comparisons. The initial blood samples provided a baseline reading on each animal, showing the frequency at which micronucleated cells enter the blood cell pool under normal physiological conditions. If mice are treated with saline, the influx of MN ceils should continue at or near the background level, resulting in only a slight increase in MN ceils over the experimental time frame. Stimulation of the erythropoietic system, a n d / o r changes in hormone and cytokine levels may account for the slight rise in the background frequency of MN cells over time. In any case, data presented here indicate that bleeding mice for periodic MN readings does not significantly affect the sponta-
Table 3 Colchicine response: calculation of three slopes per mouse Mouse Sex Treatment MN cells a/106 erythrocytes (day 0)
slope b
(day 4)
slope
(day 8)
slope
(day 12)
1 2 3 4
Female Female Female Female
Saline Saline Saline Saline
(2000) (2111) (1918) (2082)
-2 -4 55 44
(1991) (2096) (2136) (2256)
23 48 - 36 36
(2084) (2287) (1993) (2401)
122 - 6 67 45
(2570) (2265) (2262) (2579)
9 10 11 12
Female Female Female Female
Colch c Colch Colch Colch
(2168) (2221) (1903) (2212)
77 81 111 114
(2477) (2544) (2345) (2666)
169 219 298 246
(3154) (3420) (3538) (3650)
204 81 442 30i
(3970) (3745) (5307) (4852)
MN cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c Colch, colchicine(1 /zg/g b.w.; i.p. injection). The difference between means (four animals or 24 slopes per treatment group) was compared with a one-tailed t-test: p value = 0.0001; degrees of freedom = 22. a
16
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
neous frequency of MN cell formation. On the other hand, a larger divergence from background values can result during clastogen exposure. The analysis of multiple blood samples by hand scoring has been previously reported (Luke et al., 1988a,b), and demonstrated a clastogen dependent accumulation of MN cells in the peripheral blood. In this case, the numbers of MN NCEs at the different time points were pooled (summed) and the 'temporal average' was calculated for comparison purposes. In contrast, in these flow cytometry experiments the numbers of MN RBCs at each time point were treated as separate and discrete readings in order to assess the current status of MN cells in each mouse. Although 'temporal averages' might be calculated, it seemed more informative to use high speed FCM data acquisition to measure the kinetics of MN cell influx following clastogen exposure. Since MN cells are not cleared from the peripheral blood pool, any change in their frequency can be detected in the RBCs if enough cells are analyzed. In fact, millions of cells can be scored for MNs by taking advantage of the processing speed of modern flow cytometers. (Note that these experiments are only possible because resolution and scoring accuracy have been optimized.) FCM based analysis of multiple samples provides more information regarding the impact of a chemical on each individual biological system and thus differs significantly from conventional micronucleus assays whereby a single blood or bone marrow sample is analyzed. These experiments also differ from our previous studies in which only two blood samples were obtained at day 0 and day 4, and the net change in MN cells (AMN) was determined for each mouse. Fig. 1 shows the response of five male mice to three cycles of the bleeding/dosing regimen. This figure provides important insights into multiple sample analysis, showing that the rate of MN cell influx during cycles two and three was as high as or higher than during cycle one. In the absence of PCE toxicity, a continuation of the first cycle rise in MN cells would be expected (i.e. continuity of response). This figure also suggests that bleeding can enhance the response to clastogens, since the number of MN cells is accumulating faster in the
third cycle than in the first and second cycles. This effect might be expected since bleeding stimulates erythropoiesis, which sensitizes erythroblasts to clastogens (Hirai et al., 1991). Finally, it is apparent from Fig. 1 that each mouse in the set is reflecting the same general sensitivity to MMS treatment. Thus multiple dose/multiple bleeding protocols as described in this presentation provide significantly more information regarding the response of each mouse and may prove to be statistically more powerful. With multiple samples, all of the data from each mouse can be used in the evaluation of a clastogenic response (in this case four data points). Regression analysis has been particularly useful for processing multiple sample data since the resulting regression lines can signal cytogenetic damage for a single mouse, or sets of mice. In the present examples, when mice were dosed with clastogens or a spindle poison, regression line slopes reflected the elevated MN cell level and were 8-12 times above background values. By calculating slopes, one essentially compresses the data from multiple samples into a rate of MN cell influx corresponding to the specific dosing conditions. A comparison of the rates of MN influx (i.e. slopes) for individual animals (Fig. 2) provides an overview of biological variation in mice dosed with saline or CP and is helpful for analyzing the consistency of a positive response. In these studies, all of the chemically dosed mice exhibited a rise above the background level, even when the response was relatively weak as with colchicine (Table 2). In theory, multiple samples from each mouse should provide a more accurate and detailed assessment of a clastogenic response compared to the conventional single sample per animal analysis. Each measurement of MN cell influx should be evaluated for consistency with respect to preceding or following rates. For example Fig. 3 shows an increase in accumulated MN cells with each repeat of the dose/bleeding cycle. With the present three cycle protocol, each animal also provides three independent measurements of the rate of MN cell influx. Since clastogen induced MNs will have entered the peripheral blood pool by the 48-h mark, each 4-day time interval can be
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
considered a separate independent experiment to test the effect of treatment on MN formation. Accordingly, each blood sample can be used as the basis of comparison for evaluating the change in MNs upon subsequent readministrations of test material. For each mouse, these successive slopes should reflect the overall characteristics of the regression line (e.g. Fig. 3), but will contribute more degrees of freedom to statistical analysis procedures. For example, when slope analysis was carried out on successive readings for four saline and four colchicine animals (i.e. change in MNs for days 0-4, days 4-8, days 8-12), the degrees of freedom were increased from 6 to 22, and the confidence interval was markedly enhanced. Since each mouse contributed three independent slopes reflecting the accumulation of MN ceils between four successive blood readings, a larger number of discrete observations were supplied to'the statistical operation. By increasing the degrees of freedom in this manner, the confidence interval resulting from an intergroup comparison of mean slopes was enhanced (i.e. the p value of a onetailed t-test was lowered approximately 10-fold). With multicycle analysis (e.g. three slopes/mouse; Table 3), four mice were able to provide a level of statistical confidence that required eight animals per treatment group in Table 2. Bleeding and dosing regimens should be designed to highlight any increase in MN cells that might occur. Based on a large number of experiments, a 48-h dosing and 96-h bleeding schedule seems quite adequate to allow MN cells to accumulate in the peripheral blood. The methods that we have outlined are compatible with other dosi n g / b l e e d i n g schedules, so long as bleeding is performed after sufficient time has elapsed to allow chemically induced MN cells to enter the peripheral blood pool. The choice of dosing and bleeding protocols should be guided by the toxicity of the test substance and the rate of accumulation of MN cells into the blood pool under the specific experimental conditions. The considerations of dosing and sacrifice time that are of critical importance for the analysis of MN PCEs are not pertinent to the analysis of accumulated MN cells in the total red blood cell pool. Bleed-
17
ing and dosing regimens involving extended treatment times allow a longer period for MN cells to accumulate if a clastogenic response is occurring. Thus, the time a cell spends in the short-lived PCE population is not critical to the analysis of accumulated MN events in peripheral blood. We recommend that an initial sample should be obtained before any dosing begins. We continue to analyze one million total blood cells for each sample in order to score sufficient rare MN cells in a large cell pool. In the present experiments, data analysis was limited to 4 000 000 cells per mouse which provided a suitable demonstration of the capabilities of flow cytometry to determine the rate of influx of MN cells due to various chemical treatments. The general method described here provides a means for precisely and rapidly evaluating the kinetics of chemically induced MN formation and readily complements subchronic studies.
Acknowledgement This research was supported by Grant 2 R44 G M 35043-02 from the National Institutes of Health.
References Collaborative Study Group for the Micronucleus Test (1992) Micronucleus test with mouse peripheral blood erythrocytes by acridine orange supravital staining, Mutation Res., 278, 83-98. Grawe, J., G. Zetterberg and H. Amneus (1992) Flow-cytometric enumeration of micronucleated polychromatic erythrocytes in mouse peripheral blood, Cytometry, 13, 750758. Hayashi, M., S. Sutou, H. Shimada, S. Sato, Y.F. Sasaki and A. Wakata (1989) Difference between intraperitoneal and oral gavage application in the micronucleus test, Mutation Res., 223, 329-344. Hayashi, M., H. Norppa, T. Sofuni and M. Ishidate Jr. (1992a) Mouse bone marrow micronucleus test using flow cytometry, Mutagenesis, 7, 251-256. Hayashi, M., H. Norppa, T. Sofuni and M. Ishidate Jr. (1992b) Flow cytometric micronucleus test with mouse peripheral erythrocytes, Mutagenesis, 7, 257-264. Heddle, J.A. (1973) A rapid in vivo test for chromosome damage, Mutation Res., 18, 187-190.
18
A.M. Tometsko et al. / M u t a t i o n Research 334 (1995) 9-18
Heddle, J.A., M. Hite, B. Kirkhart, K. Mavournin, J.T. MacGregor, G.W. Newell and M.F. Salamone (1983) The induction of micronuclei as a measure of genotoxicity, Mutation Res., 123, 61-118. Heddle, J.A., M.C. Cimino, M. Hayashi, F. Romagna, M.D. Shelby, J.D. Tucker, Ph. Vanparys and J.T. MacGregor (1991) Micronuclei as an index of cytogenetic damage: past, present, and future, Environ. Mol. Mutagen., 18, 277-291. Hirai, O., Y. Miyamae, Y. Fujino, H. Izumi, A. Miyamoto and H. Noguchi (1991) Prior bleeding enhances the sensitivity of the in vivo micronucleus test, Mutation Res., 264, 109-114. Hudson, B., W.B. Upholt and J. Divinny (1969) The use of an ethidium analogue in the dye-buoyant density procedure for the isolation of closed circular DNA: The variation of the superhelix density of mitochondrial DNA, Proc. Natl. Acad. Sci. USA, 62, 813-820. Hutter, K.J. and M. Stohr (1982) Rapid detection of mutagen induced micronucleated erythrocytes by flow cytometry, Histochemistry, 75, 353-362. Krishna, G., D. Brott, G. Urda, M. McKeel, J. Zandee and J. Theiss (1993) Comparative micronucleus quantitation and pre- and post-column fractionated mouse bone marrow by manual and flow methods, Mutation Res., 302, 119-127. Latt, S.A. and G. Stetten (1976) Spectral studies on 33258 Hoechst and related bisbenzimidazole dyes useful for fluorescent detection of deoxyribonucleic acid synthesis, J. Histochem. Cytochem., 24, 24-33. Luke, C.A., R.R. Tice and R.T. Drew (1988a) The effect of exposure regimen and duration on benzene-induced bone-marrow damage in mice. I. Sex comparison in D B A / 2 mice, Mutation Res., 203, 251-271. Luke, C.A., R.R. Tice and R.T. Drew (1988b) The effect of exposure regimen and duration on benzene-induced bonemarrow damage in mice. II. Strain comparisons involving B6C3F1, C57B1/6 and D B A / 2 male mice, Mutation Res., 203, 273-295. MacGregor, J.T., C.M. Wehr and D.H. Gould (1980) Clastogen-induced micronuclei in peripheral blood erythrocytes, Environ. Mutagen., 2, 509-514. Matter, B. and W. Schmid (1971) Trenimon-induced chromo-
somal damage in bone-marrow cells of six mammalian species evaluated by the micronucleus test, Mutation Res., 12, 417-425. Mavournin, K.H., D.H. Blakey, M.C. Cimino, M.F. Salamone and J.A. Heddle (1990) The in vivo micronuclues assay in mammalian bone marrow and peripheral blood. A report of the U.S. Environmental Protection Agency Gene-Tox Program, Mutation Res., 239, 29-80. Salamone, M.F. and J.A. Heddle (1983) The bone marrow micronucleus assay: rationale for a revised protocol, in: F.J. de Serres (Ed.), Chemical Mutagens 8, Plenum, New York, pp. 111-149. Salamone, M.F., J. Heddle, E. Stuart and M. Katz (1980) Towards an improved micronucleus test. Studies on 3 model agents, mitomycin C, cyclophosphamide, and dimethyl benzanthracene, Mutation Res., 74, 347-356. Schlegel, R. and J.T. MacGregor (1982) The persistence of micronuclei in peripheral blood erythrocytes: detection of chronic chromosome breakage in mice, Mutation Res., 104, 367-369. Schmid, W. (1975) The micronucleus test, Mutation Res., 31, 9-15. Tometsko, A.M., D.K. Torous and S.D. Dertinger (1993a) Analysis of micronucleated cells by flow cytometry. 1. Malaria as a biological model, Mutation Res., 192, 129133. Tometsko, A.M., S.D. Dertinger and D.K. Torous (1993b) Analysis of micronucleated cells by flow cytometry. 2. Scoring rare events with high fidelity, Mutation Res., 192, 137-143. Tometsko, A.M., D.K. Torous and S.D. Dertinger (1993c) Analysis of micronucleated cells by flow cytometry. 3. Advanced technology for detecting clastogenic activity, Mutation Res., 192, 145-153. Tucker, J.D., J.A. Heddle, J.T. MacGregor and F.B. Oleson (1989) Automated micronucleus scoring workshop report, Environ. Mol. Mutagen., 14, 62-63. Wallen, C.A., R. Higashikubo and L.A. Dethlefsen (1982) Comparison of two flow cytometric assays for cellular RNA-acridine orange and propidium iodide, Cytometry, 3, 155-161.
ELSEVIER
Mutagenesis
Mutation Research 334 (1995) 9-18
Analysis of micronucleated cells by flow cytometry. 4. Kinetic analysis of cytogenetic damage in blood Andrew M. Tometsko *, Stephen D. Dertinger, Dorothea K. Torous Litron Laboratories, 1351 Mt. Hope Ave., Rochester, NY 14620, USA
Received 15 October 1993;revision received 7 March 1994; accepted 18 March 1994
Abstract Micronucleated cells (MN cells) generated spontaneously or by clastogen action accumulate in the peripheral blood of the mouse, and their presence reflects the level of chromosome damage. Traditionally, micronucleated cells have been scored by visual inspection. With the development of flow cytometry based scoring procedures, vast numbers of cells can be analyzed, making it possible to determine the change in the number of MN cells in the total peripheral blood pool. This report describes experiments whereby initial blood samples were obtained before dosing, providing mouse-specific controls for measuring subsequent changes in MN cells. Mice were then dosed with saline (solvent control), methyl methanesulfonate, cyclophosphamide or colchicine every 48 h and bled every 96 h for 12 days. For each blood sample, one million fixed erythrocytes (RBCs) were interrogated for the presence of micronuclei, and regression analysis was used to determine the rate of MN cell influx per day for each animal or sets of animals. To evaluate the effect of treatment on MN induction, the mean slopes of solvent and chemically treated animals were compared using t-tests. The results of these experiments indicate that the kinetics of MN induction continues near the background frequency for saline dosed mice, whereas clastogenic agents or spindle poisons cause a significant influx of MN events into the blood. The results suggest that some studies may benefit from a flow cytometry based analysis of multiple blood samples, especially when the number of mice is limited, or when a weak clastogen is being investigated. Keywords: Micronucleus; Flow cytometry; Clastogens; Chromosome damage; Kinetics
I. Introduction Since its introduction, the mouse micronucleus assay has been in a state of evolution with improvements to every aspect of the method. The assay (Matter and Schmid, 1971; Heddle, 1973; Schmid, 1975) is based on the fact that clastogens
* Corresponding author.
or spindle poisons cause genetic damage to erythroblasts. The resulting accumulation of chromosomes and chromosome fragments can be scored in the red blood cell population which is normally D N A deficient. This assay was originally restricted to bone marrow, where the number of micronucleated cells (MN ceils) could easily be counted by visual inspection of immature polychromatic erythrocytes (PCEs) (Salamone et al., 1980; Heddle et al., 1983; Salamone and Heddle,
0165-1161/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0165-1161(94)00026-L
10
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
1983; Mavournin et al., 1990). However, it was found that micronucleated erythrocytes are not cleared from the blood in the mouse (MacGregor et al., 1980; Schlegel and MacGregor, 1982), thus allowing the analysis to be carried out with blood samples (Mavournin et al., 1990; Collaborative Study Group for the Micronucleus Test, 1992). With manual scoring methods a limited number of cells are analyzed for each mouse (i.e. 10002000 total PCEs per blood or bone marrow sample). Since micronuclei are present as rare events at a frequency of about 0.2% at background levels, analyzing only 1000 cells adversely impacts on assay statistics and results in a large coefficient of variance (_+70%) (Tometsko et al., 1993b). It has been realized for some time that the next step in the evolution of the assay should involve automated methods for objectively and accurately scoring larger numbers of micronucleated erythrocytes to improve assay statistics (Tucker et al., 1989). Over the years, flow cytometry (FCM) has received considerable attention as a possible solution to objectively analyzing micronucleated cells, and a variety of methods have been proposed (Hutter and Stohr, 1982; Grawe et al., 1992; Hayashi et al., 1992a,b; Krishna et al., 1993; Tometsko et al., 1993a,b,c). These methods differ considerably in their approaches to scoring and analyzing micronucleated cells and should be studied carefully for resolution of MN cells and elimination of 'noise' during data acquisition (Heddle et al., 1991). With flow cytometry, the rate of data acquisition is increased by orders of magnitude over hand scoring, and processing speeds of 1 000000 cells per minute have been achieved (Tometsko et al., 1993b,c). The sheer volume of data provided by flow cytometry makes quality control and scoring accuracy important concerns. Therefore, chemical, biological and instrument parameters must be carefully optimized and proper control of the associated variables should be demonstrated as a prelude to a flow cytometry based analysis procedure. With this in mind, the authors focused previous reports on the resolution of MN cells from other cell populations and the development of methods for assessing scoring accuracy. It was only subsequent to
these resolution and accuracy studies that the analysis of a biological response due to chemical treatment was carried out (Tometsko et al., 1993a,b,c). This article describes the analysis of multiple blood samples by flow cytometry in order to determine the kinetics of MN cell influx into the peripheral blood. Multiple sample analysis had been previously reported in conjunction with hand scoring (Luke et al., 1988a,b). Although initial blood samples were obtained, the 0-h data were excluded from the analysis of the treatment groups where data for each mouse were averaged over time to yield a 'temporal average'. In the present method, the initial blood sample plays a central role in the analysis by serving as a mouse specific control (internal standard) for evaluating the net change in MN cells. The bleeding and dosing regimen described herein is based on previous work (Tometsko et al., 1993c) in which mice were dosed at 0 h and 48 h and were bled initially (0 h; before dosing) and at 96 h. For continuity, this previous bleeding and dosing protocol was extended through two additional cycles, terminating with a blood sample obtained on day 12. The results suggest that the analysis of multiple samples by flow cytometry may provide an important advantage when increased sensitivity is required in the micronueleus assay (e.g. studies involving a weak clastogen), or when the number of mice is limited (e.g. studies involving transgenic mice). The methods and procedures presented here are easily adaptable to subchronic studies by periodically obtaining blood samples.
2. Material and methods Biological and chemical considerations
Adult male and female mice (strain BALB/c) from a randomly bred closed colony were purchased from Charles River Breeding Laboratories, Inc., Wilmington, MA. A commercial diet and water were available to the mice ad libitum. An experiment with methyl methanesulfonate involved one treatment group consisting of five male mice, whereas the cyclophosphamide and colchicine studies included both solvent and
A.M. Tometskoet al. / Mutation Research334 (1995)9-18 chemical treatment groups with four male and four female mice in each set. On days 0, 2, 4, 6, 8 and 10, animals were dosed with saline (solvent control), cyclophosphamide (CP: CAS No. 605519-2: Sigma Chemical Co.; 100 /~g/g b.w.) (Tinwell et al., 1990; MacGregor et al., 1980; Salamone et al., 1980; Schlegel and MacGregor, 1982), methyl methanesulfonate (MMS: CAS No. 66-27-3: Eastman Kodak Co.; 100 ~ g / g b.w.) (Jenssen and Ramel, 1976, 1978; Tsuyoshi et al., 1989) or colchicine (CAS No. 64-86-8: Sigma Chemical Co.; 1/~g/g b.w.) (Hayashi et al., 1989). Each of the solutions was prepared in saline. The mice were bled initially (i.e. 0 h; before dosing) and on days 4, 8 and 12. Micronuclei were stained with Hoechst 33258 (HST: CAS No. 23491-44-3: Sigma Chemical Co.) which is specific for DNA (Latt et al., 1976). Propidium iodide (PI: CAS No. 25535-16-4) (Hudson et al., 1969; Wallen et al., 1982) was used to quantitate the PCE population.
Preparation of blood samples for FCM analysis In these studies, tail vein blood samples (5 drops; about 200 ~1)were collected in heparin (0.5 ml; 500 USP units/ml) initially and every 4 days. The blood cells were fixed by adding 180/~1 of a blood suspension to 2 ml of ultracold ( - 70°C) methanol (Tometsko et al., 1993a). All of the samples for each set of mice were stored at - 7 0 ° C until the completion of the dosing regimen. At ultralow temperatures, fixed blood samples are stable indefinitely. At the time of analysis, cells were diluted with sodium chloride/sodium bicarbonate buffer (8 ml) (sodium chloride (9 g) plus sodium bicarbonate (0.445 g) per liter). The cells were collected by centrifugation and were stored at 4°C until analysis. A sample of the pellet (5-10 /zl) was diluted into Tris-borateEDTA (TBE) buffer (1 ml) containing Hoechst 33258 (84 ng/ml TBE). Quantitative analysis of cell populations by flow cytometry The number of MN cells present per one million erythrocytes was determined for every blood sample in this study using a FACStar Plus flow cytometer (Becton Dickinson). The DNA-specific fluorochrome Hoechst 33258 was used to differ-
11
entially stain red blood cells with and without micronuclei. The cells were interrogated with a UV laser beam from a 5-W Innova 90 Argon Ion Laser (Coherent). The fluorescence emission from each cell was collected by photomultiplier tubes (PMTs) and the number of MN cells in the analysis area was tabulated. The FACStar Plus flow cytometer is equipped with four detectors, which are used to sense forward scatter, side scatter, blue fluorescence and yellow fluorescence signals. The blue PMT registered fluorescence emission between 420 nm and 555 nm, whereas the yellow PMT measured emission greater than 580 nm. Data were acquired in the dual parameter mode. The dimensions of the micronucleus analysis window were set using red blood cells containing the malaria parasite Plasmodium berghei as a model. By interrogating one million total blood cells per sample, a coefficient of variance of < 3% should be obtained for rare events present at a 0.2% frequency when analyzed under optimum conditions (Tometsko et al., 1993a,b). Since new MN events must pass through the PCE population, toxicity at the erythroblast level can be detected as a decrease in the number of PCEs present in the blood. FCM can provide a rapid assessment of toxicity as a prelude to micronucleus analysis. In these experiments, the P C E / N C E ratio was determined by staining blood samples with propidium iodide which causes PCEs to fluoresce red when excited by a UV laser beam. Erythroblast toxicity was minimal under the dosing regimens as described.
Slope analysis The experimental design of these studies makes use of initial and subsequent blood samples so that the induction of MN events is carefully monitored over time for each animal. With these multiple data points, it is possible to calculate a regression line for each animal or each set (Cricket-Graph T M Software) which reflects the influx of MN cells per million erythrocytes per day (i.e. the kinetics of MN formation). It should be apparent that a number of statistical methods may be employed to test whether the slopes of these lines are significantly different or whether they might be estimating the same population. For this
A.M. Tometsko et al. /Mutation Research 334 (1995) 9-18
12
study, we chose to analyze the effect of chemical treatment on mean slopes with one-tailed t-tests (Statview II). An additional analysis was performed on colchicine data where the response of each female mouse was transformed into a series of three slopes covering the range 0 - 4 days, 4 - 8 days, and 8-12 days. This treatment is based on experiments which indicate that the influx of new MN cells returns to the spontaneous background level 48 h after dosing. In one such experiment, the ability of the system to reset to background frequencies was evaluated with four male mice that were dosed with MMS ( 1 0 0 / x g / g b.w.) at 0 and 2 days only, and were bled at 0, 4 and 8 days. As expected, the influx was 353 _+ 85 MN cells/ day for the first 4-day cycle. However, the rate of influx was - 2 5 _+ 77 MN cells/day for the second 4-day cycle, showing that MN cell influx returned to background frequencies within 48 h of dosing. Therefore, each cycle represents a sep-
arate and independent experiment, and the starting point for a subsequent cycle can be reset each time a mouse is bled. By calculating the slope for each 4-day time interval as a separate experiment (as discussed for colchicine later), a three-fold increase in the number of independent readings was available for statistical evaluation (one-tailed t-test).
3. Results
As a result of cytogenetic damage, MN cells accumulate in the peripheral blood of mice where they can be analyzed relative to background levels. The initial blood sample (day 0) represents a baseline reading on the spontaneous background frequency of MN cells for each mouse. Subsequent samples therefore reflect any accumulation of MN cells resulting from clastogen exposure. Newly formed MN cells require about 48 h to
Table 1 Cyclophosphamide response Mouse
Sex
Treatment
MN cells
a/106 erythrocytes
Slope b
day0
day4
day 8
day l2
1 2 3 4 5 6 7 8 Average SD
Female Female Female Female Male Male Male Male
Saline Saline Saline Saline Saline Saline Saline Saline
2485 2718 2825 2360 2649 2359 2630 2485 2563 169
2813 2762 2463 2347 2665 2675 2709 2619 2632 155
2608 2734 2890 2479 2808 2717 2679 2731 2706 124
3002 2910 2781 2957 2805 2608 2674 2634 2796 150
34 14 7 48 15 20 3 14 19.4 14.8
9 10 11 12 13 14 15 16 Average SD
Female Female Female Female Male Male Male Male
CP c CP CP CP CP CP CP CP
3010 2375 2624 2420 2560 2539 2594 2569 2586 191
3361 3262 3412 3271 2741 3415 3398 3322 3273 223
5069 3983 4015 4540 4239 3893 4170 4051 4245 388
6210 5212 6335 5179 5760 4729 5246 5357 5504 552
283 231 293 239 277 176 218 227 243.0 39.3
MN cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c CP, cyclophosphamide (100 # g / g b.w.; i.p. injection). The difference between means (eight animals per treatment group) was compared with a one-tailed t-test: p value = 0.0001.
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
reach the bloodstream (MacGregor et al., 1980). In previous studies (Tometsko et al., 1993), mice were dosed with MMS at 48-h intervals and were bled initially and at 96 h. For this study the dosing/bleeding regimen was extended to 12 days (three cycles) during which time five male mice were treated with methyl methanesulfonate. Fig. 1 shows the response of each mouse to multiple doses of MMS. Since this regimen provided multiple blood samples per mouse, a measurement of the kinetics of MN cell influx into the peripheral blood could be readily obtained by regression analysis. For the MMS dosed mice shown in Fig. 1, the regression equation is: y = 1917 + 1126x R 2 = 0.94 This equation describes an influx rate of 1126
13
20000 18000 -
cycle 1
I
cycle 2
]
cycle 3
I
16000 14000
-
12000 MN
cells 10 0 0 0 8000 6000 4000 2000 0 0
i
i
2
4
•
i
,
6 Time
i
•
8
i
r
10
12
(Days)
Fig. 1. The n u m b e r of M N cells per million erythrocytes was plotted as a function of time for five male B A L B / c mice dosed with methyl methanesulfonate (100 t x g / g b.w.) every 48 h and bled every 4 days. Each 4-day period defined one cycle of the b l e e d i n g / d o s i n g regimen. Notice that each of the five mice displayed similar responses to MMS treatment.
Table 2 Colchicine response Mouse
Sex
Treatment
M N cells a/106 erythrocytes
Slope b
day 0
day 4
day 8
day 12
1 2 3 4 5 6 7 8 Average SD
Female Female Female Female Male Male Male Male
Saline Saline Saline Saline Saline Saline Saline Saline
2000 2111 1918 2082 2642 2594 2366 2438 2269 277
1991 2096 2136 2256 2549 2238 2599 2879 2343 303
2084 2287 1993 2401 2780 2376 2766 2414 2388 282
2570 2265 2262 2579 2731 2593 2773 2820 2574 213
45 16 22 41 12 3 35 17 23.9 14.9
9 10 11 12 13 14 15 16 Average SD
Female Female Female Female Male Male Male Male
Colchicine c Colchicine Colchicine Colchicine Colchicine Colchicine Colchicine Colchicine
2168 2221 1903 2212 2087 2236 2519 2423 2221 190
2477 2544 2345 2666 3166 3198 3602 2944 2868 433
3154 3420 3538 3650 3546 4194 4461 3101 3633 474
3970 3745 5307 4852 dead 4685 6233 3217 4572 1022
152 136 285 223 182 209 300 63 193.8 78.4
a M N cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c 1 / x g / g b.w., i.p. injection. T h e difference between m e a n s was compared with a one-tailed t-test: p value = 0.0001. The difference between m e a n s (four female mice per treatment group) was compared with a one-tailed t-test: p value = 0.0015; degrees of freedom = 6.
A.M. Tometsko et al. /Mutation Research 334 (1995) 9-18
14
MN cells per million erythrocytes per day during the MMS exposure time frame. In a second multiple dose experiment, mice were treated with saline (i.e. solvent) or CP, and the number of MN cells per million erythrocytes was obtained for each sample. The means reflect an increase due to CP treatment (see Table 1). The mean slopes obtained by regression analysis describe an influx rate of 19.4 MN cells/day and 243 MN cells/day for saline and CP treated mice, respectively. These results indicate that CP caused a 12.5-fold increase in the rate at which MN cells accumulated in the peripheral blood pool under these exposure conditions. In order to provide a graphic overview, the rate of influx (MN cells/million R B C s / d a y ) for each saline and CP dosed mouse was plotted as a bar graph in Fig. 2. This figure shows that the rate of micronucleus formation for all saline mice was low, whereas an elevated response was obtained for mice dosed with cyclophosphamide. This type of graphic presentation is useful for highlighting the biological variation (i.e. rate of MN cell in-
350
300
250
200 Rate MN
of Influx 15o
100
0 I
2
3
4
5
6
7
8
9
l0
11
12
13
14
15
16
Mouse N u m b e r
Fig. 2. In this experiment, mice were bled every 4 days and injected with 0.9% saline or cyclophosphamide (CP; 100 p~g/g b.w.) every 2 days for 12 total days. Regression analysis was used to determine the rate at which micronucleated cells entered the peripheral blood pool for each mouse over the experimental time frame. The rate of MN cell influx (i.e. MN cells/day) was obtained from the slope of the regression line and was plotted as a bar graph for each mouse, providing an overview of the biological variation for the saline and CP dosed mice.
5000 []
Saline
y - 2249.5 + 24.000x • .
45OO
=
.
R'2 - 0.910 .
~
=
.
.
40OO
Average Number o f MN cells
3500
3000
2500,
20o0
•
i 2
,
i
,
4
i
i
6 Time
8
10
12
(Days)
Fig. 3. The average n u m b e r of MN cells was plotted as a function of time for eight mice that were bled every 4 days and dosed with saline or colchicine (1 ~ g / g b.w.) every second day for 12 days. Each blood sample was fixed, stained with Hoechst 33258, and was quantitatively analyzed for the presence of MN cells by high speed flow cytometry. Regression analysis was performed on the mean MN cell values and the regression equations yielded slopes of 24 MN cells/day and 195 MN cells/day for saline and colchicine dosed mice, respectively.
flux) of individual mice due to clastogen treatment. In a spindle poison experiment, mice were also dosed with colchicine and the results are presented in Table 2. In this experiment regression analysis produced a mean slope for the saline animals of 24 MN cells/day while colchicine treatment resulted in a slope of 194 MN cells/day. Therefore, the effect of colchicine at this dose was an eight-fold increase in the influx of MN cells above the background level. The graphic view of this effect (Fig. 3) was obtained by plotting the average MN value per million RBCs for each treatment group versus time. This presentation shows that for this low level of genotoxic activity, a longer dosing and bleeding regimen is preferred since it allows the treatment profile to diverge from the background profile. While solvent controls show a slight rise over time, clastogen treatment results in a more rapid elevation (accumulation) of MN events in the blood. Additional analysis of the colchicine data was performed to illustrate how multiple sample analysis may offer a means for improving the sensitivity of the assay or possibly reducing the number
15
A.M. Tometsko et aL / Mutation Research 334 (1995) 9-18
of animals required. For this demonstration, it was necessary to reduce the number of animals per set from eight to four so that the p values would be higher and would be more sensitive to changes in the degrees of freedom. Accordingly, only data from the female mice (Table 2) were used for this analysis and each dosing/bleeding cycle was processed as a separate and discrete experiment, resulting in a three-fold increase in the degrees of freedom. With this adjustment, it was possible to obtain a p value with four mice that was obtained when the treatment group consisted of eight animals each. Operationally, the data for each of the four saline and colchicine animals were transformed into a series of three successive slopes corresponding to days 0-4, days 4 - 8 and days 8-12, respectively (see Table 3). Given the time frame in which clastogen induced MN cells enter the peripheral blood pool, each successive slope reflects a separate application of test material, and each analysis can be thought of as an independent observation. The statistics of the assay are affected considerably by the three-fold increase in statistical sample size. For example, when the slopes were compared for four saline dosed and four colchicine dosed females, a p value of 0.0015 was obtained (one-tailed t-test; degrees of freedom = 6). This is in contrast to a p value of 0.0001 when the data were processed as three
independent slope readings per mouse (degrees of freedom = 22). These results indicate that by increasing the degrees of freedom, multiple independent readings improve the confidence level of statistical analysis, even when small numbers of animals are used.
4. Discussion The purpose of this presentation is to demonstrate that FCM based methods and procedures can be used to assess the kinetics of MN cell influx into the peripheral blood. In these studies, an initial blood sample was obtained from each mouse and was used for subsequent comparisons. The initial blood samples provided a baseline reading on each animal, showing the frequency at which micronucleated cells enter the blood cell pool under normal physiological conditions. If mice are treated with saline, the influx of MN ceils should continue at or near the background level, resulting in only a slight increase in MN ceils over the experimental time frame. Stimulation of the erythropoietic system, a n d / o r changes in hormone and cytokine levels may account for the slight rise in the background frequency of MN cells over time. In any case, data presented here indicate that bleeding mice for periodic MN readings does not significantly affect the sponta-
Table 3 Colchicine response: calculation of three slopes per mouse Mouse Sex Treatment MN cells a/106 erythrocytes (day 0)
slope b
(day 4)
slope
(day 8)
slope
(day 12)
1 2 3 4
Female Female Female Female
Saline Saline Saline Saline
(2000) (2111) (1918) (2082)
-2 -4 55 44
(1991) (2096) (2136) (2256)
23 48 - 36 36
(2084) (2287) (1993) (2401)
122 - 6 67 45
(2570) (2265) (2262) (2579)
9 10 11 12
Female Female Female Female
Colch c Colch Colch Colch
(2168) (2221) (1903) (2212)
77 81 111 114
(2477) (2544) (2345) (2666)
169 219 298 246
(3154) (3420) (3538) (3650)
204 81 442 30i
(3970) (3745) (5307) (4852)
MN cells, micronucleated peripheral blood erythrocytes. b Slope, change in micronucleated cells per day. c Colch, colchicine(1 /zg/g b.w.; i.p. injection). The difference between means (four animals or 24 slopes per treatment group) was compared with a one-tailed t-test: p value = 0.0001; degrees of freedom = 22. a
16
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
neous frequency of MN cell formation. On the other hand, a larger divergence from background values can result during clastogen exposure. The analysis of multiple blood samples by hand scoring has been previously reported (Luke et al., 1988a,b), and demonstrated a clastogen dependent accumulation of MN cells in the peripheral blood. In this case, the numbers of MN NCEs at the different time points were pooled (summed) and the 'temporal average' was calculated for comparison purposes. In contrast, in these flow cytometry experiments the numbers of MN RBCs at each time point were treated as separate and discrete readings in order to assess the current status of MN cells in each mouse. Although 'temporal averages' might be calculated, it seemed more informative to use high speed FCM data acquisition to measure the kinetics of MN cell influx following clastogen exposure. Since MN cells are not cleared from the peripheral blood pool, any change in their frequency can be detected in the RBCs if enough cells are analyzed. In fact, millions of cells can be scored for MNs by taking advantage of the processing speed of modern flow cytometers. (Note that these experiments are only possible because resolution and scoring accuracy have been optimized.) FCM based analysis of multiple samples provides more information regarding the impact of a chemical on each individual biological system and thus differs significantly from conventional micronucleus assays whereby a single blood or bone marrow sample is analyzed. These experiments also differ from our previous studies in which only two blood samples were obtained at day 0 and day 4, and the net change in MN cells (AMN) was determined for each mouse. Fig. 1 shows the response of five male mice to three cycles of the bleeding/dosing regimen. This figure provides important insights into multiple sample analysis, showing that the rate of MN cell influx during cycles two and three was as high as or higher than during cycle one. In the absence of PCE toxicity, a continuation of the first cycle rise in MN cells would be expected (i.e. continuity of response). This figure also suggests that bleeding can enhance the response to clastogens, since the number of MN cells is accumulating faster in the
third cycle than in the first and second cycles. This effect might be expected since bleeding stimulates erythropoiesis, which sensitizes erythroblasts to clastogens (Hirai et al., 1991). Finally, it is apparent from Fig. 1 that each mouse in the set is reflecting the same general sensitivity to MMS treatment. Thus multiple dose/multiple bleeding protocols as described in this presentation provide significantly more information regarding the response of each mouse and may prove to be statistically more powerful. With multiple samples, all of the data from each mouse can be used in the evaluation of a clastogenic response (in this case four data points). Regression analysis has been particularly useful for processing multiple sample data since the resulting regression lines can signal cytogenetic damage for a single mouse, or sets of mice. In the present examples, when mice were dosed with clastogens or a spindle poison, regression line slopes reflected the elevated MN cell level and were 8-12 times above background values. By calculating slopes, one essentially compresses the data from multiple samples into a rate of MN cell influx corresponding to the specific dosing conditions. A comparison of the rates of MN influx (i.e. slopes) for individual animals (Fig. 2) provides an overview of biological variation in mice dosed with saline or CP and is helpful for analyzing the consistency of a positive response. In these studies, all of the chemically dosed mice exhibited a rise above the background level, even when the response was relatively weak as with colchicine (Table 2). In theory, multiple samples from each mouse should provide a more accurate and detailed assessment of a clastogenic response compared to the conventional single sample per animal analysis. Each measurement of MN cell influx should be evaluated for consistency with respect to preceding or following rates. For example Fig. 3 shows an increase in accumulated MN cells with each repeat of the dose/bleeding cycle. With the present three cycle protocol, each animal also provides three independent measurements of the rate of MN cell influx. Since clastogen induced MNs will have entered the peripheral blood pool by the 48-h mark, each 4-day time interval can be
A.M. Tometsko et al. / Mutation Research 334 (1995) 9-18
considered a separate independent experiment to test the effect of treatment on MN formation. Accordingly, each blood sample can be used as the basis of comparison for evaluating the change in MNs upon subsequent readministrations of test material. For each mouse, these successive slopes should reflect the overall characteristics of the regression line (e.g. Fig. 3), but will contribute more degrees of freedom to statistical analysis procedures. For example, when slope analysis was carried out on successive readings for four saline and four colchicine animals (i.e. change in MNs for days 0-4, days 4-8, days 8-12), the degrees of freedom were increased from 6 to 22, and the confidence interval was markedly enhanced. Since each mouse contributed three independent slopes reflecting the accumulation of MN ceils between four successive blood readings, a larger number of discrete observations were supplied to'the statistical operation. By increasing the degrees of freedom in this manner, the confidence interval resulting from an intergroup comparison of mean slopes was enhanced (i.e. the p value of a onetailed t-test was lowered approximately 10-fold). With multicycle analysis (e.g. three slopes/mouse; Table 3), four mice were able to provide a level of statistical confidence that required eight animals per treatment group in Table 2. Bleeding and dosing regimens should be designed to highlight any increase in MN cells that might occur. Based on a large number of experiments, a 48-h dosing and 96-h bleeding schedule seems quite adequate to allow MN cells to accumulate in the peripheral blood. The methods that we have outlined are compatible with other dosi n g / b l e e d i n g schedules, so long as bleeding is performed after sufficient time has elapsed to allow chemically induced MN cells to enter the peripheral blood pool. The choice of dosing and bleeding protocols should be guided by the toxicity of the test substance and the rate of accumulation of MN cells into the blood pool under the specific experimental conditions. The considerations of dosing and sacrifice time that are of critical importance for the analysis of MN PCEs are not pertinent to the analysis of accumulated MN cells in the total red blood cell pool. Bleed-
17
ing and dosing regimens involving extended treatment times allow a longer period for MN cells to accumulate if a clastogenic response is occurring. Thus, the time a cell spends in the short-lived PCE population is not critical to the analysis of accumulated MN events in peripheral blood. We recommend that an initial sample should be obtained before any dosing begins. We continue to analyze one million total blood cells for each sample in order to score sufficient rare MN cells in a large cell pool. In the present experiments, data analysis was limited to 4 000 000 cells per mouse which provided a suitable demonstration of the capabilities of flow cytometry to determine the rate of influx of MN cells due to various chemical treatments. The general method described here provides a means for precisely and rapidly evaluating the kinetics of chemically induced MN formation and readily complements subchronic studies.
Acknowledgement This research was supported by Grant 2 R44 G M 35043-02 from the National Institutes of Health.
References Collaborative Study Group for the Micronucleus Test (1992) Micronucleus test with mouse peripheral blood erythrocytes by acridine orange supravital staining, Mutation Res., 278, 83-98. Grawe, J., G. Zetterberg and H. Amneus (1992) Flow-cytometric enumeration of micronucleated polychromatic erythrocytes in mouse peripheral blood, Cytometry, 13, 750758. Hayashi, M., S. Sutou, H. Shimada, S. Sato, Y.F. Sasaki and A. Wakata (1989) Difference between intraperitoneal and oral gavage application in the micronucleus test, Mutation Res., 223, 329-344. Hayashi, M., H. Norppa, T. Sofuni and M. Ishidate Jr. (1992a) Mouse bone marrow micronucleus test using flow cytometry, Mutagenesis, 7, 251-256. Hayashi, M., H. Norppa, T. Sofuni and M. Ishidate Jr. (1992b) Flow cytometric micronucleus test with mouse peripheral erythrocytes, Mutagenesis, 7, 257-264. Heddle, J.A. (1973) A rapid in vivo test for chromosome damage, Mutation Res., 18, 187-190.
18
A.M. Tometsko et al. / M u t a t i o n Research 334 (1995) 9-18
Heddle, J.A., M. Hite, B. Kirkhart, K. Mavournin, J.T. MacGregor, G.W. Newell and M.F. Salamone (1983) The induction of micronuclei as a measure of genotoxicity, Mutation Res., 123, 61-118. Heddle, J.A., M.C. Cimino, M. Hayashi, F. Romagna, M.D. Shelby, J.D. Tucker, Ph. Vanparys and J.T. MacGregor (1991) Micronuclei as an index of cytogenetic damage: past, present, and future, Environ. Mol. Mutagen., 18, 277-291. Hirai, O., Y. Miyamae, Y. Fujino, H. Izumi, A. Miyamoto and H. Noguchi (1991) Prior bleeding enhances the sensitivity of the in vivo micronucleus test, Mutation Res., 264, 109-114. Hudson, B., W.B. Upholt and J. Divinny (1969) The use of an ethidium analogue in the dye-buoyant density procedure for the isolation of closed circular DNA: The variation of the superhelix density of mitochondrial DNA, Proc. Natl. Acad. Sci. USA, 62, 813-820. Hutter, K.J. and M. Stohr (1982) Rapid detection of mutagen induced micronucleated erythrocytes by flow cytometry, Histochemistry, 75, 353-362. Krishna, G., D. Brott, G. Urda, M. McKeel, J. Zandee and J. Theiss (1993) Comparative micronucleus quantitation and pre- and post-column fractionated mouse bone marrow by manual and flow methods, Mutation Res., 302, 119-127. Latt, S.A. and G. Stetten (1976) Spectral studies on 33258 Hoechst and related bisbenzimidazole dyes useful for fluorescent detection of deoxyribonucleic acid synthesis, J. Histochem. Cytochem., 24, 24-33. Luke, C.A., R.R. Tice and R.T. Drew (1988a) The effect of exposure regimen and duration on benzene-induced bone-marrow damage in mice. I. Sex comparison in D B A / 2 mice, Mutation Res., 203, 251-271. Luke, C.A., R.R. Tice and R.T. Drew (1988b) The effect of exposure regimen and duration on benzene-induced bonemarrow damage in mice. II. Strain comparisons involving B6C3F1, C57B1/6 and D B A / 2 male mice, Mutation Res., 203, 273-295. MacGregor, J.T., C.M. Wehr and D.H. Gould (1980) Clastogen-induced micronuclei in peripheral blood erythrocytes, Environ. Mutagen., 2, 509-514. Matter, B. and W. Schmid (1971) Trenimon-induced chromo-
somal damage in bone-marrow cells of six mammalian species evaluated by the micronucleus test, Mutation Res., 12, 417-425. Mavournin, K.H., D.H. Blakey, M.C. Cimino, M.F. Salamone and J.A. Heddle (1990) The in vivo micronuclues assay in mammalian bone marrow and peripheral blood. A report of the U.S. Environmental Protection Agency Gene-Tox Program, Mutation Res., 239, 29-80. Salamone, M.F. and J.A. Heddle (1983) The bone marrow micronucleus assay: rationale for a revised protocol, in: F.J. de Serres (Ed.), Chemical Mutagens 8, Plenum, New York, pp. 111-149. Salamone, M.F., J. Heddle, E. Stuart and M. Katz (1980) Towards an improved micronucleus test. Studies on 3 model agents, mitomycin C, cyclophosphamide, and dimethyl benzanthracene, Mutation Res., 74, 347-356. Schlegel, R. and J.T. MacGregor (1982) The persistence of micronuclei in peripheral blood erythrocytes: detection of chronic chromosome breakage in mice, Mutation Res., 104, 367-369. Schmid, W. (1975) The micronucleus test, Mutation Res., 31, 9-15. Tometsko, A.M., D.K. Torous and S.D. Dertinger (1993a) Analysis of micronucleated cells by flow cytometry. 1. Malaria as a biological model, Mutation Res., 192, 129133. Tometsko, A.M., S.D. Dertinger and D.K. Torous (1993b) Analysis of micronucleated cells by flow cytometry. 2. Scoring rare events with high fidelity, Mutation Res., 192, 137-143. Tometsko, A.M., D.K. Torous and S.D. Dertinger (1993c) Analysis of micronucleated cells by flow cytometry. 3. Advanced technology for detecting clastogenic activity, Mutation Res., 192, 145-153. Tucker, J.D., J.A. Heddle, J.T. MacGregor and F.B. Oleson (1989) Automated micronucleus scoring workshop report, Environ. Mol. Mutagen., 14, 62-63. Wallen, C.A., R. Higashikubo and L.A. Dethlefsen (1982) Comparison of two flow cytometric assays for cellular RNA-acridine orange and propidium iodide, Cytometry, 3, 155-161.